DPhilin Cancer Science 2024 Project Booklet V6

DPhil in Cancer Science
University of Oxford
2024 Intake Project Book
DPhil in Cancer Science
2024 Intake Project Book
Introduction
This handbook provides an overview for prospective students looking to study for a DPhil in
Cancer Science starting in 2024 at Oxford University. The Programme provides research based
doctoral training for cancer researchers from clinical, biological, engineering, mathematics, and
statistics background. Students will receive a world-leading research training experience that
integrates an education initiative spanning cancer patient care, tumour biology and research
impact; on- and post-programme mentorship; and a specialised, fundamental, subject-specific
training tailored to individual research needs. Students participating in the scheme will be
offered:
• a choice of interdisciplinary cutting-edge cancer research projects.

• the ability to gain a working in-depth knowledge of the fundamentals of cancer biology
and cancer patient care through advanced level seminars.
• a world-renowned research environment that encourages the student’s originality and
creativity in their research.
• opportunities to develop skills in making and testing hypotheses, in developing new
theories, and in planning and conducting experiments.
• an environment in which to develop skills in written work, oral presentation and
publishing the results of their research in high-profile scientific journals, through
constructive feedback of written work and oral presentations.
At the end of their DPhil course, students should:

• have a thorough knowledge of the basic principles of cancer research including the
relevant literature and a comprehensive understanding of scientific methods and
techniques applicable to their research.
• be able to demonstrate originality in the application of knowledge, together with a
practical understanding of how research and enquiry are used to create and interpret
knowledge in their field.
• have developed the ability to critically evaluate current research and research
techniques and methodologies.
• be able to act autonomously in the planning and implementation of research.
• have the grounding for becoming an influential cancer researcher of the future.
Selection Criteria & Eligibility
There are four tracks in the programme as described below, meaning that non-clinicians,
undergraduate medical students and post-graduate medical trainees are all eligible to apply
for the fully funded (at home rate) studentships.
Application Track 1 – Clinical Trainees. Qualified doctors at all stages of training from the
foundation training to higher specialist training.
Application Track 2 – Medical Undergraduates. Medical students who are currently
undertaking a primary medical qualification (MBBS, MBChB or equivalent). At entry, we will
be looking for evidence of completion of at least the first two years of a primary medical
qualification and achievement at the level of an upper-second or first-class honours degrees
(or iBSc).
Application Track 3 – Non-Clinical/Fundamental Scientist. Science graduates that hold (or be
predicted to achieve) the equivalent of a first-class or strong upper second-class
undergraduate degree with honours in biological, medical, or chemical science, as
appropriate for the projects offered.
Application Track 4 – Non-Clinical/Fundamental Scientist. Science graduates that hold (or be
predicted to achieve) the equivalent of a first-class or strong upper second-class
undergraduate degree with honours in engineering, mathematical/data, or physical science,
as appropriate for the projects offered.
All applicants will be judged on the following:
• commitment and passion to a career in cancer research
• evidence of motivation for and understanding of the proposed area of study
• commitment to the subject, beyond the requirements of the degree course
• preliminary knowledge of relevant research techniques
• capacity for sustained and intense work
• reasoning ability and academic curiosity.
Funding
All offered places are fully funded at the home rate. This includes salary/stipend,
University/College fees, and a research consumables budget of ~£13k p.a.
Salary and stipend provisions are summarised below:
• Application Track 1: 3 years of salary at Grade E63 or E64 Clinical Researcher rate.
• Application Track 2: 3 years of stipend at the flat rate of £21,000 per annum.
• Application Tracks 3 & 4: 4 years of stipend at the flat rate of £21,000 per annum.
1
International applicants are eligible; however, funding is limited to the Home level for this
programme and therefore international applicants would need to either source further funding
or support themselves financially for the remaining fees.
Notable Scholarships
Black Academic Futures Scholarships
These awards offer UK Black and Mixed-Black students scholarship funding to pursue graduate
study at Oxford, alongside a programme of on course mentoring and support. The Medical
Sciences Division has guaranteed places across its DPhil courses (including the DPhil in Cancer
Science). For more information, visit the Black Academic Futures website.
To receive a Black Academic Futures Scholarship, submit your application to the DPhil in Cancer
Science Programme by the December deadline. All those that include eligible ethnicity will
automatically be considered. You do not need to submit any additional documents and there
is no separate scholarship application form for these awards.
How to Apply
A detailed summary on how to apply can be found here. In brief, prospective students apply
with a prioritised list of three projects selected from this booklet by Friday 1st December 2023.
Shortlisted students will be invited to interview in January. If successful, students will be
allocated a project on the basis of their ranking during the review process. It is strongly
suggested that students contact supervisors of projects they are interested in applying for prior
to application.
2
Projects
Projects are listed below in the following structure “Title Eligible Application Tracks – Primary Supervisor
Page number.”
Clicking on the project title below will take you to the project page.
Project Name Page Number
Comprehensive Proteomics Study of Oncogenic JAK2 as a Basis for Improved Therapies
Against Myeloproliferative Neoplasms 1,2,3,4 – Dr Adán Pinto-Fernández 7
Targeting radio-ligand therapy to protonated CEACAM6 epitopes induced in acid-resistant
cells under improved oxygen – release from biochemically rejuvenated transfusions 2,3,4 -
Prof. Pawel Swietach, Prof. Katherine Vallis and Prof. Simon 9
Understand the developmental origin of Clear Cell Sarcoma using long-read single cell
sequencing .1,3,4 – Assoc Prof. Sarah Snelling 11
How does macrophage marker F4/80 control peripheral tolerance during tumour
progression? 1,2,3,4 – Prof. Kim Midwood 13
Chronic infection, host immunity, and cancer risk 3,4 – Ling Yang 15
Myeloid cells in sustained intestinal inflammation and colorectal cancer: the role of IRF5
1,2,3– Prof. Irina Udalova 17
Cancer vaccination for precision prevention of Lynch Syndrome 1,2,3– Prof. Simon Leedham
& Assoc Prof. David Church 19
Therapeutic manipulation of reactive oxygen species in ATRX-deficient cancers 1,2,3 – Prof.
Ester Hammond 21
Identifying ploidy-dependent mitotic vulnerabilities during DNA damage repair and
chromosome segregation 1,2,3 – Prof. Ulrike Gruneberg 23
The role of mutational order in colorectal cancer evolution, epigenetics and response to
therapy 1,2,3 – Prof. Simon Buczacki 25
Spatial resolution of the human transcriptome during gastrointestinal tumorigenesis –
1,2,3,4 Dr. Francesco Boccellato 27
ADP-ribosyl hydrolase as a biomarker for PARP inhibitor sensitivity/resistance 1,2,3,4 – Prof.
Ivan Ahel 29
Improving Immunotherapeutic Efficacy in Colorectal Cancer Using Ultrasound-activated
Nanoparticles and Image-guided Drug Delivery 1,2,3,4 – Prof. Tim Elliot 31
Uncovering the regulation and functions of supermeres in colorectal cancer - 1,2,3,4 – Prof.
Clive Wilson 34
Improving CAR-T cells for B-ALL1,2,3,4 – Prof. Omer Dushek 37
Deciphering spatial differences in Histopathological subtype of colorectal cancer liver
metastasis (CRLM) 1,2,3,4 – Dr. Alex Gordon-Weeks 39
Targeting innate immunity for intestinal injury recovery - 1,2,3 Dr. Monica Olcina 41
Investigating an interplay between autoimmunity and cancer 1,2,3 – Prof. Xin Lu 43
3
An integrated systems biology approach to investigate the spatial Myeloma tumour
microenvironment 1,2,3,4 – Prof. Udo Oppermann 45
Do mutations in cancer arise through histone post-translational modifications? 1,2,3,4 –
Assoc Prof. Peter Sarkies 47
Characterizing the tumour microenvironment of oesophageal cancer to uncover key
factors in response to immunotherapy 1,2,3 – Carol Leung & Prof. Benoit Van den Eynde 49
Tertiary Lymphoid Structures in Lung Cancer 1,2,3 - Dr. Isabela Pedroza-Pacheco 51
Interrogating the fibroblast phenotype of DNA repair deficient cancer 1,2,3 – Assoc Prof.
Eileen Parkes 53
Exploring the biological drivers of 1q21+ high risk Myeloma by using multi-OMICS analysis
of patient derived tumour, immune cells and bone marrow aspirates1,2,3,4 – Prof. Anjan
Thakurta 55
Development of Spatially Fractionated Radiotherapy Techniques 1 – Assoc Prof. Geoff
Higgins 57
Investigating the role of hypoxia for the FLASH effect by combining FLASH Radiation with
hypoxia-modulated anticancer drugs,1,3,3,4 – Kristoffer Petersson 59
Using Long-read Sequencing to Advance Personalised Decision Making in Multiple
Myeloma 1,2,3,4 – Assoc Prof. Adam Cribbs 61
The role of oxygen chemosensitivity in tumourgenesis 1,2,3,4 – Prof. Richard White 64
Exploiting synthetic defects in metabolism and DNA repair to improve the treatment of
glioma and AML 1,2,3 - Prof. Peter McHugh 66
BLOod Test Trend for cancEr Detection (BLOTTED): an observational and prediction model
development study using English primary care electronic health records data 1,2,3,4 – Dr.
Pradeep S. Virdee 68
Tackling Cancers Defective of High-Fidelity DNA Repair Mechanisms 1,2,3,4 – Fumiko Esashi
70
Predicting Response to Therapy in Oesophageal Cancer 4 – Prof. Jens Rittscher 73
Developing single-cell transcriptomics tools for PARP inhibitor resistance in BRCA1/2-
deficient cells and tumours1,2,3,4 – Prof Madelena Tarsounas 75
Harnessing measurements of the tumour microenvironment to improve the early detection
of prostate cancer. 1,3,4 – Professor Richard Bryant 77
Investigating Hypoxic Adaptation in Glioblastoma (GBM) Stem Cells through Pooled
Kinome-Wide CRISPR-Cas9 Knockout Screen 1,2,3 – Dr. Sneha Anand 79
Dietary modification to augment colorectal cancer treatment 1,2,3 – Dr Dimitrios
Koutoukidis 81
Investigating the adaptive immune responses to tumour neo-antigen and the impact on
patient disease course. 1,2,3,4 –Assoc Prof. Rachael Bashford-Rogers 83
Epigenetic control of cancer cell phenotypes via nuclear F-actin based chromosome
motility. 1,2,3 – Prof. Eric O'Neil 85
4
Modulation of tumour immunogenicity by IGF’s in Prostate Cancer – 1 Dr Valentine
Macaulay 87
Understanding the mechanisms by which molecular and phenotypic heterogeneity of MLL-
rearranged infant ALL affects clinical outcome– 1 - Assoc Prof. Anindita Roy 90
Describing T Cell recognition of tumours by machine-learning and statistical models 4 –
Assoc Prof. Hashem Koohy 93
Urological cancers beyond the microscope; novel multiomic analysis of features associated
with DNA instability and the tumour immune micro-environment 1,2,3,4 – Assoc Prof. Clare
Verrill 95
Investigating the role of the ubiquitin ligase BIRC6 in aneuploid glioblastoma cell survival
1,2,3 – Dr Paul Elliott 97
Characterising the NK/Myeloid crosstalk during tumour immune escape 3 – Assoc Prof.
Audrey Gérard 99
The impact of Hypoxia on HLA-E surface expression and peptide presentation3 – Assoc Prof.
Geraldine Gillespie 101
An interdisciplinary approach to understand how interactions between proliferating and
invasive melanoma cells can promote metastasis 4 – Prof. Ruth Baker 103
Exploring different modalities of growth factor inhibitor to treat cancer.2,3 – Dr. Jon Elkins
105
The inactivation of p97 system in colorectal cancer therapy 1,2,3,4 – Prof Kristijan Ramadan
107
Assessment of Oesophageal cancer patient responses to immunotherapy via human tissue
avatars.1,2,3 – Prof. Richard Owen 109
The T-cell receptor landscape of adult diffuse gliomas, a non-invasive tool for tumour
detection and classification? 1,2,3,4 – Assoc Prof. Olaf Ansorge 110
Take epigenetic diagnostics “from bench to bedside”1,2,3,4 –Prof. Prof Benjamin Schuster-
Böckler 112
Spatial interrogation of low grade prostate cancer to identify genomic events responsible
for driving indolent not aggressive disease,1,2,3,4 – Dr. Alistair Lamb 114
New Immune Therapies for Acute Myeloid Leukaemia (AML) And Myeloid Blood Cancers
1,2,3,4 – Prof. Paresh Vyas 116
Mapping the ovarian cancer ascites ecosytem for discovering novel therapeutic
approaches 1,2,3 – Dr. Nicola Ternette 118
Elucidating the role of trans-lesion synthesis DNA polymerases in mutational processes and
therapy resistance,1,2,3,4 - Marketa Tomkova 120
Multi-cancer detection testing in clinical practice. 1,2,3,4 - Dr. Brian Nicholson 122
Understanding STING regulation in cancer and the crucial role of ubiquitination in the ER –
1,3 Prof. John Christianson 124
5
Data Driven approaches for addressing inequalities in Cancer Outcomes - 1,2,3,4 Prof Eva
Morris 126
Molecular enhancement of vaccine-induced immune responses for cancer therapy - 1,2,3
Prof. Paul Klenerman 128
Peptide specific natural killer cell receptors in cancer – 2,3 Malcom Sim 130
Modulating inhibitory receptor signalling to enhance the T-Cell response in Cancer – 2,3,4
Ricardo Fernandes 132
Engineered oncolytic herpes virus therapy for neuroblastoma – 1,2,3,4 Dr Alistair Easton 134
Pre-Clinical modelling of oncolytic virus therapy for hepatic Colorectal Cancer metastases
using ex-vivo, perfused, hepatectomies 1,2,3,4 Dr Kerry Fisher 136
Building Patient specific digital twins for cancer drug development and cancer treatment -
1,2,3,4 Assoc Prof. Rachael Bashford Rogers 138
The Oxford Cancer Vaccine Project – Move from neoantigens to state of the art mRNA-
Lipid nanoparticles (LNP) - 1,2,3,4 Cesar Lopez-Camacho & Lennard Lee 140
Influence of viral based delivery methods on distinct pathways of antigen uptake and
presentation: implications for precision vaccine development - 1,2,3,4 Prof. Sarah Rowland-
Jones 142
Bioinformatics and machine learning approaches for early cancer detection in a resource
restricted health system using whole genome nanopore sequencing of cell-free DNA. - 1,2,3,4
Prof. Anna Schuh 144
Spatial analysis of the bone marrow in myeloproliferative neoplasms DNA. -4 Prof. Helen
Byrne 147
6
Comprehensive Proteomics Study of Oncogenic JAK2 as a Basis for
Improved Therapies Against Myeloproliferative Neoplasms 1,2,3,4 – Dr
Adán Pinto-Fernández
Primary Supervisor: Dr. Adán Pinto-Fernández
Additional Supervisors: Prof. Benedik Kessler
Eligibility: All tracks are eligible to apply for this project.
Abstract of the project

Myeloproliferative neoplasms (MPN) are progressive blood cancers where the bone marrow overproduces mature white blood
cells. Of interest, an activating mutation in a cellular protein called JAK2 is found in most MPN cases. Ruxolitinib (Jakafi) is a
highly potent and selective JAK2 inhibitor approved for MPN in 2011 and has demonstrated major benefits to patients. However,
most patients with mutated JAK2 do not respond to this drug and, therefore, the identification of more efficient therapies
remains critical. (More info in references 1 and 2, at the end of the proposal). This PhD project aims to understand why current
therapies targeting JAK2 for the treatment of myeloproliferative neoplasms do not work in patients carrying the above-
mentioned mutation. Understanding the functional consequences of this mutation will provide novel insights for overcoming
resistance to therapy and eventually improve the survival and life quality of MPN patients. We performed a comprehensive mass
spectrometry-based interactomic, proteomic, and phospho-proteomic analysis to identify proteins associated with activated
JAK2 in five different blood cancer cell lines harbouring wildtype JAK2 (TF1 and F36P) and the mutation associated with MPN:
JAK2 V617F (HEL, SET2, and UKE1). The cells were treated with Erythropoietin (EPO) at different time points (0, 5 min., 15 min.,
and 30 min.) in the presence/absence of two JAK2 inhibitors, 1824 (Ruxolitinib) or 532. Interactome data has been fully analysed
and proteome and phospho-proteome data have been acquired and preliminarily analysed. The proposed research involves the
validation of the interactome data, the analysis of the proteome and phospho-proteome data and the study of JAK2 roles in
MPN resistance to immunotherapy. Project overview in Figure 1.
Figure 1: Discovery of JAK2 wildtype versus V617F mutant dependent molecular pathways. Proteomic workflow
including interactome, phospho-proteome, and total proteome analysis of five different blood cancer cell lines.
Proposed research includes the validation of the JAK2 interactome, the comprehensive bioinformatic analysis of
the total proteome and phosphor-proteome data and the study of JAK2 in cancer immunogenicity.
Analysis of the JAK2 interactome unveiled the association of wildtype JAK2 with mitochondrial components upon activation
with EPO. Notably, JAK2 inhibitors did prevent EPO-dependent binding of JAK2 WT to such mitochondrial components. Most
importantly, this regulation seemed to be lost in cells harbouring the activating V617F mutation (Figure 2), unveiling a
potential resistance mechanism in the cells with the mutation.
Our results represent novel biology associated with JAK2 signalling and EPO activation cascades and could possibly provide a
potential link to oxygen sensing and energy metabolism, affecting cellular proliferation and differentiation. Supporting these
observations, previously published data described that interfering with mitochondrial functions induced cell death in MPN cells
containing the activating mutation in JAK2. In summary, the appointed PhD student, with the guidance of experts in cancer
research and applying advanced biomedicine and bioinformatics tools, will aim to validate and expand the above-described
findings in order to better understand how mutated JAK2 works and to identify novel, more efficient, treatments for MPN.
7
Figure 2: JAK2 inhibition prevents
association of mitochondrial
components to wildtype (circled in
blue), but not to V617F JAK2.
Comparative scatter plots showing
the JAK2 interactomes in WT cells
(left) and JAK2V617F cell lines (right).
Conditions EPO 30 minutes treatment
in the presence/absence of JAK
inhibitor 1824 (Ruxolitinib) were
plotted. The bait is labelled in blue
and selected interactors in red.
Research objectives and proposed outcomes
• To study the co-localisation/interaction between JAK2 (mutant and WT), upon EPO stimulation (+/- JAK inhibitor and +/-
JAK2 genetic inhibition), and the identified mitochondrial components by co-immunoprecipitation and confocal
microscopy. These experiments aim to validate the proteomics data.
• To measure mitochondrial activity reflected by oxygen consumption rate (OCR) and extracellular acidification rates
(ECAR) in the same conditions using a Seahorse Real-Time Cell Metabolic Analyser and ROS production using molecular
probes in a live cell imaging system.
• Evaluate whether combined use of JAK2 inhibitors and ROS inducers such as LAAO, CCCP, Rotenone, Cisplatin, enhances
MPN anti-tumour therapy.
• Comprehensive bioinformatic analysis of the remaining data (phospho-proteome and total proteome) and further
validation.
• Study of JAK2-dependent tumour antigenicity in T cell activation/T cell killing assays (in collaboration with Prof. Tao
Dong).
Translational Potential
With this research plan, we expect to uncover novel JAK2 biology that could explain resistance mechanisms to current therapy
and allow the characterisation of additional targets for the treatment of MPN. With the first three objectives of the proposal, we
want to validate our interactomics data suggesting that JAK2, upon activation, has important roles in mitochondrial function and
cellular metabolism. For instance, it has been shown that mutated JAK2 induced the accumulation of reactive oxygen species
(ROS) in patients with MPN, highlighting key roles for the JAK2 pathway in mitochondrial activity and intracellular ROS
generation (reference 3). The authors of this study suggested that targeting ROS accumulation might prevent the development
of JAK2V617F MPN. Further supporting this hypothesis, another study demonstrated that a ROS inducer exhibited cytotoxicity
and induced apoptosis in JAK2V617F MPN cell lines in a ROS production-dependent manner (reference 4). Based on this, we will
determine whether ROS inducers enhance the anti-cancer effects of JAK2 inhibition. Importantly, additional JAK2 biology will be
explored using the remaining proteomic analysis, potentially providing additional insights into resistance to JAK inhibition.
Finally, the interplay between JAK2 status/inhibition and its implications to resistance to current immunotherapy have been
described (reference 5) and will be investigated using relevant T cell models.
Summarise the training opportunities.

The student will have opportunities to gain expertise in advanced bioinformatics, proteomics, metabolic studies (Seahorse and
mass-spectrometry), cellular biology (immunofluorescence, tissue culture, live cell imaging, gene knockdown/knockout etc),
immunology (T cell activation and T cell killing) and biochemistry (immunoprecipitation and immunoblotting).
Relevant References
1.Schieber M, Crispino JD, Stein B. Myelofibrosis in 2019: moving beyond JAK2 inhibition. Blood Cancer J. 2019;9(9):74.
2.Staerk J, Constantinescu SN. The JAK-STAT pathway and hematopoietic stem cells from the JAK2 V617F perspective. JAKSTAT. 2012;1(3):184-90.
3.Marty C, Lacout C, Droin N, Le Couedic JP, Ribrag V, Solary E, et al. A role for reactive oxygen species in JAK2 V617F myeloproliferative neoplasm progression.
Leukemia. 2013;27(11):2187-95.
4.Machado-Neto JA, Traina F. Reactive oxygen species overload promotes apoptosis in JAK2V617F-positive cell lines. Rev Bras Hematol Hemoter.
2016;38(3):179-81.
5. Horn S, Leonardelli S, Sucker A. et al. Tumor CDKN2A-Associated JAK2 Loss and Susceptibility to Immunotherapy Resistance. J Natl Cancer Inst. 2018;
1;110(6):677-681
Return to Projects list

8
Targeting radio-ligand therapy to protonated CEACAM6 epitopes induced
in acid-resistant cells under improved oxygen – release from
biochemically rejuvenated transfusions 2,3,4 - Prof. Pawel Swietach, Prof.
Katherine Vallis and Prof. Simon
Primary Supervisor: Prof. Pawel Swietach
Additional Supervisors: Prof. Katherine Vallis and Prof. Simon Newstead
Eligibility: Track 2, 3 and 4 students are eligible to apply for this project.

All under-perfused tumours accumulate acidic end-products of metabolism. Successful cancer cells adapt to acidosis and
eliminate competition from acid-sensitive neighbours by releasing more acid. The outcome of this process is acid-driven
Darwinian evolution, which we propose to derail by specifically killing acid-resistant colorectal cancer (CRC) cells using
radioligand therapy (RLT) that targets acid-induced epitopes in their protonated conformation. Our CRISPR/Cas9 screen
determined that Complex I genes are essential for survival at low pH, consistent with a greater reliance on respiration when
acidosis blocks glycolysis. In a panel of colorectal cancer cell lines, we described CEACAM6 as a surface-expressed marker of
such acid-resistant phenotypes. We will target this epitope using synthetic peptides inspired by CEACAM-binding dr-adhesins
produced by E. coli. To target the protonated confirmation, peptides will be generated at low pH. Validated peptides will be
linked with radionuclides to selectively kill acid-resistant cells in spheroid and xenograft models. It is unclear whether tumour
hypoxia is problematic for RLT efficacy in CRC patients who often become anaemic (e.g. rectal bleeding). Transfusions increase
haematocrit but stored red blood cells (RBCs) have reduced O 2-release kinetics and may not improve tumour oxygenation.
However, we demonstrated how O2-unloading can be improved by in-bag biochemical rejuvenation, which we propose to
combine with RLT to improve efficacy in vivo. This will be tested in xenograft models established in mice receiving standard-
storage or rejuvenated blood. The project capitalizes on recent advances in acid-driven oncogenic progression to develop a
new line of therapy that exploits tumour-specific features.
Objective 1: Verifying the target (months 0-9): We have characterized the mechanisms of acid-adaptation and identified
CEACAM6 as the surface-expressed marker of acid-resistant phenotypes. Moreover, acid-resistant lines induce further
CEACAM6 gene expression within 12h exposure to acidity. We will verify CEACAM6 as a marker of acid-resistant cells in
spheroids and xenografts by correlating with its distribution for pHLIP, a peptide that integrates into membranes in acidic
niches. Next, we will map the acid-induction signalling pathway. A candidate negative-regulator is the transcription factor
CDX1, which we will test by immunofluorescence and promoter binding (ChIP-qPCR). We believe that these mechanisms may
be general in cancer, e.g. our pilot data in PDAC, a cancer that develops from ductal epithelia that are basolaterally acidic.
Thus, there is potential for protonated CEACAM6 to be a broad-spectrum cancer target, which we will test in a panel of cell
lines. Our group has expertise in studying acid-evoked processes in vitro and in animal models. This part of the project is low-
risk because it is supported by strong pilot data, and will build the student’s confidence.
Objective 2: Refining and testing the ligand (months 10-24; preliminary work prior to transfer viva): We will develop ligands,
suitable for development as theranostic agents, that bind preferentially to protonated CEACAM6. Candidate peptides are
based on bacterial dr-adhesins that bind in a highly conformation-sensitive manner. Whilst antibodies to a related isoform,
CEACAM5, have been raised by others, our aim is to seek ligands that selectively bind at low pH to minimize off-target actions
in non-acidic tissues or on circulating CEACAM6 fragments. For this, we will adopt a strategy that we developed for generating
synthetic nanobodies (sybody), involving phage display. Optimized sequences will be used to prepare ligands for conjugation
with fluorophores for testing accuracy of targeting in CRC spheroids. Successful ligands will be conjugated with radionuclides
via bifunctional metal chelators, focusing initially on Lutetium-177 for therapy and SPECT imaging, and Gallium-68 for PET
imaging. Novel peptide- and sybody-based ligands will be evaluated in vitro in competitive binding and cytotoxicity assays,
and their radiolabelling yields (RLY) determined. In vivo microPET/SPECT-CT imaging, biodistribution/PK and tumour/normal
organ dosimetry studies will be performed in tumour-bearing mice. The cancer cell killing efficacy and selectivity of
radiolabelled ligands will be tested initially in spheroid and then xenograft co-cultures established from ad-mixtures of acid-
resistant and acid-sensitive CRC cell lines. Relevant readouts include histological measures of the balance between acid-
resistant and acid-sensitive cells, overall tumour growth and survival. These experiments will enable dosage optimization for
further testing in future experiments on other animal models of cancer, such as orthotopic or genetically-engineered mice.
This part of the project will involve screening peptides for target binding and has an element of risk, which we mitigate by
using bacterial proteins as a starting point.
9
Objective 3: Radio-sensitisation to enhance RLT anti-cancer effects (months 25-36): Many tumours are hypoxic, but the effect
of reduced oxygenation on RLT efficacy is unclear. Based on the favourable effect of raising tissue O2 availability on external
beam therapy (EBRT), it is plausible that RLT efficacy is sub-optimal in hypoxic microenvironments, which also harbour the
acid-resistant targets. Oxygen availability can be raised by blocking the electron transport chain (e.g. atovaquone), but this
may have off-target effects. We will test a novel approach that stems from our work on oxygen-handling by RBC. Our research
has found that O2-unloading from RBCs is slower than previously estimated. This arises from restricted gas diffusion across
the ‘crowded’ RBC cytoplasm. The size of this diffusional barrier can cause O 2 supply to tissues to become diffusion-limited.
In blood-units stored for transfusion, RBCs undergo spherical remodelling which lengthens intracellular pathlength and slows
O2 release. We showed that kinetically-slow RBCs resulted in poor O2 extraction in perfused kidneys. Anaemic CRC patients
are predisposed to hypoxia but transfusions with kinetically-compromised blood may not improve tissue oxygenation. We
refined an in-bag biochemical rejuvenation process that accelerates O 2 release and postulate that transfusions with such
bloods can improve tumour oxygenation, without affecting other parameters such as acidity. To test this intervention, we will
wean mice on an iron-deficient diet which causes anaemia. Once haematocrit stabilizes, subcutaneous xenografts will be
established and mice will be transfused with stored bloods randomized to receive rejuvenation. The efficacy of this
intervention will be measured histologically by pimonidazole staining. Our findings may impact other cancer treatments, such
as EBRT. This proof-of-principle part of the project may use any form of RLT, and does not necessary require objective 2 to be
met
Acid-driven disease progression is a well-established concept in cancer that is rooted in Darwinian evolution. The current
strategy for interfering with acid-selection is to remove tumour acidosis (the selection pressure) by using buffers (e.g. oral
bicarbonate) or by urase enzymes that produce buffer. Despite excellent preclinical results, acid-buffering is not without
problems in humans, such as compliance to take large amounts of bicarbonate. We propose an alternative approach that
attenuates acid-driven progression by eliminating the acid-adapted cells. By targeting the protonated conformation, we
anticipate good selectively over non-cancer cells expressing similar proteins. Mathematical modelling predicts that eliminating
the fittest cells should permanently swerve the disease trajectory from its most malignant form. Our second translational
deliverable is the use of rejuvenated blood products to improve tumour oxygenation, which we predict may improve RLT and
EBRT. Attempts have been made to raise oxygen availability by blocking mitochondria but this can be problematic systemically.
Our method is less invasive and addresses a niche in the area of transfusion medicine, which is a major part of patient care.
Indeed, anaemia is a common co-morbidity in cancer patients.
Training opportunities
Our host institution has state-of-the-art facilities, resources, and expertise to undertake the project. This includes cancer
physiology, RLT, nuclear imaging, biophysics and nanobody production. We collaborate with clinical partners and the blood
service, and our cross-disciplinary approach opens new lines of discovery science. We offer training across three Oxford
departments: in acid-base physiology and oxygen transport, oncology and RLT, and peptide biochemistry. Our team has a
track-record in CRC research, with unique access to ~100 CRC lines and corresponding spheroid and xenograft models. RLT is
an emerging area of oncology, with much optimisation ahead and therefore with ample opportunities to make impact. The
research project implements risk-mitigating steps to ensure that the work-packages are not entirely contingent on success in
prior objectives. Overall, the student will have experienced a range of methods from synthetic protein chemistry to animal
studies; from basic-research to translational deliverables. This project uniquely offers opportunities to collaborate with
England’s blood service, NHS Blood & Transplant, to improve transfusion products specifically for cancer patients. This area is
novel because it stems from discoveries made by our labs in the last three years.
References
• CRISPR-Cas9 screen identifies oxidative phosphorylation as essential for cancer cell survival at low extracellular pH.
Michl J, Wang Y, Monterisi S, Blaszczak W, Beveridge R, Bridges EM, Koth J, Bodmer WF, Swietach P. Cell Rep. 2022
Mar 8;38(10):110493. doi: 10.1016/j.celrep.2022.110493.
• What do cellular responses to acidity tell us about cancer? Blaszczak W, Swietach P. Cancer Metastasis Rev. 2021
Dec;40(4):1159-1176. doi: 10.1007/s10555-021-10005-3.
• Evolutionary dynamics of carcinogenesis and why targeted therapy does not work. Gillies RJ, Verduzco D, Gatenby
RA. Nat Rev Cancer. 2012 Jun 14;12(7):487-93. doi: 10.1038/nrc3298.
10
Understand the developmental origin of Clear Cell Sarcoma using
long-read single cell sequencing .1,3,4 – Assoc Prof. Sarah Snelling
Primary Supervisor: Assoc Prof. Sarah Snelling
Additional Supervisors: Assoc Prof. Adam Cribbs
Abstract
Clear Cell Sarcoma (CCS) of tendons, is a rare malignant soft tissue sarcoma, typically derived from neural crest cells. It
usually presents in the distal lower extremities of young adults, frequently attached to tendons. It behaves like a high-
grade soft tissue sarcoma and is associated with poor overall survival due to spreading to other parts of the body with
recurrence after treatment also being very common. CCS neoplastic cells express the EWSR1-ATF1 fusion gene in most
cases, with EWSR1-CREB1, EWSR1-CREM or EWSR1-DDIT3 fusion genes comprising a smaller subset of cases.
Prof Snelling leads the Tendon Seed Network chapter of the Human Cell Atlas, whose aim is to investigate the cell
architecture of healthy human tendons. We have performed single-cell sequencing on several anatomically different
tendon tissues across several healthy donors. Tendons have historically been thought of as acellular, however we have
shown that tendons are composed of a diverse and rich cellular microenvironment. We are now applying these
technologies to generate healthy tissue atlases of other joint-resident soft tissues including synovium and ligament
Considering that very little is known about the developmental origin of CSS, we hypothesise that CSS may develop from
cells residing within the tendon. As such, we will leverage the reference maps of healthy musculoskeletal tissues with
single-cell sequencing data of CSS that we will generate as part of this proposed project. We will cross reference these
datasets of CSS and healthy soft tissues to identify whether the cellular origin of CSS is tendons or another tissue type.
We will then determine the oncogenic drivers of CSS with the long-term goal of utilising this data to enable identification
of novel therapeutics to treat this currently incurable cancer.
In collaboration with the Snelling group, work in the Cribbs lab focuses on developing novel single-cell technology and
computational analysis frameworks that empower new modes of treatment for disease. Recently we have developed
scCOLOR-seq1, a method to overcome low basecalling accuracy making long-read single-cell transcriptomic sequencing
highly accurate. This new technology enables us to measure translocations, alternative splicing, and allows variant
calling. We have begun to apply this technology to understand the development of drug resistance in Sarcomas.
Our aim is to apply long-read single-cell sequencing technology to primary CCS patient samples and then generate
computational models that will help us to identify the developmental origin and oncogenic drivers of this cancer.
Work package 1: Apply scCOLOR-seq to Clear Cell Sarcoma tumours.

The student will apply scCOLOR-seq to investigate the gene expression, isoform expression, translocations, and copy
number variation within 10 CSS primary tumours. Specifically, the aim of this work package will be to identify genomic
signatures that can provide us with a better understanding of the developmental origin on CCS.
Work package 2: Development of a computational analysis strategy to improve long-read single-cell sequencing.
Working simultaneously alongside work package 1, bespoke computational analysis pipelines will be written to help
process the long-read sequencing data. The student will work alongside Dr Cribbs, who will provide extensive
computational training, to develop skills in python and R programming, as well as software development. We have
already developed generic long-read single-cell sequencing workflows (ref). The student will expand the development of
this code with an emphasis on cross comparison analysis of our healthy tendon, ligament, synovium and soft tissue
datasets and the long-read single-cell data generated in work package 1.
Work package 3: Develop and apply machine learning models to understand the genomic features that are important
for developing Clear Cell Sarcoma.
Evaluate the accuracy and utility of a variety of unsupervised and supervised classification and machine learning
algorithms (e.g. k-means/hierarchical clustering, linear discriminant analysis, support vector machines, Neural Networks
and others) to identify features that are important for CCS pathogenesis. Specifically, we will develop a classifier model
11
using data (structural variation, isoform expression, gene expression) generated from the long-read sequencing
experiments. Knowledge from this model will be used to identify features that drive the development of CSS and then
generate a priority list of potentially druggable targets for functional validation.
The stated aim of this project is to study the developmental origin of Clear Cell Sarcoma and identify drug targets for
therapy. By its very definition, this work is likely to identify novel therapeutic intervention points within the development
of Clear Cell Sarcoma.
We have extensive collaborations with several pharmaceutical partners, and we will utilise these interactions to explore
the translational potential of targets.
Although CSS is rare, samples are already banked (50 frozen, 35 FFPE) which guarantees analysis can take place and we
have successful pilot data showing high nuclei yields and robust sequencing. Our collaborators in London have been
banking samples over a number of years and this ensures numbers are adequately powered to inform the study.
Furthermore, we are continuing to collect CSS samples to provide additional validation samples. This methodology can
be applied to other cancers and is a long-term goal from the work, but using CSS as an initial example due to the burden
of disease and lack of diagnostic or treatment pathway.
The student will receive training in cellular, molecular, and epigenetic biology for this project. This will involve wet-lab
workflows for generating long-read single-cell sequencing data. Extensive training in computational biology will be
provided so that the student can analyse their own data. Specifically, this will include software development, data
analytics, statistics and computational pipeline development. Outside the lab, the student will be expected to attend
regular seminars with high profile external speakers, journal clubs and training in presentation skills, scientific writing,
and data management. As part of this project, you will collaborate and be co-supervised by Prof Pillay, a clinical
pathologist at UCL who will provide samples for this study.
References
1. Philpott, M. et al. Nanopore sequencing of single-cell transcriptomes with scCOLOR-seq. Nat Biotechnol
(2021).
2. Baldwin, M.J., Cribbs, A.P., Guilak, F. et al. Mapping the musculoskeletal system one cell at a time. Nat Rev
Rheumatol 17, 247–248 (2021).
12 Return to Projects list

How does macrophage marker F4/80 control peripheral tolerance
during tumour progression? 1,2,3,4 – Prof. Kim Midwood
Primary Supervisor: Prof. Kim Midwood
Additional Supervisors: Professor Siamon Gordon and Dr Anja Schwenzer
Abstract
Our immune system detects tumors and activates cytotoxic T cells to destroy them. However, tumors can de-activate
these cells, switching the immune response instead to become tolerant, helping cancer growth and spread. Drugs that
reactivate cytotoxic T cells have revolutionized the treatment of people with cancer. However, this approach does not
work for many patients, nor all types of tumor, and can be associated with severe side effects caused by the destruction
of healthy tissues. Our recent data show how macrophages contribute to tumor progression, using the cell membrane
marker F4/80 to communicate with, and switch on, tolerant T cells. However, whilst F4/80 is essential for tumors to
survive and thrive, nothing is known about how this molecule works. We will investigate how F4/80 controls the
immune axis in tumors. This will lead to a better understanding of how cancer hijacks the immune response for its own
gain, and enable the design of new therapies that block this pathway, which are safer and effective in more patients
than existing drugs.

Macrophages are terminally differentiated migratory leucocytes which sense physiologic and pathologic changes in
their microenvironment through a variety of plasma membrane receptors that regulate their responses to maintain
homeostasis (Fig.1A,B)[1]. In addition to specialised phagocytic receptors to recognize and clear senescent, apoptotic
and necrotic cells and microbes, they are potent secretory cells able to nourish or destroy abnormal host cells and
microbial targets, depending on their functional activation state. Tissue resident and recruited macrophages express
adhesion molecules such as the F4/80 antigen, also known as EMR1 or ADGRE1, a widely used biomarker and the
founder member of a family of 7 transmembrane G protein coupled receptors (GPCRs) that possess large extracellular
epidermal growth factor- like (EGF) repeats. Human myeloid cells express a closely related EMR2 receptor, which is
absent in the mouse, as well as an EMR1 homologue with distinct properties. In spite of extensive studies of its specific
expression in the mouse, the role of F4/80 remains unknown. However, a major breakthrough in our understanding of
its immunological function derived from studies using F4/80 knockout (KO) mice, which revealed this molecule to be
dispensible for macrophage development, but responsible for peripheral tolerance. Required for Treg generation in
vivo in response to a model antigen introduced into the anterior chamber of the eye [2], F4/80 expression is also
essential for organ transplantation [3] and engraftment of Lewis lung cancer cells (Fig.1C-E). This project will explore
the role of F4/80 in tumour-host interactions in vivo and identify cellular and extracellular binding partners of this
adhesion GPCR.
Aim 1 will investigate the cellular and molecular basis of F4/80-mediated tolerance using a syngeneic, orthotopic breast
cancer grafting model (well established at KIR)[4] in wild type or F4/80 null mice (maintained at the Dunn School).
Immune profiling of tumor, spleen and lymph nodes during tumor growth using multicolour spectral flow cytometry
(Aurora, KIR) will reveal F4/80 dependent changes in abundance, and activation/polarization status, across myeloid
and lymphoid compartments. Multiplexed immunofluorescent imaging (GE CellDive, KIR) will reveal cell interaction
partners of F4/80+ macrophages, and the tissue localization of these cellular networks. scRNA seq of sorted CD45+
cells from wild type and F4/80 knockout tumor bearing mice will reveal cell type-specific transcriptional changes
associated with loss of F4/80, and pathway/interaction network analysis will highlight candidate effector molecules.
Cellular interactions, and their contribution to F4/80-mediated tolerance, will be validated in vivo (e.g. by cell type or
effector molecule depletion/blockade), and in vitro (e.g. using tumor-immune cell co-cultures) [4].
Aim 2 will identify extracellular binding partners of F4/80. More than 30 adhesion GPCRs exist, utilizing their sizable
extracellular domains to form multimeric protein complexes of signalosome-like structures. Despite identification of
extracellular ligands for other adhesion GPCRs, including the interaction of chondroitin sulphate (CS) and CD55 with
close family members EMR2 [5] and CD97 [6] respectively, F4/80 remains an orphan receptor. F4/80 expression is
elevated in specific murine breast tumor- associated macrophage subpopulations with enriched APC capabilities.
These cells localize to extracellular matrix tracks that support cell infiltration into the tumor, whilst F4/80 low cells are
restricted to the periphery (Fig.1F). Multiplexed imaging using a panel of matrix markers will identify tumor
13
components that interact with F4/80+ cells in these tracks. In parallel, the adhesion of fetal liver macrophage cell lines
from wild type and F4/80 knockout mice [7](Dunn School), to purified matrix molecules (e.g. tenascin-C (TNC),
fibronectin, collagen type I/IV, osteopontin, CS)(KIR) as well as more complex matrices (e.g. matrigel, tumor derived
cell free matrices) which better recapitulate the 3D environment of the tumor, will be assessed. Binding sites within
F4/80 for ligands will be mapped as for CS-EMR2 [6] and downstream signaling examined.
A B C CD169
F4/80
D
E 900
F
*** F 4 /8 0
tu m o r v o lu m e (m m )
3
750 WT
600 ***
450
***
300
**
150 TNC
F4/80lo CXCL1+
0
0 5 10 15 20 25 30 F4/80hi MHC-II hi
days
Translational potential
Macrophages are an integral component of tumours, responding to, and in turn, influencing the malignant cells, as well
as all stromal and other immune cellular and extracellular components of the tumour microenvironment. Their
interactions with the extracellular matrix could affect their own growth, adhesion, migration and activation, as well as
of the malignant cells and metastases. We need to learn more about their polarization, and the switch from trophic to
cytotoxic potential, crucial in the tumour-host interaction. This project will establish some of the basic principles for
further translation to human cancer. Among the EGF-TM7 family CD97 and EMR2 have been investigated in a range of
human cancers, with, for example, changes in the cellular compartmentalization of EMR2 correlating with poor
prognosis in breast cancer [8]. F4/80 has been neglected hitherto. Evidence that it mediates peripheral tolerance,
which can be abrogated by targeting F4/80, makes this a compelling model for further research in mouse models of
cancer. Moreover, study of the autoproteolyic cleavage of these adhesion GPCRs (Fig1B), may also be relevant to
shedding of soluble receptor in vivo, which could contribute to cancer pathogenesis and to diagnosis through its
presence in plasma or other body fluids [9].
The student will be trained in techniques including: in vivo tumor models, immune cell isolation/activation, 2D & 3D
matrix adhesion assays, multiplexed tissue imaging, spectral flow cytometry, RNA seq dataset generation and
bioinformatic analysis in mouse and human pathology
References
[1] Gordon, S. and A. Plüddemann, Tissue macrophages: heterogeneity and functions. BMC Biology, 2017. [2] Lin, H.H., et al., The macrophage F4/80
receptor is required for the induction of antigen-specific efferent regulatory T cells in peripheral tolerance. J Exp Med, 2005. [3] Conde, P., et al., DC-
SIGN(+) Macrophages Control the Induction of Transplantation Tolerance. Immunity, 2015. [4] Deligne C et al, Matrix-Targeting Immunotherapy
Controls Tumor Growth and Spread by Switching Macrophage Phenotype. Cancer Immunol Res. 2020. [5] Stacey, M., et al., The EGF–like domains of
the human EMR2 receptor mediate cell attachment through chondroitin sulfate glycosaminoglycans. Blood, 2003. [6] Aust, G., L. Zheng, and M.
Quaas, To Detach, Migrate, Adhere, and Metastasize: CD97/ADGRE5 in Cancer. Cells, 2022. [7] Fejer, G., et al., Nontransformed, GM-CSFdependent
macrophage lines are a unique model to study tissue macrophage functions. PNAS, 2013. [8] Davies, J.Q., et al., Leukocyte adhesion-GPCR EMR2 is
aberrantly expressed in human breast carcinomas and is associated with patient survival. Oncol Rep, 2011. [9] Boucard, A.A., Self-activated adhesion
receptor proteins visualized. Nature, 2022.

14
Chronic infection, host immunity, and cancer risk 3,4 – Ling Yang
Primary Supervisor: ling Yang
Additional Supervisors: Christiana Kartsonaki
Abstract
Worldwide, chronic infection of certain pathogens (i.e. virus, bacteria and parasite) is responsible for more than 2 million new
cancer cases each year, with about a third of these cases occur in China. 1 Although the aetiological roles of several common
pathogens (e.g. H. pylori, HBV, HCV, and HPV) for certain cancers are well established, questions remain about their relevance
for other cancers and the role of other pathogens (e.g. human herpes virus) in cancer development. Previous studies have
been often constrained by small sample sizes and investigation of a single pathogen with a single cancer. Consequently,
uncertainty remains about the roles of common chronic infections in cancer development and progression.2 It is also unclear
as to whether there are important interactions between infections and other lifestyle/behavioural risk factors for cancer, and
how genetic variants in both host (e.g. human leukocyte antigen; HLA) and pathogen and their interactions influence host
immune response and susceptibility to and progression of infection, as well as subsequent cancer risk. For example, variability
in HLA class I or II type antigens that affect the cellular immune response to HPV may partially explain why only a minority of
infected women develop cervical cancer.3
The China Kadoorie Biobank (CKB) included 0.5 million

Chinese people during 2004-08 from 10 different areas
across China, with different types of information collected
by questionnaire and physical measurements, and with
long-term follow-up of health outcomes for each
participant by the linkage with national death, disease
registries and nationwide health insurance systems. So far,
we have collected more than 70,000 deaths and over 1.5
million hospitalization events for >5000 different diseases.
In subsets of CKB participants, we also had genetic and
other biomarkers (e.g. inflammation) measured.
Based on this CKB cohort, we are conducting a large case-

cohort study including all incident site-specific cancer cases
in China (>30,000) and a randomly selected sub-cohort of
participants (~10,000) to assess and quantify the role of
different infectious pathogens in various cancers
development. In collaboration with UK Biobank and the
German Cancer Research Centre (DKFZ), a CKB custom-
designed multiplex serology 4 panel has been developed
and been used to detect and quantify 47 antibodies for 17
infectious pathogens (see details in Table), selected for
their cancer relevance and will enable population-specific variation in infections to be assessed in relation to different cancers.
Moreover, in collaboration with the Wellcome Centre for Human Genetics, we are sequencing HBV and HCV viral genomes
from ~5,000 CKB HBV infected participants, half of whom who developed incident HCC during follow-up and the other half
who did not. Genome-wide association studies (GWAS) and genome-to-genome analysis will be used to investigate host-virus
interactions and genotype-phenotype relationships.5
These, together with available lifestyle, physical measurements, medical history, other biomarkers and health outcome data,
will enable comprehensive assessment of causal roles of multiple chronic infections in aetiology of site-specific cancers and
certain other diseases.
The proposed DPhil project will be based on this ongoing infection and cancer programme and the specific research proposal
will be developed according to the student’s interests and aptitude.
15
Research objectives
The specific lines of investigation covered by this DPhil project will depend on the student’s interests and previous training,
and may be involved within following key areas of work: (1) to examine the associations of chronic infection of particular
pathogens with risks of all or certain site-specific cancers and estimate the infection-related cancer burden; (2) to explore
the role of host immune system genetics (e.g. HLA) in susceptibility to specific types of chronic infection and cancer; (3) to
establish the value of serological markers, in combination with other lifestyle and genetic risk factors, in predicting the risk
of infection-related cancers.
By the end of the DPhil, the student will be competent to review the literature, to plan, undertake and interpret analyses of
large-scale data, and to report research findings, including 3-5 publications in peer-reviewed journals and presentation at
conferences.
This project will lead to improved understanding about the aetiological role of infectious pathogens in cancer development,
help identify high risk individuals for early detection/mass-screening and targeted cancer treatment and inform policy makers
to develop and implement suitable strategies for cancer prevention locally and globally.
The student will be based within the CKB research group in Nuffield Department of Population Health. There are excellent
facilities and a world-class community of population health, data science and genomic medicine researchers. There will be in-
house training opportunities in epidemiology, statistics, genetics, and bioinformatics and if necessary attendance at relevant
courses.
References
1. Plummer M, de Martel C, Vignat J, Ferlay J, Bray F, Franceschi S, et al. Global burden of cancers attributable to
infections in 2012: a synthetic analysis. Lancet Glob Health. 2016;4:e609-16.
2. IARC. Biological agents. Volume 100 B. A review of human carcinogens. IARC Monographs on the Evaluation of
Carcinogenic Risks to Humans 2012; 100(Pt B):1-441
3. Zoodsma M, Nolte IM, Schipper M, Oosterom E, van der Steege G, de Vries EG, et al. Analysis of the entire HLA
region in susceptibility for cervical cancer: a comprehensive study. J Med Genet. 2005;42(8):e49.
4. Waterboer T, Sehr P, Michael KM, Franceschi S, Nieland JD, Joos TO, et al. Multiplex human papillomavirus serology
based on in situ-purified glutathione s-transferase fusion proteins. Clin Chem. 2005;51(10):1845-1853.
5. Ansari MA, Pedergnana V, L C Ip C, Magri A, Von Delft A, Bonsall D, et al. Genome-to-genome analysis highlights the
effect of the human innate and adaptive immune systems on the hepatitis C virus. Nat. Genet. 2017; 49:666–673.

16
Myeloid cells in sustained intestinal inflammation and colorectal
cancer: the role of IRF5 1,2,3– Prof. Irina Udalova
Primary Supervisor: Prof. Irina Udalova
Additional Supervisors: Prof. Fiona Powrie
Abstract
Dysregulation of the intestinal immune system can lead to inflammation, which may progress to colorectal cancer (CRC)n 1. Myeloid
cells are playing a central role in maintenance of homeostasis, initiation of inflammation, restoration of tissue upon injury and
mediation of chemoresistance in tumours. Integrating cues of their immediate tissue microenvironment. Interferon Regulatory
Factor 5 (IRF5) has been identified to act as a master transcription factor in myeloid cells that controls both acute and chronic
inflammation and is protective in pathogen clearance 2,3,4. IRF5 has been shown to regulate macrophage transcription of pro-
inflammatory mediators, monocytes differentiation into pro-inflammatory macrophage phenotype, and more recently, neutrophil
effector responses without compromising neutrophil maturation processes or their ability to enter into the tissue 5. Our preliminary
data suggest that IRF5 function in myeloid cells is indeed a key element in controlling the onset of intestinal inflammation and
pathology, but its role in resolution of inflammation or cancerogenesis is not well understood. Of relevance, IRF5 has also been
identified as a DNA-damage sensor, highlighting a potential beneficial role in CRC 6. Both tumour-associated macrophages (TAMs)
and tumour-associated neutrophils (TANs) appear to play a major role in cancer progression or hinderance 7. Therefore, dissecting
the molecular mechanisms orchestrating neutrophil, monocyte and macrophage function is crucial for identification of treatment
options for both inflammatory bowel diseases (IBD) and CRC.
Research Objectives
First, we will use advanced imaging and spatial transcriptomic analyses to compare myeloid cell heterogeneity in the colon in
resolution of inflammation using the Helicobacter hepaticus and anti-IL10R (Hh + aIL10R) colitis model 2 with that in microbe driven
colitis associated cancer (CAC) models 8, characterised in the Powrie group; as well as during the onset and peak of inflammation,
characterised previously in our lab. It was suggested that localisation of macrophages and neutrophils within the tumour
microenvironment might be a crucial determinant of their function 9. Therefore, using already generated reporter strains (e.g.
CX3CR1-GFP x Ly6G-Tomato) we will assess the localisation of different macrophage and neutrophil subsets and their interaction
cell-cell contacts, which could also provide further information about their function and potential targeting. This will help identify
molecular targets in shaping their phenotype and directing towards resolution rather than progression of inflammation into cancer
development.
Second, we will assess the role of IRF5 in myeloid cells. Previous work in the lab has profiled IRF5-dependent inflammatory CD11c+
macrophages at peak of inflammation using scRNA-seq 2. New data suggest the role for IRF5 in driving neutrophil-dependent
inflammation. Based on this work, we aim to utilise already generated targeted mouse models (CX3CR1-cre ER2 IRF5 fl/fl; CCR2-
mKate-cre ER2 IRF5 fl/fl; Ly6G-cre IRF5 fl/fl) to understand the beneficial vs pathologic role of IRF5 in resolution of inflammation.
We hypothesize that lack of IRF5 in macrophages may be beneficial as macrophages are being polarised towards a tissue-
regenerating phenotype. Furthermore, the T cell pool was shifted at peak of inflammation, which might be a result of different T
cell priming by IRF5-proficient and deficient macrophages. The lack of IRF5 in neutrophils may be detrimental as neutrophils can
be involved in priming barrier cell repair and/or regenerating the extra cellular matrix around the perturbances.
17
Third, the role of TAMs and TANs in CRC is yet unclear with various studies suggesting both detrimental and beneficial effects.
Investigation of the differences in T cell priming of IRF5-deficient macrophages and/or neutrophils will also offer insight of the
interplay of IRF5 in innate immune cells with the adaptive immune system in both resolution and CRC. Thus, in addition to inhibition
of IRF5 (as above) we would also consider stimulating IRF5 specifically at tumour sites might improve anti-cancer immunity 10. This
could be achieved by targeted delivery of adenoviral vector expressing IRF5 (overexpression) or inhibition of IRF5 activation through
phosphorylating kinase inhibition 11 in CAC models.
T cell immunity, which is beneficial in tumours, is undermined by immunosuppressive myeloid cells, of which a subset of TREM2+
macrophages have been identified as a potential target in tumours 12. Understanding the role of macrophages as pivotal cells in the
resolution of inflammation as well as progression of inflammation into CRC will help shaping specific therapies targeting
macrophages. IRF5 also plays a crucial role in mediating monocyte recruitment and their differentiation into pro-inflammatory
macrophages, as well as in effector neutrophil functions, during intestinal inflammation and may therefore be central during
resolution and cancer development. Moreover, the inhibitor of IRF5 activation pathway may prove beneficial for inflammation-
induced cancer.
The student will be trained in the Hh + aIL10R colitis and CAC models as well as in basic immunology techniques like flow cytometry,
RT-qPCR and in vitro cultures to analyse the outcomes. Furthermore, insights and potential guided analysis of single-cell RNA
sequencing as well as cutting-edge microscopy and spatial transcriptomics (GeoMx and CosMx Nanostring platforms) to define the
localisation of macrophage subsets within the tumour microenvironment will be made available.
References
1. Mantovani, A et al. Cancer-related inflammation. Nature 454, 436–444 (2008). 2. Corbin, A. L. et al. IRF5 guides monocytes
toward an inflammatory CD11c + macrophage phenotype and promotes intestinal inflammation. Sci. Immunol. 6085, 1–16 (2020).
3. Weiss, M. et al. IRF5 controls both acute and chronic inflammation. Proc. Natl. Acad. Sci. U. S. A. 112, 11001–11006 (2015). 4.
Pandey, S. P., Yan, J., Turner, J. R. & Abraham, C. Reducing IRF5 expression attenuates colitis in mice, but impairs the clearance of
intestinal pathogens. Mucosal Immunol. 12(4):874-887 (2019). 5. Khoyratty, T. E. & Ai, Z. Distinct transcription factor networks
control neutrophil-driven inflammation. Nat. Immunol. 22(9):1093-1106 (2021). 6. Hu, G. et al. Signaling through IFN regulatory
factor-5 sensitizes p53-deficient tumors to DNA damage-induced apoptosis and cell death. Cancer Res. 65, 7403–7412 (2005). 7.
Quail, D. F. et al. Neutrophil phenotypes and functions in cancer: A consensus statement. J. Exp. Med. 219, 39 (2022). 8. Kirchberger,
S. et al. Innate lymphoid cells sustain colon cancer through production of interleukin-22 in a mouse model. J. Exp. Med. 210, 917–
931 (2013). 9. Caprara, G., Allavena, P. & Erreni, M. Intestinal Macrophages at the Crossroad between Diet, Inflammation, and
Cancer. Int. J. Mol. Sci. 21, 4825 (2020). 10. Byrne, A. J. et al. A critical role for IRF5 in regulating allergic airway inflammation.
Mucosal Immunol. 10, 716–726 (2017). 11. Ryzhakov et al. Defactinib inhibits PYK2 phosphorylation of IRF5 and reduces intestinal
inflammation. Nat Commun 12(1):6702 (2021). 12. Molgora, M. et al. TREM2 Modulation Remodels the Tumor Myeloid Landscape
Enhancing Anti-PD-1 Immunotherapy. Cell 182(4), 886-900 (2020).

Cancer vaccination for precision prevention of Lynch Syndrome 1,2,3–
Prof. Simon Leedham & Assoc Prof. David Church
Primary Supervisor: Prof. Simon Leedham & Assoc Prof. David Church
Additional Supervisors: Professor Sarah Blagden
Eligibility: Track 1, 2 and 3 students are eligible to apply for this project
Abstract
Lynch syndrome is the most common Mendelian inherited cancer syndrome, affecting 175,000 patients in the UK alone.
Very high lifetime gastrointestinal and uro-gynaecological cancer risks mean that patients with this condition need to
undertake lifelong intensive, intrusive and expensive surveillance programs. Lynch syndrome results from a inherited
failure of DNA mismatch repair resulting in epithelial cell hypermutation and neoantigen generation. Lynch syndrome is
thus an exciting disease target for vaccination as the generation of neoantigens by the hypermutation process is sufficient
to stimulate an endogenous immune response which could be enhanced with vaccination. In established cancers the
development of immune evasion - the evolved ability of the cancer to bypass immune surveillance - may prevent
vaccination from eliminating cancer growth. However, there is good reason to speculate that targeting lesions earlier in
the disease process will be more successful through enhancement of endogenous immune-surveillance to kill mutated
cells at the very earliest stage of Lynch syndrome lesion initiation (known as immune editing). Improving immune system
surveillance of mutated epithelial cells through vaccination could thus provide a novel cancer prevention strategy in this
common population. This project will explore key questions around the impact of early mutation selection in Lynch
syndrome patients to generate key scientific outcome measures for a future vaccination precision prevention trial.
Figure 1. Lynch syndrome lesion initiation and progression and the proposed impact of novel vaccination strategy
Research objectives.
While most cancers result from a combination of lifestyle factors and variants in multiple genes, some are caused by a
single faulty gene which is passed down through generations in affected families. In the UK the most common such
condition is called Lynch Syndrome (LS), which is estimated to affect 175,000 people. Individuals with LS have lifetime
risk of colorectal, endometrial and urological cancers of up to 80%. Consequently, they are recommended to undergo
regular surveillance by colonoscopy, and counselling for symptoms of other malignancies. However, such surveillance is
costly, subject to geographical variation, and fails to detect all cancers, and consequently Lynch syndrome is estimated
to be the cause of about 3-5% of colorectal cancers and 2% of endometrial cancers worldwide.
In Lynch Syndrome, faulty gene(s) lead to failure of DNA mismatch repair (MMR), which is a key safeguarding process
that checks DNA has been copied accurately before cells divide. The fault means that dangerous changes in the DNA go
un-checked and can lead to widespread mutation development (known as hypermutation) which results in precursor
19
lesion formation with eventual progression to cancer. Regular aspirin halves bowel cancer risk, but does not modify risk
of other cancers. Thus, there is a clear unmet need for novel cancer prevention strategies in this common population.
The increased cancer risk with LS has led researchers to investigate whether vaccination can be used to reduce this.
However, to date such vaccines have targeted features in the cancers which are not-specific to LS, and have not shown
evidence of meaningful patient benefit. We propose a new approach that takes advantage of the fact that cancers in
individuals with LS carry predictable mutations in their DNA and protein as a consequence of the defect in MMR. These
mutations generate neoantigens within the tumour cells, which can be recognised by the immune system. In established
cancers the development of immune evasion - the evolved ability of the cancer to bypass immune surveillance - may
prevent vaccination from eliminating cancer growth. However, there is good reason to speculate that targeting lesions
earlier in the disease process will be more successful through enhancement of endogenous immune-surveillance to kill
mutated cells at the very earliest stage of Lynch syndrome lesion initiation (known as immune editing). This postulate is
supported by the fact that pre-cancerous polyps in individuals with LS have been reported to spontaneously regress,
consistent with elimination by a functional anti-tumour immune response. In this project, we will focus on LS-associated
lesions in the colon, given their frequency and accessibility; however, the insights this work generates will be transferable
to other organs affected by LS carcinogenesis (eg endometrium).
This project will address 3 key questions surrounding early Lynch lesion biology to allow us to identify the optimum
antigenic targets for vaccine therapy and to assess the impact of future vaccination:
1. Can we spatially assess initiating mutations in Lynch patients and use these to measure impact of vaccination on
immune editing capacity? We have identified some key initiating mutations in Lynch lesions and will design
custom spatial biology tools to identify cells with these mutations in the colons of Lynch patients. We will use
this to measure clonal competition and look for dynamic change in competition metrics as a key scientific output
measure for vaccination efficacy.
2. Do initiating mutations that impact key homeostatic signalling pathways cause derangement in epithelial-
microenvironmental crosstalk that promotes clonal expansion of mutated cells? Some of the identified early
mutations are in key signalling pathways that coordinate immune and stromal cell function in early lesions. We
will use genomic and spatial biology tools to assess the impact of these mutations on intercompartmental
crosstalk in human tissue and appropriate mouse models.
3. When do immune escape alterations occur in LS-associated carcinogenesis and how do they modulate the anti-
tumour immune response? Preliminary data in mouse models of LS indicate that immune escape occurs at the
transition of an aberrent crypt to an established polyp, however corresponding human data are lacking. We will
define this, and its immunological correlates using similar methods to those applied in (2).
Vaccination against cancer is not a new concept, but interest in the potential of this strategy has been supported by the
success of immunotherapy in some solid tumours, including mismatch repair deficient colorectal cancer, and pandemic-
related innovations in vaccine technology and infrastructure. Lynch syndrome is an exciting disease target for vaccination
as the generation of neoantigens by the hypermutation process is sufficient to stimulate an endogenous immune
response which could be enhanced with vaccination. We aim to pursue a strategy of vaccination against recurrent
immunogenic mutations in Lynch syndrome, with the aim of enhancing endogenous immune editing to prevent
precursor lesion initiation and propagation. We will work with the Oxford Vaccine group to develop a novel vaccine
against appropriate antigens and then will aim to translate this through to a precision prevention clinical trial in Lynch
patients through the Oxford Early Detection/Precision Prevention clinical trial unit. We believe this precision prevention
approach could profoundly benefit Lynch syndrome families by decreasing multi-organ lesion generation at the point of
initiation and thus prevent or reduce the need for intensive, uncomfortable and expensive surveillance regimens.
The student will join an Oxford-based multidisciplinary team of gastroenterologists, oncologists, clinical trialists,
immunologists and vaccinologists, so the position would be very suitable for clinical or non-clinical candidates. Training
will be provided in human and mouse tissue curation and genomic and spatial biology interrogation of this tissue.
Students will undertake training to undertake bioinformatic and spatial biology analysis and will work with collaborators
for more advanced analysis. Students will be required to undertake training to obtain a home office license and will be
responsible for looking after a small colony of mice. Clinical sessions in oncology or gastroenterology could be arranged
for clinical candidates.

20
Therapeutic manipulation of reactive oxygen species in ATRX-deficient
cancers 1,2,3 – Prof. Ester Hammond
Primary Supervisor: Prof. Ester Hammond
Additional Supervisors: Dr Anna M. Rose
Abstract
Telomere maintenance is an essential cancer hallmark, allowing malignant cells to divide without limit. One major
telomere maintenance mechanism is called Alternative Lengthening of Telomeres (ALT). The ALT-pathway is particularly
prevalent in aggressive brain cancers (such as high-grade glioma (>40%) and astrocytoma (90%)), as well as cancers of
mesenchymal origin, such as osteosarcoma [1]. The central genetic event underpinning ALT-pathway activation is loss of
ATRX [2]. In addition to ATRX loss, ALT-pathway activation requires another factor. Our recent work demonstrated that
this second factor is excessive accumulation of reactive oxygen species (ROS) in the tumour microenvironment, which
could be due to hypoxia, concurrent gene mutation and/or redox gene dyrsregulation. Elevated ROS levels lead to trapping
of DNA-interacting proteins, which subsequently causes replication fork stalling and ALT-pathway activation [3]. ATRX
protein is essential for fork re-start and so, in the absence of ATRX, there is aberrant downstream processing of stalled
forks [4,5]. This aberrant processing produces DNA double-strand breaks, the genetic substrate for ALT-telomere
elongation (Figure A).
Crucially, however, there is a “Goldilocks zone” of ROS for

cells lacking ATRX: too little ROS, and the threshold of
ALT-pathway activity is not met, and telomere synthesis
is not triggered. Excessive ROS levels, though, lead to
genetic catastrophe and cell death, as ATRX-null cells
have intrinsic vulnerability to DNA damage due to the
many roles of ATRX in genome integrity. There is,
however, a level at which ROS are “just right”, where
ATRX-deficient cells can leverage ROS-induced trapping
of proteins to allow telomere maintenance and limitless
cell division (Figure B). Therapeutic manipulation of ROS
levels to outside of the “Goldilocks zone” provides a novel
approach to treatment for ATRX-deficient ALT cancers.
The central aim of this project is to exploit the ROS-

dependency of ALT cancer cells for therapeutic benefit in
ATRX-deficient tumours. In particular, we will explore the
potential of novel redox modulating agents i.e. those that
either increase ROS directly or, indirectly by decreasing
anti-oxidant capacityin sensitization of ATRX-deficient
cancer cell lines and animal models to standard of care
chemotherapy and radiotherapy. As well as assessing the
efficacy of these agents, we will study the molecular
mechanisms and underpinning cell biology, allowing a
deeper understanding of the role of ROS in ATRX-
deficient ALT cancers.
21
Research objectives
The rationale for this study is that whilst ATRX-mutation drives the evolution of ALT-cancers, it also provides a unique
therapeutic opportunity: ATRX is ubiquitously-expressed with roles in many cellular processes and so cells lacking ATRX
have vulnerabilities which can be therapeutically exploited. In this project, we will explore the genetic and cellular
perturbations in ALT-cancers, in particular, the role of abnormal reactive oxygen species (ROS) metabolism. Prof.
Hammond is an expert in redox and hypoxia biology in cancer. Dr Rose is an expert in ATRX cell biology and the ALT
pathway, with clinical expertise in paediatric oncology. Our pilot work demonstrated that ALT-positive gliomas have strong
dysregulation of redox pathway genes, highlighting the potential role of ROS. Further, treatment of ATRX-deficient cells
with ROS-generating agents induced ALT-pathway activity in non-ALT cell lines. Excessive ROS led to trapping of proteins
on DNA, leading to the formation of abnormal DNA lesions. Further, we found that both the accumulation of trapped
proteins and induction ALT-pathway was dependent on the accumulation of R-loops, RNA:DNA hybrid structures. Our
preliminary data also suggested that pre-treatments which elevate ROS (such as silencing of SOD1 gene) sensitised ATRX-
deficient cells to camptothecin, through induction of hyper-ALT. Camptothecin derivatives – such as irinotecan/topotecan
– are widely used in the treatment of brain cancers. In this project, we will greatly expand this preliminary data by using
ROS-generating treatments to sensitise a range of ALT brain cancer cell lines to various chemotherapeutic agents (e.g.
etoposide, camptothecin-derivatives, PARP-inhibitors). Through the collaboration with Dr Monica Olcina, we will also
assess the role of such pre-treatments in radiosensitisation, including in pre-clinical murine models, as appropriate.
Radiotherapy itself is known to generate ROS and, as such, this could represent a novel approach to hyper-ALT induction
Development of novel therapeutics for ATRX-deficient cancers is an urgent area of clinical unmet need [6,7]. The outcomes
for ALT-cancers is very poor, with little progress made in survival in over 50 years. The work in this project is hypothesis-
driven and will generate pre-clinical data that will be critical in informing future clinical trials and translational work. The
insights into gene dysregulation, telomere dysfunction and genome stability will clarify the pathways involved in ALT-
cancer biology, which is the first critical step in developing targeted therapies. The project will involve the opportunity to
conduct cell and animal studies using novel ROS-inducing agents, assessing their ability to sensitise ATRX-deficient cells to
traditional standard-of-care chemotherapy and radiotherapy. These assays will hopefully lead to future early-phase clinical
trials of the novel agents. Further, our work into chemosensitisation and radiosensitisation will align directly with our
groups wider work into oncolytic virus delivery of synthetic lethal shRNA molecules to ATRX-deficient cells.
Prof. Hammond, Dr Rose and Dr Olcina have worked together collaboratively for the past 2-years, developing a new and
exciting interdisciplinary collaboration. They have a strong track record for supervising DPhil, MSc and BSc students. This
project offers the opportunity to join a well-funded, collaborative and interdisciplinary team. The student will be based
jointly in the Hammond lab (Department of Oncology) and Rose group (Department of Paediatrics), with strong links to
the Olcina group. The student will have the opportunity to learn a wide range of molecular and cell biology techniques
including tissue culture, protein analysis, qRT-PCR, c-circle assay, immunoflouresnce and microscopy, RADAR assay, gene
silencing and overexpression, use of radiation sources, hypoxia chambers, and in vivo mouse work.
References
[1] Heaphy CM, et al. (2011) The American journal of pathology;179(4):1608-1615. [2] Clatterbuck-Soper SF & Meltzer PS
(2023) Genes, 14(4):790. [3] Rose AM et al. (2023) Nucleic acids research, gkad150. Advance online publication. [4] Huh
MS et al. (2016) Cell death & disease;7(5):e2220. [5] Lu R& Pickett HA (2022) Open biology:12(3):220011. [6] Rong L et al.
(2022) CR;41(1):142. [7] George SL et al. (2020) EBioMedicine; 59:102971. [8] Bartholf-DeWitt S, et al. (2022) JCI
Insight;7(17):e151583.
22
Identifying ploidy-dependent mitotic vulnerabilities during DNA damage
repair and chromosome segregation 1,2,3 – Prof. Ulrike Gruneberg
Primary Supervisor: Prof. Ulrike Gruneberg
Additional Supervisors: Prof. Monika Gullerova
Eligibility: Track 1,2 and 3 students are eligible to apply for this project.
Abstract
Faithful cell division ensures the correct segregation of the genetic material over multiple generations. Failure of this
process may result in cells of abnormal ploidy, such as aneuploidy or tetraploidy, caused by whole-genome doubling. Ploidy
changes have been implicated in enabling cancer evolution to take place and are thus considered a driving force for
tumorigenesis. On the other hand, ploidy changes may impart specific vulnerabilities to tumour cells. Recent reports
identified the spindle assembly checkpoint, the key quality control mechanism during mitosis, as well as DNA damage
repair dealing with replication stress, as such vulnerabilities 1-4. These insights could potentially be exploited for therapeutic
use. We propose to examine the precise mechanisms and effects of disrupting DNA damage repair, the spindle assembly
checkpoint and other mitotic control mechanisms in non-transformed and transformed diploid and tetraploid cells with
the aim of identifying and characterising specific ploidy-dependent susceptibilities.
Research objectives
Tetraploidy and aneuploidy have both been shown to promote tumorigenesis 5,6. To understand the molecular biology
underpinning these properties, we will compare euploid non-transformed MCF10A breast and hTert-immortalised retinal
epithelial cells (hTert-RPE1) with near-diploid tumour cells HCT116 and highly aneuploid tumour HeLa cells, as well as
tetraploid derivatives of hTert-RPE1, MCF10A and HCT116. These cell lines constitute an accepted system to probe the
relationship between ploidy and dependence on mitotic control mechanisms 2. Tetraploid cells will be generated by
induction of cytokinesis failure and FAC sorting for tetraploid DNA content, as already established in the lab. Recent
research in both yeast and mammalian cells has shown that tetraploid and aneuploid cells are more sensitive to replication
stress during S-phase as well as mitotic aberrations during the cell division process, giving rise to the idea that these
vulnerabilities could be exploited therapeutically 2-4.
Together with our collaborator Monika Gullerova, we will use our combined expertise in spindle assembly checkpoint and
DNA damage repair analysis to carry out a targeted comparative analysis of perturbing these processes in cells of different
ploidies with the aim of identifying disturbances which affect aneuploid or tetraploid cells more than their euploid
counterparts. In particular, our previous research in the Gruneberg lab into the regulatory roles of distinct kinase-
phosphatase modules at different mitotic transitions is likely to identify experimental situations which affect aneuploid
cells more negatively than euploid cells 7-10. To characterise these situations, we will compare the sensitivity of cell lines
with different ploidies to different phosphatase depletions/degron-tags (siPP1 & PP1dTag, siPP2A-B56, siPP2A-B55) and
analyse timing and success of cell cycle progression, error correction proficiency and spindle assembly checkpoint
competence using established assays. For DNA damage repair analysis, we will test the repair competence of cells with
different ploidies using various experimental approaches such as non-homologous-end-joining and homologous
recombination reporter cell lines, comet assay, clonogenics and gH2AX clearance 11,12. Any differences between aneuploid
and diploid cells will be followed up by a detailed analysis of the molecular biology underpinning these, using fluorescent
markers that we have already in the lab, including the key spindle checkpoint kinase MPS1, the attached-kinetochore
marker astrin, and fluorescently-tagged cyclin B as a general marker of cell cycle progression, to assess the success of the
different stages of chromosome segregation.
23
Aneuploidy and tetraploidy have long been recognised as drivers of tumorigenesis. As these are states that are largely
specific to cancer cells, there is an impetus to use the vulnerabilities that are created by these abnormal ploidies to
specifically eradicate tumour cells. We postulate that specific interference with DNA damage repair or alterations to the
kinase-phosphatase balance orchestrating mitotic progression will be lethal to aneuploid tumour cells but not their euploid
untransformed counterparts. Our characterisation of the effect of these disturbances on cells with abnormal ploidy will
thus be of significant therapeutic value.
Students will receive comprehensive training in molecular biology, classical protein biochemistry, assays to analyse DNA
damage repair and spindle assembly checkpoint proficiency, cutting edge fixed and live cell imaging techniques as well as
quantitative image analysis methods.
References
Passerini, V., and Storchova, Z. (2016). Too much to handle - how gaining chromosomes destabilizes the genome. Cell Cycle
15, 2867-2874. 10.1080/15384101.2016.1231285.
2. Cohen-Sharir, Y., McFarland, J.M., Abdusamad, M., Marquis, C., Bernhard, S.V., Kazachkova, M., Tang, H., Ippolito,
M.R., Laue, K., Zerbib, J., et al. (2021). Aneuploidy renders cancer cells vulnerable to mitotic checkpoint inhibition. Nature
590, 486-491. 10.1038/s41586-020-03114-6.
3. Quinton, R.J., DiDomizio, A., Vittoria, M.A., Kotynkova, K., Ticas, C.J., Patel, S., Koga, Y., Vakhshoorzadeh, J.,
Hermance, N., Kuroda, T.S., et al. (2021). Whole-genome doubling confers unique genetic vulnerabilities on tumour cells.
Nature 590, 492-497. 10.1038/s41586-020-03133-3.
4. Gemble, S., Wardenaar, R., Keuper, K., Srivastava, N., Nano, M., Mace, A.S., Tijhuis, A.E., Bernhard, S.V., Spierings,
D.C.J., Simon, A., et al. (2022). Genetic instability from a single S phase after whole-genome duplication. Nature 604, 146-
151. 10.1038/s41586-022-04578-4.
5. Knouse, K.A., Davoli, T., Elledge, S.J., and Amon, A. (2017). Aneuploidy in Cancer: Seq-ing Answers to Old Questions.
Annu Rev Canc Biol 1, 335-354. 10.1146/annurev-cancerbio-042616-072231.
6. Fujiwara, T., Bandi, M., Nitta, M., Ivanova, E.V., Bronson, R.T., and Pellman, D. (2005). Cytokinesis failure generating
tetraploids promotes tumorigenesis in p53-null cells. Nature 437, 1043-1047. 10.1038/nature04217.
7. Hayward, D., Roberts, E., and Gruneberg, U. (2022). MPS1 localizes to end-on microtubule-attached kinetochores
to promote microtubule release. Curr Biol 32, 5200-5208 e5208. 10.1016/j.cub.2022.10.047.
8. Hayward, D., Bancroft, J., Mangat, D., Alfonso-Perez, T., Dugdale, S., McCarthy, J., Barr, F.A., and Gruneberg, U.
(2019). Checkpoint signaling and error correction require regulation of the MPS1 T-loop by PP2A-B56. J Cell Biol 218, 3188-
3199. 10.1083/jcb.201905026.
9. Hayward, D., Alfonso-Perez, T., and Gruneberg, U. (2019). Orchestration of the spindle assembly checkpoint by
CDK1-cyclin B1. FEBS Lett 593, 2889-2907. 10.1002/1873-3468.13591.
10. Hayward, D., Alfonso-Perez, T., Cundell, M.J., Hopkins, M., Holder, J., Bancroft, J., Hutter, L.H., Novak, B., Barr, F.A.,
and Gruneberg, U. (2019). CDK1-CCNB1 creates a spindle checkpoint-permissive state by enabling MPS1 kinetochore
localization. J Cell Biol 218, 1182-1199. 10.1083/jcb.201808014.
11. Ketley, R.F., Battistini, F., Alagia, A., Mondielli, C., Iehl, F., Balikci, E., Huber, K.V.M., Orozco, M., and Gullerova, M.
(2022). DNA double-strand break-derived RNA drives TIRR/53BP1 complex dissociation. Cell Rep 41, 111526.
10.1016/j.celrep.2022.111526.
12. Burger, K., Schlackow, M., and Gullerova, M. (2019). Tyrosine kinase c-Abl couples RNA polymerase II transcription
to DNA double-strand breaks. Nucleic Acids Res 47, 3467-3484. 10.1093/nar/gkz024.

The role of mutational order in colorectal cancer evolution, epigenetics
and response to therapy 1,2,3 – Prof. Simon Buczacki
Primary Supervisor: Prof. Simon Buczacki
Additional Supervisors: Dr. Dan Woodcock
Abstract
Cancer development is characterised by the sequential acquisition of somatic mutations in DNA. Leveraging our expertise in gene
engineering (Chan et al. 2023, Chan et al. 2023) of human intestinal organoids (3D mini guts), our group has recently shown for the
first time in solid organ cancers, that the order in which mutations are acquired results in unexpected changes to phenotype, gene
expression and response to drug therapy (Fig 1)
(Chan et al. Manuscript Submitted). These
heretical data identify that alterations in DNA
structure (chromatin) and accessibility appear key
in determining the functional consequences of
changes to mutational order. Our findings
establish mutational order as a novel route to both
tumour heterogeneity and resistance to targetted
agents. This project maximises on our group’s
expertise in gene engineering, organoid culture
and multiomics to advance our recent findings to
develop clinically applicable multiomic signatures
of mutational order (SMOs) that may better help
predict tumour biology and response to therapy.
Research objectives
WP1 – Production and multiomic characterisation of CRISPR-engineered switched colon organoid models containing the main
truncal mutations of colorectal cancer in distinct orders. Using protocols established in our group, the student will generate
CRISPR-Cas9 engineered organoids containing mutations in APC, KRAS, P53, SMAD4 and PIK3CA
(Fig 2). Isogenic organoid models will be generated with mutations placed in reverse and
alternate orders. Confirmation of targetting will be performed using Sanger sequencing and
western blotting. Organoid models will be analysed using RNA sequencing (gene expression) and
ATAC sequencing (chromatin accessibility). Live cell imaging, immunofluorescence and electron
microscopy (scanning and transmission) will be performed to characterise phenotypes. Selected
models will also be anlaysed using single cell RNA sequencing to establish changes in cell fate and
identity.
WP2 – Drug screen analysis of CRISPR-engineered switched organoid models. Our group has
an established pipeline for drug screening of cancer organoids. The student will ultilise this
pipeline to perform drug screening using both standard-of-care colorectal cancer treatments and
Figure 2. CRISPR-Cas9 novel compounds in engineered organoids and isogenic-switched models. Organoids will be
mutated human colon assessed for viability and characterised using a live cell imaging platform.
organoids
WP3 – Integration of experimental multiomic analyses to generate signatures of mutational
order (SMOs). Data generated in WP1 will be integrated using advanced bioinformatic techniques to generate SMOs. SMOs will
be validated on publically available data sets to identify tumours where mutational order can be inferred from SMO presence.
SMOs will be applied to identify differential response to adjuvant treatment regimes and variations in clinical outcomes.
Until very recently due to technological constraints, the ability to study effects of alterations in mutational order have been near
impossible. Our pioneering work in this new area of cancer biology has identified that changes in mutational order unexpectedly
generate not only profound variations in cell behaviour and stem cell dynamics but also lead to differential responses to
chemotherapies. Whilst there are multiple routes to drug resistance and sensitivities, the impact of mutational order on this
25
remains completely unexplored. This project therefore has the potential to generate novel mechanistic insights and new putative
drug targets to improve chemotherapy efficacy.
The student will be directly supervised by Prof. Simon Buczacki (Tumour Evolution and Cell Identity Laboratory). All students in the
group have a weekly one to one with Prof. Buczacki. We have a weekly lab meeting where the student will present their data and
discuss findings to the wider group and collaborators. There is common cross-fertilisation of ideas and expertise within the group
to foster a caring and supportive lab environment. The group’s philosophy is to constantly challenge traditional dogma and address
the ‘big’ questions in cancer and stem cell biology. The student will be trained in a wide range of wet and dry lab techniques by
members of the group expert in these areas. The group is fortunate to have an embedded post-doctoral bioinformatician with
significant expertise in multiomic analyses (Dr Amit Mandal). The second supervisor, Dr Dan Woodcock is an expert in cancer
evolution and multiomic data integration. Dr Woodcock’s expertise as second supervisor will be integral to many aspects of the
project especially WP3.
Wet lab: Human tissue processing, organoid culture (normal and tumour), CRISPR-Cas9 gene engineering, NGS (RNAseq,scRNAseq,
ATACseq), RT-PCR, western blotting, Sanger sequencing, flow cytometry and microscopy (confocal, live cell and EM).
Dry lab: Multiomic data (processing, analysis, integration), coding (R, Python) and big data handling.
References
Chan DKH, Collins SD, Buczacki SJA (2023). Generation and immunofluorescent validation of gene knockouts in adult human colonic
organoids using multi-guide RNA CRISPR-Cas9. STAR Protocols. 4(1):101978.
Chan DKH, Mandal A, Hester S, Yu Z, Higgins G, Kessler B, Fischer R, Buczacki SJA (2023). Biallelic FBXW7 knockout induces AKAP8-
mediated DNA damage in neighbouring wildtype cells. Cell Death Discovery. In press
Chan DKH et al. Mutational order and epistasis regulate the transcriptional consequences of FBXW7 mutations during early
colorectal cancer. In submission

Spatial resolution of the human transcriptome during gastrointestinal
tumorigenesis – 1,2,3,4 Dr. Francesco Boccellato
Primary Supervisor: Dr. Francesco Boccellato
Additional Supervisors: Jan Bornschein
Abstract
The gastrointestinal mucosa is organised in invaginations called glands in the stomach and crypts in the colon. How cancer
arises from these invaginations is still unclear, but before cellular transformation occurs, the analysis of early mucosal
aberrations in biopsies enables the detection of pre-cancerous conditions. Tissue response to stress, toxic dietary
compounds, infections and inflammation, might alter the microenvironment posing the mucosa at risk of malignant
transformation. We plan to analyse pre-cancerous conditions of the gastrointestinal tract to understand why they have a
different tissue configuration compared to the normal.
Growth factors and morphogens shape the tissues during embryogenesis, and they are probably responsible of its
homeostasis in the adult. We hypothesize than an alteration of the morphogen signalling microenvironment is the driver
for the altered configuration of the tissue observed in pre-cancerous conditions of the gastrointestinal tract. The project’s
main objective is to use spatial transcriptomics to map the molecular triggers dictating morphological and cellular
composition changes in the gastro-intestinal diseased mucosa. We will focus on the detection of morphogens and growth
factors involved in cellular regeneration and differentiation, and we plan to harness our established mucosoid cultures, an
evolution of organoids, to test the different growth factors combinations in vitro.
We use fixed samples of gastrointestinal biopsies from healthy individuals and from patients with pre-cancerous conditions
or lesions. We are particularly interested in the etiogenesis of oesophageal, stomach and colorectal cancer. Three patients
per condition will be analysed using the GeoMX whole genome spatial transcriptomic profiling provided by NanoString®.
Transcriptomic data from this experiment are analysed to extract information about the expression of morphogenic
signals, their receptors and downstream target genes. The activity of morphogenic pathways is tested on mucosoid
cultures using synthetic morphogens and corresponding pathway inhibitors. Upon stimulation with morphogens, cells in
the mucosoids cultures can regenerate and differentiate into the different stomach lineages as they do inside the organ.
Mapping growth factors directly in the original human tissue and testing their function on relevant human-derived cultures
promises to be a robust strategy to understand mechanisms of carcinogenesis
Research Objectives
Aim1) Generating a spatial map of the morphogenic signals during gastrointestinal disease progression.
We will profile the transcriptome of different parallel regions of healthy and pre-cancerous gastrointestinal mucosa (eg:
Barrett’s oesophagus, intestinal metaplasia, colon polyps). Our clinical collaborator and co-supervisor Dr. Jan Bornschein
is involved in the identification of the patients for this project. Prof Fadi Issa, runs Nanostring at the department of surgical
science at University of Oxford. We will focus on the detection of genes related to morphogen signalling pathways involved
in cellular regeneration and differentiation. The expression of these genes, receptors and related transcription factors will
be mapped in the different region of interest of the epithelium, stroma and cell of the immune system. The comparative
analysis of the transcriptomic profiles will be performed using the software “SignaLink3” [1] to identify the gene
interactome and the putative pathways crosstalk. The software was developed by T. Korcsmaros, a collaborator in this
project.
Aim II) Assessing the role of morphogens in driving epithelial differentiation, regeneration and proliferation.
The pre-cancerous conditions of the gastrointestinal mucosa are characterised by a different morphology but also by a
disbalance in the cell population lineages. The morphogen signalling ligands identified in Aim 1 will be tested on mucosoid
cultures originated from the same biopsies. Mucosoids are a patented [2] development of the organoid cultures; cells are
cultivated in a monolayer forming an epithelial barrier which is very similar to the gastrointestinal epithelium [3, 4]. Cells
within the mucosoids can differentiate upon stimulation [5]. By adding ligands or pathway inhibitors in the cultivation
cocktail of the mucosoid cultures it is possible to determine their role in epithelial cell regeneration, proliferation and
differentiation using different published functional or biochemical assays [3, 5]
27
Although there is a strong focus on understanding the microenvironment of cancer and the contribution of neighbouring
non-transformed cells to the disease, little is known about the microenvironment of pre-cancerous conditions, and an
unbiased approach to map all the morphogens has never been attempted. We aim to find dysregulations in specific
morphogen signalling cascades that are predictive for disease progression. The gold standard for the detection pre-
cancerous conditions is endoscopy and tissue imaging. Alternative serological analysis is accurate, but have a low
sensitivity. A combination of ligands or proteins involved in morphogen signalling pathways could be use as surrogate of
those conditions to develop diagnostic tests for pre-cancerous conditions and to predict risk of progression
Day-to-day supervision and training will be provided by Francesco Boccellato and from post-docs in the lab. The student
will have the opportunity to learn cutting edge technologies such as spatial-transcriptomic and organoid and mucosoid
cultures. We expect the student to become proficient into data analysis and we will support this by encouraging the
attendance to bioinformatic courses. Imaging with confocal microscopies and standard biochemical assays are also part of
the basic training.
References
1. Csabai, L.; Fazekas, D.; Kadlecsik, T.; Szalay-Bekő, M.; Bohár, B.; Madgwick, M.; Módos, D.; Ölbei, M.; Gul, L.;
Sudhakar, P.; Kubisch, J.; Oyeyemi, O. J.; Liska, O.; Ari, E.; Hotzi, B.; Billes, V. A.; Molnár, E.; Földvári-Nagy, L.; Csályi, K.;
Demeter, A.; Pápai, N.; Koltai, M.; Varga, M.; Lenti, K.; Farkas, I. J.; Türei, D.; Csermely, P.; Vellai, T.; Korcsmáros, T.,
SignaLink3: a multi-layered resource to uncover tissue-specific signaling networks. Nucleic acids research 2022, 50, (D1),
D701-d709.
2. Boccellato, F.; Meyer, T. F. Generation, proliferation and expansion of epithelial cells from primary tissue into
mucosoid cultures. 2019, 2019.
3. Boccellato, F.; Woelffling, S.; Imai-Matsushima, A.; Sanchez, G.; Goosmann, C.; Schmid, M.; Berger, H.; Morey, P.;
Denecke, C.; Ordemann, J.; Meyer, T. F., Polarised epithelial monolayers of the gastric mucosa reveal insights into mucosal
homeostasis and defence against infection. Gut 2019, 68, (3), 400-413.
4. Sepe, L. P.; Hartl, K.; Iftekhar, A.; Berger, H.; Kumar, N.; Goosmann, C.; Chopra, S.; Schmidt, S. C.; Gurumurthy, R. K.;
Meyer, T. F.; Boccellato, F., Genotoxic Effect of Salmonella Paratyphi A Infection on Human Primary Gallbladder Cells. mBio
2020, 11, (5).
5. Wölffling, S.; Daddi, A. A.; Imai-Matsushima, A.; Fritsche, K.; Goosmann, C.; Traulsen, J.; Lisle, R.; Schmid, M.; Reines-
Benassar, M. D. M.; Pfannkuch, L.; Brinkmann, V.; Bornschein, J.; Malfertheiner, P.; Ordemann, J.; Link, A.; Meyer, T. F.;
Boccellato, F., EGF and BMPs Govern Differentiation and Patterning in Human Gastric Glands. Gastroenterology 2021, 161,
(2), 623-636.e16.

ADP-ribosyl hydrolase as a biomarker for PARP inhibitor
sensitivity/resistance 1,2,3,4 – Prof. Ivan Ahel
Primary Supervisor: Prof. Ivan Ahel
Additional Supervisors: Prof. Ahmed Ahmed
Abstract
To protect the genome from damage organisms have evolved a cellular defence mechanisms termed the DNA damage
response (DDR). The DDR includes a diverse set of signal transduction pathways and effector proteins that act to sense
DNA lesions and effectively repair the damage, limiting the propagation of genomic instability. Exploiting DDR pathways
to specifically target and kill cancer cells has become an attractive therapeutic avenue within cancer research. This is
exemplified by the synthetic lethal interaction between PARP inhibition and BRCA1 or BRCA2-deficient tumours1. Ivan
Ahel (co-supervisor on this project) laboratory recently identified HPF1 protein as a novel interactor and critical
regulator of PARP1 ADP-ribosylation activity upon DNA damage2. Functionally, HPF1 suppresses DNA damage-induced
hyper auto-modification of PARP1 and promotes in trans ADP-ribosylation of histones and many other proteins
involved in regulation of genome stability. They further demonstrated that HPF1 is a critical specificity factor that allows
modification of target proteins by PARP1 on serine residues (Ser-ADPr)3,4. Crucially, the work also identified ARH3 as a
hydrolase which specifically removes Ser-ADPr5 and further showed that Ser-ADPr is the major form of ADP-ribosylation
following DNA damage6. Taken together, the insights surrounding Ser-ADPr open a large, exciting, and novel area of
research into the fundamental understanding of the pathways regulated by this modification. Strikingly, our recent
data show that ARH3 knockout in model cell lines associates with PARP inhibitor (PARPi) resistance, while ARH3
overexpression is associated with PARPi sensitivity7. Based on these results, we hypothesize that ARH3 activity and
protein levels affect sensitivity to PARPi, thus representing; i) a predictor for the success of these therapies and, ii) a
novel target for further drug development. Currently, PARP inhibitors are used to treat ovarian cancer and several
other cancers, and we therefore propose to test the hypothesis that ARH3 expression might be a useful diagnostic tool
with which to stratify cancer patients into sub-groups that will be sensitive/resistant to PARPi treatment with a
particular focus on ovarian cancer. The mechanism of sensitivity/resistance of cells with deregulated ARH3 expression
cells to PARPi is unknown, and elucidating this mechanism will be another goal of this proposed work.
Research objectives
Objective 1. Characterise the effect of ARH3 under- and overexpression in a series of model and primary cancer cell
lines on PARP inhibitor sensitivity/resistance. We will collect and test a variety of ovarian cancer cell lines, profiling
them for ARH3 protein expression levels and then treating with several different PARPi of varying PARP-trapping
capabilities (olaparib, talazoparib, veliparib). To determine the impact of ARH3 protein levels on PARPi vulnerability,
we will not only assess drug sensitivity and levels of PARP1, PARG, and ARH3 across a panel of ovarian cancer cell lines,
but also assess the impact of systemically varying ARH3 by knockdown, knock out and inducible overexpression in
HGSOC lines of defined genotype, including Ovcar8 (BRCA1/2 wt, PARPi resistant), PE01 (BRCA2-mutant, PARPi
sensitive), Kuramochi (BRCA2-mutant, PARPi partially sensitive) and COV362 (BRCA1-mutant, PARPi sensitive). Rescue
experiments with wild type vs. catalytically inactive ARH3 will assess the suitability of ARH3 as a target for the
development of inhibitors.
Objective 2. To determine the frequency of ARH3 gene alterations in a larger set of HGSOC samples, we will: i)
interrogate data of an ongoing whole exome sequencing study of 504 ovarian cancers searching for ARH3 and PARG
copy number alterations and mutations; and ii) perform semi-quantitative detection of ARH3, as well as of PARG, PARP1
and PAR, by immunohistochemistry (IHC) on two independent sets of tissue microarrays (TMAs) containing a total of
1200 ovarian cancers. To augment these analyses, which will be limited by the small number of tumors treated with
PARPi, we will also evaluate levels of ARH3, PARG, PARP1 and PAR in patient-derived xenograft (PDX) models that have
been assayed for response to single-agent PARPi, including ones that have a high HRD score but did not respond. This
objective will be performed in co-supervisor (Prof Ahmed Ahmed) laboratory at the Nuffield Department of Women's
& Reproductive Health, University of Oxford.
Objective 3. Elucidating the mechanistic basis for the sensitivity/resistance of cells with deregulated ARH3 expression
cells to PARPi (modulation of the PARP-trapping, regulation of DNA repair pathway choice, regulation of the chromatin
structure/epigenetic marks). For these studies we will use largely cell biology/biochemical and genomics approaches.
29
Our data suggest that ARH3 protein expression levels in cancer patients might be a marker that confers
sensitivity/resistance of the tumour to PARPi, providing a rationale for using PARPi for certain patients. In longer term,
understanding the mechanisms of DNA repair and PARPi resistance through studies of ARH3 protein, may reveal new,
unexpected avenues for treatments in the future.
The student will have opportunities to train in diverse set of methods including cell biology/cell culture approaches for
structure/function analyses, well-established cell survival assays that we be applicable for wide range of cell toxicity
studies, immunohistochemistry methods and patient-derived xenograft (PDX) models.
References
1. Bryant et al (2005) Specific killing of BRCA2-deficient tumours with inhibitors of poly(ADP-ribose)

polymerase. Nature 434, 913-917.
2. Gibbs-Seymour, I., Fontana, P., Rack, J.G., and Ahel, I. (2016) HPF1/C4orf27 Is a PARP-1-Interacting Protein
that Regulates PARP-1 ADP-Ribosylation Activity. Mol Cell 62, 432-442.
3. Bonfiglio, J.J., Fontana, P., Zhang, Q., Colby, T., Gibbs-Seymour, I., Atanassov, I., Bartlett, E.J., Zaja, R., Ahel,
I.*, and Matic, I.* (2017) Serine ADP-ribosylation depends on HPF1. Mol Cell 65, 932-940. (*Corresponding
authors)
4. Suskiewicz, M.J., Zobel, F., Ogden, T.E., Fontana, P., Ariza, A., Yang, J., Zhu, K., Bracken, L., Hawthorne,
W.J., Ahel, D., Neuhaus, D., and Ahel, I. (2020) HPF1 completes the PARP active site for DNA-damage
induced ADP-ribosylation. Nature 579, 598-602.
5. Fontana, P., Bonfiglio, J.J., Palazzo, .L, Bartlett, E., Matic, I., and Ahel, I. (2017) Serine ADP-ribosylation
reversal by the hydrolase ARH3. Elife Jun 26;6. pii: e28533.
6. Palazzo, L., Leidecker, O., Prokhorova, E., Dauben, H., Matic, I., and Ahel, I. (2018) Serine is the major
residue for ADP-ribosylation upon DNA damage. Elife Feb 26;7. pii: e34334.
7. Prokhorova, E., Zobel, F., Smith, R., Zentout, S., Gibbs-Seymour, I., Schützenhofer, K., Peters, A.,
Groslambert, J., Zorzini, V., Agnew, T., Brognard, J., Nielsen, M.L., Ahel, D., Huet, S., Suskiewicz, M.J., and
Ahel, I. (2021) Serine-linked PARP1 auto-modification controls PARP inhibitor response. Nat Commun 12,
4055.

Improving Immunotherapeutic Efficacy in Colorectal Cancer Using
Ultrasound-activated Nanoparticles and Image-guided Drug Delivery
1,2,3,4 – Prof. Tim Elliot
Primary Supervisor: Prof. Tim Elliot

Additional Supervisors: Associate Prof. Robert Carlisle
Abstract
Cancer immunotherapy using immune checkpoint inhibitors (ICI) targeting the programmed cell death-1 receptor (PD-1) and
its ligand (PD-L1) are routinely used in clinic for treating solid tumours1. However, >95% of colorectal cancers are
immunologically ‘cold’, microsatellite stable and DNA mismatch repair proficient tumours that do not currently benefit from
immunotherapy2. Treatment failure is often due to factors e.g. low neoantigen burden, loss of MHC-I, dysfunctional antigen
processing and presentation, lack of tumour-infiltrating lymphocytes and local immunosuppression. Furthermore, the
presence of dense stroma and dysregulated tumour blood vessels can serve as physical and functional barriers to drug
delivery. There is an unmet need for the development of therapeutic
and image-guided approaches to enhance drug delivery into and
throughout tumours and transform a ‘cold’ immune-excluded tumour
into a ‘hot’ immune-inflamed tumour.
Figure 1. Ultrasound-activated nanoparticles for delivery of drugs.
(A) Schematic diagram for ultrasound-guided delivery of nanoparticle encapsulated

treatment antibodies. (B) TEM of novel ultrasound-responsive solid particles
manufactured using immunostimulatory proteins. (C) and (D) Improved delivery of
oncolytic adenoviruses (‘PC-Ad’) using microbubbles (‘SV’) and ultrasound. Green
(virus), red (blood vessel), blue (DAPI). Images were adapted from [5] and [7].
Focused ultrasound has emerged as a promising approach for immuno-modulation of tumours through thermal or mechanical
perturbation3. Ultrasound-mediated cavitation using novel, biocompatible nanoparticles has been used to enhance the
delivery of chemotherapy and oncolytic viruses into tumours4,5 (Figure 1). Ultrasound itself is a relatively low-cost system used
in clinical imaging of soft tissues, tissue stiffness and tumour blood flow, and can be used for real-time monitoring of drug
delivery. The Elliott Group at the Centre for Immuno-Oncology investigates the mechanisms of antigen processing and
presentation involved in T cell response to cancer and during immunotherapy using biochemical, computational, and physical
science methods. The Biomedical Ultrasonics, Biotherapy and Biopharmaceuticals Laboratory (BUBBL) at the Institute of
Biomedical Engineering specialises in the development of novel ultrasound-activated nanoparticles for image-guided drug
delivery. This multidisciplinary project is a partnership between immuno-oncology and biomedical engineering.
Research objectives
The aim of this project is to develop and validate ultrasound-activated microbubbles (MB) and solid sonosensitive particles
(SP) to actively transport and enhance the distribution of ICI in tumours via mechanical cavitation. The treatment efficacy of
ultrasound-activated cavitation agents as propellants or vehicles for ICI delivery and the mechanistic effects of cavitation on
antigen presentation, T cell avidity and function will be investigated. Preclinical colorectal cancer models already
comprehensively characterised by the Elliot group will be used 6,7.
Work Package 1: Development of ultrasound-activated nanoparticles for targeted drug delivery

PD-1 and/or PD-L1 monoclonal antibodies of different doses will be co-delivered with SP or formulated into SP or MB. The
physicochemical properties of the SP and MB formulations, doses of PD-1/PD-L1 monoclonal antibodies delivered with, or
formulated into, each cavitation agent and the acoustic parameters for tumour delivery will be optimised, and correlated to
the tumour volume, microvessel density and stromal thickness. For microscopic evaluation of the drug spatial distribution,
PD-1 and/or PD-L1 monoclonal antibodies will be bioconjugated with amine-reactive reagents for fluorescent labelling with
near infrared dyes. WP1 will be performed in collaboration with Prof. Constantin Coussios and Prof. Eleanor Stride.
Work Package 2: Establishment of a preclinical colorectal cancer model for novel treatment and mechanistic studies
31
The BALB/c mouse model of colorectal carcinoma, CT26 is a microsatellite stable and DNA mismatch repair proficient tumour
which shares molecular features with aggressive, undifferentiated, and refractory human colorectal cancer. CT26 is one of the
most extensively investigated syngeneic tumour model in preclinical studies and has been used to validate most
immunotherapeutics currently in the clinic or under clinical trials, with >500 studies in literature. SP and MB formulations with
optimised PD-1 or PD-L1 doses will be administered into mice bearing subcutaneous CT26 tumours and actively delivered into
the tumours via ultrasound-mediated cavitation. Tumour growth monitoring will be conducted to determine treatment
efficacy and response classification into progressors or regressors. Tumours and secondary lymphoid organs will be harvested
at the study endpoint for functional analyses using multiparametric flow cytometry, transcriptomics, immunofluorescence
staining and multiplex imaging. In particular, biomarkers related to MHC Class I antigen presentation, antigen processing e.g.
tapasin and calreticulin, CD8+ and CD4+ T cell, antigen-presenting cells such as dendritic cells, immunosuppressive cells, cancer-
associated fibroblasts, extracellular matrix remodelling, vascular inflammation and tumour hypoxia will be examined closely.
Experiments will be designed to distinguish the immunological effects of cavitation alone from the effects of enhanced
antibody delivery instigated by cavitation. WP2 will be conducted in collaboration with Dr. Doreen Lau and A/Prof. Joanna
Hester.
Work Package 3: Understanding the effects of novel treatment on antigen presentation and T cell dynamics in cancer
The mechanical and immuno-modulatory effects of the optimised nanoparticle formulation and focused ultrasound on
antigen presentation and T cell dynamics in cancer will be examined with immuno-profiling, biophysical measurements and
live imaging. Tumour antigen-specific CD8+ T cells of high versus low functional avidity will be identified based on pMHC
tetramer staining and image-based biophysical measurements of overall pMHC:TCR binding strength using acoustic force
spectroscopy. These T cell populations will be fluorescently labelled and adoptively transferred into CT26 mice prior to
treatment with this novel approach. Non-invasive imaging of tumour stiffness using ultrasound elastography and vascular
permeability and perfusion using contrast-enhanced ultrasound will be conducted before and after treatment to derive
clinically relevant imaging biomarkers of stromal and vascular response to treatment. Imaging at the single-cell level will be
performed using two-photon microscopy to examine the migration and cellular kinetics of tumour-specific T cells across
physical barriers (tumour stroma and vasculature). This will be conducted on vibratome-sliced tumours in perfusion chambers
with dye-labelling of the tumour blood vessels for visualisation and second harmonic generation imaging of the stromal
collagen fibres at the tumour invasive margin. Image processing and analysis of the fluorescent T cell tracks, velocities, and
spatial confinement within different tumour compartments (peritumoral versus intratumoral), T cell spatial distribution and
distance to tumour cells, tumour stroma and blood vessels will be conducted 8. WP3 will be done in collaboration with Dr.
Doreen Lau and Prof. Eleanor Stride.
This work will develop and validate novel ultrasound-activated nanoparticles for image-guided delivery of ICI in preclinical
colorectal cancer models. The aim is to investigate whether drug penetration can be optimised using this approach and better
understand its mechanistic effects on antigen presentation, T cell avidity and function in tumours. An Oxford spin-out
company, OxSonics Therapeutics, in which a similar cavitation technology (‘Nanocups’) is based, has already reached clinical-
stage and is currently in Phase I/II clinical trial for ultrasound-guided delivery of anti-cancer agents in metastatic colorectal
cancer patients. The difference in the present proposal is the introduction of a sonosensitive particle that is capable of not
only enhancing antibody delivery and distribution, but also in promoting the local immune response through the introduction
of immuno-stimulatory proteins adjacent to the cavitation process. A long-term plan for this project would be to translate the
technology and discoveries into clinics to optimise the efficacy of immunotherapeutics.
This project is ideally suited for non-clinical candidates with a background in pharmacology, biomedical sciences or
bioengineering and is also suitable for clinical candidates wishing to gain research experience in preclinical immunotherapy
modelling and novel drug delivery in an interdisciplinary setting. Training on chemical engineering, tumour immunology,
cellular imaging with two-photon microscopy and acoustic force spectroscopy, as well as the use of therapeutic and diagnostic
imaging ultrasound will be provided. More general research, communication, teaching, innovation, and career development
skills training will be given by the Medical Sciences Division.
References:
1.Ribas, A., and Wolchok, J.D. (2018). Cancer immunotherapy using checkpoint blockade. Science (1979) 359, 1350–1355.
10.1126/science.aar4060.
32
2.Ganesh, K., Stadler, Z.K., Cercek, A., Mendelsohn, R.B., Shia, J., Segal, N.H., and Diaz, L.A. (2019). Immunotherapy in
colorectal cancer: rationale, challenges and potential. Nature Reviews Gastroenterology & Hepatology 16, 361–375.
10.1038/s41575-019-0126-x.
3. Joiner, J.B., Pylayeva-Gupta, Y., and Dayton, P.A. (2020). Focused Ultrasound for Immunomodulation of the Tumor
Microenvironment. The Journal of Immunology 205, 2327–2341. 10.4049/jimmunol.1901430.
4. Grundy, M., Bau, L., Hill, C., Paverd, C., Mannaris, C., Kwan, J., Crake, C., Coviello, C., Coussios, C., and Carlisle, R.
(2021). Improved therapeutic antibody delivery to xenograft tumors using cavitation nucleated by gas-entrapping
nanoparticles. Nanomedicine 16, 37–50. 10.2217/nnm-2020-0263.
5. Carlisle, R., Choi, J., Bazan-Peregrino, M., Laga, R., Subr, V., Kostka, L., Ulbrich, K., Coussios, C.-C., and Seymour, L.W.
(2013). Enhanced Tumor Uptake and Penetration of Virotherapy Using Polymer Stealthing and Focused Ultrasound. JNCI:
Journal of the National Cancer Institute 105, 1701–1710. 10.1093/jnci/djt305.
6. Sugiyarto, G., Prossor, D., Dadas, O., Arcia-Anaya, E.D., Elliott, T., and James, E. (2021). Protective low-avidity anti-
tumour CD8+ T cells are selectively attenuated by regulatory T cells. Immunotherapy Advances 1. 10.1093/immadv/ltaa001.
7. Sugiyarto, G., Lau, D., Hill, S.L., Arcia-Anaya, E.D., Boulanger, D.S., Parkes, E.E., James, E., and Elliott, T. (2023).
Reactivation of low avidity tumor-specific CD8+ T cells associates with immunotherapeutic efficacy of anti-PD-1. Under
Review.
8. Lau, D., Garçon, F., Chandra, A., Lechermann, L.M., Aloj, L., Chilvers, E.R., Corrie, P.G., Okkenhaug, K., and Gallagher,
F.A. (2020). Intravital Imaging of Adoptive T-Cell Morphology, Mobility and Trafficking Following Immune Checkpoint
Inhibition in a Mouse Melanoma Model. Front Immunol 11, 1514. 10.3389/fimmu.2020.01514.

Uncovering the regulation and functions of supermeres in colorectal
cancer - 1,2,3,4 – Prof. Clive Wilson
Primary Supervisor: Prof. Clive Wilson

Additional Supervisors: Prof. Adrian Harris and Prof. Chris Cunningham
Abstract
Defective communication between cells is a key factor in cancer formation and progression. Traditionally, such signals are
considered to involve single molecules like growth factors, but recently, more sophisticated multimolecular assemblies, such as
extracellular vesicles (EVs) and non-vesicular nanoparticles, which can harbour a wide range of activities, have emerged as
alternative mediators. Supermeres are a new type of protein:RNA complex produced by many cancer cell types1. They are
enriched with cargos that are upregulated in a wide range of cancers, for example, glycolytic enzymes, TGFBI, miR-1246, MET,
GPC1 and AGO2. Most extracellular RNA is reported to be associated with supermeres and not EVs. Cancer-derived supermeres
increase lactate secretion, transfer cetuximab resistance and reduce hepatic lipids and glycogen in vivo. However, the cellular
origin of supermeres and their links with other secreted multimolecular complexes are unknown, hindering their detailed
analysis. We previously used a Drosophila prostate-like cell with highly enlarged endosomal compartments to unpick the
biology of exosomes, small EVs generated in these compartments. This work revealed a previously unidentified, evolutionarily
conserved exosome subtype called Rab11a-exosomes made in recycling endosomes, which appear to control cancer
progression2. Although they represent only a small fraction of secreted EVs, they are major mediators of several EV-associated
functions3. We have now identified protein aggregates in our fly system with a similar protein signature to human supermeres,
shown that their biogenesis is interlinked with Rab11a-exosome formation, and identified genetic manipulations of highly
conserved molecules that block their assembly3-6 (and see below). We hypothesise that human supermeres are generated in
recycling endosomes via similar mechanisms to those in the fly and that these compartments play a central role in cancer cell
communication. In this project, the student will test this idea in colorectal cancer (CRC) cells, where we first identified Rab11a-
exosomes. Ongoing molecularly stratified trials and collection of serial blood samples from patients will allow the specific
functions of supermeres and their potential as biomarkers to be tested. The student will spearhead the analysis of this new
form of cancer cell signalling, working with basic scientists and clinicians who have diverse skills and expertise.
Research objectives
The project has four research objectives and proposed outcomes:
2.1 Determine the role of cellular stress responses in supermere release: In our fly model, supermere-like structures coalesce in
Rab11a-positive recycling endosomes to form very large central dense-core granules3-6. Like human supermeres1, TGFβ-induced
(TGFBI, a highly conserved secreted molecule) is the most abundant protein in these complexes (Fig. 1A), which are also
enriched in other supermere cargos, such as GPI-anchored proteins, proteases6 and cleaved membrane proteins, like Amyloid
Precursor Protein (APP). Our studies reveal that complex formation is dependent on TGFBI (Fig. 1) and Rab11a-exosomes3-5. In
recent unpublished findings, we have shown that other oncogenic proteins control this process, and that APP-like proteins play
a key role in separating Rab11a-exosomes from these aggregates. This separation regulates the activities of both the exosomes
and the aggregates.
To
determine whether human supermeres are selectively produced in Rab11a-compartments, we will test whether supermere
release is affected in CRC cell lines, eg. HCT116, SW620, by stresses that promote Rab11a-exosome secretion, eg. nutrient
depletion, hypoxia, drugs that block growth factor signalling. Supermeres will be isolated by ultracentrifugation; if they are
abundant in non-stressed, as well as stressed, cells, their composition will be assessed by western/miRNA analysis, to determine
whether, like exosomes, their structure is altered by their origin.
Proposed outcome: Determine the origin and composition of supermeres generated under different stress conditions.
34
2.2 Identify regulators of supermere biogenesis: We will genetically manipulate CRC cells under both nutrient-depleted and -
replete conditions, knocking down genes involved in the formation of supermere-like structures in flies, eg. TGFBI, ESCRTs (to
suppress exosome biogenesis), etc, and test how supermere release is affected. Our current data suggest that TGFBI
knockdown (kd) will selectively block supermere formation and this will change the activities of the cancer cell secretome; the
resulting preparations, depleted of supermeres, will be analysed by proteomics and transcriptomics, to more fully determine,
through a process of elimination, the composition of these structures.
Proposed outcome: Define mechanisms controlling supermere generation in CRC cells and identify specific supermere cargos.
2.3 Determine cancer-related functions of supermeres: Cancer supermeres are readily internalised within target cells and have
multiple functions in metabolism and drug resistance1. We will confirm this using supermere preparations from nutrient-
depleted and -replete cells to determine whether supermere signalling changes under stress. We will also test the functions of
conditioned medium from cells +/- nutrient-depletion, +/- kd (eg. TGFBI kd) that blocks supermere release to assess supermere
function in the presence of other signals, eg. exosomes. We will include assays for metabolic changes 1, drug resistance1,
including oxaliplatin and 5FU, and effects on endothelial2 and other stromal cells.
Proposed outcome: The functions of supermeres in the presence or absence of other signals will be defined.
2.4 Assess the therapeutic implications of supermeres in CRC patients: Recent unpublished studies indicate that some Rab11a-
exosome markers from CRC cell lines are detected in EV preparations from CRC patients’ plasma and these may be differentially
expressed in patients that respond differently to neoadjuvant chemoradiotherapy (nCRT). We will isolate supermeres from
plasma of these and other patients, both pre- and post-therapy, and from blood-bank controls, confirming integrity after
freeze-thaw, testing effects of different anticoagulants, etc; protein and miRNA content will be assessed in relation to
therapeutic response.
Proposed outcome: These preliminary studies will assess whether supermeres carry biomarkers that relate to CRC or response
to nCRT, and pave the way for more patient-focused studies to block supermere function in the future.
2.5 Academic value of research: Signalling by multimolecular complexes is emerging as a key, but poorly understood,
mechanism by which normal and cancer cells can completely reprogramme target cells and their microenvironment. Wilson’s
lab has made a number of fundamental science discoveries that have opened up new ideas and approaches that allow the
analysis for these processes. The group has a strong track record in discovery science with relevance to cancer biology 2-7, then
developing ideas translationally through collaborations (see 2.6). The student will play a central role in taking a similar approach
to characterise cancer supermeres.
2.6 Collaborations involved and how these will be facilitated by the award: Profs Adrian Harris and Clive Wilson, working with
Assoc. Prof Deborah Goberdhan have a strong, long-standing track record of collaboration in tumour cell biology, in which
Drosophila has informed cancer studies, eg [2,8]. Prof Chris Cunningham has collaborated with Goberdhan and Harris to study
exosomes in CRC patients. This project brings these teams together to work in a completely new area, supermere biology, so
that fundamental discoveries in flies can again be exploited in clinically related studies. The project will facilitate these
interactions and increase the opportunities for subsequent translation.
There is increasing evidence that multimolecular complexes, like supermeres, play central roles in cancer signalling, but defining
those roles and assessing the translational potential requires advanced cell biological and genetic analysis. Our fly studies have
opened up the opportunity to gauge the importance of supermere signalling in CRC and other cancers and determine the potential
relevance of supermeres to early detection, patient stratification, prognosis and therapy, guided by the expertise of the
collaborating supervisors (see 2.6 above). This project will provide the proof-of-principle work for future patient-focused, CRUK-
funded studies in this area.
The student will use a range of cancer-relevant techniques, including advanced cell culture and biochemical methods, cell biology,
molecular genetics, high-resolution fluorescence imaging and bioinformatics. They will also develop expertise in cancer signalling,
working at the interface of the collaborative environment we have established. Overall, this work should open up a new field in
cancer biology, relevant to CRC and other cancers.
35
References
1. Zhang Q, et al. (2021) Supermeres are functional extracellular nanoparticles replete with disease biomarkers and therapeutic
targets. Nat Cell Biol. 23:1240-1254; 2. Fan S-J et al. (2020) Glutamine deprivation alters the origin and function of cancer cell
exosomes. EMBO J. 2020, e1030093 and N&V, van Niel G, Théry C. (2020) EMBO J. 2020, 39:e105119; 3. Marie, P.P., et al. (2023)
Accessory ESCRT-III proteins are conserved and selective regulators of Rab11a-exosome formation. J Extracell Vesicles 12:e12311;
4. Dar GH, et al. (2021) GAPDH controls extracellular vesicle biogenesis and enhances the therapeutic potential of EV mediated
siRNA delivery to the brain. Nat Commun. 12:6666; 5. Wells, A. et al. (2023) A Rab6 to Rab11 transition is required for dense-core
granule and exosome biogenesis in Drosophila secondary cells. BioRxiv:
https://www.biorxiv.org/content/10.1101/2023.04.04.535541v1. 6. Redhai S, et al. (2016) Regulation of dense-core granule
replenishment by autocrine BMP signalling in Drosophila secondary cells. PLoS Genet. 12:e1006366; 7. Wainwright SM, et al.
(2021) Drosophila Sex Peptide controls the assembly of lipid microcarriers in seminal fluid. PNAS 118:e2019622118; 8. Fan, S-J et
al. (2016) PAT4 levels control amino acid sensitivity of rapamycin-resistant mTORC1 from the Golgi and affect clinical outcome in
colorectal cancer. Oncogene 35, 3004-15.

36
Improving CAR-T cells for B-ALL1,2,3,4 – Prof. Omer Dushek
Primary Supervisor: Prof. Omer Dushek
Additional Supervisors: Prof. Ronjon Chakraverty
Abstract
T cells patrol the body in search of antigens derived from infectious organisms or cancer cells. They use their T cell antigen
receptors (TCRs) to recognise peptide antigens on major-histocompatibility-complexes (pMHC). T cells have remarkable
antigen sensitivity; they can become activated when recognising only a single pMHC. This high sensitivity is important
because infectious organisms and cancers deploy evasion mechanisms to reduce the amount of antigen presented to T
cells. T cells are now engineered to recognise cancer antigens using chimeric antigen receptors (CARs). This therapy is
approved to treat B cell cancers. However, many patients relapse with B cells that express low levels of the target antigen.
It is now clear that CARs have a profound defect in antigen sensitivity; CARs require 100-1000-fold more antigen than
the TCR to activate T cells. There is an urgent need to increase the sensitivity of CARs to prevent these relapses. The high
antigen sensitivity of the TCR is partly a result of their adhesion receptor CD2 binding to its ligand CD58. In recent work,
we have found that CARs fail to efficiently exploit the CD2/CD58 adhesion interaction. Here, we will use our basic
understanding of how CD2 functions to generate novel CARs that efficiently exploit CD2 and test them using in vivo
relapse models of B cell acute lymphoblastic leukaemia (B-ALL). This will improve existing CAR-T cell treatments for B cell
cancers and should allow for new treatments to eliminate cancers expressing low levels of target antigens.
Background. The T cell antigen receptor (TCR) has remarkable antigen sensitivity; it is able to recognise even a single
peptide antigen on target cells (1). It achieves this high sensitivity in part by exploiting the T cell adhesion receptor CD2
binding to its ligand CD58. Recent studies have shown the CD2/CD58 interaction to play critical roles in T cell function,
including recognition of cancerous and infected cells (1).
Adoptive cell transfer (ACT) of genetically engineered T cells expressing Chimeric Antigen Receptors (CARs) is a clinically
approved cancer therapy for haematological malignancies (2). CARs are synthetic receptors that are generated by the
fusion of an antibody-derived, antigen-binding single-chain variable fragment (scFv) with intracellular signalling motifs
from the cytoplasmic tails of the TCR complex. Although CAR-T cells targeting the B cell surface antigens such as CD19
are initially highly effective, a large fraction of patients relapse when malignant cells emerge with reduced expression
the target antigens (2). It is now appreciated that CARs are relatively insensitive, requiring 100 to 1000-fold higher
antigen densities to induce T cell activation compared to the native TCR (2). However, the mechanism underlying their
defect in antigen sensitivity is presently unknown. There is now an urgent need to improve the antigen sensitivity of CAR-
T cells in order to treat patients with relapse and to avoid relapse in the first place.
Research objectives
In recently completed work, we have identified that the mechanism underlying the antigen sensitivity of CARs is their
inability to efficiently exploit the CD2/CD58 interaction (3). We found that CD2 increased the antigen sensitivity of the
TCR by 125-fold but only <5-fold for CARs. Here, we will use our basic understanding of how CD2 improves antigen
sensitivity of the TCR to produce novel synthetic antigen receptors that efficiently exploit CD2.
Objective1. Produce a highly sensitive synthetic antigen receptor targeting CD19.
Objective 2. Use pre-clinical models of B-ALL relapse to benchmark the in vivo activity of our novel synthetic receptor
(see Objective 1).
Objective 2. Use in vivo findings (Objective 2) to inform on improved design (Objective 1).
The Dushek and van der Merwe laboratories have a long-standing history of studying antigen recognition by the T cell
receptor, including understanding the role of accessory receptors such as CD2. Recently, they have developed an
experimental system to analyse the antigen sensitivity of CARs relative to the TCR. They will lead on producing highly
sensitive CARs targeting CD19 (Objective 1). The Chakraverty laboratory is focused on using murine models to understand
the mechanisms of relapse to approved CD19 CARs. They will lead on in vivo testing (Objective 2) and have recently
establish a model in their laboratory based on a previous report (4). There are now examples where the surface
expression of CD19 is not only reduced but is completely abolished. If synthetic antigen receptors with higher sensitivity
37
are insufficient to avoid this cancer evolution process, the study of relapse can identify preserved surface proteins that
can be targeted at the outset alongside CD19 (Objective 3).
Currently, CARs underlie approved cancer treatments and a large number of promising treatments (e.g. using not only T
cells but NK cells and macrophages). It follows that fundamental improvements to their antigen sensitivity can
immediately be adopted by the large community of academic and industrial researchers.
T cell isolation, tissue culture, synthetic design and genetic modification (CRISPR, Lenti-transductions), flow cytometry,
mathematical modelling (to understand mechanism), pre-clinical models of B cell cancers.
References
1. Siller-Farfán, J. A., & Dushek, O. (2018). Molecular mechanisms of T cell sensitivity to antigen. Immunological
Reviews, 285(1), 194–205. https://doi.org/10.1111/imr.12690
2. Majzner, R. G., & Mackall, C. L. (2018). Tumor antigen escape from car t-cell therapy. Cancer Discovery, 8(10), 1219–
1226. https://doi.org/10.1158/2159-8290.CD-18-0442
3. Burton, J., Siller-Farfán, J. A., Pettmann, J., Salzer, B., Kutuzov, M., van der Merwe, P. A., & Dushek, O. (2023).
Inefficient exploitation of accessory receptors reduces the sensitivity of chimeric antigen receptors. PNAS.
4. Jacoby, E., Nguyen, S., Fountaine, T. et al. CD19 CAR immune pressure induces B-precursor acute lymphoblastic
leukaemia lineage switch exposing inherent leukaemic plasticity. Nat Commun 7, 12320 (2016).
https://doi.org/10.1038/ncomms12320

38
Deciphering spatial differences in Histopathological subtype of
colorectal cancer liver metastasis (CRLM) 1,2,3,4 – Dr. Alex Gordon-
Weeks
Primary Supervisor: Dr. Alex Gordon-Weeks
Additional Supervisors: Prof. Simon Buczacki
Abstract
CRLM occur in 50% of colorectal cancer patients and metastasis is the most common cause of colorectal cancer death.
Surgical resection is considered the most successful treatment option, but recurrence occurs in up to 75% of patients
within 5 years of attempted curative surgery. However, the outcomes following surgery are highly heterogenous. The
strongest predictor of good versus poor outcome is the histological growth pattern. CRLM grow either in a replacement
pattern (cancer cells at the invasive edge invade the hepatic parenchyma, co-opting hepatic vasculature) or a desmoplastic
pattern (the invasive edge is characterised by a rim of inflammatory tissue consisting of fibroblasts and immune cells). The
Figure 1.
Through multiplexed
immunohistochemistry (GE Cell-Dive) of
CRLM from replacement and desmoplastic
morphologies, biologically distinct niches are
defined. Niche cellular composition is
compared between growth patterns to
identify differences in cellular ecology.
Niches of interest (those with significantly
different cellular composition between
growth patterns) are chosen for spatial
proteomics to interrogate differences in cell-
cell communication networks and deposited
ECMs within the niches. This approach will
provide an enhanced understanding of the
biological differences between the growth
patterns identifying potential therapeutic
targets.
Co-culture experiments combining CRLM

patient-derived organoids (PDOs) of known
histological subtype with primary liver
fibroblasts will enable us to interrogate the
cellular interactions that provide a basis for
the multiplexed IHC and proteomics
observations. Downstream analysis will
include assessment of PDO morphology and
stem cell dynamics through single-cell RNA-
seq and the growth pattern-specific
fibroblast response through analysis of
fibroblast-derived extracellular matrix. Here,
pharmacological methods to switch
plastically between pathological subtype will
be explored.
39
replacement subtype has a significantly worse overall survival when compared with the desmoplastic subtype, but very
little is known about the biological drivers linking histological growth pattern to outcome.
Research objectives
1. Perform mIHC and downstream image analysis on CRLM samples from patients with replacement and
desmoplastic tumours.
2. Perform spatial proteomics on identified niches of interest within CRLM specimens and work with
bioinformaticians within Professor Roman Fischer’s group to analyse the data outputs.
3. Identify replacement-dependent signalling mechanisms by comparing PDO-fibroblast co-cultures to identify

tractable therapeutic targets for interchanging between histological subtypes.
Outcomes
1. Gain a completely novel understanding of the CRLM cellular and proteomic microenvironment and
understand for the first time how this differs between CRLM histological subtypes.
2. Identify mechanisms through which histopathological subtype can be manipulated, providing a potential
patient-tailored therapeutic solution.
1. Patient-derived organoid isolation, organotypic culture, CRISPR-Cas9 gene editing, single cell
sequencing of organoids and extracellular matrix biology (Alex Gordon-Weeks/Simon Buczacki)
2. Mass spectrometry, proteomics, metabolomics, data analytics/integration (Roman Fisher)1,2
3. Multiplexed IHC performance and image analysis (Mark Coles)
40
Targeting innate immunity for intestinal injury recovery - 1,2,3 Dr.
Monica Olcina
Primary Supervisor: Dr. Monica Olcina
Additional Supervisors: Prof. Simon Buczacki
Eligibility: Track1, 2, and 3 students are eligible to apply for this project.
Abstract
Effective radiotherapy requires a suitable therapeutic window allowing appropriate tumour control without excessive
toxicity to healthy tissues. Further improvements in our understanding of the mechanisms underlying treatment-
induced normal tissue injury may lead to the identification of therapeutic targets that reduce normal tissue toxicity
(ideally while simultaneously improving tumour response). We are interested in understanding the role of druggable
innate immunity proteins that, when targeted, simultaneously improve tumour response and reduce normal tissue
toxicity. Our previous work has identified complement receptor, C5aR1 as a key modulator of both tumour response
and radiation-induced bowel toxicity 12. While C5aR1 is well-known for its role in the immune compartment, we find
that C5aR1 is also robustly expressed on malignant epithelial cells, highlighting potential tumour-cell specific functions.
We identify that C5aR1 primarily regulates cell fate in malignant cells, and that C5aR1 targeting results in increased NF-
kB-dependent apoptosis specifically in tumours and not normal tissues. Crucially, targeting C5aR1 improves tumour
response while reducing normal tissue toxicity following irradiation in the abdominal cavity (Figure 1)
Figure 1. Working model for the effects of C5aR1 and C5aR1 blockade on tumour and normal tissue radiation (RT)
responses.
The aim of this project is to investigate the role of C5aR1 in modulating recovery from bowel injury, including the
mechanisms underlying the intriguing opposite effect of C5a/C5aR1 signalling in regulating cell fate in transformed and
untransformed cells. Alongside these mechanistic studies we propose collecting patient samples to identify potential
predictive markers correlating with normal tissue toxicity (and treatment response). There is increasing interest in
investigating complement products as potential biomarkers, especially given that as soluble factors they can be easily
detected from patient plasma. We hypothesize that C5a/C5aR1 signalling is key to recovery from normal tissue injury
as well as tumour response to cytotoxic treatment such as radiotherapy. Therefore, levels of C5a could be used to
predict predisposition to treatment/associated toxicity. Using patient plasma samples, we will explore whether C5a, as
well as a range of other markers, are correlated with acute normal tissue toxicity and tumour response in plasma patient
samples collected in collaboration with Dr Muirhead (Consultant Clinical Oncologist, Oxford University Hospitals).
Research objectives
The overall objectives of this project are to: 1) investigate mechanisms underlying recovery following treatment-induced
intestinal injury; 2) Identify potential plasma markers correlating with susceptibility to treatment-induced injury.
Mechanistically, the student will investigate whether C5a/C5aR1 signalling regulates stem cell
regeneration/differentiation either directly or indirectly through modulation of an inflammatory milieu. Understanding
how C5aR1 signalling regulates intestinal injury will help us understand how to apply future C5a/C5aR1 targeting
41
therapies to reduce treatment-related side effects.
Tools employed in this part of the project will include intestinal organoid culture (expertise available in the Buczacki
lab). The student will also analyse spatial transcriptomics datasets from intestinal crypts of WT and C5aR1-/- mice
(generated by the Olcina lab). Any changes observed will be validated using techniques such as qPCR,
immunohistochemistry and functional assays (including in in vivo models). The student will be based across both groups
to enhance collaborative working.
To identify potential markers of response, we will initially focus on two patient groups: 1) patients receiving stardard of
care chemoradiotherapy for the treatment of primary tumours in the pelvic region. 2) Patients receiving SABR to
different sites for the treatment of primary or oligomestatatic disease (in the context of continued immune checkpoint
treatment). This will allow us to investigate markers of toxicity at a variety of sites and in the context of combined
immune checkpoint and radiotherapy treatments. Additional samples from patients undergoing other treatments could
be considered in the future.
The Olcina lab has expertise in collecting and processing patient plasma samples. The collaboration with the Buczacki
lab will be greatly facilitated by having a joint student pursuing this project. Having a co-supervisor with expertise in
intestinal stem cell biology will be key to the success of this project. The collaboration with Dr Muirhead will be critical
to allow streamline patient recruitment and sample collection.
Translational potential of the project

Widening the therapeutic window of cancer treatments through normal tissue protection would be a critical step in
treatment dose escalation, potentially increasing the chances of tumour control and improved patient quality of life.
For example, diarrhoea induced following irradiation delivered for the treatment of tumours in the pelvic region is a
dose-limiting toxicity for some abdominal tumours. This study will investigate the mechanisms underlying the intriguing
opposite effect of C5a/C5aR1 signalling in regulating cell fate in transformed and untransformed cells. Such mechanistic
detail will be critical for the most effective use of future C5aR1 inhibitors to increase the therapeutic window of
standard-of-care cancer treatments. This work also has the potential to uncover biomarkers of normal tissue injury
which could indicate which patients are at higher risks of toxicity. While the project will build on our previous work
focusing on radiotherapy we will investigate whether these mechanisms also apply in the context of other treatments,
including immunotherapy.
Training Opportunities
By undertaking this project, the student will benefit from working in two laboratories with complementary expertise in
innate immunity and radiobiology (Olcina) and intestinal stem cell biology (Buczacki). Furthermore, the student will
have the opportunity to develop expertise in patient sample handling, collection and processing in addition to a range
of laboratory techniques including how to work with organoid cultures, flow cytometry, immunohistochemistry and
standard cell and molecular biology techniques. The student will also have access to spatial transcriptomics data and
will receive training in how to analyse these data. The student will likely also acquire experience in using colorectal
cancer models with clinically relevant molecular phenotypes while working both independently as well as in a team.
Career development opportunities will be offered, including attendance of national and international conferences and
transferable skills training courses.
The collaboration with Dr Muirhead will provide an opportunity for the student to spend time in a clinical setting with
the wider radiotherapy multidisciplinary team as well as patients to learn about clinical oncology, all facilitating a truly
translational aspect to the project.
References
1.Olcina MM, et al. Targeting C5aR1 Increases the Therapeutic Window of Radiotherapy. bioRxiv (2020)
doi:https://doi.org/10.1101/2020.10.27.358036.
2.Beach C, et al. Innate immune receptor C5aR1 regulates cancer cell fate and can be targeted to improve
radiotherapy in tumours with immunosuppressive microenvironments. bioRxiv (2023)
doi:doi.org/10.1101/2023.01.10.521547.

42
Investigating an interplay between autoimmunity and cancer 1,2,3 –
Prof. Xin Lu
Primary Supervisor: Prof. Xin Lu
Additional Supervisors: Prof. Lynn Dustin
Abstract
Reinvigoration of host immune systems to eliminate tumours is one of the most exciting developments in cancer
therapy. Despite of its potential to achieve durable responses, only small percentage of patients respond to
immunochemotherapy. A key challenge in immune-oncology field is to understand why some patients benefited from
the treatment whereas others failed. Oesophageal cancer is the 6th most common cancer and incidence is rising rapidly
and mortality is close to incidence. A major clinical challenge is to develop novel and effective therapies for oesophageal
cancer. In 2021 FDA approved checkpoint blockade-based immunotherapy to treat inoperable oesophageal
adenocarcinomas in combination with chemotherapy. It is therefore timely to identify molecular mechanisms that
confer tumour sensitivity or resistance to immunochemotherapy.
To address this challenge, we will take the full advantage of our recently reported Lud2015-005 trial of oesophageal
adenocarcinoma patients treated with anti-PDL1 alone for four weeks and followed with immunochemotherapy
(Carroll et al, Cancer Cell, July, 2023). It is increasingly clear that tumour-associated B cells and their antibody products
contribute to improved clinical outcomes and successful immunotherapy in diverse tumours. Antibodies may target
normal self-antigens, tumour neoantigens, oncogenic viral proteins, and even endogenous retroviral gene products
(Ng et al, Nature, April 2023). In many cases, autoantibody levels correlate with clinical benefit in patients treated with
checkpoint blockade-based immunotherapy (CBI) either alone or in combination with chemotherapy or radiotherapy.
Notably, patients in the Lud2015 trial with clinical benefit also tend to produce increased numbers of autoantibodies.
We have cloned hundreds of autoantibodies from a number of Lud2015-005 trial patients with known clinical
outcomes. This project will build on our recently published results and preliminary findings. The objective of this
proposal is to test whether the observed increased autoantibody production, a common feature of autoimmune
diseases, is simply a side effect of increased immune activity, or if the autoantibodies produced by tumour-associated
B cells may contribute to cancer cell killing including antibody dependent cell-mediated cytotoxicity (ADCC). In this
project, we will use a combination of cutting-edge molecular assays, organoids, co-culture techniques and single-cell
sequencing. This work will form an essential component of our overall aim to improve future treatment of
gastrointestinal cancers and inform the broader implementation of immunotherapy.
Research objectives
Background: Reinvigoration of host immune systems to eliminate tumours is one of the most exciting developments in
cancer therapy. Therapies are being developed to inhibit pathways that tumours use to evade immune surveillance.
Antagonists of the CTLA-4 and PD-1/PD-L1 immune checkpoint pathways (i.e. antibodies to CTLA-4, PD-1 or PD-L1)
unleash previously suppressed T-cells to eliminate tumour cells. This strategy - termed immune checkpoint targeting
therapy (ICT) - has achieved durable overall survival in patients with highly metastatic tumours. However, only a subset
of tumours responds to ICT and understanding why many are resistant to ICT and ICT-related combination therapies is
a major scientific challenge. Examining cellular responses before and after interventions is the key to address this
problem. In 2012 it caused ~400,000 deaths worldwide and incidence is rising rapidly. Mortality has remained closely
related to oesophageal cancer incidence, with a 5 year survival of <15%, indicating that many oesophageal cancer cells
are resistant to existing therapies.
Objectives: The overall aim is to identify molecular differences before and after immune checkpoint therapy and
between responders and non-responders. The objective of this proposal is to test whether the observed increased
43
autoantibody production, a common feature of autoimmune diseases, is simply a read out of increased immune
response or the autoantibodies produced by tumour associating B cells may positively contribute to cancer cell killing
including antibody dependent cell-mediated cytotoxicity (ADCC).
Approaches: Key approaches include co-culture techniques, antibody-antigen interaction, organoid technology to
explore interactions between tumour cells and immune cells. Single-cell sequencing may be used to dissect the cellular-
level response to altered interactions. Additionally, autoantibody cloning, antibody-antigen interaction and antibody
dependent cell-mediated cytotoxicity (ADCC) assays will be performed
This project is poised to have major implications for guiding future clinical decision making for patients with
oesophageal cancer. Specifically, the connections made by the student between molecular characteristics and
responses to therapy in the trial of immunotherapy in oesophageal cancer will be vital for developing new clinical
stratification models.
The overall strategy for the project and the laboratory research will be supervised by Prof. Xin Lu, Director of the Ludwig
Institute for Cancer Research in Oxford, who has extensive experience of mentoring clinical and non-clinical DPhil
students. There are opportunities to tailor the exact direction of this project to the interests and background of the
trainee, with a focus on important emerging tools such as organoids, co-culture, single cell sequencing, TAPs
sequencing technology to detect DNA methelome of cfDNA and autoantibody detection in liquid biopsies. Prof Lynn
Dustin also has extensive mentoring experience and will provide expertise in autoimmunity. This project will enable
the student to benefit from expertise and technologies at both the Ludwig Institute for Cancer Research and Kennedy
Institute. The student will have opportunities to integrate with the wider scientific and clinical communities in Oxford
through established collaborative networks, and with the national and international communities at conferences. The
student will benefit from the training and career development programme at the Ludwig, which includes: regular oral
presentations, journal clubs, and skill development in writing, data management and public engagement.

An integrated systems biology approach to investigate the spatial
Myeloma tumour microenvironment 1,2,3,4 – Prof. Udo Oppermann
Primary Supervisor: Prof. Udo Oppermann
Additional Supervisors: Assoc Prof. Adam Cribbs
Abstract
Multiple Myeloma (MM) is a plasma cell derived malignancy with the tumour residing within the bone marrow environment.
Work on solid tumours provides clear evidence that the complex interactions of the cells constituting the tumour
microenvironment play a pivotal role in mediating survival, proliferation, drug resistance, and progression of the cancer by
shaping adaptive and innate immunity 1,2. In regard to MM, the immune responses to therapy have been investigated using
flow cytometry or single-cell NGS based platforms. Here, significant progress has been obtained to understand the immune
composition and activation states of bone marrow aspirate-derived myeloma samples. However, a drawback of this
approach has been the lack of spatial information that can be obtained. In order to fill this evident gap in understanding
myeloma and its microenvironment the DPhil candidate will apply a novel and systematic approach that integrates multiple
layers of information (i.e. targeted proteomics, transcriptomics, genomics) within a spatial context. This would significantly
advance our understanding of myeloma phenotypes and interactions with immune and stromal cells and will help to better
understand responses to therapy. We will use MM trephine biopsy samples obtained from routine clinical diagnostics and
clinical trials and apply ‘omics’ approaches to these tissue sections. Following experimental data collection, the analysis will
include the application of machine learning algorithms and will result in an integrated view of the myeloma tumour and its
microenvironment.
Research objectives
Suggested overall workflow and clinical samples: In the first phase of the DPhil project (over the first 12-18 months) the
candidate will acquire necessary basic skills such as sectioning of formalin-fixed paraffin embedded (FFPE) trephine biopsies,
laser-capture microscopy, proteomics, multiplexed imaging, next-generation sequencing using in-house and collaborator
facilities. Within this period basic statistical and computational skills will be developed (if required) by making use of the
excellent training facilities at Oxford. The remainder of the DPhil will be used to apply these skill sets to longitudinal trephine
samples obtained from clinical trials, such as immune therapies or novel reagents in early phase clinical trials. Biobanked
samples will be obtained through close collaboration with clinical collaborator Ramasamy at OTMC.
In situ analysis of MM trephine biopsies and targeted protein marker detection: To this end we have created and validated
a targeted proteomic-based affinity reagent panel for deployment in cycling immunofluorescence and/or imaging mass
cytometry (IMC) studies. The panel is based on our previous marker selections to characterise the immune environment
(myeloid, lymphoid compartments; immune activation, exhaustion) and in addition allows to identify key myeloma factors
(such as targets of immune and/or IMID therapies); in addition it provides the ability to determine stromal cell types (such
as fibroblasts, endothelial and mesenchymal cells, osteoblasts, adipocytes) in formalin-fixed paraffin embedded (FFPE)
trephine biopsies. Data analysis will be carried out by using software packages such as HALO, successfully used in the group.
The outcome of this workflow is the characterisation of regions of interest (i.e. regions that attract or repulse immune cell
infiltration and activation) in trephine samples; subsequent histological sections of the same sample will then be used to
perform laser caption dissection followed by targeted genome sequencing and proteomic analysis of the region of interest.
Targeted genome sequencing analysis using the recently developed myeloma genome panel developed by co-applicant
Thakurta3 will be applied by the DPhil candidate on DNA from matching FFPE sections using laser capture microdissection
using an extraction-free library preparation technique developed by collaborator Rao 4. Depending on the number of tumour
cells, we will subdivide the tumour into subregions and individually barcode them. We will thus be able to obtain copy
number alteration and single nucleotide variant information from spatially distinct regions of the biopsies, which we will
integrate with proteomic data from adjacent sections.
Mass spectrometry based proteomic analysis of the myeloma tumour and its environment will be in collaboration with
Professors Kessler and Fischer at the Centre for Medicine Discovery (CMD), who have co-developed and applied in situ mass
spectrometry workflows5. They will supervise the student in laser capture microscopy and mass spectrometry techniques as
well as data processing and analysis.
45
Integration and data analysis – we anticipate that by combining the information and multiple datasets with the spatial
distribution of immune and bone cells relative to the tumour cells plus the clinical data will provide significant novel insights
into multiple myeloma pathobiology. Co-supervisor Cribbs has developed workflows for dataset integration and will work
with the DPhil candidate to apply machine learning techniques in order to identify patterns that define tumour –
microenvironment interactions and therapeutic response.
We expect that this project will significantly increase insights from our previous genomic and transcriptomic analyses of
myeloma bone marrow aspirates by applying spatial omics analyses, including genomics and global proteomics. To our
knowledge this is a pioneering attempt in the myeloma field and we foresee that a systems-based approach will enable us
to answer the following research questions: what are the cellular components (immune, stromal cells) that infiltrate or
surround the tumour? What is the activation/exhaustion status of the immune cells that are found at the tumour site? Does
the genomic background play a role in the immune response or the spatial parameters? Can therapeutic responses and
mode of action be gained from spatial information?
The translational value lies in (i) a better understanding of the heterogeneity of MM and how this is related to spatial
interactions and tumour phenotypes (such as proliferation, immune exhaustion, evasion), and (ii) the possibility to identify
therapeutic intervention points that can be utilized to inform on adequate therapies to achieve desired responses, especially
in relapsed/refractory settings. Furthermore, we expect that (iii) LCM of the MM tumour and its immune and stromal
environment will provide essential and novel information about how tumour cells communicate with their non-tumour
environment, which possibly provides novel avenues in immuno-oncology research.
The DPhil candidate will receive training in wet lab activities covering multiple of aspects of modern cell and molecular
biology (next generation sequencing, state of the art proteomics, metabolomics, imaging (cycling immunofluorescence,
imaging mass cytometry) as well as training in computational and statistical methods. The Oxford Translational Myeloma
Centre is an ideal place for interested candidates either with a medical training (with a required interest in haemato-
oncology) wishing to deep-dive into state-of-the-art multiomics and computational techniques or molecular biologists
wishing to expand their skill sets and apply these to myeloma research. Training of next-generation fellows is a key aspiration
of OTMC. Ample opportunity is provided to successful candidates to interact with pharmaceutical industry, due to multiple
existing research projects of the investigators and collaborators of this proposal.
References
1. Binnewies, M., et al. Understanding the tumor immune microenvironment (TIME) for effective therapy. Nat Med 24,
541-550 (2018).
2. Tang, T., et al. Advantages of targeting the tumor immune microenvironment over blocking immune checkpoint in
cancer immunotherapy. Signal Transduct Target Ther 6, 72 (2021).
3. Sudha, P., et al. Myeloma Genome Project Panel is a Comprehensive Targeted Genomics Panel for Molecular Profiling
of Patients with Multiple Myeloma. Clin Cancer Res 28, 2854-2864 (2022).
4. Rao, S., et al. Intra-prostatic tumour evolution, steps in metastatic spread and histogenomic associations revealed by
integration of multi-region whole genome sequencing with histopathological features. BioRXive (2023).
5. Davis, S., Scott, C., Ansorge, O. & Fischer, R. Development of a Sensitive, Scalable Method for Spatial, Cell-Type-
Resolved Proteomics of the Human Brain. J Proteome Res 18, 1787-1795 (2019).

Do mutations in cancer arise through histone post-translational
modifications? 1,2,3,4 – Assoc Prof. Peter Sarkies
Primary Supervisor: Assoc Prof. Peter Sarkies
Additional Supervisors: Assoc Prof. Benjamin Schuster-Böckler
Abstract
Mutations in the sequence of DNA are the fundamental source of cancer, leading to activation of oncogenes and
inactivation of tumour suppressor genes[1]. Almost all cancers demonstrate genomic instability with mutation rates
greatly elevated above healthy cells; indeed, genomic instability is widely accepted as an enabling characteristic supporting
cancer growth by increasing the phenotypic diversity of cells within a cancer and providing ways in which cancers can
become more aggressive by evading the immune system, or initiating metastasis [2]. Over the last 10 years, high
throughput sequencing of primary human cancers has provided a spectacularly detailed characterisation of the mutations
that define cancers and identified a number of mutational “signatures” that together can be used to classify mutational
processes [3]. Some mutational signatures can be assigned to specific cellular processes [4] or exogenous agents such as
tobacco smoke[5]. However, the vast majority are still unexplained [6]. It is therefore key to our understanding of cancer
to develop improved insight into the sources of these mutational signatures. Mutations often arise as a consequence of
damage to DNA. The Sarkies lab discovered that DNA methylation, a key form of epigenetic regulation in human cells, also
has the potential to damage DNA through off-target alkylation of cytosine[7,8]. Left unrepaired, this can lead directly to
mutations[9]. Epigenetic regulation beyond DNA methylation, in particular histone post-translational modifications, takes
place close to the DNA and, like DNA methylation, uses metabolically active co-factors. We therefore hypothesise that
histone post-translational modifications could be a hitherto unknown source of DNA damage and thus account for some
of the unclassified mutational signatures in cancer.
The aim of this joint project is to test this hypothesis by combining expertise from the two groups. The first stage will take
place in the Sarkies laboratory, which specialises in using evolutionary approaches to understand epigenetic regulation.
We will use co-evolution analysis across eukaryotic species as well as co-expression analysis from human cancers and
healthy tissues to identify potential associations between histone post-translational modification rates and specific DNA
repair pathway activity. This will give clues as to how histone post-translational modification could damage DNA. In the
second part of the project, using the expertise of the Schuster-Böckler laboratory (e.g. [10]), the student will test whether
the links between histone post-translational modifications and DNA repair pathways can explain mutational signatures
using state-of-the-art machine learning methods to analyse high-throughput sequencing data. Finally, we will test our
hypotheses in the laboratory by using experimental introduction of histone modifications into specific sites in human Acute
Myeloid Leukaemia (AML) cells to test if this incurs increased mutations in the same region. Together this will enable us
to identify novel sources of endogenous DNA damage due to epigenetic processes. Crucially, this will not only enable us
to better understand how cancer arises, but also to predict which cancers will be particularly vulnerable to specific
chemotherapeutic agents, as cells with high levels of histone post-translational modification induced DNA damage will be
more vulnerable to agents that introduce the same type of damage.
Research objectives
Academic value
Mutational signatures underpin how cancers develop. Understanding where different mutational signatures come from,
however, is much harder than classifying them. Our hypothesis that histone post-translational modifications might be
responsible for some of the many “orphan” signatures therefore has the potential to provide important new insight into
cancer development. Additionally, our research may provide new insights into why epigenetic pathways are so frequently
perturbed in cancer. Together, these will provide new avenues to explore the fundamental basis of cancer development
and progression.
Collaborative value
The Sarkies lab is based in the Biochemistry Department and is focussed on using evolutionary methods to understand
epigenetic regulation. The Schuster-Böckler lab is based in the Ludwig Department of Cancer and focusses on
computational analysis of molecular processes in cancer. This will therefore bridge two different departments and link
two labs with disparate interests and expertise. Sarkies and Schuster-Böckler are two early career group leaders, and have
47
never collaborated together. Supervising a joint student is a direct way to promote collaboration and establish a new and
lasting collaboration between the two research groups.
The key translational outcome is in the arena of better targeting of chemotherapies. Crucially, many chemotherapy
treatments work by introducing DNA damage. If we can better understand sources of DNA damage due to epigenetic
pathways such as histone post-translational modifications it will enable us to predict, from the activity of these pathways,
which cancers will be more vulnerable to particular DNA damaging agents. If we discover that a particular histone post-
translational modification promotes DNA alkylation damage, tumours with high levels of this modification could be
targeted for treatment with alkylating agents. Since measuring histone post-translational modification levels is
straightforward, such a targeted approach would be an attractive and feasible outcome of the research described in this
proposal.
The project will provide training in comparative genomics across species and gene expression network construction and
analysis in cancer. It will provide training in machine-learning methods and modelling approaches for mutational signature
identification. It will also provide training in basic molecular biology techniques, including cell culture, CRISPR and high-
throughput sequencing.
References
1. Hanahan D, Weinberg RA. Hallmarks of cancer: The next generation. Cell. 2011;144: 646–674.
doi:10.1016/j.cell.2011.02.013
2. Hanahan D. Hallmarks of Cancer: New Dimensions. Cancer Discov. 2022;12: 31–46. doi:10.1158/2159-8290.CD-21-
1059
3. Koh G, Degasperi A, Zou X, Momen S, Nik-Zainal S. Mutational signatures: emerging concepts, caveats and clinical
applications. Nat Rev Cancer. 2021;21: 619–637. doi:10.1038/s41568-021-00377-7
4. Alexandrov LB, Jones PH, Wedge DC, Sale JE, Campbell PJ, Nik-Zainal S, et al. Clock-like mutational processes in
human somatic cells. Nat Genet. 2015;47: 1402–1407. doi:10.1038/ng.3441
5. Alexandrov LB, Ju YS, Haase K, Van Loo P, Martincorena I, Nik-Zainal S, et al. Mutational signatures associated with
tobacco smoking in human cancer. Science (80- ). 2016;354: 618–622. doi:10.1126/science.aag0299
6. Alexandrov LB, Kim J, Haradhvala NJ, Huang MN, Tian Ng AW, Wu Y, et al. The repertoire of mutational signatures
in human cancer. Nature. 2020;578: 94–101. doi:10.1038/s41586-020-1943-3
7. Rošić S, Amouroux R, Requena CE, Gomes A, Emperle M, Beltran T, et al. Evolutionary analysis indicates that DNA
alkylation damage is a byproduct of cytosine DNA methyltransferase activity. Nat Genet. 2018;50: 452–459.
doi:10.1038/s41588-018-0061-8
8. Dukatz M, Requena CE, Emperle M, Hajkova P, Sarkies P, Jeltsch A. Mechanistic Insights into Cytosine-N3
Methylation by DNA Methyltransferase DNMT3A. J Mol Biol. 2019;431: 3139–3145. doi:10.1016/j.jmb.2019.06.015
9. Sarkies P. DNA Methyltransferases and DNA Damage. Adv Exp Med Biol. 2022;1389: 349–361. doi:10.1007/978-3-
031-11454-0_14
10. Tomkova M, Tomek J, Kriaucionis S, Schuster-Böckler B. Mutational signature distribution varies with DNA
replication timing and strand asymmetry. Genome Biol. 2018;19. doi:10.1186/s13059-018-1509-y
48
Characterizing the tumour microenvironment of oesophageal cancer
to uncover key factors in response to immunotherapy 1,2,3 – Carol
Leung & Prof. Benoit Van den Eynde
Primary Supervisor: Carol Leung & Prof. Benoit Van den Eynde
Additional Supervisors: Assoc Prof. Eileen Parkes
Abstract
With the advent of immune checkpoint blockade as standard of care in many cancers, such as oesophageal adenocarcinoma
(OAC), there is now a pressing need to identify new combination strategies for the very high proportion of immune checkpoint
blockade-resistant or non-responding diseases. OAC is a solid cancer of unmet needs, bound to affect increasing numbers of
people and overall poorly characterised in terms of tumour microenvironment. Unlike other solid tumours, OAC offers relatively
easy access to the primary tissue, making it an ideal cancer type for microenvironment studies. OAC is characterised by
chromosomal instability, with highly chromosomally unstable (CIN-high) OAC carrying a particularly poor prognosis, associated
with treatment resistance. CIN-high OAC has high expression of genes involved in organisation and remodelling of the
extracellular matrix, which determines the structure of the tumour and induces a distinct fibroblast phenotype within the
microenvironment. How this in turn affects the immune compartment of CIN-high OAC cancers remains to be characterised. In
addition, the link between CIN-status, tumour microenvironment and response to immunotherapy has still to be explored.

Here we will investigate the tumour microenvironment of CIN-high and CIN-low OAC using fresh tissue biopsies and blood
collected from OAC patients before and after treatment with immune checkpoint blockade combinations. The aims are:
• Profile the immune components in the tumour microenvironment of CIN-high and CIN-low OAC. Using multiparameter
flow cytometry, fresh and untreated OAC tissue and blood samples will be processed and analysed, to phenotype the
tumour microenvironment, both in terms of immune cells and fibroblasts. We have previously optimised flow
cytometry panels to be employed in this aim. To complement this, a functional assessment of the myeloid
compartment within the microenvironment will be performed, in order to determine the pro-tumour or anti-tumour
role of such a plastic subset of cells (i.e. their ability to suppress T cell proliferation). This will be the first time the
tumour microenvironment in OAC has been profiled and phenotyped in such depth. Collaborative partners in this aim
include OUH surgeons and gastroenterologists, as well as the tissue banking team.
Outcome: Characterisation of the tumour microenvironment in OAC and identification of distinctive immune
phenotypes in CIN-high vs. CIN-low disease.
• OAC tumour microenvironment 3D modelling. Organoids and matched fibroblasts were previously established from
patient samples (Parkes lab). Building on this, we aim to obtain 3D co-cultures to include the myeloid compartment
previously characterised (in aim 1) and represent a more accurate OAC tumour microenvironment. This will help
dissect the complex crosstalk between the tumour cells, the fibroblasts and the immune cells. Healthy monocytes and
granulocytes will be added to organoid + fibroblast cultures and the following questions will be asked: is the CIN-type
and fibroblast phenotype inducing polarisation of the myeloid cells to become tumour-promoting, i.e. are these cells
immunosuppressive after co-culture? What mechanisms underlie this polarisation event (e.g. cytokines secretion,
contact dependent)?
Outcome: Novel organoid-based co-culture assays to characterise the interaction between myeloid cells, fibroblasts
and OAC tumour cells.
• Interrogate the changes in tumour microenvironment upon treatment with immune checkpoint blockade. Pre-
treatment biopsies as well as matched post-treatment samples will be compared in terms of tumour
microenvironment profile and transcriptomics, in order to assess the changes occurred within the tumour
microenvironment as a result of immune checkpoint-blockade therapy, with particular attention to checkpoint
blockade-resistant OAC. The overarching goal is the identification and validation of tumour microenvironment
pathways to target alongside standard of care as a foundation for combination early phase studies.
49
This proposal will identify and characterise new therapeutic opportunities in the immuno-oncology space by dissecting the
interactions between tumour and its microenvironment immune checkpoint blockade-resistant OAC. Targeting the tumour
microenvironment addresses an urgent barrier to successful immune recognition of cancer cells, and as such supports the future
development of combination immuno-oncology strategies. Moreover, the characterisation of the microenvironment of CIN-high
OAC will inform on approaches potentially relevant to other CIN-high solid cancers.
The successful applicant will be co-supervised by Dr. Carol Leung, Dr. Eileen Parkes and Prof. Benoit Van den Eynde. The
student/clinical researcher will be given training in immunology and molecular biology techniques including flow cytometry, 3D
culturing of organoid models, primary tissue processing, cell sorting, digital pathology analysis as well as bioinformatic training
for RNAseq analysis. The Van den Eynde and Parkes labs are established teams including postdoctoral fellows, DPhil students as
well as technical support. Hands on training and support will be provided for all the techniques outlined. Students will also have
the opportunity to attend bioinformatics training to learn or advance their bioinformatic skills. In addition to the scientific aspect
of the research project, the student/clinical researcher will benefit enormously from the career development programme at the
Ludwig Institute.
Cancer Centre: This project has a distinct focus on immuno-oncology, a key theme in cancer research. This includes novel
collaborations involving two research groups, in the Ludwig Institute for Cancer Research, and the Department of Oncology,
promoting a cancer centre-wide approach. This project focuses on oesophageal adenocarcinoma, a digestive tract cancer of
unmet need, with the potential to translate benefits onto other solid cancers
References
Li, Duran…Parkes, Izar, Bakhoum. Metastasis and immune evasion from extracellular cGAMP hydrolysis. Cancer Discov.
11(5):1212-1227; 2021.
McAuliffe…Leung and Van den Eynde. Heterologous prime-boost vaccination targeting MAGE-type antigens promotes tumor T-
cell infiltration and improves checkpoint blockade therapy. J Immunother Cancer. 9(9): e003218; 2021.
50
Tertiary Lymphoid Structures in Lung Cancer 1,2,3 - Dr. Isabela Pedroza-
Pacheco
Primary Supervisor: Dr. Isabela Pedroza-Pacheco
Additional Supervisors: Assoc Prof. Eileen Parkes
Abstract
Organised collections of immune cells, called Tertiary Lymphoid Structures (TLS) are identified in multiple cancer types. In
lung cancer, B cell-rich TLS near tumours are strong predictors of immune checkpoint blockade response in lung
adenocarcinomas. TLS mimic the organisational and structural features of secondary lymphoid organs. TLS are thought to
act as a local source of immune cell recruitment for rapid immune response and therefore, play an important role in
promoting and maintaining anti-cancer immune responses. Studies of TLS in established cancers have suggested that TLS
can be induced by chemotherapy or vaccination – although why TLS are induced in some patients and not others, or the
consequent role they may play in orchestrating immune responses, remains unclear. Here we use a novel in vivo model in
which we can reliably generate both lung tumours and the tertiary lymphoid structures to interrogate the relationship of
TLS with early tumorigenesis.
Research objectives
• Investigation of the timing of TLS formation alongside lung cancer development

Here, a time course experiment will be performed. We will use our established methods to establish mice treated to develop
cancer and/or organised tertiary lymphoid structures. Left lungs will be profiled using multiplex immunofluorescence, using
our established panel. Spatial relationship of immune cells and TLS to developing tumours as well as the composition of TLS
over time will be analysed using digital pathology with features identified in association with tumour growth/regression.
Remaining right lobes will be digested and profiled using state-of-the-art high-dimensional spectral flow cytometry analysis
for high-throughput protein-based screening of myeloid, T-cells, B-cell/antigen presenting cell subsets, and natural killer
cells. Mediastinal lung-draining lymph nodes from each mouse will be profiled similarly. This 38-multiplex spectral flow
cytometry panel has been developed and validated uniquely in-house by Dr Isabela Pedroza-Pacheco and is a low-cost high-
throughput immunophenotyping approach.
Outcomes: This approach will address the role of TLS in early tumorigenesis and identify factors driving TLS neogenesis that
could be exploited therapeutically. Immunophenotypes and novel TLS signatures associated with effective and sustained
immune responses will be defined.
• Characterisation of tumour-controlling TLS to understand factors driving effective anti-cancer immunity

Mature TLS with a GC are reportedly most effective at promoting tumour immune recognition in established tumours. Our
preliminary data demonstrates that there is marked variation between the number, size, and appearance of TLS associated
with tumour control. TLS characteristics that could promote recognition and elimination of precancerous/cancerous cells
are currently unknown. Therefore, in this work package, spatial transcriptomics will be used to identify distinguishing
features of TLS that develop in the absence/reduction of tumour growth and those which occur alongside tumour growth.
We will use fresh frozen samples from mice at endpoint (~28 weeks post urethane injection) and profiled using the
NanoString GeoMx platform. TLS for profiling will be selected using features of size, organisation, presence of germinal
centres and proximity to tumours for elution and whole transcriptome sequencing. Up to 40 regions of interest across each
of 5 conditions will be profiled.
Outcomes: Defining TLS features that are associated with tumour control, regression or prevention will enable development
of future therapeutic approaches promoting effective and sustained immune responses.
• Define the spatio-temporal distinctions in antigenic landscape driving TLS formation and maintenance
Whilst it is thought that effective antigen presentation and anti-tumour responses are facilitated by TLS, it is not known if
antigen presentation within TLS, or in the presence of TLS, differs from antigen presentation in tumour tissue. Moreover,
51
detailed study of antigens presented during tumorigenesis and TLS formation using longitudinal samples has not previously
been possible. Defining this changing landscape would enable identification of antigens that may be potently and distinctly
capable of driving TLS formation and effective immune responses in early stages of cancer. Mass spectrometry (MS)-based
immunopeptidomics is currently the only method that enables the direct identification of MHC-bound peptide antigens.
Here, we will use a novel in-house established timsTOF workflow for increased sensitivity interrogating the MHC-I and MHC-
II antigen landscape and temporal relationship with cancer in TLS-rich and TLS-low/absent tumour lung tissue.
Outcomes: These data will identify the antigens presented over the course of tumour development as well as in TLS-poor
and TLS-rich lung cancers. Therefore, the distinct immunopeptidome of TLS can be discerned and relationship to effective
anti-cancer immunity pinpointed.
To date, TLS have been studied in established cancers, not in the context of tumorigenesis. Therefore, the factors driving
TLS formation, maturation and maintenance and their role in enhancement of immune response in early tumour
development are not known. This study will use a novel and tractable in vivo modelling approach that facilitates the
development of TLS alongside lung cancer. In-depth characterisation will mechanistically dissect predictive cellular and
molecular factors driving effective anti-tumour immune responses. This will identify TLS-specific factors which enable
tumour control or escape, with future therapeutic relevance.
In vivo modelling, Flow cytometry analysis, multiplex immunofluorescence, digital spatial profiling, bioinformatics
techniques include R for analysis of transcriptomic data, immunepeptidomic analysis.
References
Bradley, T., Peppa, D., Pedroza-Pacheco, I., …Borrow, P. & Haynes, B. F. RAB11FIP5 Expression and Altered Natural Killer
Cell Function Are Associated with Induction of HIV Broadly Neutralizing Antibody Responses. Cell 175, 387-399.e17 (2018).
Moody, M. A., Pedroza-Pacheco, I., … Borrow, P. & Haynes, B. F. Immune perturbations in HIV-1–infected individuals who
make broadly neutralizing antibodies. Sci Immunol 1, aag0851–aag0851 (2016).

Interrogating the fibroblast phenotype of DNA repair deficient cancer
1,2,3 – Assoc Prof. Eileen Parkes
Primary Supervisor: Assoc Prof. Eileen Parkes

Additional Supervisors: Prof. Kim Midwood
Abstract
Understanding the tumour microenvironment of BRCA1-mutant breast cancer is key to developing new therapeutic
strategies to improve durable responses to treatment. Cancer-associated fibroblasts (CAFs) are not merely by- standers
in the tumour microenvironment, but crucial in treatment response and metastasis. Using models of this disease subtype
developed in our lab we have identified a unique CAF phenotype potentially associated with inhibitor resistance. In this
proposal we will investigate how CAF phenotype may impact response to PARP inhibition, and further investigate the
CAF phenotype in BRCA1-mutant breast cancer.
CAFs are the dominant component of the stromal tumour microenvironment and contribute to cancer progression as
well as therapeutic response. CAFs contribute to an immune-excluded and immunosuppressed tumour
microenvironment as well as driving pro-tumorigenic macrophage polarisation. Recently it has been reported that
tumorigenesis in BRCA1-mutant cancer is partly driven by a distinct fibroblast population. Despite this crucial role, the
phenotype of CAFs in BRCA-deficient breast cancer and their contribution to therapeutic response/resistance,
particularly to PARP inhibition, is poorly understood.
Research objectives
• Investigation of response to PARP inhibitor treatment and impact of fibroblast phenotype in BRCA1-mutant
and wildtype in vivo models
Our preliminary data demonstrates the distinct CAF phenotype in Brca1-deficient treatment-naïve tumours. This aim will
further characterise the impact of PARP inhibition on fibroblast phenotype using FACS immune profiling. Using the 4T1
Brca1 isogenic models, we will perform in vivo treatment with the PARP inhibitor niraparib. Tumours will be resected
and half the tumour dissociated and digested for subsequent FACS analysis, using lymphoid (CD3, CD4, CD8, CD19, NKp46
and CD279), myeloid (CD80, MHS II, Ly-6G, Ly-6C, CD11b, CD11c, CD206, F4/80) and fibroblast (panCK, CD31, PDGFRα,
FSP1, SMA, Podoplanin, MHCII) antibody panels.
Outcomes: These experimental approaches will further address the distinct CAF phenotype in the 4T1 shBrca1 model as
well as identifying the impact of PARP inhibition on CAFs in the tumour microenvironment, an important unanswered
question.
• Characterisation of fibroblast phenotype in human breast cancer using spatial transcriptomics

While our preliminary data strongly supports a distinct CAF phenotype in the murine models we have studied, we do not
yet know the translational relevance of this to human breast cancer. In collaboration with the Breast Cancer Now biobank
we have received preliminary approval (see letter of support) to obtain formalin-fixed paraffin- embedded samples of
triple negative breast cancer from patients with BRCA1 mutation (n = 4) and from BRCA1 wildtype tumours (n = 4).
Tumours will be matched for size (T) and nodal status (N) stage as well as histological subtype (ductal cancer) to minimise
confounding variables. Whole face sections (a single 4 µm section from each tumour) will be stained with DAPI, CD45+
(immune cells), SMA (CAFs) and panCK (tumour cells). Regions of interest will be selected using the GeoMx platform.
Downstream Illumina NGS capturing whole transcriptome will then be performed on distinct regions of interest, with in
depth bioinformatic analysis performed.
Outcomes: Identifying the ability of murine models to model CAFs in BRCA-mutant breast cancer will address the
translatability of the findings of this study – whether in vivo models are sufficient to investigate potential combination
strategies or alternative approaches (e.g. PDX, patient-derived organoids and co-culture) are needed to accurately model
and optimise clinical impact.
53
• Modelling fibroblast phenotype to investigate how fibroblasts impact tumour growth, therapeutic response
and metastasis in breast cancer
From our preliminary data analysis (RNAseq of fibroblasts from 4T1 Brca1 isogenic tumours) and further data obtained
in aim (2), consensus hits between mouse and human tumours will be identified and ranked. Top hits identified will be
genetically manipulated on GFP-labelled 3T3 Balb/c fibroblasts (in house) using lentivirus CRISPR Cas9 (LentiCRISPR v2,
in house). We will use conditioned media from 4T1 and 4T1 shBrca1 cells in vitro and perform 10 – 14 days co- culture
with fibroblasts to determine the direct impact of chemokines and conditioned media on fibroblast plasticity.
Conditioned media from genetically manipulated fibroblasts will additionally be co-cultured with bone marrow derived
monocytes, derived from bone marrow flushes of humanely killed Balb/c mice, to determine impact of fibroblast
phenotype on myeloid cell differentiation and phenotype, analysed using FACS.
Outcomes: Together, these data will identify the role of distinct fibroblast population in tumour growth and spread, and
potential therapeutic opportunities to directly target fibroblasts to improve patient outcomes in BRCA-mutant breast
cancer.
The direct comparison of hits from an in vivo model and human breast cancer samples will be impactful in determining
how efficiently mouse models can be used to develop therapeutic strategies around CAFs and what limitations there
might be of this approach. Overall, it is intended that this work will lead to discovery of potential novel CAF-related
targets that could be further validated for clinical use, and accelerate translation to the clinical setting by identifying
potential CAF phenotypes that predict response/resistance to PARP inhibition and thus patient selection for alternative
or combination strategies.
In vivo modelling, Flow cytometry analysis, multiplex immunofluorescence, digital spatial profiling, CRISPR/Cas9, cell
culture, bioinformatics techniques include R for analysis of transcriptomic data.
References
van Vugt MATM, Parkes EE. When breaks get hot: inflammatory signaling in BRCA1/2-mutant
cancers. Trends Cancer. 2022;8(3):174-189. doi:10.1016/j.trecan.2021.12.003

Exploring the biological drivers of 1q21+ high risk Myeloma by using
multi-OMICS analysis of patient derived tumour, immune cells and bone
marrow aspirates1,2,3,4 – Prof. Anjan Thakurta
Primary Supervisor: Prof. Anjan Thakurta
Additional Supervisors: Assoc Prof. Karthik Ramasamy
Abstract
Multiple Myeloma is an incurable malignancy of the clonally expanded plasma cells and is genomically heterogeneous. Up to
40% of newly diagnosed (NDMM) and 70% of relapsed and refractory Multiple Myeloma (RRMM) patients carry gain or
amplification in chromosome 1q21 locus. Copy number gain (3 copies) or amplification (4+ copies) together designated as
1q21+ are poor prognostic biomarkers in NDMM and feature prominently in driving disease progression in the relapse
settings. Various groups have previously analysed large datasets to define the prognosis of 1q21 MM. Notable among these,
the analysis from the Myeloma Genome Project (MGP), co-led by the applicant, identified amplification of 1q21 in the
background of ISS3 (international staging system) as a ‘double hit’ MM with very poor clinical outcome.1 Another analysis of
1905 patients from the Myeloma UK group interrogated a combined dataset from Myeloma X and XI trials and confirmed the
poor prognosis of 1q21. This analysis also demonstrated co-occurrence of genomic features to add adverse prognosis of this
group of MM patients.2 Recently, by paired tumour sample analysis, we have found genetic lesions of 1p32LOH (loss of
heterozygosity) with 1q21 gain co-evolve in driving clinical resistance in Myeloma (Ansari-Pour, Thakurta, unpublished results).
There are many postulated biological drivers and therapeutic targets of MM within the 1q21 locus that are being analysed
and pursued. Previously, by unsupervised analysis of the genome and transcriptome of 514 NDMM patients, we identified a
biological subset of patients (designated as MDMS8) with poor prognosis3. While diverse high-risk genomic features (including
~ a third with 1q21 gain) were identified, we found a common transcriptional program among the patients of this subset. The
tumour cells in this group showed significant loss (~ 10%) of genome, damaged DNA repair pathways, dysregulated cell cycle
as key genomic and transcriptomic features of high-risk disease. Master regulators were identified as key drivers of the
transcription program. However, due to lack of availability of samples, a detailed proteomic analysis was not performed to
interrogate the functional and metabolic drivers of high-risk MM.
Analysis of Myeloma proteome (and metabolome) is thus far limited but some published proteomics literature suggests its
feasibility and the potential utility. For example, by analysing 5 paired patient samples Ng et al identified CDK6 as a driver of
resistance to immunomodulatory drugs4. Proteomic analysis of bortezomib sensitive and resistant patient samples similarly
led to the discovery of a potential biomarker of resistance5. Another group explored comparative proteomic analysis of bone
marrow aspirates and serum to identify new biomarkers for MM5. Metabolic analysis of MM has not been published yet,
offering a new opportunity to initiate initial exploration in the study of a high-risk disease subset such as 1q21+ MM. Our
hypothesis is that a systematic proteomic analysis of 1q21+ MM cell lines and patient samples may add to the overall
understanding of key proteins and biological processes or their regulation that functionally drive the disease. In a separate
project, we are exploring genomic drivers of 1q21+ MM for identifying therapeutic targets. Here, the proposed project takes
a complementary approach and explores a proteomics-based analysis of patient tumour cells, immune cells and the bone
marrow aspirates to establish proteomic and metabolomic profiles of 1q21 MM and validate key functional drivers of high-
risk disease.
Research objectives
The project will be divided in terms of experimental work done at Botnar Lab (in collaboration with Prof Oppermann) and
Centre for Medicines Discovery, Oxford (with Prof Benedikt Kessler’s group), at Metabolon (in collaboration with Dr
Sarangarajan) and computational analysis will be done within OTMC (with support from a computational science team). We
will access cell lines (Sarah Gooding, Oxford Oppermann, Oxford, commercial sources), and patient bone marrow aspirates
(Oxford Biobank supervised by Prof Ramasamy) for 1q21+ and 1q21- (standard risk patients) and separate CD138+ and CD138-
fractions and use them for proteomic/metabolomic analysis. In agreement with the collaborators, some data or materials
from a companion project could be shared for this project work (Dr Erin Flynt). This project will be run in three phases: Phase
I: In the initial phase, pilot experiments will be set up to study Myeloma cell lines representing normal, 1q21 gain and 1q21
amplified myeloma cells. Alongside, proteomics analysis of immune cell types (CD4, CD8 cells and NK cells) and serum will be
performed from peripheral blood from healthy donors.
55
Total proteome, phospho proteome and metabolic profiles (in collaboration with Metabolon) will be generated from the cell
lines to develop the methods for experimental pre-analytics and bioinformatic analytics. A datahub in OTMC will be set up for
storage and analysis of data. Various experimental conditions will be tested to mimic real life conditions of patient derived
bone marrow aspirates. CytoF based analysis will be conducted for comparing some markers from the immune subsets and
Myeloma cell lines using panels already in use in the Botnar lab. Analysis of the cell line data will be used as a benchmark
when analysis of patient samples is tested. Once the pilot experiments are completed, based on the analysis, a subset of cell
lines will be chosen to perform drug treatments to understand the impact on the proteome and metabolome. Standard of
care therapeutic agents such as steroids, immunomodulatory drugs, proteasome inhibitors and anti-CD38 antibodies will be
used alone and in clinically used combinations. The final outputs of this work will be the identification of protein(s) affected
in 1q21+ Myeloma and/or via drug treatment with potential biological or clinical impact. Phase 2: This phase is based on the
analysis of 30-40 patient samples in each control or 1q21 + group and further analysis of preclinical data and correlative
analysis of potential biomarkers. The key types of questions we will pursue are: What are the specific proteomic/phospho-
proteomic and metabolic profiles of 1q21+ Myeloma in CD138+ cells, immune subsets or in bone marrow aspirates compared
to normal counterparts? What are the biological drivers of this segment of vs those without (in previously identified genomic
high risk MDMS8 segment with 1q+ vs 1q- counterpart). Biobanked and genomically characterized patent samples will be used
for this analysis. Phase 3: Validation of key proteins or metabolites identified from preclinical and clinical samples will be the
focus in this phase. A subset of the most promising candidates from CD138+, bone marrow aspirate or immune cells will be
selected. For preclinical validation, cell lines will be modified to create knock downs or CRISPR mediated knock outs, for
expression of specific mutants, based on targeted genes and the phenotypic and biochemical effects will be measured. Target
feasibility/ targetability analysis is outside the scope of this project.
This project is forward and reverse translational in nature whereby the student will explore preclinical data analysis coupled
with retrospective analysis of patient data to derive insights for the development of biomarkers or therapeutic targets relevant
for the 1q21+ high-risk Myeloma. The project is planned to run side by side the genomics oriented collaborative project to
cross fertilize both projects. The role of 1q21 in relapse in Myeloma is well documented and there is also a clear need to find
actionable insights for detection and characterisation of 1q associated biomarkers. There is clear potential to develop a
personalised treatment approach to test translationally driven clinical concepts within the OTMC Myeloma Master protocol.
This project will bring in an approach so far not addressed by genomics and other approaches.
The student will learn basic molecular techniques, cell biology methods, some OMICS techniques, data generation and
computational analysis of large datasets from multiple sources. The student will also learn the relevant methods of diagnosis
and treatment of myeloma and other related OTMC’s projects (such as single cell analysis, long read sequencing etc.)
References:
1. Walker BA, et al. A high-risk, Double-Hit, group of newly diagnosed myeloma identified by genomic analysis. Leukemia. 2019.
33(1):159-170.
2. Shah V, et al. Prediction of outcome in newly diagnosed myeloma: a meta-analysis of the molecular profiles of 1905 trial patients.
Leukemia. 2018;32(1):102-110
3. Ortiz Estevez, M., et al. Integrative multi omics identifies high-risk multiple myeloma subgroup associated with significant DNA
loss and dysregulated DNA repair and cell cycle pathways. BMC Medical Genomics; 2021; 14:295-302
4. Chen, Y., et al. 2021; J. Proteome Res. 2021, 20, 5, 2673–268.
5. Ng, Y., et al. Nature Comm. 2022; 13:1009 https://doi.org/10.1038/s41467-022-28515-1
6. Chanukuppa, V., et al. A Comprehensive Study Using Bone Marrow Interstitial Fluid and Serum Samples, 2021 Frontiers in
Oncology, vol 10 doi: 10.3389/fonc.2020.566804

Development of Spatially Fractionated Radiotherapy Techniques 1 –
Assoc Prof. Geoff Higgins
Primary Supervisor: Prof. Geoff Higgins
Additional Supervisors: Kristoffer Petersson
Eligibility: Track 1 students are eligible to apply for this project.
Abstract
Conventional fractionated radiotherapy involves the homogenous delivery of radiation across the entire tumour in 2 Gy
fractions. Recently, there has been significant interest in delivering non-uniform “spatially fractionated” radiotherapy
(SFRT) particularly for bulky, locally advanced tumours that usually do not respond well to conventional radiotherapy,
without increasing the side-effects of treatment. SFRT typically involves delivering extremely high doses of radiotherapy
(up to 20 Gy) to multiple, small spherical subregions within the tumour, whilst ensuring that the dose to the periphery of
the tumour, and to the surrounding organs at risk are within standard limits. It has been postulated that the high tumour
control rates that have been reported with SFRT might be due to anti-tumour immune responses, triggered bystander
effects, or changes in the tumour vasculature caused by the peak radiation doses.
The advent of Magnetic Resonance Linear Accelerators (MRL) has increased the ability to accurately deliver radiotherapy
to moving tumours and facilitate real time treatment adaptation. This proposal seeks to implement SFRT using a MRL
machine. In addition, we will seek to further develop the capabilities of MRL technology by assessing whether low field
strength MR imaging is able to effectively identify regions of tumour hypoxia, using a relatively new technique called
oxygen-enhanced MR imaging.
Research objectives
The Oxford Institute for Radiation Oncology has a well-established partnership with GenesisCare that provides access to
a clinical Viewray MRL which combines a 0.35 Tesla MR scanner with a 6MV photon linear accelerator and offers a
potentially ideal opportunity to accurately implement SFRT. This project will seek to assess the capability of delivering
SFRT to bulky tumours using the Viewray MRL (Fig 1). In addition, we will investigate the functional imaging capability of
the MRL in detecting hypoxic regions within the tumour using oxygen enhanced MR imaging (OE-MRI). This technique
has been developed as a non-invasive approach to measure the dynamics of tumour oxygenation. Previous attempts at
boosting the dose of radiotherapy to hypoxic regions have not been able to obtain real time assessment of tumour
hypoxia, and have only been able to deliver very modest increases in radiotherapy dose. The potential to utilise modern
MRL delivery techniques to cause a far greater increase in radiation doses might represent a step change in our ability to
overcome hypoxia mediated radio resistance.
The first part of the project will involve radiotherapy planning studies to develop the optimal methodology to deliver MRL
based SFRT and will include analysis of the position, size and dose of high dose “spheres”.
Subsequently, we will open a clinical study combining SFRT and OE-MRI in patients with locally advanced disease that
cannot be treated with radical radiotherapy. This would be expected to include patients with large volume non-small cell
lung cancer who are currently treated with high-dose palliative radiotherapy (39 Gy in 13 fractions). This study will also
incorporate exploratory imaging biomarkers of early response to treatment, in addition to circulating markers of
inflammation, immune activation and tumour response.
This project will centre on an academic collaboration between experts in radiation oncology (Geoff Higgins) and medical
physics (Kristoffer Petersson) and will involve detailed radiotherapy planning studies and clinical implementation.
57
Fig. 1. Illustration of a SFRT
treatment plan. Clockwise
from top left; axial, model,
coronal and sagittal views
of colour wash dose
distribution of 3 Gy around
the tumour periphery and
18 Gy in the three 1.5 cm
diameter dose peak
spheres within the
tumour. Adapted from (3).
The development of MRL machines has advanced the capability to deliver radiotherapy in ways that have not previously
been possible. SFRT has the potential to improve radiotherapy outcomes for patients with large tumours that are not
amenable to conventional radical radiotherapy and is particularly suitable for MRL based delivery. Combining SFRT with
an OE-MRI approach to target hypoxic regions might result in improvements for cancer patients treated with
radiotherapy.
This project is suitable for a clinical oncology trainee interested in pursuing a career as a clinical academic. In addition to
the successful candidate gaining expertise in MR guided radiation therapy delivery, they will be also be taught advanced
radiotherapy planning methodology, functional MR image analysis and clinical trial experience.
References
1) Yan et al. Spatially fractionated radiation therapy: History, present and the future. CTRO 20: 30-38 (2020).
2) Wi et al. The Technical and Clinical Implementation of LATTICE Radiation Therapy (LRT). Radiat Res 194(6): 737-746
(2020).
3) Amendola et al. Improved outcome of treating locally advanced lung cancer with the use of Lattice Radiotherapy. CTRO
9: 68-71 (2018).
4) Salem et al. Oxygen-enhanced MRI Is Feasible, Repeatable, and Detects Radiotherapy-induced Change in Hypoxia in
Xenograft Models and in Patients with Non-small Cell Lung Cancer. Clin Can Res 25(13): 3818-3829 (2019).
5) O’Connor et al. Oxygen-Enhanced MRI Accurately Identifies, Quantifies, and Maps Tumor Hypoxia in Preclinical Cancer
Models. Cancer Res. 76(4): 787-95.

58
Investigating the role of hypoxia for the FLASH effect by combining
FLASH Radiation with hypoxia-modulated anticancer drugs,1,3,3,4 –
Kristoffer Petersson
Primary Supervisor: Kristoffer Petersson
Additional Supervisors: Assoc. Prof Geoff Higgins
Abstract
FLASH radiotherapy is a novel radiotherapy delivery methodology using ultra-high dose rates (>30-40 Gy/s), ~1000 times
higher than what is used clinically1. Recent global preclinical research has indicated that FLASH can reduce radiation-induced
damage in healthy tissues with similar anti-tumour effects as conventional radiotherapy, the so-called “FLASH effect”.
Though FLASH clinical trials have started, we do not yet understand the radiobiological mechanism(s) responsible for the
FLASH effect. One hypothesis is that FLASH radiation consumes oxygen at a much higher rate than can be re-supplied,
causing hypoxia and thereby increased radio-resistance in cells/tissue2. Why this is specifically sparing normal tissue has
been debated, with one idea being that the tumour cells that are hard to kill with radiation are often already hypoxic, i.e.
little change in response to FLASH vs conventional irradiation1.To test this hypothesis and at the same time test the response
of FLASH radiation as a combination treatment with clinical anticancer drugs, we will combine FLASH radiation with
Tirapazamine (activated in hypoxia) and Atovaquone (increases tumour oxygenation) in 2D (Tirapazamine, hypoxia and
normoxia) and 3D (Spheroids, Tirapazamine and Atovaquone) cell cultures 3. If the FLASH effect is caused by hypoxia, the
FLASH radiation + Tirapazamine should be more toxic than conventional radiation + Tirapazamine. Similarly, the FLASH
sparing effect should be reduced when combined with Atovaquone, resulting in similar response as conventional radiation
+ Atovaquone. To summarise, the response in tumour cells should increase when FLASH is combined with these drugs.
However, for a therapeutic benefit it will be important to understand how normal tissue/cells respond to the treatment. For
that purpose, we will also expose 3D cell cultures (organoids) to the treatment and evaluate the radiation-induced toxicity.
The FLASH-drug combinations that show therapeutic benefit will subsequently be evaluated in mice to further investigate
its potential for clinical translation.
Research Objectives
The research project aims to elucidate if the oxygen consumption hypothesis is the main contributor of the FLASH sparing
effect seen for normal tissue. By using 2D and 3D cell cultures with controlled levels of oxygen and comparing the response
to FLASH vs conventional irradiation with and without drugs that are known to modulate oxygen levels or that activates in
hypoxia, we will be able to quantify the importance of oxygen concentration for the FLASH effect. In addition, the FLASH-
drug combination that show the highest therapeutic potential will be evaluated in vivo (mice). This will give an indication of
the potential of this combined therapy for clinical translation
FLASH radiotherapy has received a lot of research focus recently due to its ability to spare normal tissue, while keeping the
tumour response comparable to conventional radiotherapy. Without knowing much about the radiobiological mechanisms
responsible for these beneficial characteristics of FLASH, clinical trials have already started. Likely, such studies are not
designed in an optimal way as we do not yet know when we see a FLASH effect, how large it is, and why we see it. Hence,
we need more preclinical data to support the design of clinical trials and to guide us in the clinical translation of this
promising treatment technique. This project will tell us about the role of hypoxia for the FLASH effect and could potentially
serve to answer all the questions stated above. In addition, our FLASH-drug combination treatment could help to further
widen the therapeutic window beyond what FLASH alone is able to achieve, i.e. further enhancing the biological benefit of
the FLASH effect.
The student will be trained in various in vitro assays, such as Western Blot, DDR analysis, and clonogenic assay. They will
also be trained and work with more advanced in vitro models, e.g. spheroids, organoids and 3D bio-printed tissue models.
In addition, the student will be trained on the use of in vivo (mouse) models, evaluating normal tissue toxicity and tumour
response to treatment. Correct dosimetry is essential in order to get useful and accurate preclinical results from our studies.
Dosimetry also becomes more challenging at ultra-high dose rate. Hence, there will be some training opportunities in
radiation dosimetry.
59
References:
1. J.D. Wilson, E.M. Hammond, G.S. Higgins, K. Petersson. “Ultra-High Dose Rate (FLASH) Radiotherapy: Silver Bullet or
Fool's Gold?” Front. Oncol. 9, 1563 (2020). 10.3389/fonc.2019.01563
2. E.J. Moon*, K. Petersson* & M.M. Olcina*. “The importance of hypoxia in radiotherapy for the immune response,
metastatic potential and FLASH-RT”, International Journal of Radiation Biology, 98:3, 439-451(2022) *Equal
contribution.10.1080/09553002.2021.1988178
3. G. Adrian, JL. Ruan, S. Paillas, C.R. Cooper, K. Petersson. (2022). “In vitro assays for investigating the FLASH effect”.
Expert Reviews in Molecular Medicine 24, e10, 1–12 (2022). https://doi.org/10.1017/erm.2022.5

60
Using Long-read Sequencing to Advance Personalised Decision Making
in Multiple Myeloma 1,2,3,4 – Assoc Prof. Adam Cribbs
Primary Supervisor: Assoc Prof. Adam Cribbs
Additional Supervisors: Sarah Gooding
Abstract
Work within the Cribbs lab focuses on developing novel technology and computational analysis frameworks that empower new
modes of treatment for disease. Recently we developed several single-cell and bulk long-read (LR) sequencing approaches to
improve their utility and accuracy1, 2. This technology allows for the measurement of translocations, variant calling, and alternative
splicing in unprecedented detail. We have begun to apply this technology to understand the development of drug resistance in
Multiple Myeloma (MM).
We propose that leveraging LR sequencing's multi-modal data can unlock a deeper grasp of functional genomics, refining our
diagnosis and prediction of MM with greater precision. Thus, we will employ state-of-the-art machine learning methodologies to
identify and integrate the modalities which best predict outcomes. By doing so, we aspire to unearth not only novel diagnostic
methods but also discover potential therapeutic targets, thus propelling us further towards our goal of personalised MM patient
care.
LR technology
Short-read (SR) sequencing methods face challenges in
fully elucidating the intricacies of both DNA and RNA
rearrangements. Owing to their capacity to cover
thousands of base pairs, long-reads possess the ability to
detect mutations in regions where short-reads falter.
They excel in identifying complex chromosomal
rearrangements, especially those involving numerous
chromosomes and heavily repetitive regions3.
At the RNA level, the challenge lies in accurately

documenting alternative splicing events, which can
generate a multitude of substrate combinations from the
same gene, resulting in thousands of isoforms. Isoform
shifts can change drug responses, as we have shown in
the response to IMiD drugs in myeloma4. Existing SR-
sequencing methods are inefficient at capturing this
complexity, necessitating the use of LR-sequencing
(Fig.1). LR enables end-to-end mRNA sequencing,
effectively circumventing these limitations.
Finally, LR sequencing provides the added advantage of

simultaneously detecting DNA modifications, such as Figure 1: A comparison of the advantages and
methylation, in a single experiment, thereby enhancing disadvantages of long-read technologies
our understanding of genetic information at multiple

levels.
Research objectives
Our ultimate aim is to apply our long-read RNA and DNA technology to primary MM patient samples and then generate
computational models that help us to understand the molecular mechanisms of drug resistance in patients with relapsed MM.
61
Overall aim: Use LR sequencing data to discover new myeloma characteristics and potential treatment targets.
With the goal of uncovering unique insights into myeloma biology and pinpointing potential therapeutic targets, we will leverage
LR sequencing data to understanding the intricacies of the disease, illuminating novel hallmarks, and thus empowering us to
target myeloma more effectively and precisely.
Hypothesis: LR sequencing will give unprecedented detail on the splicesome of MM, largely unstudied, and that can lead to novel
drug targeting
We will sequence 600 RNA samples isolated from several ongoing clinical trial samples with external collaborators at UCL, London,
at baseline and following relapse. We have already generated pilot data on RNA and DNA isolated from 21 MM patients and
identified several novel translocations not identified previously by SR-RNA sequencing (Fig2). We have a well-established wet-
lab workflow with which we will generate data on 600 MM patients. It is anticipated that over half of the samples will have been
sequenced by the start of the DPhil.
Outputs
- Apply workflow to 400 newly diagnosed and 200 follow-up myeloma patients in a clinical trial.
- Assess the role of alternative splicing in plasma cells in causing disease variation and treatment response.
- Gain new insights into disease mechanisms and treatment responses.
- Potentially influence changes in therapeutic intervention.
Work package 1: What do structural changes of MM look like with LR RNA sequencing?
Analysis of LR-RNAseq: The student will adapt and further develop our established computational analysis workflows
(https://github.com/cribbslab/TallyTriN) to facilitate efficient
identification of genomic alterations and changes in the
transcriptome landscape.
We will use workflows described above to analyse our data

focusing on:
• Evaluating changes in Isoform expression before and after
relapse.
• Measuring structural changes, mutations and SNPs.
• Exploring relationships in gene expression with copy
number variation (CNV)
Work package 2: Developing a splicosome map of MM?

Using the data collected from LR-RNA-Seq, we will develop a
comprehensive map of the spliceosome's interactions and RNA
modifications across different MM stages. This could provide
insights into the heterogeneity and evolution of the disease.
• This analysis will be done in a large phase III MM clinical trial

recruiting 1500 patients across the UK. We will focus on
early progressors, representing the most unmet medical
need.
• Identify potential splicing-driven resistance mechanisms,
for example in relation to IMiD exposure.
Translational potential Figure 2: We have identified several novel translocations in RNA

isolated from patients with MM.
The stated aim of this project is to study the molecular mechanisms of Multiple Myeloma drug resistance and evaluate drug targets
for therapy. By its very definition, this is likely to identify novel therapeutic intervention points within the development of Multiple
Myeloma. We have extensive collaborations with several pharmaceutical partners, and we will utilise these interactions to explore
the translational potential of targets.
62
The student will receive training in the necessary cellular, molecular, and epigenetic biology for this project. This will involve wet-
lab workflows for generating LR sequencing data. Extensive training in computational biology will be provided so that the student
can analyse their own data. Specifically, this will include software development, data analytics, statistics and computational
pipeline development. Outside the lab, the student will be expected to attend regular seminars with high profile external
speakers, journal clubs and training in presentation skills, scientific writing, and data management.
References
1. Sun, J. et al. Correcting PCR amplification errors in unique molecular identifiers to generate absolute numbers of
sequencing molecules. bioRxiv, 2023.2004.2006.535911 (2023).
2. Philpott, M. et al. Nanopore sequencing of single-cell transcriptomes with scCOLOR-seq. Nat Biotechnol (2021).
3. Rausch, T. et al. Long-read sequencing of diagnosis and post-therapy medulloblastoma reveals complex rearrangement
patterns and epigenetic signatures. Cell Genom 3, 100281 (2023).
4. Gooding, S. et al. Multiple cereblon genetic changes are associated with acquired resistance to lenalidomide or
pomalidomide in multiple myeloma. Blood 137, 232-237 (2021).

The role of oxygen chemosensitivity in tumourgenesis 1,2,3,4 – Prof.
Richard White
Primary Supervisor: Prof. Richard White
Additional Supervisors: Professor Sir Peter Ratcliffe and Assoc Prof. Tammie Bishop
Abstract
Cellular hypoxia impinges on the biology of cancer in multiple ways. One of the least well researched is the role of oxygen
chemosensitivity, a classical response to hypoxia that involves depolarization and calcium entry in excitable cells and which
is centrally involved in cardiopulmonary and respiratory control 1. Despite decades of research, understanding of the mode
of hypoxic signal generation and transduction is still incompletely understood. Nevertheless, several new lines of evidence
point to a role of oxygen chemosensitivity in specific types of cancer:
• activation of a specific component of the transcriptional response to hypoxia mediated by hypoxia inducible factor
2 (HIF-2) leads to acquisition of oxygen chemosensitivity and a phenotype resembling a rare form of cancer
(paragangliomas) in certain tissues
• in Type I cells of the carotid body (a common site of paragangliomas) activation of oxygen chemosensitivity leads
to cell proliferation2.
• small cell lung cancer (SCLC) appears to evolve from neuroepithelial cells of the lung, a relatively uncommon lung
cell type that displays oxygen chemosensitivity3
Despite these advances, our understanding remains limited, in part because of the difficulty of studying mammalian systems
that display oxygen chemosensitivity. This project capitalizes on evidence that the basic oxygen chemosensitivity pathway
is conserved in the zebrafish (Danio rerio)4. This is a model organism optimized for high-resolution imaging coupled with
ease of genetic manipulation, and has been increasingly used as a model for cancer. We will exploit the imaging potential of
zebrafish to enable rapid molecular genetic analysis of the mechanisms underlying oxygen chemosensitivity and their
interface with cell growth and proliferation.
Research objectives
The past several decades have witnessed the elucidation of several key mechanisms by which cells sense oxygen levels.
Central to this are HIF-1 and HIF-2, which are hydroxylated at proline residues by prolyl-hydroxylases, allowing their
recognition and ubiquitination by the VHL E3 ubiquitin ligase5,6. While this response allows for chronic adaptation to hypoxia,
organisms also exhibit a rapid acute response to hypoxia. This is a classical response to hypoxia that involves depolarization
and calcium entry in certain excitable cells (e.g. those in the carotid body), which results in a rapid integrated
cardiopulmonary response to restore oxygen levels. The sensors underlying this acute chemosensitivity to oxygen remain
poorly understood.
Not all cells exhibit acute oxygen chemosensitivity, which may be due to the intrinsic cellular context or due to interactions
with cells in the local microenvironment. Recent work from the Bishop and Ratcliffe labs have demonstrated that activation
of HIF-2 (encoding EPAS1) leads to acquisition of oxygen chemosensitivity in otherwise insensitive cells such as those in
the adrenal medulla, suggesting that this factor is sufficient to endow cells with this phenotype. Interestingly, activation of
HIF-2 is also tightly linked to the development of certain types of cancer known as paragangliomas, a rare tumour that
often harbours mutations in numerous genes affecting the oxygen sensing pathway, including gain of function mutations in
HIF-2 itself, as well as loss-of-function mutations in the main negative regulators of HIF: VHL, PHD2 and PHD1 7. A high
proportion of these mutations occur either as germline alleles or due early somatic mosaic events, indicating there may be
a critical developmental window for the development of later disease.
It remains unclear whether the oxygen chemosensitive phenotype is mechanistically linked to the development of these
tumours. Understanding this requires a model system that allows us to interrogate both phenotypes, but one that is also
amenable to rapid genetic manipulation of genes that are likely to be involved. While mice have been a mainstay of such
studies, they are time-consuming and it is difficult to test multiple genes in a scalable manner. For these reasons, this project
aims to utilize the unique capabilities of the zebrafish (Danio rerio) to study this problem. The zebrafish is a small, transparent
vertebrate that is especially well-suited to imaging and rapid genetic manipulation of specific cell types within their native
64
microenvironment. The basic oxygen chemosensitivity pathway is conserved in fish, especially in certain neuroendocrine
cells (i.e. catecholaminergic cells associated with pharyngeal arch blood vessels or neuroepithelial cells of gills) 4,8. In addition,
the zebrafish is highly amenable to modelling of many different types of cancers using genetic approaches9.
The first aim of this project will be to study acute oxygen chemosensitivity using GCaMP6f reporter zebrafish. This is a
fluorescent transgenic line in which depolarization leads to an increase in intracellular calcium, which can be quantified
across the entire body of the animal. We will expose fish to acute hypoxia and use the GCaMP6f reporter as a readout of
their response, and then use CRISPR to test candidate genes which we hypothesize may be involved in this response10-12.
These genes include hif-2 itself, G-protein coupled receptors (rgs5/rgs4), atypical mitochondrial subunits (cox4i2, higd1c,
ndufa4l2) and potassium channels (task1, task3). The second aim of the project will be to assess whether developmental
timing of Hif-2 activation affects the propensity to form paragangliomas in the fish, and whether the other candidate genes
above modify the tumorigenic phenotypes. While we initially will focus on the neuroendocrine cell types known to be
sensitive to oxygen, a long-term goal is to use the fish to also discover new and unexpected cell types that may harbour such
sensitivity.
Identification of proteins underlying acute oxygen chemosensitivity will provide new therapeutic targets in pseudohypoxic
tumours such as paraganglioma. While rare, these are notoriously difficult to treat, which will have a large impact upon the
field. The mechanisms we discover may also provide therapeutic links to tumours that arise from other oxygen sensitive cell
types such as small cell lung cancer.
The student will join a vibrant group of researchers at Ludwig Cancer Research, and benefit from the combined expertise of
the Ratcliffe and Bishop groups (oxygen sensing) and the White group (zebrafish genetics and cancer). Specific skills they
will gain are in vivo imaging, creation of transgenic lines, biochemical analysis, and CRISPR manipulation (Cas9/Cas13).
References:
1 Ortega-Saenz, P., Moreno-Dominguez, A., Gao, L. & Lopez-Barneo, J. Molecular Mechanisms of Acute Oxygen Sensing
by Arterial Chemoreceptor Cells. Role of Hif2alpha. Front Physiol 11, 614893 (2020).
https://doi.org:10.3389/fphys.2020.614893
2 Platero-Luengo, A. et al. An O2-sensitive glomus cell-stem cell synapse induces carotid body growth in chronic
hypoxia. Cell 156, 291-303 (2014). https://doi.org:10.1016/j.cell.2013.12.013
3 Youngson, C., Nurse, C., Yeger, H. & Cutz, E. Oxygen sensing in airway chemoreceptors. Nature 365, 153-155 (1993).
https://doi.org:10.1038/365153a0
4 Pan, W. et al. Single-cell transcriptomic analysis of neuroepithelial cells and other cell types of the gills of zebrafish
(Danio rerio) exposed to hypoxia. Sci Rep 12, 10144 (2022). https://doi.org:10.1038/s41598-022-13693-1
5 Bishop, T. & Ratcliffe, P. J. Signaling hypoxia by hypoxia-inducible factor protein hydroxylases: a historical overview
and future perspectives. Hypoxia (Auckl) 2, 197-213 (2014). https://doi.org:10.2147/HP.S47598
6 Ratcliffe, P. J. Harveian Oration 2020: Elucidation of molecular oxygen sensing mechanisms in human cells:
implications for medicine. Clin Med (Lond) 22, 23-33 (2022). https://doi.org:10.7861/clinmed.ed.22.1.harv
7 Favier, J., Amar, L. & Gimenez-Roqueplo, A. P. Paraganglioma and phaeochromocytoma: from genetics to
personalized medicine. Nat Rev Endocrinol 11, 101-111 (2015). https://doi.org:10.1038/nrendo.2014.188
8 Hockman, D. et al. Evolution of the hypoxia-sensitive cells involved in amniote respiratory reflexes. Elife 6 (2017).
https://doi.org:10.7554/eLife.21231
9 White, R., Rose, K. & Zon, L. Zebrafish cancer: the state of the art and the path forward. Nat Rev Cancer 13, 624-636
(2013). https://doi.org:10.1038/nrc3589
10 Timon-Gomez, A. et al. Tissue-specific mitochondrial HIGD1C promotes oxygen sensitivity in carotid body
chemoreceptors. Elife 11 (2022). https://doi.org:10.7554/eLife.78915
11 Moreno-Dominguez, A. et al. Acute O(2) sensing through HIF2alpha-dependent expression of atypical cytochrome
oxidase subunits in arterial chemoreceptors. Sci Signal 13 (2020). https://doi.org:10.1126/scisignal.aay9452
12 Zhou, T., Chien, M. S., Kaleem, S. & Matsunami, H. Single cell transcriptome analysis of mouse carotid body glomus
cells. J Physiol 594, 4225-4251 (2016). https://doi.org:10.1113/JP271936

Exploiting synthetic defects in metabolism and DNA repair to improve
the treatment of glioma and AML 1,2,3 - Prof. Peter McHugh
Primary Supervisor: Prof. Peter McHugh
Additional Supervisors: Prof. Chris Schofield
Abstract
Current treatments for gliomas, an aggressive form of brain tumour, are non-specific and do not significantly increase clinical
survival rates. Likewise, treatments for acute myeloid leukaemia (AML), especially relapsed disease, remain challenging. We
have discovered a potential synthetic lethal response in tumours which exhibit a mutation in the isocitrate dehydrogenase (IDH)
gene: IDH mutations are present in ~80% of gliomas and ~20% of AMLs. Specifically, the loss of several related DNA repair
factors and IDH mutation leads to loss of cancer cell viability. Combining mechanistic cellular studies with our ongoing studies
on small molecule DNA repair inhibitors could provide a route to treat these aggressive diseases.
Research objectives
Following the sequencing of thousands of glioblastoma samples, IDH was found to be mutated in in around 80% of tumours and
in a significant subset of AML. A specific IDH1 (R132H) mutation occurs in about 70% of glioma tumours. IDH1 and IDH2 play
key roles in redox metabolism, catalysing the oxidative decarboxylation of isocitrate to a-ketoglutarate (α-KG) and CO2
Normal cell producing NADPH (Cohen et al., 2013). The cancer-associated mutant forms of IDH
enzymes produce 2-hydroxyglutarate (2-HG), termed an ‘oncometabolite’, a metabolic
intermediate that helps tumour cells survive and proliferate. 2-HG acts as a
ALKBH Excision repair competitive inhibitor of enzymes using a-ketoglutarate as a co-substrate, including a
Alkylating agent family of metal dependent dioxygenases, the ALKB family (Rohle et al., 2013).
ALKBH2 and ALKBH3 are DNA repair enzymes that directly remove alkylation DNA
IDH mutant cancer cell damage. ALKBH2 is predominantly involved in repairing 1-methyladenine (1-meA)
lesions on double-stranded DNA whilst a substrate for ALKBH3 is 3-methylcytosine (3-
meC) lesions on single-stranded DNA (Dango et al., 2011, Fedeles et al., 2015).
ALKBH Excision repair
Alkylating agent
Following a genetic screen we observed a synthetic lethal response to ALKBH2 and
ALKBH3 loss that is induced by loss of several DNA repair excision repair factors. This,
in turn, suggested that cancer cells harbouring IDH mutations could be sensitive to
IDH mutant cell with inhibited excision repair
simultaneous loss of DNA exision repair genes by virtue of their reduced ALKB repair
activity (Fig. 1). Follow-up studies suggest that this hypothesis is correct, and that
inactivation of excision repair enzymes in IDH mutant tumour cells can be lethal.
ALKBH
Alkylating agent
X
Excision repair
We will explore the detailed biology of the synthetic relationship of DNA repair defects
with the key clinically-relevant IDH1 R132H patient mutation, as well as ALKBH2/3
Figure 1. Cells depend upon disrupted cells. By performing CRISPR-Cas9-based screens that target all know DNA
both the ALKBH enzymes and damage response (DDR) genes in IDH mutant cells, we will survey and define the full
excision repair to remove DNA range of DDR genes required for survival of IDH mutant cells and therefore idenfify
damage. Loss of both can be additional new therapeutic targets. These findings will be validated in isogenic,
lethal in tumour cells. matched glioblastoma and AML cell lines with and without IDH mutations, allowing us
to explore this potential mechanism of synthetic lethality in a relevant cancer setting.
We will characterise the nature of the DNA repair defects observed in repair defective IDH mutant (and ALKBH deficient) cells
using a wide range of well-established cellular, genetic and biochemical assays available to us. We will also define the pathway
to cell death in cancer cells mutated in IDH1 or lacking ALKBH2/3 that is synthetic with additional repair pathway loss.
Furthermore, inhibitors of IDH and ALKBH2/3 are available, developed in the group of Prof. Chris Schofield (Woon et al., 2012).
Molecules that target both wild-type IDH1 but also selectively inhibit the R132H form (as well as other clinically observed
variants) have been developed and will be used to test our hypothesis that in IDH mutated cells can be killed through DNA repair
pathway inhibition. Moreover, we will work in collaboration with our chemistry collaborators to generate improved inhibitors
of key excision repair factors that can be used to selectively target IDH deficient tumours, work which builds upon well-
66
developed work in this area in our laboratories. These tools, both genetic and chemical, will ultimately be combined by the
student to perform key proof-of-principle experiments to explore this novel approach to treating two cancers of unmet need.
This proposal addresses a key priority of the Cancer Research UK and the Oxford Centre as it uses basic science to validate novel
approaches to two difficult to treat cancers, AML and glioma.
Cell culture, genomic engineering (CRISPR-Cas9 and base/prime editing), large-scale screens, general molecular biology
methods, DNA damage and repair assays, advanced microscopy, cell sorting methods, protein purification chemical biology,
protein science/enzyme inhibition, and biochemical assays. The student will also benefit from interactions with clinical
colleagues involved in treating glioma and AML, as part of their thesis committee.
References:
• Cohen AL, et al. (2013) IDH1 and IDH2 mutations in gliomas. Curr Neurol Neurosci Rep, 13: 345
• Dango S, et al. (2011) DNA unwinding by ASCC3 helicase is coupled to ALKBH3-dependent DNA alkylation repair and
cancer cell proliferation. Mol Cell, 44: 373-84
• Fedeles BI, et al. (2015) The AlkB Family of Fe(II)/α-Ketoglutarate-dependent Dioxygenases: Repairing Nucleic Acid
Alkylation Damage and Beyond. J Biol Chem, 290(34), 20734-42
• Jalbert LE, et al. (2017) Metabolic Profiling of IDH Mutation and Malignant Progression in Infiltrating Glioma. Sci Rep, 7:
44792
• O’Connor MJ (2015) Targeting the DNA Damage response in cancer. Mol Cell, 60(4), 547-560
• Rohle D, et al. (2013) An inhibitor of mutant IDH1 delays growth and promotes differentiation of glioma cells. Science,
340: 626-30
• Woon EC, et al. (2012) Dynamic combinatorial mass spectrometry leads to inhibitors of a 2-oxoglutarate-dependent
nucleic acid demethylase. J Med Chem, 8;55(5), 2173-84

67
BLOod Test Trend for cancEr Detection (BLOTTED): an observational
and prediction model development study using English primary care
electronic health records data 1,2,3,4 – Dr. Pradeep S. Virdee
Primary Supervisor: Dr. Pradeep S. Virdee
Additional Supervisors: Prof. Eva Morris & Dr. Brian Nicholson
Abstract
Blood tests are commonly requested in NHS General Practice. Blood tests might be ordered when the patient attends their
GP with symptoms or signs, to monitor a known medical condition, or as part of a “health check”. Some clinical guidelines
for GPs include recommendations that they should investigate for cancer if a blood test is lower or higher than a normal
level. These recommendations are only helpful for a small number of cancers, such as bowel or pancreatic. Over time, a
patient can build up a sequence, or trend, of blood test results. This trend might tell GPs more information than single blood
test results. For example, a small drop from a steady trend could be more useful than waiting for the blood test to drop
below a fixed level. However, the research has not yet been done to tell us which approach is most helpful to find patients
who need referral to hospital for cancer investigation. In this research, we will explore if blood tests trend can detect patients
with cancer better than single blood tests and check which patient groups trend is more helpful in and for what cancers,
with primary focus on digestive and blood cancers.
Research Objectives
Background: A recent clinical review confirms that simple blood tests have an important role in identifying patients for
cancer investigation [1]. However, analysis of National Cancer Diagnosis Audit in Primary Care data suggests that primary
care investigations may delay referral [2]. Smarter use of blood tests to select patients for further cancer investigation could
increase cancer yield and reduce unnecessary referrals. Our recent research highlighted that trends over time in serial blood
tests could be more useful than single blood tests and non-specific symptoms to select patients for colorectal cancer
investigation, with our colorectal cancer prediction models having good predictive ability [3,4]. However, trends are subtle
so difficult to spot and may exist for various cancers.
Aim: To utilise trends in blood tests from primary care for early detection of cancer.
Objectives: There are three main objectives:
1) identify trends among repeated blood tests indicative of cancer – the student will learn of smoothing techniques, such as
LOWESS, to graphically describe trends in each blood test, both overall and by personal, clinical, and cancer characteristics
(e.g. age, sex, comorbidity, diagnosis route, site, stage). Collaborators: the Big Data Institute will collaborate on data curation
and understanding of electronic health records data.
2) assess predictive ability of blood test trends for different cancer types – the student will learn of dynamic models, which
utilise repeated measures data for assessing clinical outcomes. These include statistical models, such as joint modelling, and
machine-learning models. Collaborators: the Computational Health Informatics Lab will collaborate on the machine-learning
aspect for modelling trends and the Big Data Institute will collaborate on the interpretation of repeated measures data from
national datasets.
3) develop and test prediction models utilising blood test trend to optimise patient selection for referral – the student will
learn of the intricacies of developing and testing statistical and machine-learning prediction models and their clinical
application. Collaborators: the Computational Health Informatics Lab will collaborate on the machine-learning aspect for
developing prediction models and the Big Data Institute will collaborate on the interpretation of results from national
datasets.
Data: Data from ~30 million patients from the CPRD primary care database is available to develop the models. It includes
information on patient characteristics, deprivation, blood tests, symptoms, medications, cancer diagnosis, and other
variables over 2000-2019. It is linked to the National Cancer Registration and Analysis Service, Hospital Episode Statistics
databases, and Office of National Registration death database.
Outcomes: The main outcome will be prediction models that incorporate blood test trend for cancer risk. Outputs will
include peer-reviewed journal publications for each objective separately and conference presentations.
68
Academic value: This research will develop an evidence base for blood test trend for cancer detection and inform clinical
practice. The DPhil candidate will develop leadership and research skills in various areas, including primary care, electronic
health records data, patient and public involvement, and more. The student will grow their academic publication record and
research networks at courses and events. Collaborations in this research will provide direct access to further multidciplinary
teams to improve efficiency in conducting this research.
The major route to a cancer diagnosis is GP-referral after the patient presents with symptoms. Despite millions of referrals
each year, around half of cancer diagnoses in the UK are made late-stage, where the likelihood of survival is heavily reduced
[5]. These prediction models have the potential to highlight high-risk patients before symptoms develop [6]. Initiating cancer
investigation earlier could lead to the diagnosis of cancer at an earlier stage where likelihood of survival is increased. By
helping GPs rule-in and rule-out patients for referral, cancer yield could be increased and unnecessary investigations
reduced, minimising psychological and physical harm to patients and economic costs of unnecessary testing in the NHS.
Throughout this DPhil, skills and experience will be developed in conducting independent research, working with routinely
collected linked electronic health records data, patient and public involvement, statistical analysis, prediction modelling,
general skill development, and more. The student will be offered a comprehensive training programme and encouraged to
attend relevant courses.
Internal training: The Medical Sciences Division, Big Data Institute, and Richard Doll Building at the University of Oxford run
extensive series of courses. Throughout the DPhil, the student will attend relevant courses, such as ‘Conducting national and
international research’, ‘Good Clinical Practice’, ‘An Introduction to Patient and Public Involvement’, ‘Statistical Data Analysis
with R’, ‘Writing Skills’, and ‘Viva Preparation’. The student will be based in the Cancer Research Group, who will support
the student and provide support throughout the research, and will also work closely with other multidisciplinary groups,
such as the Medical Statistics Group and interdepartmental Computational Health Informatics Lab and Big Data Institute, to
get direct additional support into various aspects of the DPhil, such as statistical methods and analystic software learning.
External training: External courses will cover general career development, research-specific training, and active
collaboration building. These courses include ‘Machine-Learning in R Software (Royal Statistical Society)’, ‘Analysis of
Repeated Measures (Bristol)’, ‘Statistical Methods for Risk Prediction (Birmingham/Keele)’, and ‘Joint Modelling of
Longitudinal and Survival Data (Leicester/Italy)’, and CRUK DPhil courses/events. Attendence and presentation of findings
at scientific conferences will also be encouraged.
References:
[1] Watson J, et al. Blood markers for cancer. BMJ (Clinical researched). 2019;367:l5774.
[2] Rubin GP, et al. Impact of investigations in general practice on timeliness of referral for patients subsequently diagnosed
with cancer: analysis of national primary care audit data. Br J Cancer. 2015;112(4):676-87.
[3] Virdee PS, et al. Full Blood Count Trends for Colorectal Cancer Detection in Primary Care: Development and Validation of
a Dynamic Prediction Model. Cancers (2022). 14, 4779.
[4] Virdee PS, et al. Trends in the full blood count blood test and colorectal cancer detection: a longitudinal, case-control
study of UK primary care patient data. NIHR Open Research (2022). 2, 32:1-53.
[5] NICE. Suspected cancer recognition and referral: site or type of cancer (2020) [Accessed 6-Oct-2022]:
https://www.nice.org.uk/guidance/ng12

69
Tackling Cancers Defective of High-Fidelity DNA Repair Mechanisms
1,2,3,4 – Fumiko Esashi
Primary Supervisor: Fumiko Esashi

Additional Supervisors: Prof. Bass Hassan
Abstract
Recent technological advancements in cancer genomics have revealed significant cell-to-cell heterogeneity, highlighting
the role of mutability in driving cancer evolution, posing therapeutic challenges (1). A recent study has indicated that the
simultaneous impairment of two key high-fidelity DNA repair mechanisms, homologous recombination (HR) and
mismatch repair (MMR), contributes to adaptive mutability and drug resistance (2). Notably, while HR loss is lethal in
most cell types, MMR deficiency may mitigate this lethality. Our hypothesis is that MMR-defective backgrounds enable
the survival and rapid evolution of cancer cells with HR loss. To investigate the impact of MMR/HR dual deficiencies on
cancer development, we propose innovative experimental and bioinformatic approaches. Specifically, by conditionally
inactivating HR in MMR-defective cellular model systems, we will identify genetic and genomic factors affecting cell
survival. Additionally, we will perform association analyses based on somatic cancer mutations databases to uncover
potential biomarkers and therapeutic strategies for early diagnosis and treatment of these cancers.
Research objectives
Individuals with inherited mutations within genes encoding MMR or HR factors exhibit increased risk to develop a wide
range of cancers, as seen in patients with hereditary nonpolyposis colorectal cancer/Lynch syndrome (HNPCC/LS) or
hereditary breast and ovarian cancer syndrome (HBOC), respectively. It is widely described that MMR defects confer
mutator phenotypes with no lethal impact. Conversely, the biallelic mutations of genes encoding key HR regulators, such
as the breast cancer susceptibility 2 (BRCA2) and the partner and localizer of BRCA2 (PALB2), elicits lethality, although
monoallelic BRCA2 or PALB2 mutations are sufficient to increase cancer risk. Notably, a recent study suggests that the
simultaneous impairment of MMR and HR drives adaptive mutability and drug resistance (2). However, the causal
relationship of this phenomenon remains unclear. We hypothesise that MMR defective mutator background alleviates
the lethal impact of HR loss and assists rapid evolution of cancer. This project tests this hypothesis and identifies genetic
and genomic elements that are associated with MMR- and HR-defective cancers.
The genetic concept of ‘synthetic lethality’ or ‘synthetic viability’, involving the combination of mutations in multiple
genes leading to cell death or growth, respectively, has gained rising attention in recent years for its potential for
discovering new therapeutic targets in challenging cancers. Previous studies have relied on genome-wide loss-of-function
screens in knockout cell lines. However, this approach has limitations, such as phenotypic changes obscured by secondary
mutations. This project tackles these shortcomings by utilising the auxin-inducible degron (AID) technology (3) to
conditionally deplete endogenous BRCA2 or PALB2 in MMR-defective HCT116 cell lines. This allows for highly-specific
examination to uncover the direct impact of BRCA2 or PALB2 depletion in MMR-defective mutator background. Our
preliminary study shows that, indeed, the acute depletion of BRCA2 or PALB2 confers lethality (Fig 1A), as well as
increased sensitivity to a chemotherapeutic drug,
poly (ADP-ribose) polymerase inhibitor
olaparib (Fig 1B, C) as expected (4). To
identify genes that affect normal survivals
of BRCA2- or PALB2-depleted HCT116 cells,
we will leverage the CRISPR-mediated
modulation of transcription, namely
CRISPR interference/activation
(CRISPRi/a) (5). Our lab has already
established the systems combining AID and
CRISPRi/a. Using this technology, we will
identify genetic factors, down- or up-
regulation of which affects the survival of
BRCA2- or PALB2-depleted cells.
In parallel, we will directly assess genome changes that occur upon BRCA2- or PALB2 depletion in HCT116 cells. We will
isolate several clonal HCT116 cell lines which have survived upon depletion of BRCA2 or PALB2 for one month. Our
preliminary analysis indicates distinct chromosomal aberrations in these cells, arising highly repetitive centromeric
70
regions of chromosomes. This observation is particularly intriguing as these repetitive regions are known to be targeted
by MMR or HR (6, 7). To gain high resolution pictures of genome changes at these regions, we will conduct long-read
whole genome sequencing using Oxford Nanopore Technology (ONT) that has advantages over traditional short-read
sequencing. It enables the detection of alterations in repetitive sequences, as well as DNA modifications, such as CpG
methylation. This approach is expected to provide a comprehensive understanding of the observed genome changes.
Finally, we will conduct a bioinformatic assessment of publicly available somatic cancer mutation databases, including
COSMIC, to determine the prevalence of simultaneous impairment of MMR and HR pathways. We will initially focus
colon cancers, which exhibit MMR deficiency in approximately 15% of cases. We will also explore the potential
correlation between MMR/HR co-downregulation and the development of drug resistance. By examining the factors
identified in our CRISPRi/a and long-read sequencing studies, we aim to uncover their association with drug resistance
mechanisms. This integrative approach will provide valuable insights into the underlying mechanisms driving drug
resistance in these specific cancer types and inform the development of targeted therapeutic strategies.
(ii) the collaborations involved and how these will be facilitated by the award:
The proposed project holds distinct opportunities to establish novel collaborations that will be facilitated by the award.
This project stands out by venturing into new areas of research through long-read whole genome sequencing and cancer
somatic mutation analysis, which have not been previously explored within our team nor in the field. The collaboration
with Dr Adam Cribbs from the Nuffield Department of Orthopaedics, Rheumatology and Musculoskeletal Sciences
(NDORMS) provides expertise in ONT-based long-read sequencing, offering new avenues for discovery and high-
resolution analysis of genomic alterations associated with MMR and HR deficiencies. Dr. Cribbs' knowledge and resources
in this field will significantly enhance the comprehensiveness of our research. Additionally, the second supervisor Prof
Bass Hassan, a clinical research scientist at the Sir William Dunn School, brings expertise in cancer mutation analysis and
perspectives for translational research. With Prof Hassan's guidance, we will navigate and interpret the vast dataset of
cancer somatic mutations, gaining a deeper understanding of the genetic landscape and its implications for cancer
development. These collaborations not only provide access to specialised techniques and resources in their respective
expertise, but also open up new possibilities for uncovering critical insights into the genome instability, including those
at repetitive regions, often found in difficult-to-treat cancers. Furthermore, their ability to bridge the gap between
laboratory discoveries and clinical applications will significantly enhance the scientific impact and translational potential
of our work. This interdisciplinary collaboration promotes knowledge exchange, fosters practical relevance, and ensures
that our research has real-world implications in the fight against cancer.
Translational potential of the project.

The proposed project holds significant translational potential. Firstly, by identifying genetic and genomic elements that
influence the survival of HR- and MMR-deficient hypermutable cells, our research is expected to identify early diagnostic
markers and strategies for timely intervention. We can exploit vulnerabilities specific to these cancer cells, leading to
more effective treatments while minimising adverse effects. Secondly, by integrating bioinformatic analyses of cancer
genomes, this project is expected to reveal the prevalence of simultaneous impairments in MMR and HR pathways in
colon cancer and potentially identify previously unspecified cancer 'signatures' associated with dual HR/MMR deficiency.
Further assessment of correlation between MMR/HR co-downregulation and the development of drug resistance will
offer an opportunity to develop novel therapeutic strategies for these challenging-to-treat cancers. In future, similar
approach could be applied to assess other types of cancers, such as ovarian cancer and pancreatic cancer, which are
commonly observed in HNPCC/LS and HBOC patients.
Our research project offers valuable training opportunities in key areas of cancer research, including: (1) cell culture
techniques, encompassing cell line maintenance, manipulation, and experimental assays; (2) the opportunity to learn
and apply long-read sequencing techniques, including sample preparation, data generation, and analysis; and (3)
bioinformatic techniques for analysing publicly available somatic cancer mutation databases and exploring genetic
patterns. By providing training in these areas, our project equips researchers with essential skills for future scientific
endeavors in the field of cancer research. The candidates will be well supported in the Dunn school in related methods
training, including CRISPR, light microscopy imaging and flow cytometry through in-house facilities, namely the Genome
Engineering Oxford (led by Dr Joey Riepsaame), the Dunn School Bioimaging Facility (led by Dr Alan Wainman), and the
Don Mason Facility of Flow Cytometry (led by Dr Robert Hedley), respectively.
71
References
1. Loeb, L.A. (2011) Human cancers express mutator phenotypes: origin, consequences and targeting. Nat Rev
Cancer, 11:450-7. doi: 10.1038/nrc3063
2. Russo, M. et al. (2019). Adaptive mutability of colorectal cancers in response to targeted therapies. Science,
366:1473-1480. doi: 10.1126/science.aav4474.
3. Natsume, T. et al. (2016). Rapid Protein Depletion in Human Cells by Auxin-Inducible Degron Tagging with
Short Homology Donors. Cell Reports, 15, 210–218. doi: 10.1016/j.celrep.2016.03.001.
4. Hopkins, J.L. et al (2022) DNA repair defects in cancer and therapeutic opportunities Genes Dev. 36: 278–293.
doi: 10.1101/gad.349431.122.
5. Gilbert, L. A. et al. (2014). Genome-Scale CRISPR-Mediated Control of Gene Repression and Activation. Cell,
159, 647–661. doi: 10.1016/j.cell.2014.09.029.
6. Aze, A. et al. (2016). Centromeric DNA replication reconstitution reveals DNA loops and ATR checkpoint
suppression. Nat Cell Biol, 18, 684-91. doi: 10.1038/ncb3344.
7. Saayman, X. et al. (2023). Centromeres as universal hotspots of DNA breakage, driving RAD51-mediated
recombination during quiescence. Mol Cell 83, 523–538a. doi: 10.1016/j.molcel.2023.01.004.
72
Predicting Response to Therapy in Oesophageal Cancer 4 – Prof. Jens
Rittscher
Primary Supervisor: Prof. Jens Rittscher
Additional Supervisors: Elizabeth Smyth
Abstract
The incidence of oesophageal adenocarcinoma (OAC) has increased over the last three decades. Approximately 8500 new
cases of OAC are diagnosed every year in the UK. Only one in seven patients survive for more than 5 years. OAC is a cancer
that is associated with extensive treatment requirements and in consequence a considerable decline in health-related
quality of life. Curative treatment includes a combination of chemotherapy, radiotherapy, and/or immunotherapy
followed by extensive surgery, however less than half of patients are cured using this approach.
Research objectives
The goal of this project is the development of tissue-based biomarkers that can accurately predict therapy response to
save time, improve quality of life, and potentially improve survival. Here, we will be building on our recent work on an AI
based approach for predicting response to radiotherapy in colorectal cancer on the basis of digitised H&E histology images
[1]. Enhancing the interpretability of AI model predictions in such a way that they provide contextual information that
effectively supports clinical decision making is an important aspect which we aim to refine and extend. Morpho-molecular
correlates [2] that relate morphological features with disease relevant molecular traits are core to generating interpretable
contextual information. The main aims of this projects are:
● Predicting which patients benefit from radiotherapy or chemotherapy,

● Predicting response to immune checkpoint inhibitors,
● Develop histology-based biomarkers to enable precision oncology.
This project provides the opportunity of developing state-of-art deep learning methods in an interdisciplinary setting. You
will be working with oncologists, pathologists, and engineers to formulate the clinical questions and translate these into
technical requirements. To be successful in this project you will have a background in machine learning and a passion for
developing translational AI solutions in a responsible manner. In addition to developing the underlying methodology, you
will assess the inherent bias of specific clinical trial cohorts, and develop validation strategies that ensure robustness. As
part of the development of morpho-molecular correlates you will perform a multi-omic analysis of existing patient cohorts.
Dataset Type What is available
VESTIGE Phase II randomised clinical trial Biopsy for some patient
Resection for most
Adjuvant chemo vs Tumour response grading (PDL1,

immunotherapy in high risk (after MMR, planned: Claudin 18.2, HER2,
neoadjuvant chemo) FGFR2, RNAseq)
OCCAMS Prospective UK dataset WGS
Cambridge Chemotherapy plus surgery RNAseq for 250

based outcomes
Methylation
73
Limited radiotherapy IHC lacking but could be done
GSTT/RMH Prospective UK dataset No omics

dataset
Chemotherapy and surgery
oucomes
Tumour response grading
Limited radiotherapy
References
[1] Wood, Ruby, et al. "Joint prediction of response to therapy, molecular traits, and spatial organisation in
colorectal cancer biopsies" Medical Image Computing and Computer Assisted Intervention–MICCAI 2013: 24st
International Conference, Vancouver, Canada (accepted for publication)
[2] Sirinukunwattana, Korsuk, et al. "Image-based consensus molecular subtype (imCMS) classification of colorectal
cancer using deep learning." Gut 70.3 (2021): 544-554

74
Developing single-cell transcriptomics tools for PARP inhibitor
resistance in BRCA1/2-deficient cells and tumours1,2,3,4 – Prof
Madelena Tarsounas
Primary Supervisor: Prof Madelena Tarsounas
Additional Supervisors: Dr. Christiana Kartsonaki
Abstract
Cells and tumours with compromised breast cancer susceptibility genes BRCA1 or BRCA2 retain the ability to proliferate,
in spite of the severe genomic instability caused by accumulation of DNA lesions. This vulnerability is exploited by highly-
specific therapies that enhance the susceptibility of BRCA1/2-deficient tumours to DNA damaging agents, with the poly-
ADP ribose polymerase inhibitors (PARPi; e.g. olaparib) as a prominent example. In spite of clear therapeutic benefits, cure
rates for BRCA1/2-mutated cancers remain low, as patients frequently develop resistance to PARPi. Several mechanisms
of PARPi resistance have been reported. However, PARPi resistance remains a critical problem in the clinic, limiting
sustained responses to these drugs. Here, we aim to identify transcriptional signatures associated with PARPi resistance,
specifically olaparib resistance.
To identify such signatures, we will perform single-cell RNA sequencing (scRNAseq) using BRCA1/2-deficient cells in culture
and cells obtained from patient-derived xenograft (PDX) models that have acquired olaparib-resistance upon prolonged
exposure to olaparib. We will furthermore combine scRNAseq and EdUseq data to test whether replication failure at
specific genomic sites could interfere with transcription and therefore represent the underlying mechanism of the
identified transcriptome alterations.
In the longer term, this line of research can lead to predictive markers for patient response to olaparib, which could
facilitate early patient stratification and the development of personalized treatment strategies.
Research Objectives
The work proposed here will help elucidate, at single-cell resolution, the relationship between the emergence of rare,
tumour-initiating cells sub-populations within tumours lacking BRCA1 or BRCA2 genes and the response to the PARP
inhibitor olaparib. Accordingly, we will pursue two main objectives:
a. Characterize the single-cell transcriptomic landscape of olaparib-resistant BRCA1/2-deficient cells in culture. The
scRNAseq technology will enable us to generate gene expression profiles of single cells and to identify cell sub-populations
with specific transcriptional signatures. To achieve this first objective, scRNAseq will be carried out in populations of
olaparib-sensitive and -resistant BRCA1/2-deficient cells, already generated in Tarsounas lab. We will prepare libraries
from each cell line, before and after olaparib resistance onset, to sequence between 7,000 and 10,000 cells using the
standard protocol of the Chromium Single-Cell 3’ gene expression profiling solution (10x Genomics).
Unsupervised clustering approaches will be developed to classify cells into sub-groups with specific signatures (e.g.
immune response, metastasis etc.) and to monitor cell dynamics using algorithms for pseudotime analysis. We will apply
this combination of analytical approaches to the cell lines that are sensitive or become resistant to Olaparib, and anticipate
that this will allow us to identify cell clusters with unique patterns of gene expression, which could not be resolved at the
whole-cell population level. In addition, this approach will enable us to identify eventual differences between the
signatures specific to BRCA1- and BRCA2-deficient cells. Lastly, the collection of signatures identified for distinct cell
subpopulations selected by olaparib will be further explored in the large METABRIC and TCGA PanCancer Atlas breast and
ovarian cancer cohorts (5,098 samples, among which 355 and 362 carry alterations in BRCA1 and BRCA2, respectively),
specifically to assess their prognostic ability through univariable and multivariable regression models.
b. Characterize the single-cell transcriptomic landscape of olaparib-resistant BRCA1/2-deficient PDX tumours in vivo. In
addition to linking the transcriptomic signatures of olaparib-resistant cell subpopulations to tumour gene expression data
and clinical information found in databases (e.g. TCGA, METABRIC), we will recapitulate in vivo the results obtained in vitro
using cell cultures. To achieve this, scRNA-seq will be carried out in cell suspensions prepared from BRCA1- or BRCA2-
mutated (n = 3 BRCA1-/- and n = 1 BRCA2-/-) olaparib-naïve and -resistant patient-derived xenografts (PDX). In these models
75
resistance emerged after treatment with olaparib for up to 150 days, when individual tumours regrew. These models are
also already available for processing in Tarsounas lab.
(ii) This project will facilitate the collaboration between basic cell biology and bioinformatics, and the student funded here
will be trained in and benefit from both types of expertise. Importantly, this collaboration will enable the co-applicant, Dr
Christiana Kartsonaki, who is an early career researcher within the NDPH, to achieve scientific independence. Securing this
CRUK award will enable her to co-supervise a graduate student in a new area of research, and at the same time, will place
her in a stronger position for further funding applications.
In spite of initial responses to targeted therapies such as PARPi, BRCA1/2-deficient tumours develop a resistance to these
therapies. PARPi resistance often entails genomic rearrangements and mutations that trigger rewiring of the damage
response pathways within the tumour so that apoptotic responses to treatment are replaced by cell survival and
metastasis. Here we anticipate to identify new, robust transcriptional signatures associated with Olaparib resistance,
which can be used to stratify patients for PARPi therapy. In addition, these gene expression profiles will identify
vulnerabilities that can be exploited to target resistant disease. In the longer term, these approaches can be used to
develop patient screening protocols using machine learning and statistical methods.
76
Harnessing measurements of the tumour microenvironment to improve
the early detection of prostate cancer. 1,3,4 – Professor Richard Bryant
Primary Supervisor: Professor Richard Bryant
Additional Supervisors: Dr James Grist
Abstract
Approximately 190,000 invasive prostate biopsy procedures are performed each year in the UK alone. The current standard of
care recommends the use of pre-biopsy MRI scan to improve the accuracy of the prostate biopsy. Despite the success of using
the ‘Prostate Imaging Reporting And Data System’ (PI-RADS) (1) score (and similar scores such as the Likert score) in the reporting
of MRI scan images ahead of targeted prostate biopsy, around 5-15% of clinically significant cancers would potentially be missed
if individuals with an elevated age-specific PSA and normal clinical imaging did not receive a prostate biopsy. It is difficult to derive
precise figures, but this would likely amount to several thousand men with a delayed or missed diagnosis of clinically significant
prostate cancer each year in the UK alone, the clinical impact of which is currently unknown.
Whilst changes in imaging contrast due to tumour growth tend to be slow, primarily due to the inherently low sensitivity of clinical
MRI, alterations in the motion of water (due to the restriction of motion by highly cellular regions of cancerous tissue) and the
direct metabolic reprograming of cancerous cells to upregulate production of lactate to fuel
proliferation (known as the “Warburg Effect”) can be harnessed to potentially provide early detection
of prostate cancer (2). Indeed, the first in-human trial using hyperpolarised MRI demonstrated the
ability of this technique to detect oncological metabolism in the absence of conventional imaging
alterations (see Figure 1 showing cancerous metabolism in both the left and right sides of the
prostate, with conventional imaging showing changes only on the right side), and Restricted
Spectrum Imaging (RSI) has been shown to be highly sensitivity in the detection of prostate cancer
(3).
We have experience in advanced imaging techniques to study neurological and cardiovascular
disease, and now is an opportune time to expand these techniques to benefit the oncological arena,
this being an area of important unmet clinical need for prostate cancer. The DPhil student will work
with the surgical and radiological teams to recruit and image a cohort of 40 men aged between 60-
69 years (as defined by power calculations from previous studies in Oxford) being investigated for
possible prostate cancer. Participants will receive standard of care pre-biopsy MRI imaging, with the
functional MRI appended to the end of the clinical session. The biopsy process will be performed
based on the clinical imaging, as per our current practice. Of the N = 40 men in this pilot study, 10
individuals will have a normal (PIRADS 1-2) pre-biopsy MRI, 10 individuals will have an equivocal
Figure 3 (PIRADS 3) MRI, and 20 will have a ‘suspicious’ (PIRADS 4-5) pre-biopsy MRI. The inclusion of patients
with either a PIRADS 1-2 or a PIRADS 3 MRI is important given that individuals in the age range of 60-69
years often have diffuse PIRADS 3 change within the peripheral zone of the prostate gland due to their young age. All 40
individuals will undergo prostate biopsy following the performance of the clinical and functional scans, with targeted and
systematic biopsy cores being obtained for the N = 30 men with clinical lesions (PIRADS 3-5), and systematic biopsies alone being
obtained for men with a normal (PIRADS 1-2) clinical imaging, as per our current protocol. The clinical and functional MR images
will then be correlated with the final prostate biopsy pathology, to test the hypothesis that some clinically significant prostate
cancer cases may solely be visible on functional MRI and not on standard clinical imaging. This approach will improve sensitivity
of detection – with RSI acquisitions, if shown to be beneficial in the detection of prostate cancer, ready to be run in clinical practice
in every patient in our NHS trust, and beyond, and hyperpolarised MRI offered as a clinical scan in targeted patients at a later
time point (2-5 years). In turn, this will benefit many men being investigated for possible prostate cancer using MRI scans as part
of the clinical pathway.
We currently have an ethics application going through internal review to enable this project. See figure 2 for the workflow for
this project.
Figure 4
77
Research objectives
The two key objectives for this DPhil project are:
1) To establish repeatable imaging protocols for assessing functional and clinical MRI within the prostate. A sub-cohort (N =
20, 5 from each PI-RADS group) of patients and healthy controls (N = 8) will undergo test re-test imaging on the same
day to assess for the repeatability of imaging results. Collaborators required: Dr James Grist, Dr Ruth MacPherson,
Professor Damian Tyler. Outcome: Presentation of results at leading radiological conferences (European Congress of
Radiology, International Society of Magnetic Resonance in Medicine), and publication of results in a leading medical
imaging journal.
2) To perform a clinical study assessing the added value of functional MRI to the routine clinical protocol. The data from the
full cohort of patients will be analysed to assess for the sensitivity and specificity of the clinical, functional, and combined
imaging approaches for the detection of biopsy proven clinically significant prostate cancer. Collaborators required: Dr
James Grist, Dr Ruth MacPherson, Dr Richard Colling, and Professor Ian Mills. Outcome: Presentation of results at a
leading radiological conference and publication of results in a leading medical journal.
If successful, this project will enable the early detection of clinically significant prostate cancer in a cohort of patients undergoing
investigation for possible prostate cancer, who might otherwise have a delayed or missed diagnosis. However, the further clinical
translation and impact of this project will be felt from the incorporation of these additional functional MRI approaches into clinical
practice in Oxford and beyond – thus benefitting a larger population of patients with suspected prostate cancer. However, beyond
this impact there is also the potential to apply these advanced imaging methods to other cancers, to detect early therapeutic
response to chemo/radiotherapy – for example paediatric brain tumours, renal cancer, and breast cancer.
The student will be provided with training in clinical research and patient recruitment by Professor Richard Bryant, and will be
provided with training from Professor Damian Tyler and Dr James Grist in MRI physics, data acquisition, image reconstruction and
quantitative post-processing for both hyperpolarised MRI and RSI. Further training in image co-registration will be provided. Dr
Ruth MacPherson will provide training in clinical image analysis and interpretation, and Dr Richard Colling will provide training in
histology and will work with Professor Ian Mills to complement this with molecular phenotyping of cancerous tissue. By the end
of the project, the student will have a wide-ranging skill set that will be of great benefit to further their career in clinical research.
References
1. Turkbey B, Rosenkrantz AB, Haider MA, et al.: Prostate Imaging Reporting and Data System Version 2.1: 2019 Update of Prostate
Imaging Reporting and Data System Version 2. Eur Urol 2019; 76:340–351.
2. Nelson SJ, Kurhanewicz J, Vigneron DB, et al.: Metabolic Imaging of Patients with Prostate Cancer Using Hyperpolarized [1-
13C]Pyruvate. Sci Transl Med 2013; 487:109–113.
3. Brunsing RL, Schenker-Ahmed NM, White NS, et al.: Restriction spectrum imaging: An evolving imaging biomarker in prostate
MRI. J Magn Reson Imaging 2017; 45:323–336.

Investigating Hypoxic Adaptation in Glioblastoma (GBM) Stem Cells
through Pooled Kinome-Wide CRISPR-Cas9 Knockout Screen 1,2,3 – Dr.
Sneha Anand
Primary Supervisor: Dr. Sneha Anand
Additional Supervisors: Prof. Daniel Ebner
Abstract
Glioblastoma (GBM) is a highly aggressive brain tumour characterized by its resistance to treatment and invasive nature.
Central to this resistance is the existence and chemo-radiotherapeutic resistance of glioblastoma stem cells (GSC) within a
hypoxic core of the tumour. Hypoxia, a condition of low oxygen availability, is a critical feature of the tumor
microenvironment (TME) that promotes GBM progression, migration, immune evasion, and therapy resistance and could
be a major factor in the generation of recurrent GBM which leads to poor survival outcomes of GBM patients. Protein
kinases, involved in various signalling pathways, have been implicated in the response of cancer cells to hypoxia (Astrid A
Glück 2015). They are attractive targets for cancer therapy due to their critical roles in oncogenic processes and the
availability of kinase inhibitors. In the context of hypoxia, molecular crosstalk between Hypoxia Inducible factors (HIFs) and
protein kinases plays a crucial role in modulating cellular responses to low oxygen levels. Dysregulated kinase activities are
common in cancer, and hypoxia can further disrupt the cellular kinome network, enhancing the robustness of oncogenic
pathways. Targeting hypoxia-dependent kinase signalling hubs holds the potential for attenuating the survival advantages
conferred by hypoxic adaptation. In this research proposal, we aim to investigate the molecular basis underlying the
adaptation and survival of GSC cells in hypoxia. We propose to conduct a pooled kinome-wide CRISPR-Cas9 knockout screen
using a lentiviral single guide RNA (sgRNA) library that targets human kinases (John G Doench 2016). The goal is to identify
key kinases that contribute to the hypoxic adaptation of GBM cells, potentially uncovering new therapeutic targets for GBM
treatment.
Research objectives
This research project aims to investigate the molecular mechanisms underlying hypoxic adaptation in GBM using patient-
derived GSC lines. GSCs are known to play a significant role in tumor initiation and recurrence, and hypoxia promotes their
self-renewal and invasive properties. The study involves characterizing GSC behaviour in low oxygen conditions through
hypoxia chamber culture, hypoxia marker staining, and analysis of hypoxia-inducible factor (HIF) expression using various
techniques such as immunofluorescence staining, western blotting, or quantitative PCR. Following characterization, a
kinome-wide CRISPR-Cas9 knockout screen (Fig.1) will be performed using a lentiviral sgRNA library targeting human kinases.
The abundance of sgRNAs targeting kinases will be analysed through next-generation sequencing, and bioinformatics
analysis will be conducted to identify kinases exhibiting differential knockout effects in hypoxic compared to normoxic
conditions. The identified kinases will undergo functional characterization to understand their specific roles in hypoxic
adaptation. Ultimately, the study aims to uncover potential therapeutic targets for disrupting hypoxia-dependent
mechanisms in GBM.
Figure 1: Schematic illustrating the CRISPR knockout screen: Lentiviral transduction of patient-derived GSC lines with the
sgRNA library to induce kinase knockout. Following this, cells will be divided into normoxia and hypoxia groups and
maintained in their respective conditions for a specific duration. Genomic DNA extraction from cells in both normoxia and
hypoxia groups. Next-generation sequencing to analyze the abundance of sgRNAs targeting kinases. Data processing to
79
determine sgRNA frequency and identify kinases significantly affected in hypoxic conditions compared to normoxic
conditions.
Expected Outcomes:
1. Characterization of GSC Behaviour in Hypoxic Conditions: The study will provide insights into the behaviour of GSCs
under hypoxic conditions. This includes understanding the activation of specific signalling pathways in response to low
oxygen levels, such as the upregulation of hypoxia-inducible factors (HIFs) and their downstream target genes related
to angiogenesis, glycolytic metabolism, and therapy resistance.
2. Identification of Hypoxia-Related Therapeutic Targets: The characterization of GSC behaviour in hypoxic conditions can
lead to the identification of potential therapeutic targets specifically active in the hypoxic microenvironment. These
targets may offer opportunities for the development of novel therapies aimed at disrupting hypoxia-mediated
adaptations and improving treatment outcomes for glioblastoma patients.
3. Identification of Hypoxia-Responsive Kinases: Through the kinome-wide CRISPR-Cas9 knockout screen, the study will
identify kinases that play a critical role in GBM cell survival and adaptation in hypoxia. This information can provide
valuable insights into the molecular mechanisms underlying hypoxic adaptation in GBM.
4. Potential Therapeutic Targets: The study will yield a list of kinases whose inhibition could potentially attenuate the
survival advantages conferred by hypoxia in GBM. These kinases may serve as promising targets for the development
of novel anti-GBM therapies, with the aim of enhancing treatment efficacy and overcoming drug resistance associated
with hypoxia.
5. Manuscript Generation: The research project aims to generate two manuscripts from the study, highlighting the
findings, methodologies, and implications of the research. These manuscripts will contribute to the scientific literature
and would benefit other researchers and clinicians in the field.
Understanding the molecular basis of GBM cell adaptation to hypoxia is crucial for developing effective treatment strategies
against this aggressive brain tumor. The identified kinases can be further investigated for their roles in hypoxia-mediated
processes, such as angiogenesis, metabolism, and immune evasion. Additionally, functional validation studies can be
conducted to confirm the impact of specific kinases on GBM cell behavior under hypoxic conditions.
Overall, the findings from this study will contribute to a deeper understanding of GBM biology and provide potential targets
for the development of novel therapeutic strategies to improve patient outcomes in GBM treatment.
This research project offers valuable training opportunities in CRISPR-based cell screening, cell biology, molecular biology,
and bioinformatics analysis. The student will learn essential techniques for gene editing, high-throughput screening, and
cellular manipulation. They will gain hands-on experience in cell culture, molecular biology techniques, and an introduction
to bioinformatics data analysis. The project will enhance their skills in experimental design, data interpretation, literature
review, and scientific writing. Overall, this project provides a solid foundation for their DPhil studies and future research in
cancer biology and therapeutics.
References
Astrid A Glück, Daniel M Aebersold, Yitzhak Zimmer, Michaela Medová. 2015. "Interplay between receptor tyrosine kinases
and hypoxia signaling in cancer." Int J Biochem Cell Biol 101-114.
Giaccia, E B Rankin & A J. 2008. "The role of hypoxia-inducible factors in tumorigenesis." Cell Death & Differentiation 678–
685 .
John G Doench, Nicolo Fusi, Meagan Sullender, Mudra Hegde, Emma W Vaimberg, Katherine F Donovan, Ian Smith, Zuzana
Tothova, Craig Wilen, Robert Orchard, Herbert W Virgin, Jennifer Listgarten & David E Root. 2016. "Optimized sgRNA design
to maximize activity and minimize off-target effects of CRISPR-Cas9." Nature Biotechnology 184–191.

Dietary modification to augment colorectal cancer treatment 1,2,3 – Dr
Dimitrios Koutoukidis
Primary Supervisor: Dr Dimitrios Koutoukidis
Additional Supervisors: Prof. Simon Buczacki and Dr Victoria Woodcock
Abstract
Obesity is associated with worse progression-free survival among cancer survivors. One of the reasons might be that
patients with obesity are receiving lower doses of chemotherapy compared with patients without obesity. Preclinical
models show that energy restriction reduces chemotherapy toxicity and may augment chemotherapy efficacy. This is
hypothesised to be driven by improvements in glucose regulation and reductions in IGF-1 through stress resistance in
healthy cells and stress sensitisation in cancer cells. Furthermore, they suggest that standalone energy restriction is not
as efficacious as a combination of energy restriction and established cancer treatments, such as chemotherapy.1
Although fasting itself around chemotherapy seems to not raise safety concerns based on a limited number of clinical
cases and has shown some preliminary potential for increasing chemotherapy efficacy (higher radiological response and
percentage of Miller Payne 4/5), adherence is challenging, the employed protocols vary, and the substantial increase in
energy intake for 2-3-week period following the short-term (24-60h) fasting period during chemotherapy administration
may negate its benefits. Furthermore, the 24-60h fasting that has shown benefit in mice cannot be expected to have the
same effect as a 24-60h fast in humans, given the different physiology and lifespan. Instead, it is likely that a longer
period of energy restriction will be required to achieve results of similar magnitude. Additionally, intentional weight loss
may be considered counterintuitive during cancer treatment as weight loss, primarily muscle mass loss, has been
associated with worse outcomes.2 3
Existing low-energy dietary interventions in other clinical settings have been shown to affect the postulated mechanism.
Intentional weight loss among people with obesity in other clinical settings has reliably improved glucose regulation and
reduced insulin with the relationship following a dose-response manner and substantial weight loss to be required for a
clinically meaningful improvement in glucose regulation to be observed. We and others have shown high adherence and
significant weight loss with 800kcal/day diets including high-protein nutritionally replete total diet replacement with
soups and shakes and food-based low-carbohydrate diets.4-7
This DPhil project will aim to adapt one such existing intervention, test its feasibility in a randomised controlled trial, and
explore the mechanisms through which the intervention may increase chemotherapy efficacy.
Research objectives
WP1: Adapting the intervention to the setting.
Using the established Intervention Mapping approach and the MRC framework for complex interventions, together with
involvement of relevant patient representatives, clinicians, and researchers. Through an iterative process, WP1 will focus
on adapting an existing feasible and effective total diet replacement intervention.
WP2: Testing the feasibility of the intervention

WP2 will involve running a randomised controlled trial to test the feasibility of the intervention. Specifically, the feasibility
will be judged against pre-determined criteria on
- recruitment
- adherence to the diet
- retention at the last follow-up
- experience of participants following the intervention.
Population: Patients with obesity (BMI ≥ 30kg/m2) and metastatic colorectal cancer.
Intervention: The exact nature of the intervention will depend on the findings on WP1, but we envisage that it involve a
low-energy (800kcal/day) intervention through total diet replacement or through food-based low-carb diet from 5 days
before each chemotherapy cycle to 1 day after each cycle followed by advice for a healthy balanced diet for the remaining
period.
81
Figure 1: Proposed diet for the first 2 cycles of chemotherapy by day. Green indicates low-energy diet Orange indicates
healthy balanced diet, and blue indicates days of chemotherapy.
Days -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Low-energy diet Healthy balanced diet Low-energy diet Healthy balanced diet
1st cycle 2nd cycle
Comparator: Standard care (chemotherapy and existing CRUK leaflet on diet during treatment).
Sample size: With 72 patients (36 in each group), the trial will be 90% powered at one-sided 5% level based on the normal
approximation approach to detect whether the proportions for adherence and retention criteria are truly above 35%
and 65%, respectively, based on an alternative hypothesis that they will be above 60% and 85%, respectively.
Feasibility outcomes: These will include rates of recruitment, adherence, and retention.
Process outcomes: Experience and acceptability of the intervention based on semi-structured qualitative interviews with
intervention participants.
Exploratory outcomes: Relative dose intensity, toxicity, and tolerance; and health-related quality of life.
WP3: Exploring the mechanisms of the intervention

The objectives of this WP is to explore potential mechanisms through which the intervention may increase chemotherapy
efficacy.
Process: Participants will be asked to provide blood and urine samples before and after each cycle of the low-energy diet
and each cycle of chemotherapy. Samples will be analysed for
- inflammation and immune serum markers (using the Olink Explore384 and macc cytometry (CyTOF)). The Olink
panel will enable quantification of 368 low-abundance inflammatory cytokines and soluble immune mediators
and CyTOF will enable quantification and functional assessment of >50 distinct immune cell subsets.
- urine and serum metabolomics and proteomicsmTOR & pmTOR PBMC ELISA.
Developing novel therapeutics that are safe and effective with minimal side effects is a key priority for both cancer
research and patient care. This need is highest among patients with metastatic disease, which this DPhil targets. This
project will pave the way for a definitive clinical trial that will assess the effectiveness of this intervention in improving
response to chemotherapy. If successful, this is likely to reduce the burden for patients and increase their progression-
free survival.
The student will have the opportunity to be trained in an array of different methodologies, including the development
of complex interventions, recruiting to, managing, and analysis the results of a randomised clinical trial, qualitative
analysis of interviews, laboratory techniques, and analysing of complex large-scale biomarker data. This will include
targeted formal courses within and outside Oxford as well as on-the-job training. They will have access to standard
operating procedures from a relevant clinical trials unit, weekly meetings with supervisors, and access to a diverse pool
of researchers applying the above methodologies at both NDPCHS and NDS.
References
1. Nencioni. Nat Rev Cancer 2018;
2. de Groot. Nat Commun 2020
3. Vernieri. Nat Commun 2020;
4. Astbury. BMJ 2018;
5. Morris. Diabetes Obes Metab 2020;
6. Koutoukidis. Obesity (Silver Spring); 2023
7. Beeken. Obesity (Silver Spring) 2017;

Investigating the adaptive immune responses to tumour neo-antigen
and the impact on patient disease course. 1,2,3,4 –Assoc Prof. Rachael
Bashford-Rogers
Primary Supervisor: Assoc Prof. Rachael Bashford-Rogers
Additional Supervisors: Dr. Isabela Pedroza-Pacheco
Abstract
B and T cell infiltration has prognostic significance in solid tumours, and ongoing studies are investigating their phenotypes
through single cell transcriptomics and spatial imaging. Characterising the B and T cell response to tumour cells, particularly
their antigenic specificities, will be key to developing more immunologically appropriate cancer therapies. Here we propose a
DPhil studentship project to develop novel technologies to be able to bridge the gap between B and T cell population
phenotypes, migration, and antibody/TCR reactivity, thus giving a unique perspective on the development of anti-self, anti-
tumour and anti-non-self-responses. Therefore, this project will characterise the nature, function and migration of myeloid
cells in primary and metastatic tumours, blood and lymph nodes to help direct us towards understanding and potentially
modulating these cells for more targeted therapies. This project will involve the development a novel high-throughput
method to probe the antigen specificities of B and T cells, which will be coupled with single cell resolution of clonal
phenotype and single cell transcriptome. This will be used to investigate the development and role of tumour-infiltrating B
and T cells across a range of tumours with varying degrees of immunogenicity. Single-cell RNA sequencing has emerged as
a powerful tool to investigate cell-to-cell variation (Buettner et al., 2015). Having expertise in developing both novel
experimental and computational methodologies for understanding immune-cell diversity and dynamics (Bashford-Rogers et
al., 2019), as well as disease-specific pathology and genetics, we are in a unique position to combine single-cell RNA
sequencing, clonal tracking through the BCR/TCR, and functional experiments.
Research objectives
This project aims to investigate the B and T cell immune response of circulating and tumour-infiltrating B cells across a range
of tumours with different levels of immunogenicity and neo-antigen expression, including pancreatic and renal cancers. This
will involve the development of a novel platform to answer key questions in the field of tumour immunology including:
• Defining whether B and T cell migration to lymph nodes reflect the response within the primary tumour.
• Determining if tumour-associated B cell subsets produce antibodies against self- or non-self antigen, poly-reactive, or
highly specific to tumour cells, and how does tumour cell neo-antigen variation and expression level dictate antibody
response and protection.
• Determining the balance of whether B and T cells serve a pro- or an anti-tumourigenic function.
• Investigating the spatial distribution of T and B cell anti-tumour clones within tumour tissue and determine their
functional role during tumour immunosurveillance.
• Characterising the molecular mechanisms defining pancreatic B and T cell activation and regulation.
• Building computational models of immunosurveillance of tumours to help stratify patients for therapies and predict
outcome.
This study will provide a unique platform to understand the probe between tumour neo-antigen, B and T cell immune-
surveillance and specificity, and B and T cell phenotype, with the overall aim of highlighting improved therapeutic options and
patient outcomes. Furthermore, the methods developed here will not just be broadly applicable to cancer, but will have wider
applications in immunology and biotechnology. This will be achieved through the development and application of novel
experimental and computational approaches, working in partnership with a global network of clinicians, immunologists and
sample cohorts.
83
The DPhil will gain experience and training in laboratory molecular biology and single cell methods, bioinformatics and
immunology. These include:
• Cell sorting and single-cell RNA sequencing of patient samples.
• Development of novel experimental and computational pipelines for the analysis of novel single-cell datasets.
• Validation of associations using a wide range of immunological techniques.
• The project will work in partnership with a global network of clinicians, immunologists and sample cohorts.
The Bashford-Rogers laboratory has a strong track record of collaboration over the last 15 years and established systems for
co-supervision. The postdoctoral fellows and the PIs will provide daily support and supervision when needed. The student will
learn established protocols for imaging, advanced sequencing, and genomic analysis from existing postdoctoral and DPhil
fellows. The candidate will be expected to participate in the weekly laboratory meetings and encouraged to present research
at national and international meetings.
References
(1) Pseudotime dynamics of T-cells in pancreatic ductal adenocarcinoma inform distinct functional states within the
regulatory and cytotoxic T-cells (iScience, 2023) Ashwin Kumar Jainarayanan; Nithishwer Mouroug-Anand; Edward H.
Arbe-Barnes; Adam J. Bush; Rachael Bashford-Rogers; Adam Frampton; Lara Heij; Mark Middleton; Michael L. Dustin;
Enas Abu-Shah; Shivan Sivakumar
(2) Activated regulatory T-cells, dysfunctional and senescent T-cells hinder the immunity in pancreatic cancer (Cancers,
2021) Shivan Sivakumar, Enas Abu-Shah, David Ahern, Edward H Arbe-Barnes, Nagina Mangal, Srikanth Reddy, Aniko
Rendek, Alistair Easton, Elke Kurz, Michael Silva, Lara R Heij, Zahir Soonawalla, Rachael Bashford-Rogers, Mark R
Middleton, Michael Dustin
(3) Shared D-J rearrangements reveal cell of origin of TCF3-ZNF384 and PTPN11 mutations in monozygotic twins with
concordant BCP-ALL (Blood, 2020) Clara Bueno, Paola Ballerini, Ignacio Varela, Pablo Menendez, Rachael Bashford-
Rogers
(4) Analysis of the B cell receptor repertoire in six immune-mediated diseases. (Nature, 2019) RJM Bashford-Rogers, L
Bergamaschi, EF McKinney, DC Pombal, F Mescia, JC Lee, DC Thomas, SM Flint, P Kellam, DRW Jayne, PA Lyons, KGC
Smith
(5) The integrated genomic and immune landscapes of lethal metastatic breast cancer. (Cell Reports, 2019) Leticia De
Mattos-Arruda, Stephen-John Sammut, Edith M. Ross, Rachael Bashford-Rogers, Erez Greenstein, Havell
Markus,Sandro Morganella, Yvonne Teng, Yosef Maruvka, Bernard Pereira, Oscar Rueda, Suet-Feung Chin, Tania
Contente-Cuomo, Regina Mayor, Alexandra Arias, Raza Ali, Wei Cope, Daniel Tiezzi, Dan Reshef,Elena Martinez,
Vicente Peg, Santiago Ramon y Cajal, Javier Cortes, George Vassiliou, Gad Getz, Serena Nik-Zainal, Muhammed
Murtaza, Nir Friedman, Florian Markowetz, Joan Seoane and Carlos Caldas

Epigenetic control of cancer cell phenotypes via nuclear F-actin based
chromosome motility. 1,2,3 – Prof. Eric O'Neil
Primary Supervisor: Prof. Eric O'Neil
Additional Supervisors: Prof. Yang Shi
Abstract
The hippo tumour suppressor pathway regulates tissue size in development and although the contribution of this
pathway to cancer is evident from tumour models and pan-cancer transcriptomics, somatic mutations are rare4. Our
research has demonstrated how epigenetic silencing of RASSF1A is responsible for YAP activation in human tumours
and correlates with poor survival across all major solid malignancies. Such ‘epigenetic plasticity’ allows dynamic
switching between phenotypes and supports progression of lesions and the appearance of cancer stem-like cells
(CSCs) in solid tumours42. During development, increasing evidence implicates the co-factor YAP1 as a key
determinant of phenotype by supporting pluripotency or differentiation through activation of distinct transcription
programmes in response to RASSF1-hippo signalling5. Moreover, the hippo pathway transduces mechanical forces
from the microenvironment to guide proliferation, stem cell behaviour and differentiation 6. Our recent work has
identified RASSF1A and MST2 reside at the nuclear envelope to sense mechanical force and influence both chromatin
and nuclear actin. This project aims to consolidate these discoveries to understand how genomic motility between
repressive heterochromatin and phase separated transcription factories are controlled by nuclear actin to influence
cell phenotype. We will also explore how nuclear actin influences the stabilisation of cell phenotypes through re-
wiring DNA methylation, specifically 5’hmc by Tet2, to influence differentiation state and clinical outcome in
pancreatic cancer.
Research Objectives
Defining the molecular mechanisms that influence cell fate will allow us to target the epigenetic plasticity behind
tumour heterogeneity, progression and therapeutic resistance.
(ii) EON is an expert in hippo pathway signalling and epigenetics in pancreatic cancer and YS is an expert in epigenetic
control of cell-phenotype in cancer.
This project asks three questions;

• How does RASSF1A regulation of nuclear actin affect chromatin at specific loci associated with pluripotency
or differentiation?
Outcome: an understanding of how nuclear actin guides the movement of specific genes into regions of repressive
chromatin or active transcription.
• Does mechanical force impact ATR-RASSF1A signalling to influence plasticity?

Outcome: defining specific extracellular or cytoplasmic cues that can trigger gene positioning and influence cell
phenotype.
• Can targeting phenotypic plasticity improve therapy in PDAC?

Outcome: Can we promote differentiation in PDAC to improve therapeutic responses and survival?
The potential of precision cancer medicine is limited by therapeutic resistance arising from tumour heterogeneity.
Heterogeneity underpins cancer progression and results from a combination of genomic instability and epigenetic
plasticity; the dynamic alterations of the epigenome responsible for establishing cell phenotype. The tumour
microenvironment governs epigenetic plasticity but exactly how multiple states are generated and maintained
unknown7. Personalised therapies targeting driver mutations are largely circumvented by the presence of genetically
diverse resistant subclones. In contrast, epigenetic plasticity is reversible and an attractive target to prevent resistant
phenotypes appearing or to revert phenotypes of recalcitrant populations (e.g. cancer stem-like cells) to improve
overall therapeutic efficacy. Moreover, as plasticity in tumours can result in genome instability 8, the underlying
alterations may highlight specific vulnerabilities not apparent from genetics alone. To understand how plasticity
85
occurs in tumours, we need to understand how the mechanisms governing cell phenotype are influenced by
epigenetics and microenvironmental cues.
The genome kinases ATM and ATR phosphorylate RASSF1A-Ser131 to influence chromatin, transcription, and DNA
replication. We now know that this influences plasticity and have shown how a SNP in RASSF1 (rs2073498) encodes
a mutation, RASSF1AA133S, that disrupts phosphorylation at Ser131 9, blocks the formation of nuclear actin (preliminary
data) and hinders differentiation. RASSF1AA133S is prevalent in Caucasian populations with a minor allele frequency
(MAF) of ≤0.17 in European cohorts and associates with early onset tumorigenesis in multiple cancers. We generated
Rassf1A133S mice that accelerate pancreatic and colorectal tumour models, supporting the hypothesis that RASSF1A
maintains differentiation and prevents phenotypic plasticity in human tumours. This model gives us the opportunity
to direct model an emerging pathological SNP in humans, while also provide a platform for strategies to intervene in
hyperplastic phenotypic model.
In addition to standard cell culture assays the candidate will receive training in high content and real-time microscopy,
epigenetics (inc ChIPseq, bioinformatics), phase separation and transcription factories, nuclear F-actin filaments etc.
In addition, there are opportunities to explore the in vivo relevance in mouse models of pancreatic cancer.
References:
1. Wu, D. et al. Glucose-regulated phosphorylation of TET2 by AMPK reveals a pathway linking diabetes to
cancer. Nature 559, 637-641 (2018).
2. Eyres, M. et al. TET2 Drives 5hmc Marking of GATA6 and Epigenetically Defines Pancreatic Ductal
Adenocarcinoma Transcriptional Subtypes. Gastroenterology 161, 653-668 e616 (2021).
3. Chatzifrangkeskou, M. et al. RASSF1A is required for the maintenance of nuclear actin levels. EMBO J 38,
e101168 (2019).
4. Harvey, K.F., Zhang, X. & Thomas, D.M. The Hippo pathway and human cancer. Nat Rev Cancer 13, 246-257
(2013).
5. Papaspyropoulos, A. et al. RASSF1A uncouples Wnt from Hippo signalling and promotes YAP mediated
differentiation via p73. Nat Commun 9, 424 (2018).
6. Dupont, S. Role of YAP/TAZ in cell-matrix adhesion-mediated signalling and mechanotransduction. Exp Cell
Res 343, 42-53 (2016).
7. Easwaran, H., Tsai, H.C. & Baylin, S.B. Cancer epigenetics: tumor heterogeneity, plasticity of stem-like states,
and drug resistance. Mol Cell 54, 716-727 (2014).
8. Pefani, D.E. et al. RASSF1A-LATS1 signalling stabilizes replication forks by restricting CDK2-mediated
phosphorylation of BRCA2. Nat Cell Biol 16, 962-971, 961-968 (2014).
9. Yee, K.S. et al. A RASSF1A polymorphism restricts p53/p73 activation and associates with poor survival and
accelerated age of onset of soft tissue sarcoma. Cancer Res 72, 2206-2217 (2012).

Modulation of tumour immunogenicity by IGF’s in Prostate Cancer – 1
Dr Valentine Macaulay
Primary Supervisor: Dr Valentine Macaulay

Additional Supervisors: Professor Tim Elliott
Abstract
Most prostate cancers have a ‘cold’ tumour immune microenvironment (TIME) with exclusion of tumour infiltrating lymphocytes
(TILs) especially CD8+ cytotoxic T cells, defective antigen processing machinery (APM), and many immunosuppressive TILs eg
regulatory T cells (Tregs). Cancer risk is affected by serum IGF-1 (sIGF-1): subjects with very low sIGF-1 show almost complete
cancer protection, while high sIGF-1 increases prostate cancer risk. IGF-1 signals via IGF receptors (IGF-1Rs) to promote cell
survival, invasion and androgen receptor (AR) activation. In other disease models IGF-1 enhances Treg function, while IGF
blockade rescues APM components, inducing CD8 dependent anti-tumour immunity. We hypothesise that IGF-1 contributes to
cancer risk via TIME immunosuppression. Using big data, patient derived explants (PDEs) and samples already obtained from
men on a trial of IGF blockade pre-prostatectomy, we will ask whether IGF-1 affects: 1) tumour intrinsic or extrinsic TIME
properties? 2) T cell phenotype/redistribution in prostate PDEs? 3) TIL localisation or exclusion? The findings may inform therapy
and suggest novel approaches to prostate cancer risk reduction
Research objectives
The aims are informed by data (Fig 1) that the prostate cancer TIME is affected by sIGF-1 and by endogenous IGF-1 (eIGF-1) from
epithelial and stromal cells. Supporting tumour extrinsic effects, sIGF-1 associates with CD4+ TILs (Fig1A-B), IGF blockade reduces
FOXP3+ Treg TILs, and IGF:PD-1 co-inhibition increases CD8+/Treg ratio and GrzB+ CD8+ cells (Fig 1E-F). Tumour intrinsic effects
are suggested by the correlation of high eIGF-1 with low TAP1 and Class I (Fig 1C-D) and by Prostate Adenocarcinoma Firehose
Legacy data (n=500) where high eIGF-1 significantly associates with low TAP1 and TAPBP and high ERAP1 that trims peptides for
presentation by MHC-I (not shown), potentially contributing to immune evasion by altering the peptide repertoire.
Figure 1. IGF-1 influences the TIME. A. Multiplex immunofluorescence (mIF) for pancytokeratin (PCK, epithelium), cytokeratin 5 (CK5, benign
glands), immune markers. B. sIGF-1 associates with CD4+ TILs in prostatectomies of men with low (7.2-12.0 nmol/l) vs high sIGF-1 (21.7-31.5
nmol/l). C. RNAseq from same cases: differentially expressed genes (DEGs) in high eIGF-1 tumours include upregulated IGFBP5 (indicates IGF-1R
activation) and downregulated TAP1 that transports antigenic peptides. D. IGF-1 downregulated TAP1 mRNA in human DU145 and murine Myc-
CaP cells (left), and reduced Myc-CaP cell surface Class I expression after 5 days (right). IGF-1 also upregulated PD-L1 in prostate cancer cells in
vitro (not shown). E-F. Mice bearing Myc-CaP allografts treated with anti-PD-1, anti-IGF xentuzumab or combination. IGF:PD-1 co-inhibition
increased tumour necrosis (E). FOXP3 TIL positivity was suppressed by anti-IGF-1 and PD-1:IGF co-inhibition (F, upper), and co-inhibition increased
CD8+ cell/FOXP3+ Treg ratio and % granzyme B positive CD8+ TILs (F centre, lower).
87
A plausible mechanism through which high IGF-1 promotes prostate cancer risk is therefore:
•High sIGF-1 and/or eIGF-1 upregulate PD-L1 and immunosuppressive cytokines in healthy or premalignant
prostate tissue to promote Treg proliferation/function, resulting in a high tissue Treg setpoint.
•Tregs condition tissue in favour of immune suppression/exclusion, favouring progression to invasive cancer.
•CD8+ T cells are primed by tumour antigens eg those from AR ligand-binding domain (LBD). But in tissues
conditioned to be immunosuppressive, primed TILs are suppressed/excluded, less able to control new cancers.
•IGF-1 deregulates APM components in tumour cells, further contributing to immune escape.
This mechanism suggests a tractable hypothesis to explain how the TIME is shaped by IGF-1:
•Tumour extrinsic effects that maintain Treg function and infiltration, suppressing CD8+ TILs
•Tumour intrinsic effects by regulating the APM, influencing ability of CD8+ TILs to recognise AR derived epitopes
•Thus IGF-1 prevents CD8+ T cell infiltration, localisation to malignant glands and lysis of prostate cancer cells.
The project will establish a new collaboration between Professor Elliott and Dr Macaulay, who will co-supervise
the Clinical Fellow to investigate these hypotheses using the following models, samples and collaborations:
•Highly curated RNAseq data: ~1000 primary
prostate cancers (collaborator: Dr Woodcock,
Pan Prostate Cancer Group, PPCG).
•Prostate scRNAseq data from the Macaulay lab
and Dr Massie, Cambridge [1], with TCR
sequencing to identify cytotoxic subclones and
their expression of IGF axis genes.
•Patient derived explants (PDEs, ref 2) that
maintain morphology, viability for 6 days
(collaborators: Mr Lamb, Profs Verrill, Mills).
•WINGMEN samples including samples from 3
trial patients who were recruited to COMBAT
Cancer (Dr Lamb, Prof Verrill, Dr Woodcock).
Depending on the findings, further investigations could include:

•Murine model to assess TIL numbers, phenotype, intratumoral location. ‘HIT’ mice (from Shoshana Yakar, NY)
harbour a hepatic IGF-1 transgene with ~3-fold higher sIGF-1 [3]. Myc-CaP allografts in WT or HIT mice will be
analysed by mIF, TIL phenotyping and RNAseq with digital cytometry (ImmuCC) to infer immune cell populations.
•Functional assays to investigate effects of IGF-1 or IGF blockade on i) T cell killing and suppression in co-cultures
of prostate cancer cells and T cells, ii) AR antigenic priming and relative ability to prime in immunosuppressive
environment, iii) immunopeptidomics to identify IGF-induced changes in antigen processing and presentation.
Translational potential: identification of a novel MOA for IGF-1 in the TIME has clear relevance to:
•Therapy: Few prostate cancers respond to immune checkpoint inhibition (ICI). If IGF-1 enables Tregs to suppress
CD8+ TILs, IGF blockade may cause Treg suppressed CD8+ cells to be less exhausted and thus rescuable by ICI.
•Cancer risk: immunosuppressive actions of IGF-1 may contribute to the ability of high IGF-1 to enhance the risk
of prostate cancer. Understanding this effect may suggest new approaches to risk reduction.
88
Training provided
The student will be trained in analysis of big data, explant and cell culture, flow cytometry and mIF. If resources
become available, it may be possible to provide training in high-cost techniques to address our hypotheses at the
single cell (scRNAseq, CyTOF) and tissue levels (CellDive, spatial transcriptomics).
References:
1. Tuong et al. Cell Rep 2021. 37:110132.
2. Centenera et al. Mol Oncol 2018. 12:1608-1622.
3. Cannata et al. Endocrinology 2010. 151:5751-61.

89
Understanding the mechanisms by which molecular and phenotypic
heterogeneity of MLL-rearranged infant ALL affects clinical outcome–
1 - Assoc Prof. Anindita Roy
Primary Supervisor: Assoc Prof. Anindita Roy

Additional Supervisors: Philip Ancliff and Jack Batram
Eligibility: Track 1 students are eligible to apply for this project
Abstract
Although there has been much progress in treating acute lymphoblastic leukaemia (ALL) in children, there is still a subset
of ALLs that have a very poor prognosis. This is particularly true for infants (i.e., children <1 year of age). Infant ALL (iALL)
is frequently driven by translocations of the Mixed Lineage Leukaemia (MLL aka KMT2A) gene, which occur in ~80% of
the cases, and represents a very aggressive type of leukaemia that is characterised by chemotherapy resistance and
high relapse rates leading to a very poor prognosis1. The major challenge is that while iALL is aggressive, necessitating
high-intensity treatment strategies, infants are extremely vulnerable to toxicity from chemotherapy and stem cell
transplantation. It is clear that new approaches, underpinned by rigorous science, are required to improve
characterisation, treatment stratification and outcomes. Recent studies have shown the benefit of upfront CD19
directed therapy in iALL2, 3. As there is no ongoing UK clinical trial for iALL treatment, the UK NCRI Leukaemia sub-group
have designed an innovative treatment guideline using Blinatumomab to replace some of the most intensive and toxic
chemotherapy blocks used in past protocols, as well as novel treatment stratification to assign post-blinatumomab
therapy, carefully tailored to the biological features of each patient’s leukaemia. These new guidelines present a unique
and important opportunity to capture comprehensive clinical outcome data. Launching these guidelines as a national
clinical study, accompanied by cutting edge scientific efforts to understand the molecular mechanisms underpinning
the aggressive nature of iALL, will fulfil an unmet need in infant leukaemia care. We have formed a consortium of expert
clinicians, scientists and parent advocates.
Research objectives
Aim 1: Delivering a national clinical study that will provide unique insights into optimal treatment pathways for iALL, to help
inform the scientific and clinical agenda. (P Ancliff/J Bartram)
The national study opened in

2021. All newly diagnosed iALL
patients are eligible for
recruitment. There are c.12-15
cases of iALL in the UK/year.
We expect to recruit >90%
since all cases are presented at
our National Leukaemia MDT.
Expected recruitment is 60
patients over 4 years or
longer, with the total
recruitment period being 48
months. A secure central
database held at GOSH will
capture demographics,
routine diagnostics (e.g.
immunophenotyping, MRD,
genetics), treatment, toxicity,
Figure 1: Schematic of project workflow. Sample collection and experimental and clinical outcome data. All
workflow for the study. Blue boxes and arrows represent protocols that are patients will be assigned a UPN and
covered by existing diagnostic pathways. Orange boxes/arrows represent patient identifiable data will not be
protocols that require funding as part of this application. (BM: Bone marrow, made available to researchers, who
PB: peripheral blood, MRD: minimal residual disease) will only receive linked anonymised
samples. As part of this study, all
90
diagnostic tests (as NHS standard of care) will be streamlined and undertaken in centres of excellence in a standardised
manner (Fig 1).
Aim 2: Detailed characterisations of each patient’s leukaemia, including biomarkers to better understand disease biology and
response to treatment. (collaboration T Milne)
2.1 At diagnosis, all patient BM samples will undergo comprehensive characterisation as part of the diagnostic workup (Figure
1), including: i) detailed immunophenotyping to define the leukaemia-associated immunophenotype (LAIP), including the
expression of myeloid antigens (S Inglott, Oxford and GOSH). This is crucial for follow-up and monitoring, especially detection
of MRD and emergence of CD19-ve ALL relapse post-immunotherapy; ii) State of the art cytogenetics and RNA-fusion panel
for detailed molecular characterization (Newcastle); iii) Ig/TCR rearrangement status by next generation sequencing to
monitor MRD (GOSH). Whole genome sequencing (WGS) will be performed as a standard of care clinical test via the NHS
Genomics Medicine Service/Genomics England, with access to raw data via the National Genomics Research Library.
2.2 To further unravel and understand the mechanisms by which molecular heterogeneity of MLL-rearranged infant ALL
affects clinical outcome, we will perform comprehensive characterisation of diagnostic and relapse samples using multi-omic
profiling. Samples will undergo RNA-sequencing, ATAC-sequencing and ChIP-sequencing for MLL-rearranged protein complex
members, and key histone modifications such as H3K4me3, H3K27ac and H3K27me3 will be used to further define the global
transcriptomic and epigenetic profile of the cells (A Roy/T Milne, Oxford). All work will be undertaken at the Roy/Milne labs
where there is extensive experience of these techniques, including optimised protocols for small cell numbers.
2.3 Once specific molecular profiles/leukaemic subpopulations have been identified, including minor myeloid subclones that
might cause plasticity or relapse; these will be correlated with a variety of clinical parameters and outcome to ascertain
whether they are relevant prognostic biomarkers. To address the challenges of biomarker/target discovery in such a rare
disease we will leverage established collaborations with the Interfant-21 trial (joint scientific committee has been set up, and
additional funding secured through Fight Kids Cancer grant), and the HARMONY alliance (https://www.harmony-alliance.eu/)
to validate identified biomarkers.
Aim 3: Identification of key vulnerabilities that may be targeted for novel therapies. (collaboration T Milne)
We will create and utilise comprehensively characterised primary patient material, in combination with state-of-the-art
preclinical models4, to provide unique insights into novel therapeutic and prognostic targets in iALL.
3.1 Full molecular characterization of iALL will give us the opportunity to study these cases in unprecedented detail at a multi-
omics level. This will allow us to generate gene regulatory networks to identify key molecular pathways perturbed in iALL 5, 6
and how this correlates with clinical outcome. We will combine these findings with results of ongoing funded projects using
CRISPR essentiality screens and molecular and functional screens in mouse models 4 (A Roy, T Milne Oxford).
3.2 We will validate targets by manipulating the expression of genes /pathways identified in 3.1, using in vitro and in vivo
assays of primary iALL patient samples and an iALL model4 (A Roy, Oxford).
Since the last international trial for treatment of infant ALL (Interfant-06) closed in July 2016, infants diagnosed with
ALL in the UK have not been enrolled in any clinical trials. While we will continue to work closely with our European and
international colleagues, the UK is unable to participate in the next international trial (Interfant-21) as it has no
randomisation arm. This leaves a vulnerable patient population without the benefit of a structured clinical trial in one
of the poorest risk leukaemias. The current national study (2021 onwards) will ensure the implementation of risk-
adapted treatment guidelines for all infants with ALL in the UK with the most comprehensive characterisation of the
leukaemia and follow-up to date. We anticipate that patients will benefit immediately by the introduction of novel
therapy approaches in our UK-NCRI-LSG guidelines, but it is imperative to capture real-life clinical data to learn about
the efficacy and toxicity of these newer therapies compared to historic treatments. The study will allow us to monitor
these treatment responses as well as investigate the biological mechanisms, which underpin its success or failure,
including finding predictive biomarkers. This will lead to better stratification and a personalised medicine approach,
which is a longer-term goal of this study. In addition, the phenotypic and molecular dataset generated will be an
invaluable resource to clinicians and scientists treating patients or studying the biology of iALL. The most important
91
feature of this project that will benefit patients is the true integration of clinical and scientific strategies in order to
accelerate the development of effective therapeutic options.
The student will be embedded in the excellent scientific environment of the WIMM. He/she will be supported by
experienced postdoctoral researchers in the lab, as well as collaborating labs (e.g. Milne) to learn cutting edge
functional/molecular techniques such as flow cytometry, in vitro and in vivo assays, RNA-sequencing, ATAC sequencing,
ChIP-sequencing, and bioinformatics. All of these techniques are already in use in our labs; and additional training
opportunities will be available for computational analysis (via dedicated courses at the Centre for Computational
Biology, WIMM). The student will be encouraged to take up the excellent training opportunities provided within the
WIMM and the University of Oxford, to develop their research career. In addition they will gain experience in running
a national clinical study including collating all clinical outcome and toxicity data under the guidance of expert clinical
trial leads at GOSH.
References
1.Pieters, R., et al., J Clin Oncol, 2019. 37(25): p. 2246-+. 2. van der Sluis, I.M., et al., N Engl J Med, 2023. 388(17): p.
1572-1581. 3. Clesham, K., et al., Blood, 2020. 135(17): p. 1501-1504. 4. Rice, S., et al., Nat Commun, 2021. 12(1): p.
6905. 5. Godfrey, L., et al., Leukemia, 2021. 35(1): p. 90-106. 6. Harman, J.R., et al., Genome Res, 2021.

92
Describing T Cell recognition of tumours by machine-learning and
statistical models 4 – Assoc Prof. Hashem Koohy
Primary Supervisor: Assoc Prof. Hashem Koohy
Additional Supervisors: Prof. Alison Simmons
Abstract
T cell recognition of a cognate peptide-MHC (pMHC) complex presented by infected/malignant and/or specialized antigen
presenting cells is crucial for orchestrating a robust and sustained immune response. The molecular interactions between
heterodimeric T Cell Receptors TCRs and pMHC ligands dictate the nature of the subsequent adaptive immune response.
However, the underlying rules governing this engagement remain incompletely understood (1). A better understanding of
TCR:pMHC interactions would allow further harnessing of adaptive T cell immunity, to assist in the development of vaccines
and therapeutics and/or preventive strategies. This research in Koohy’s group is focused on development of machine-learning
and statistical models to help a better understanding the two key components of the interaction, namely the architecture and
composition of the immune repertoire during the course of disease or treatment, and b) identifying the immunogenic T cell
epitopes that can be used as targets for vaccines or immunomodulatory treatments. In this project, we will be investigating
the extent to which T cells are involved in driving intestinal inflammation after Immune Checkpoint Blockade (ICB), and more
importantly, the source of their antigen specificity. Additionally, we will be further exploring the statistical characteristics of T
cell targets in cancers to derive accurate in silico predictions of cancer neoantigens (2).
Research objectives
The development of Immune Checkpoint Blockade ICB treatment for cancer, has revolutionized patient care. However, not all
patients respond uniformly, and a subset experience immune-related adverse events irAEs, such as Checkpoint Induced Colitis
CIC. Gastrointestinal inflammation, including CIC, affects a substantial proportion of patients treated with ICB, leading to
treatment discontinuation. The underlying mechanisms of CIC remain poorly understood, although emerging evidence
suggests dysregulated activation of T cells, potentially triggered by self-antigens or commensal microbes. Our ongoing
collaborative project with Prof Alison Simmons and Dr Agne Antanaviciute has revealed clonally expanded CD8 T cells
trafficked from blood to inflamed tissue in CIC patients. However, the antigen specificity of these cells remains elusive.
Therefore, this proposal aims to identify dysregulated T cells and uncover the main source of their antigen specificity in CIC
patients.
In a broader context, this project also focuses on modelling and identifying biomarkers of response to ICB. Numerous factors
influence the immune response to treatment, including the mutation burden, cytotoxic T cell infiltration, and defects in
antigen processing and presentation among others. Recent advances in high-throughput sequencing technologies particularly
single cell techniques, enable the measurement of these features at the single cell level in patient samples at various time
points, spanning pre-treatment, treatment and post-treatment. To study the mechanisms underlying response heterogeneity,
we employ statistical and machine learning models trained on high thought put sequencing data encompassing genomics,
transcriptomics, T cell receptor repertoire and epigenomic data. However, due to limited data availability, many potentially
predive features have not yet been integrated into statistical and computational approaches. Consequently, the scientific
community requires more sophisticated methods capable of modelling confounding factors and leveraging comprehensive
patient data.
Within our research group, we capitalize on the expertise of in-house specialists to generate diverse types of data and develop
next generation mathematical and computational models for exploring the role of aforementioned factors in response
heterogeneity. Additionally, we aim deorphanizing T cell receptor from cancer patients and gain insights into the underlying
rules governing T cell interaction with tumours.
Although cancer immunotherapies have revolutionized cancer treatment, not all patients respond equally and about 60%
patients develop gastrointestinal inflammation. This project may inform the development of novel biomarkers predictive of
inflammation and more broadly inform the development of treatment approaches requiring tailoring of peptide TCR
interactions.
93
The project offers training in T cell immunology, single cell technologies, their applications in immunology as well as their
specific computational challenges. Successful delivery of the project will also require a good understanding of machine-
learning and statistical inference preferably Bayesian Statistics. Training for both immunology and modelling will be provided
by ourselves in the unit or within the Oxford University teaching and training schemes. We closely work with our collaborators
including: Prof Alison Simmons3,4,5 and Dr Agne Antanaviciute to leverage the data from high-throughput technologies to
investigate the molecular mechanisms of checkpoint colitis (3) and to study the heterogeneity and composition of intestinal
antigen experienced immune cells mirrored by the composition of intestinal microbiome and their relevance to the outcome
of the immune checkpoint blocked treatment. This is therefore a great opportunity to learn more about the applications of
cutting-edge experimental single cell and spatial transcriptomics in action in addressing immunological and medical problems.
References
1. Hudson D, Fernandes RA, Basham M, Ogg G, Koohy H. 2023. Can we predict T cell specificity with digital biology and
machine learning? Nat Rev Immunol: 1-11
2. CH Lee JH, PR Buckley, M Jang, MP Pinho, RA Fernandes, A Antanaviciute, A Simmons, H Koohy. 2022. A robust deep
learning plafform to predict CD8+ T-cell epitopes. BioRxiv
3. Corridoni D, Antanaviciute A, Gupta T, Fawkner-Corbett D, Aulicino A, Jagielowicz M, Parikh K, Repapi E, Taylor S,
Ishikawa D, Hatano R, Yamada T, Xin W, Slawinski H, Bowden R, Napolitani G, Brain O, Morimoto C, Koohy H, Simmons A.
2020. Single-cell atlas of colonic CD8(+) T cells in ulcerative colitis. Nat Med 26: 1480-90
94
Urological cancers beyond the microscope; novel multiomic analysis
of features associated with DNA instability and the tumour immune
micro-environment 1,2,3,4 – Assoc Prof. Clare Verrill
Primary Supervisor: Assoc Prof. Clare Verrill
Additional Supervisors: Prof. Ian Mills
Abstract
Pathology is the study of disease and histopathologists analyse tissue samples to make diagnoses and assessments of how
aggressively cancers might behave. Despite extensive specialist expertise and international efforts to optimise these
assessments, they are inherently limited by the performance of human based observers. Current human based grading
systems are imperfectly predictive and prognostic of disease behaviour. In addition, despite scientific advances stressing the
importance of the tumour immune microenvironment, this feature is not routinely assessed as part of the diagnostic process,
leaving large amounts of valuable information untapped.
Combinations of novel and powerful technologies such as spatial transcriptomic and AI image-guided whole genome
sequencing have the potential to unlock new insights into disease biology. Some urological cancers lag behind many other
cancer types in terms of understanding of disease behaviour, microenvironment and molecular drivers, consequently routine
molecular testing is generally not currently undertaken as with other tumour types e.g. lung (ALK-1, EGFR).
Kidney and bladder cancers (renal cell carcinoma and urothelial carcinoma respectively) receive relatively little research
investment and have poorer outcomes than many other cancers. These cancers are now eligible for immunotherapies such
as PDL-1 inhibitors, but with variable success, which may in part be due to a lack of understanding of the tumour immune
environment.
In this project we will undertake highly detailed multi-omic profiling (Nanostring GeoMx, AI image guided whole genome
sequencing) of up to 20 kidney and bladder cancers. This discovery data set will be mined for the most scientifically valuable
areas of the tumour and immune microenvironment and the most useful markers of particular populations of immune cells.
In spite of their utility in providing biological insights, detailed multi-omic profiling is not feasible at scale. Hence we need to
translate these into pipelines that can be resourced routinely, such as immunohistochemistry or even from morphological
changes detected by AI on readily available H&E sections. We aim to test the most promising image analysis and
immunohistochemical targets in a larger cohort of up to 500 cancers.
Ultimately this will enable us to look at the relationship between genomically unstable cancer cells, other cancer cells and the
immune environment.
Research objectives
Academic value
i) Discovery - Detailed genomic, transcriptomic and image analysis AI profiling of up to 20 kidney and bladder cancers (renal
cell carcinoma and urothelial carcinoma respectively) with a particular emphasis on morphology and the immune
microenvironment. Identify the scientifically most valuable areas within cancer cells and the immune cells themselves.
ii) Testing - Select from the discovery big data set, a panel of the scientifically most useful immunohistochemical immune or
other tumour markers that indicate which particular populations are present and apply to up to 500 kidney and bladder
cancers.
iii) Prediction and validation - Create AI based image analysis signatures (proxies) from H&E alone of the scientifically most
valuable morphological changes and immune populations and test in a new validation cohort.
Cohorts with clinical outcomes, samples, patient consent and ethical approvals are all available and ready to use via previous
initiatives.
Outcomes
i) Big data discovery set. Up to 20 kidney and bladder cancers with detailed -omic profiling using the Nanostring spatial
transcriptomic platform, image guided low-pass whole genome sequencing (for copy number unstable cancer cells) and AI
95
based image analysis. This will guide the selection of a targeted panel of immunohistochemistry that will be undertaken on
a larger scale (up to 500 cases) and analysed together with the routinely available morphology from H&E stains.
ii) From (i) we identify the scientifically most valuable regions of interest in the cancers and including in the immune
microenvironment and evaluating the inter-play between cancer and immune cells. We then aim to translate this into H&E
based morphological signatures derived by image analysis AI that could be translated into clinical use at scale.
Collaborators
In this project, the DPhil candidate will work with a unique team of urological histopathologists (Verrill), scientists (Mills,
Rao), data scientists (Woodcock) and engineers with international expertise. The team have a number of established
collaborative projects in this field and have created an exciting programme of work around histogenomic associations linking
novel AI based image analysis with molecular based genomic and transcriptomic sequencing (bioRxi). Issa and Hester (TRIG)
bring expertise in the tumour microenvironment. Mills is a highly experienced DPhil supervisor.
The programme leverages significant investment made via the AI Imaging Centre of Excellence ‘’PathLAKE’’ with Oxford
hosting one of the UK’s first fully digital NHS histopathology laboratories with live clinical AI technologies [ref press release].
The project also builds on cohorts and collaborative work with industry partner Janssen Biotech Inc. (Cartography) and
pump priming funds from the University of Oxford Medical and Life Sciences Translational Fund.
The project will create novel objective and quantitative ways of analysing urological cancers above and beyond that which can
be achieved with human-based pathological assessments which are largely qualitative and inherently subjective. In addition,
there will be particular emphasis on the immune microenvironment which is currently not analysed routinely in these cancers.
Ultimately the findings will be evaluated by the end of the DPhil for the most promising which in the medium to longer term
can be further developed and leverage further funding via market opportunities or funding bodies.
The successful candidate will gain a detailed understanding of urological cancer histopathology (Verrill). They will be able to
gain laboratory experience within histopathology (sectioning, immunohistochemical staining), whole genome sequencing and
bioinformatic analysis (Rao), spatial transcriptomic experimental design and analysis (Issa, Hester), integration of diverse
datasets and machine learning (Woodcock) and deep learning for image analysis.
References
1) Srinivasa Rao, Verrill Clare. et al: Intra-prostatic tumour evolution, steps in metastatic spread and histogenomic
associations revealed by integration of multi-region whole genome sequencing with histopathological features.
bioRxiv 2023.02.27.530113; doi: https://doi.org/10.1101/2023.02.27.530113
2) Chatrian A, Colling RT, Browning L, Alham NK, Sirinukunwattana K, Malacrino S, Haghighat M, Aberdeen A, Monks A,
Moxley-Wyles B, Rakha E, Snead DRJ, Rittscher J, Verrill C. Artificial intelligence for advance requesting of
immunohistochemistry in diagnostically uncertain prostate biopsies. Mod Pathol. 2021 Sep;34(9):1780-1794. doi:
10.1038/s41379-021-00826-6. Epub 2021 May 20.
3) Haghighat M, Browning L, Sirinukunwattana K, Malacrino S, Khalid Alham N, Colling R, Cui Y, Rakha E, Hamdy FC, Verrill C,
Rittscher J. Automated quality assessment of large digitised histology cohorts by artificial intelligence. Sci Rep. 2022 Mar
23;12(1):5002. doi: 10.1038/s41598-022-08351-5. PMID: 35322056; PMCID: PMC8943120.
4) Prostate cancer AI diagnosis tool begins evaluation in Oxford - Oxford University Hospitals (ouh.nhs.uk)
5) Oxford expands the Cartography collaboration with Janssen — University of Oxford, Medical Sciences Division
96
Investigating the role of the ubiquitin ligase BIRC6 in aneuploid
glioblastoma cell survival 1,2,3 – Dr Paul Elliott
Primary Supervisor: Dr Paul Elliott
Additional Supervisors: Dr Vincenzo D’Angiolella
Abstract
Glioblastoma is a tumour with a dismal prognosis and few therapeutic options. The tumour is characterised by profound
heterogeneity, fuelled in part by the presence of rampant genomic instability. Chromosomal instability is very frequent in
glioblastoma cells and, so far, targeting aneuploid cells has proven difficult. A recent study identified co-essentiality of the
ubiquitin ligases UBA6 and BIRC6 in a subset of tumours with high aneuploidy through a role in the integrated stress response
and cell death (1). However, detailed understanding of mechanisms underpinning this dependency on BIRC6 and UBA6 are
not known.
BIRC6 is a large (~5000 amino acid) E2/E3 ubiquitin ligase of the Inhibitor of Apoptosis family. Our recent structural and
biochemical work has uncovered BIRC6 directly restricts executioner caspase-3 and -7 and through working exclusively with
the non-canonical E1 UBA6, ubiquitinates caspase-3,-7 and – 9 and other substrates (2). Thus, BIRC6 seems to be a crucial
sensor of aneuploidy to balance cell death decisions in aneuploid cells. The DPhil project will investigate how BIRC6 supports
aneuploidy cancer cell survival particularly focussing functional biological studies on glioblastoma (3). The project will
uncover the BIRC6–substrate interactome in glioblastoma and will biochemically, structurally and functionally characterise
these interactions providing key insights into how BIRC6 enables aneuploid glioblastoma cell survival, crucial for informing
novel future glioblastoma therapeutic strategies.
Research objectives
Objective 1: Uncovering BIRC6 regulated proteome in glioblastoma cell lines
To gain a mechanistic understanding of how BIRC6 enables aneuploid glioblastoma cell survival, we will determine the
proteome regulated by BIRC6 in four glioblastoma cell lines chosen for differing levels of BIRC6 dependency and aneuploidy
level. Through cutting-edge proteomics combined with Tandem Mass Tag (TMT) labelling, we will quantitatively compare
complete proteomes of these cell lines upon knock-down of BIRC6. Gene ontology analysis will be used to identify pathways
in which differentially regulated proteins are involved and thus regulated by BIRC6 in glioblastoma. Proteins that display
most significant upregulation in BIRC6-dependent glioblastoma will then be the focus for the downstream objectives in
which we will establish how each substrate identified contributes to regulate survival of glioblastoma cells.
97
Objective 2: Validation of BIRC6 substrates in glioblastoma cell lines
The most significantly upregulated proteins identified in Objective 1 will be assessed for capacity as BIRC6 substrates. This
will be achieved through cycloheximide chase assays testing either endogenous substrates or epitope tagged substrates
ectopically expressed at near endogenous levels to confirm substrate stability upon BIRC6 depletion. Candidate substrates
identified will then be verified for interaction with BIRC6 in cellulo through immunoprecipitation methods. Further to these
experiments, substrate degradation via ubiquitination will be confirmed through use of Tandem Ubiquitin Binding Entities
(TUBEs) to immunoprecipitate ubiquitinated substrates followed by ubiquitin linkage type identification using established
chain selective deubiquitinase profiling methods employed in the Elliott Lab. The substrates identified will also be analysed
for their role in controlling DNA damage, chromosomal mitotic segregation and apoptosis. While we have chosen
transformed cancer cell lines as initial models, we will confirm the findings in glioblastoma stem cell models with basal
aneuploidy and/or induced aneuploidy, well established in the D’Angiolella Lab. Non-transformed neural stem cells will also
be used to establish dependency of BIRC6 in normal tissues.
Objective 3: In vitro characterisation of BIRC6 substrates
Confirmed substrates will then be cloned, recombinantly expressed and purified through either E.coli or Sf9 expression
systems. Recombinant substrates will be tested for ubiquitination by BIRC6 in established in vitro substrate ubiquitination
assays using recombinant E1 ubiquitin ligase (UBA6) and fluorescently-labelled ubiquitin. The interactions between BIRC6
and novel substrates will then be characterised through structural biology methods including cryoEM and biophysical
techniques well established in the Elliott Lab. Structure-function insights will be explored further through use of existing
BIRC6 mutations where relevant or design of new structure-guided mutations.
Outcomes: This DPhil project will determine the BIRC6 substrate interactome in glioblastoma and will uncover how BIRC6
substrate specificity is achieved and how ubiquitination of these substrates enables glioblastoma cell survival. Together this
project will provide the first detailed mechanistic insights into BIRC6 cellular function in aneuploid glioblastoma survival
This project will provide crucial detailed insights into how BIRC6 supports aneuploid glioblastoma cell survival. We have the
longer-term vision of translating our cellular, structural and biochemical knowledge obtained to develop small molecule
compounds that disrupt BIRC6 interaction with identified substrates for therapeutic use. Furthermore, considering most
cancers feature aneuploidy, BIRC6 and its substrate interactome may be key molecular anti-cancer targets; therefore,
through this project, the detailed cellular and molecular insights into understanding how BIRC6 is exploited in aneuploid
glioblastoma will be important for informing cancer therapies more broadly.
This DPhil project provides an exciting opportunity to combine cancer cell biology methods with mass spectrometry and
biochemistry, biophysics and structural biology techniques. The student will join a dynamic team and will develop a broad
range of skills including: proteomics analysis methods, protein purification techniques, in vitro biochemical ubiquitin-related
assays, structural biology (cryoEM) and, through working closely with Dr D’Angiolella’s Laboratory, glioblastoma cancer cell
biology methods. The student will also have multiple opportunities to present their findings at inter-departmental seminar
series and national and international conferences.
References
1. Cervia, L. D. et al. A ubiquitination cascade regulating the integrated stress response and survival in carcinomas. Cancer
Discov (2022) doi:10.1158/2159-8290.cd-22-1230.
2. Dietz, L. et al. Structural basis for SMAC-mediated antagonism of caspase inhibition by the giant ubiquitin ligase BIRC6.
Science 379, 1112–1117 (2023).
3. Chen, Z. et al. A Human IAP-Family Gene, Apollon, Expressed in Human Brain Cancer Cells. Biochem Bioph Res Co 264,
847–854 (1999).

98
Characterising the NK/Myeloid crosstalk during tumour immune
escape 3 – Assoc Prof. Audrey Gérard
Primary Supervisor: Assoc Prof. Audrey Gérard
Second Supervisor: Claudia Monaco
Abstract
Cancer cells can be recognised and killed by our immune system. However, tumours developed strategies to evade the immune
system by inhibiting the main immune cell type that can kill them, called CD8 T-cells. In recent years, immune therapeutics
emerged to counteract this, called checkpoint blockade. This treatment aims to reinvigorate CD8 T-cells and has shown
unprecedented success in treating aggressive cancers. However, some patients who initially responded to checkpoint blockade
become refractory. One of the reasons is because cancer cells once again evade the attack induced by the CD8 T-cells reactivated
by checkpoint blockade. Often, they do so by mutating and blocking the effect of a specific immune mediator, IFNγ. This project
aims to understand whether the immune system can adapt to those mutated tumour clones, and whether we can leverage this
to reinvigorate the immune system yet again. Our data show that immune cells other than CD8 T-cells become activated
following this secondary escape. They are monocytes and NK cells, and importantly, they also have the potential to kill cancer
cells. We will study the crosstalk between monocytes and NK cells to understand if it is beneficial or detrimental to anti-tumour
immunity. To do this, we will recreate IFNγ-dependent tumour escape and explore when the NK/monocyte crosstalk arises,
where it happens and the effect of manipulating this crosstalk on the anti-tumour response. The goal is to understand how to
leverage this newly generated immune response to provide alternative strategies to control tumours.
Research objective
Tumours actively escape the immune system by inducing an immunosuppressive state where intra-tumoural CD8 T-cells are
“exhausted”, lacking effector functions. Checkpoint blockade is a therap designed to reinvigorate those exhausted T-cells. It has
shown unprecedented success in treating aggressive cancers such as metastatic melanoma 1. But the response rates are still only
15 to 30%, in part because tumours have developed escape strategies checkpoint blockade. The most reported pressure in this
context is induced by CD8 T-cells. Patients who progressed after initially responding to checkpoint blockade often carry tumours
deficient in IFNγ signalling pathway2. IFNγ exerts cytotoxic or cytostatic effects on tumours and induces Major histocompatibility
complex (MHC) expression. MHC expression is necessary for CD8 T-cells to recognise and kill tumour cells. Therefore, down-
regulation of the IFNγ pathway leads to tumour escape by inhibiting T-cell recognition. How the immune system responds to this
has been largely overlooked. They however have the capacity to do so. For example, we recently found that CD8 T-cells sense
Figure 1: Immune adaptation to IFNγ-dependent tumour escape. A- Control (WT-Zsgreen) and IFNγRKO (KO-mCherry) B16 melanoma tumour
cells, were ad-mixed, as indicated, and engrafted in WT mice. (B-C) After 12 days, tumours were harvested, and single cell suspension
obtained. Graph shows the ratio between the different tumour types, quantified by flow cytometry in WT (B) and CD8-deficient (C) mice.
Every dot is a mouse. D- We performed single cell RNA sequencing on immune cells from WT and KO tumours. The graph displays the putative
strength of interactions between immune cell subsets. Cells linked by blue lines represent communication observed in WT tumours. Cells
linked by red lines represent communications observed in KO tumours. E- Ad-mixed WT and KO tumours were fixed and imaged by confocal
microscopy. Macrophages (purple) were stained with F4/80 antibodies.
IFNγ present in the tumour microenvironment, which restricts their anti-tumour response3. The paradigm is that immune cells,
in particular T-cells, are inhibited or simply can no longer recognise cancer cells. To what extent immune cells adapt to tumour
escape and whether this can be leveraged, is unknown. In this context, NK cells have emerged as a potential immune target to
trigger an immune response against tumour cells that downregulated the expression of MHC molecules. But we don’t know how
cancer cells and the immune microenvironment affect NK cell fitness.
99
To recapitulate escape, we use a simple system where we engraft mice with an ad-mix WT tumour clones and “escapee” tumour
clones that do not express the IFNγR (KO). Each clone expresses a different fluorescent protein (ZsGreen or mCherry) to track
them (Fig.1A). In this model, KO tumour clones indeed escape and overtake WT clones (Fig.1B). This escape does not happen in
mice that do not produce CD8 T-cells (Fig.1C), as CD8 T-cells kill WT tumours rather than KO tumours. However, NK cells, which
have been suggested to take over and kill KO tumours, do not seem to do so efficiently. Sequencing of the immune
microenvironment of WT and KO tumours demonstrates that escape is characterised by the emergence of crosstalk between NK
and Macrophages (red lines, Fig.1D). Most macrophages in tumours are inhibitory, and we previously visualised their pro-tumour
function during metastasis4. The overall hypothesis is that the NK/macrophage crosstalk inhibits NK cell fitness, and that we can
find strategies to target this crosstalk to improve anti-tumour immunity against escape mutants.
Research Objectives:
i) Explore the emergence of the NK/Macrophage interplay as escape occurs and its potential implication on escape efficiency.
You will harvest tumours at different time points and correlate the phenotype of NK and myeloid cells with the “extent” of
tumour escape. Tumour escape can be quantified by tumour growth and the ratio between WT and KO tumours either by flow
cytometry or using IVIS spectrum in vivo imaging. Cell phenotype can be assessed by high-dimensional flow cytometry. You will
also deplete NK cells or myeloid cells and study how they influence tumour escape.
ii) Investigate the co-dependency and mediators of the NK/Myeloid crosstalk. Using ex vivo functional assays, you will study NK
cell fitness over time, as escape occurs, and the relevance of myeloid cells for NK functions. Using our already generated
transcriptomics dataset, you will identify and test candidate pathways that can interfere with NK differentiation and function.
For example, you will test known mediators of NK/Myeloid cell communication such as IL12/18/15, based on our preliminary
analysis of transcriptomics data pointing to IL18 as a mediator of NK/Myeloid cell communication in IFNγRKO tumours.
iii) Explore the spatial relationship between tumour escapees and the immune response to understand how it fuels tumour
escape. Newly generated data in the lab demonstrate that IFNγ-dependent escape leads to segregation between WT and KO
cancer cells. We hypothesise that KO and WT regions are populated by distinct immune states, potentially contributing to KO
tumour escape. Preliminary analysis shows that Macrophages are located at the margin of the tumour (Fig.1D). In addition, our
scRNAseq data suggest that NK cells and Macrophages attract each other and reside in the same niche, excluding NK cells from
the core of the tumours. We will define the localisation of NK cells and different myeloid populations in WT, KO, or admixed
tumours. To analyse immune cells of interest in different regions, you will use confocal microscopy and light-sheet microscopy,
available at the Kennedy Institute. We hypothesise that NK cells are colocalised with Macrophages at the periphery of tumours,
restricting their action.
Outcome: This project will unravel the relevance of the NK/Macrophage crosstalk for tumour escape, identifying key pathways
by which immune cell sense and adapt to changes in the tumour landscape. This will uncover potentially targetable new
mechanisms by which tumours respond to immune pressure.
This project will help us understand the relevance of the NK/myeloid interplay on tumour escape and whether we can leverage
our acquired knowledge on this pathway to counteract IFNγ-dependent tumour escape. The recent discovery that down-
regulation of the IFNγ pathway is a major escape mechanism secondary to strong immune pressure makes this project timely
and relevant. Understanding how IFNγ reshapes anti-tumour responses will be key to offer new strategies to overcome immune
escape. NK cells are often found in small numbers in tumours, and therefore have not received much attention. But they emerge
as key to control MHClow tumours, and we need to better understand how to increase their homing to tumours and which
signals lead to their inhibition. This is important, and multiple cellular therapies aim to use NK cells.
The student will be based at the Kennedy Institute of Rheumatology which is a world-renowned institute and is housed in a
state-of-the-art research facility. This project provides broad training in cancer biology and immunology covering a range of
cellular, molecular and functional immune assays. Students have access to cutting-edge technologies such as disease mouse
models of cancer, NK/Monocytes co-cultures systems, multiplex imaging.
References
1. Iwai, Y., Hamanishi, J., Chamoto, K. & Honjo, T. Cancer immunotherapies targeting the PD-1 signaling pathway. J Biomed Sci 24, 26 (2017).
2. Gao, J. et al. Loss of IFN-gamma Pathway Genes in Tumor Cells as a Mechanism of Resistance to Anti-CTLA-4 Therapy. Cell 167, 397-404 e399
(2016). 3. Mazet, J.M. et al. IFNgamma signaling in cytotoxic T cells restricts anti-tumor responses by inhibiting the maintenance and diversity
of intra-tumoral stem-like T cells. Nat Commun 14, 321 (2023). 4. Headley, M.B. et al. Visualization of immediate immune responses to pioneer
metastatic cells in the lung. Nature 531, 513-517 (2016).
100
The impact of Hypoxia on HLA-E surface expression and peptide
presentation3 – Assoc Prof. Geraldine Gillespie
Primary Supervisor: Assoc Prof. Geraldine Gillespie
Second Supervisor: Prof. Jane McKeating
Abstract
Human leukocyte antigen E (HLA-E) is a non-classical MHC protein whose primary function involves the regulation
immunity via an NK cell axis 1. This system involves HLA-E binding to conserved 9 amino acid leader sequence peptides
(termed ‘VL9’) from classical major histocompatibility complex class I (MHC Ia) proteins 2. Once expressed at the cell
surface, HLA-E-VL9 complexes are recognised by inhibitory CD94-NKG2A receptors expressed on NK cell. When MHC Ia
expression is absent or reduced, during viral infections or following malignant transformation to evade T cell recognition,
limited VL9 leader peptide availability reduces HLA-E-VL9 surface expression, causing loss of NKG2A binding, and
subsequent NK cell-mediated attack. However, there are contexts where HLA-E can present unusually diverse, non-VL9
peptides to CD8+ T cells, suggesting an unanticipated role for HLA-E in the context of T cell immunity 3,4. This previously
undiscovered role is exciting and offers potential for ‘universal’ vaccine development as only two functional allotypes of
HLA-E in humans have been reported in contrast to the enormous diversity of MHC Ia allotypes. Should HLA-E share the
same ability to target broad, sequence diverse peptide repertoires, it offers an alternative and exciting platform for novel
therapeutic strategies, especially for tumours where HLA-E is up-regulated. One of the many important factors affecting
the physiology of cancer cells is the local oxygen tension. Hypoxia or low oxygen levels are a hallmark of solid tumours
and associate with poor prognosis. A hypoxic microenvironment enhances the proliferation and invasiveness of tumour
cells, impairs drug delivery and promotes evasion of host immune responses. Hypoxia can negatively regulated MHC Ia
expression, however, the impact on HLA-E has not been explored.
Research objective
The tumor microenvironment comprises the tissue surrounding and interacting with the tumor such as extracellular
matrix, vasculature, stromal cells, and immune cells, as well as the oxygen concentration and pH levels. Hypoxia develops
as a proliferating tumor outgrows its surrounding vasculature leading to reduced oxygen levels. Hypoxia-inducible factors
(HIFs) regulate oxygen homeostasis and activate the transcription of many genes involved in angiogenesis, metastasis, and
immune suppression. To be recognized by CD8+ T cells, antigens need to be processed by the antigen processing and
presentation machinery and presented on the cell surface in complex with a MHC Ia molecule. Hypoxia has been reported
to regulate MHC Ia in a cancer type-specific manner with some studies reporting an upregulation in melanoma and
colorectal cancer cell lines 5,6. In contrast other studies report a hypoxic mediated down-regulation of MHC-I in sarcoma
and pulmonary tumor mouse models 7, non-small-cell lung cancer models 8 and multiple myeloma 9. A single report
reported that hypoxia increased HLA-E 10, highlighting a need for further studies. We hypothesize that a hypoxic tumour
microenvironment will affect the cell surface expression and/or alter the peptide presentation profiles of HLA-E on
cancer cells.
This project will examine the effect of hypoxia on HLA-E expression in various cancer cell lines such as human papilloma
virus associated cervical cancer cell lines, hepatitis B virus infected human hepatoma cells with viral integrants. In
collaboration (Prof. N. Ternette) we will explore the impact of hypoxia on the HLA-E peptide repertoire using mass
spectrometry. Parallel experiments will study the efficacy of pre-defined HLA-E restricted T cells recognition of tumour cell
lines (for which we have previously defined cancer epitopes) under hypoxic conditions. We have access to a range of
pharmacological agents and licensed drugs that modify HIF activity to define underlying mechanisms. An understanding
of the relationship between HLA-E and hypoxia, and the molecular biology underlying altered expression would provide a
significant advance in our understanding of a topic that is currently neglected, but crucially important from a cancer-
targeting perspective.
The near monomorphic nature of HLA-E combined with its dysregulation on certain cancer cells highlights its potential as
an immunotherapeutic target. Understanding the impact of hypoxia on HLA-E and the peptide repertoire presented could
facilitate targeting by peptide specific T cells or by antibodies that recognise peptide-HLA-E (TCR ‘mimics’). If instead, HLA-
E is downregulated, this information could advance targeting of cancer cells via NK cell-based approaches. Either of these
approaches could be combined with pharmacological agents that modify HIF activity to amplify the relevant immune-
101
therapeutic targeting. Elucidating the effect of hypoxia on HLA-E surface expression and antigen presentation provides
new therapeutic opportunities to harness and regulate different arms of the immune response.
This project will be co-supervised by Geraldine Gillespie and Jane McKeating who provide complementary expertise in
HLA-E biochemistry, structural biology, T cell immunology, molecular virology and hypoxia biology. The interdisciplinary
project will provide a unique training environment in cancer immunology and range of techniques will be offered including
molecular, biochemical and in vitro models of T cell suppression of cancer cells. Transferable skills including oral
presentations at joint lab meetings, critical reviewing of published scientific literature by contributing to journal clubs and
scientific writing by reviewing and drafting manuscripts for publication.
References
1 Braud, V. M. et al. HLA-E binds to natural killer cell receptors CD94/NKG2A, B and C. Nature 391, 795-799,
doi:10.1038/35869 (1998).
2 Braud, V., Jones, E. Y. & McMichael, A. The human major histocompatibility complex class Ib molecule HLA-E binds
signal sequence-derived peptides with primary anchor residues at positions 2 and 9. Eur J Immunol 27, 1164-1169,
doi:10.1002/eji.1830270517 (1997).
3 Joosten, S. A. et al. Mycobacterium tuberculosis peptides presented by HLA-E molecules are targets for human
CD8 T-cells with cytotoxic as well as regulatory activity. PLoS Pathog 6, e1000782, doi:10.1371/journal.ppat.1000782
(2010).
4 van Hall, T., Oliveira, C. C., Joosten, S. A. & Ottenhoff, T. H. The other Janus face of Qa-1 and HLA-E: diverse peptide
repertoires in times of stress. Microbes Infect 12, 910-918, doi:10.1016/j.micinf.2010.07.011 (2010).
5 Kukita, K. et al. Cancer-Associated Oxidase ERO1-alpha Regulates the Expression of MHC Class I Molecule via
Oxidative Folding. J Immunol 194, 4988-4996, doi:10.4049/jimmunol.1303228 (2015).
6 Kajiwara, T. et al. Hypoxia augments MHC class I antigen presentation via facilitation of ERO1-alpha-mediated
oxidative folding in murine tumor cells. Eur J Immunol 46, 2842-2851, doi:10.1002/eji.201646525 (2016).
7 Sethumadhavan, S. et al. Hypoxia and hypoxia-inducible factor (HIF) downregulate antigen-presenting MHC class
I molecules limiting tumor cell recognition by T cells. PLoS One 12, e0187314, doi:10.1371/journal.pone.0187314 (2017).
8 Koukourakis, I. M., Giatromanolaki, A., Mitrakas, A. & Koukourakis, M. I. Loss of HLA-class-I expression in non-
small-cell lung cancer: Association with prognosis and anaerobic metabolism. Cell Immunol 373, 104495,
doi:10.1016/j.cellimm.2022.104495 (2022).
9 Yu, Z. et al. An emerging prognosis prediction model for multiple myeloma: Hypoxia-immune related
microenvironmental gene signature. Front Oncol 12, 992387, doi:10.3389/fonc.2022.992387 (2022).
10 Sasaki, T. et al. Microenvironmental stresses induce HLA-E/Qa-1 surface expression and thereby reduce CD8(+) T-
cell recognition of stressed cells. Eur J Immunol 46, 929-940, doi:10.1002/eji.201545835 (2016).

An interdisciplinary approach to understand how interactions between
proliferating and invasive melanoma cells can promote metastasis 4 –
Prof. Ruth Baker
Primary Supervisor: Prof. Ruth Baker
Second Supervisor: Prof. Helen Byrne and Prof. Richard White
Abstract
Analysis of patient biopsies shows that melanomas are highly heterogeneous at both genetic and epigenetic levels. Single-
cell sequencing has demonstrated at least five co-existing transcriptional cell states, including distinct proliferative versus
invasive states. However, little is known about how these states coexist within tumours, whether they interact to promote
metastasis and tumour progression, and/or whether cells switch between these states. The overarching aim of this project
is to combine mathematical models with experimental data collected in the White Lab to provide new insights into these
challenging questions.
Research objective
Background: Melanomas are heterogeneous, with multiple cellular phenotypes characterised by distinct transcriptional
states [2,4]. At the extreme ends of these states are proliferative versus invasive subpopulations usually linked to the “go
or grow” hypothesis. Recent in vitro experiments performed in the White Lab have shown that during metastasis,
melanoma cells from these proliferative and invasive subpopulations spontaneously form spatially structured clusters,
with invasive cells surrounded by an outer rim of proliferating cells. Additional in vivo experiments show that
heterogeneous clusters, comprising proliferative and invasive cells, form secondary tumours, or metastases, at rates which
are significantly higher than those associated with each subpopulation.
The mechanisms giving rise to these seemingly counter-intuitive observations are not well understood. Yet a better
understanding of the process of metastasis will be key to the development of novel cancer therapeutics.
Aims and objectives: As such, the aim of the project is to develop mathematical models to analyse, interpret and integrate
dynamic imaging data at the tissue scale, and high-resolution spatial transcriptomic data, from both in vitro and in vivo
experiments. Repeated rounds of model-driven hypothesis generation and experimental validation will provide new
understanding of how melanoma cell clusters form, and how their ability to metastasise depends on both the extent of
heterogeneity and the spatial structure of the clusters. The project will initially consider just two cellular phenotypes,
proliferative and invasive [2]. Subsequent extension of the models to include the five co-existing transcriptional cell states
detailed in [4] will enable comparison of how the extent of phenotypic heterogeneity impacts metastasis. The objectives
are:
1. To develop models of cluster formation based on the coagulation-fragmentation framework [5], and to calibrate
models to quantitative data collected in the White Lab using computational Bayesian statistics approaches, to
understand the mechanisms driving observed cluster size distributions.
2. To develop spatially resolved, cell-based models of cluster formation [1,3], in which individual cellular properties
(such as rate of cell cycle progression and invasiveness) can be specified and, through careful calibration of the
model to experimental data, to explore the mechanisms that give rise to the observed spatial structure of clusters.
103
3. To use the models developed in Objective 2 to predict how metastatic potential changes with cluster composition,
and the extent of cooperation between the different cellular subpopulations.
Academic value of the research and funding justification: Funding for a DPhil student will enable us to transform
understanding of how melanoma cell heterogeneity impacts metastatic potential. Progress on this challenging problem
requires a multidisciplinary team with expertise across mathematical modelling, computational statistics, experimental
biology, and image analysis. The funds will cover the costs of a graduate student who will carry out research in the
Mathematical Institute.
Collaborations: The project will initiate a new collaboration between Professors Baker, Byrne and White. The student will
make frequent visits to the White Lab where they will interact with lab members investigating melanoma metastasis. This
will enable the student to learn the relevant melanoma biology, experimental techniques, and data analysis methods, and
to contribute to experimental design. The team will meet with the student on a weekly basis.
This project will provide new insights into the processes driving melanoma metastasis, through the interrogation of large-
scale data sets using mathematical modelling and computational statistics. It will generate new methodologies for
analysing multi-model data and strengthen expertise in multidisciplinary approaches to tackling cancer. As such, it will
contribute to the scientific themes “Cancer big data” and “Early cancer detection”.
The student will receive training in mathematical modelling using differential equations, stochastic processes and cell-
based models, as well as computational Bayesian statistics, image analysis, experimental design, and multidisciplinary
research. Within the Mathematical Institute, the student will be part of the Wolfson Centre for Mathematical Biology,
where they will take part in weekly mathematical oncology focus meetings and research skills training sessions, and attend
formal weekly seminars. The student will also spend time in the White Lab, attending group meetings and relevant
seminars at the Ludwig Institute for Cancer Research. Here, they will also have the opportunity to learn how to perform in
vitro cell-based assays, if desired.
References:
1. J. A. Bull and H. M. Byrne (2023). Quantification of spatial and phenotypic heterogeneity in an agent-based model
of tumour-macrophage interactions. PLOS Computational Biology 19:e1010994.
2. N. R. Campbell et al. (2021). Cooperation between melanoma cell states promotes metastasis through heterotypic
cluster formation. Developmental Cell 56:2808-2825.
3. F. R. Cooper et al. (2020). Chaste: cancer, heart and soft tissue environment. Journal of Open Source Software
5:1848.
4. D. Lumaquin-Yin et al. (2023). Liquid droplets are a metabolic vulnerability in melanoma. Nature Communications
14:3192.
5. J. A. D. Wattis (2006). An introduction to mathematical models of coagulation–fragmentation processes: A discrete

deterministic mean-field approach. Physica D 222:1-20.

Exploring different modalities of growth factor inhibitor to treat
cancer.2,3 – Dr. Jon Elkins
Primary Supervisor: Dr. Jon Elkins
Second Supervisor: Prof. Valentine Macaulay
Eligibility: Track 2 and 3 students are eligible to apply for this project.
Abstract
Various cancers are promoted by the binding of growth factors to receptor proteins, and are also associated with higher
levels of the growth factors. One strategy to treat cancer is to block the binding of the growth factor to the receptor. This
can be done by a drug that binds to the receptor or by a drug that binds to the growth factor itself, such as bevacizumab
which is a monoclonal antibody that binds the growth factor VEGFA and is used to treat multiple types of cancer.
Antibodies can be discovered that bind to growth factors, but they are expensive to produce and require intravenous
delivery. Cyclic peptide or small molecule inhibitors of growth factors could offer advantages over antibodies by being
cheaper, less likely to be immunogenic and given orally. Many cyclic peptides have already been approved as drugs [1].
This project aims to accelerate the discovery of cyclic peptide and small molecule binders of growth factors, by examining
the interfaces of antibodies and cyclic peptides when bound to them. The cyclic peptides discovered during the project
may also become lead candidates in a drug discovery programme. The project will combine the discovery of cyclic
peptides, the discovery of nanobodies (small single-chain antibodies that are very suitable for crystallisation), and the
structural comparison of these cyclic peptides and nanobodies when bound to growth factors.
By comparing structures, we aim to identify the features of nanobodies that give strong binding to the growth factor,
and apply that knowledge to the design of improved cyclic peptides. After an initial focus on insulin-like growth factor 1
(IGF1), for which materials are already available, the knowledge and process can be applied to other growth factors such
as EGF or VEGFA. The output will be, for each growth factor studied, a cyclic peptide inhibitor and an understanding of
the structural features necessary for drug design.
Research objective
General methodology -Cyclic peptides (CPs) will be selected in the Kawamura lab from large libraries by mRNA-display /
peptide microarray technology, and the top enriched hits that bind to IGF1 but not to insulin will be synthesised. An
initial set of CP binders of IGF1 has already been generated as a proof-of-concept. Later in the project other target
proteins can be used as bait in the peptide selection experiments.
The top hits from the screens will be validated for IGF1 binding by biophysical assays such as Biolayer Interferometry
(BLI), Surface Plasmon Resonance (SPR) or Isothermal Titration Calorimetry (ITC). The peptides with the highest binding
affinity will be used in co-crystallisation studies.
In parallel in the Centre for Medicines Discovery a nanobody selection by ribosome display can be performed analogously
to the selection of CPs. Top enriched nanobodies will be validated as for the CPs and also used for crystallisation studies.
Computational analysis of the structures will guide the design of improved CPs in subsequent iterations. Among the
factors to consider in the analysis will be potency of binding, selectivity (over insulin in the case of CPs targeting IGF1)
and physicochemical properties such as solubility and predicted stability in plasma. Knowledge of how to generate potent
CPs against growth factors will be a big step forward for the drug discovery community, and this knowledge could
potentially be applied to small molecules as well.
The most potent CPs with nanomolar affinity can be tested for blocking IGF1 signalling in cancer cell assays in the
Macaulay lab. It is anticipated that the best CPs will form the basis of a translational drug discovery project
Background to the choice of IGF1 as an initial target - Insulin-like growth factors-1 and 2 (IGF1/2) are small proteins that
are similar in sequence to insulin and are part of a system of signaling that cells use to interact with their environment.
IGF1/2 proteins are secreted into serum and high levels of IGF1 in serum are causally associated with risk of developing
prostate, breast and colorectal cancer and with prostate cancer mortality [2–5]. IGF1/2 are also implicated in the
association of tall height with cancer incidence and high-grade prostate cancer [6,7].
105
The receptors for IGF1/2 are the IGF receptor (IGF1R) and insulin receptor (INSR-A). When the target cells are cancerous
the effect of signals transmitted by the IGF receptor can be to promote tumour growth, invasion and resistance to cell
death (apoptosis), which confers resistance to chemotherapy and radiotherapy [8]. Small molecule IGF1R inhibitors
showed promise in the clinic but ultimately did not succeed as drugs, in some cases due to difficulty in inhibiting IGF1R
without affecting the related INSR either directly, or indirectly via downregulation of IGF1R:INSR hybrid receptors, or
because IGF1R inhibition caused hyperglycaemia [8].
An alternative strategy would be to target IGF1 or IGF2 directly, blocking (or neutralising) their interactions with the
receptors. Boehringer Ingelheim have developed a monoclonal antibody, xentuzumab, that binds IGF1 and IGF2.
Xentuzumab inhibits >90% of serum IGF bioactivity and has been tested in phase II clinical trials in breast and prostate
cancer where it showed insufficient activity in unselected patients, although subgroups are potentially responsive [9]. A
clinical trial in Oxford ‘WINGMEN’ (NCT05110495, Section 9) involving Macaulay may identify responsive subgroups. An
orally-active drug is therefore urgently needed, in particular if WINGMEN results support trialing longer-term IGF
blockade.
While a small-molecule IGF1 inhibitor would be ideal, IGF lacks a good binding site for a typical small molecule. Cyclic
peptides possess the benefits of both antibodies and small molecules. CPs often show greater target affinity, bioactivity
and biostability than their linear analogues, with advantageous properties over antibodies including synthetic
tractability, low immunogenicity, cheaper production and potential for oral dosing, such as for cyclosporin. CPs can bind
to their target with sub-nM affinity and >100-fold selectivity, even for challenging protein targets, including other growth
factors such as HGF [10].
Team: The two supervisors on the project bring together expertise in structural biology and inhibitor discovery (Elkins)
with expertise in cancer biology and IGF biology (Macaulay). This project is a collaboration with the lab of Prof. Akane
Kawamura at Newcastle University, who is a world leading expert in cyclic peptide discovery by mRNA-display and
peptide microarray technologies, and has a strong track record discovering cyclic peptide inhibitors of proteins of
therapeutic interest.
The new cyclic peptides discovered may lead to clinical programmes, while the knowledge gained from analysis of binding
motifs will accelerate the discovery of cyclic peptide based drugs.
The student will experience a broad range of experimental science. In the Elkins lab the structural analysis of growth
factor complexes will involve techniques such as molecular cloning, protein purification and crystallisation, X-ray
structure determination and advanced biophysical measurements. In the Macaulay lab the various types of cellular assay
necessary to evaluate the optimised cyclic peptides for their anti-cancer potential. The student will also have the
opportunity to visit the Kawamura lab to learn about cyclic peptide discovery and synthesis. Overall, an education and
training in structure-guided drug design in the cancer therapeutic area.
References
1 Zhang, H. and Chen, S. (2022) Cyclic peptide drugs approved in the last two decades (2001–2021). RSC Chem. Biol. 3, 18–31.
2 Guevara-Aguirre, J. et al. (2011) Growth Hormone Receptor Deficiency Is Associated with a Major Reduction in Pro-Aging Signaling,
Cancer, and Diabetes in Humans. Sci. Transl. Med. 3, 70ra13.
3 Travis, R. C. et al. (2016) A Meta-analysis of Individual Participant Data Reveals an Association between Circulating Levels of IGF-I
and Prostate Cancer Risk. Cancer Res. 76, 2288–2300.
4 Murphy, N. et al. (2020) Insulin-like growth factor-1, insulin-like growth factor-binding protein-3, and breast cancer risk:
observational and Mendelian randomization analyses with ∼430 000 women. Ann. Oncol. 31, 641– 649.
5Watts, E. L et al. (2023) Circulating insulin-like growth factors and risks of overall, aggressive and early-onset prostate cancer: a
collaborative analysis of 20 prospective studies and Mendelian randomization analysis. Int. J. Epidemiol. 52, 71–86.
6Nunney, L. (2018) Size matters: height, cell number and a person’s risk of cancer. Proc. R. Soc. B Biol. Sci. 285, 20181743.
7Perez-Cornago, A. et al. (2017) Tall height and obesity are associated with an increased risk of aggressive prostate cancer: results
from the EPIC cohort study. BMC Med. 15, 115.
8Simpson, A. et al. (2017) Insulin-Like Growth Factor (IGF) Pathway Targeting in Cancer: Role of the IGF Axis and Opportunities for
Future Combination Studies. Target. Oncol. 12, 571–597.
9Schmid, P. et al. (2021) A phase Ib/II study of xentuzumab, an IGF-neutralising antibody, combined with exemestane and everolimus
in hormone receptor-positive, HER2-negative locally advanced/metastatic breast cancer. Breast Cancer Res. 23, 8.10 Sakai, K. et al.
(2019) Macrocyclic peptide-based inhibition and imaging of hepatocyte growth factor. Nat. Chem. Biol. 15, 598–606.

106
The inactivation of p97 system in colorectal cancer therapy 1,2,3,4 – Prof
Kristijan Ramadan
Primary Supervisor: Prof Kristijan Ramadan
Second Supervisor: Prof. Simon Leedham
Abstract
Colorectal carcinoma (CRC) currently represents the third most common type of cancer worldwide, accounting for
approximately 10% of all diagnosed cases, with 1.85 million new cases annually [1]. The systemic treatment of
microsatellite stable (MSS) colorectal cancers, which comprise around 80% of all cases, has remained largely unchanged
since the early 2000s. Currently, two main therapies, FOLFIRI and FOLFOX, both exhibiting similar efficacy, continue to be
the most commonly used first-line treatments, particularly for colon cancers [2, 3].FOLFIRI is a chemotherapy regimen
based on irinotecan/topoisomerase inhibitor, which induces the formation of Topoisomerase 1 cleavage complexes
(TOP1ccs), a specific type of enzymatic DNA-protein crosslink (DPC) where TOP1 is covalently bound to DNA [4]. In recent
years, our research has highlighted the significance of p97 ATPase/unfoldase, SPRTN protease, and TEX264 in the repair of
TOP1ccs and subsequent cell survival [5, 6]. TEX264, acting as a p97-cofactor, serves as the receptor for TOP1ccs. It binds
to TOP1ccs and recruits the p97-SPRTN complex to the site of DNA damage (Fig. 1), facilitating the proteolytic degradation
of TOP1ccs through p97 unfoldase activity and subsequent SPRTN protease activity. This mechanism allows cancer cell
survival in response to TOP1 inhibitors.
Figure 5. The role of p97-SPRTN- TEX264

machinery in repair/processing of cytotoxic
TOP1ccs.
We have validated the importance of p97 ATPase,

SPRTN protease and TEX264 cofactor in determining
the efficacy of TOP1 inhibitors such as camptothecin
and irinotecan in various human cancer cell lines in our
laboratory (Fig. 2, data not shown), as well as in CRC patients (Fig. 3). Furthermore, we have established the p97-SPRTN-
TEX264 complex as the first biomarker for predicting the response to FOLFIRI in CRC patients.
Figure 2. SPRTN depletion

by auxin (IAA) inducible
degradation increases the
amount of DNA damage (-
H2AX; 53BP1) in HCT116
cells treated with
camptothecin (CPT) and
delays its repair.
To inactivate this p97-SPRTN-TEX264 complex and thus kill cancer cells, we aim to use the second generation of orally
available p97 inhibitor CB-5339 developed by our industry partner, Cleave Therapeutics. CB-5399 is currently undergoing
clinical trials for various blood and solid malignancies, albeit not for colorectal cancer [7]. CB-5339 has successfully
completed a phase I clinical toxicity study and is now progressing into phase II clinical trials.
107
With the support of Cleave Therapeutics (by providing the inhibitor), our objective is to investigate the molecular
mechanisms of CB-5339 both in vitro and in an orthotopic CRC mouse model. This research will provide valuable insights
Figure 3. The overexpression of SPRTN

correlates with a worse prognosis in patients
treated with irinotecan.
into how p97, and thus the entire p97-SPRTN-TEX264

complex, contributes to the survival of CRC cells treated
with CB-5339 alone or in combination with standard
chemotherapies such as FOLFIRI and FOLFOX.
Research objectives
• Characterise the sensitivity of different CRC cell lines to the p97 inhibitor CB-5339
• Investigate the role of the p97-SPRTN-TEX264 complex in the repairment of TOP1ccs induced by topoisomerase
inhibitors in CRC with the focus on DNA replication and S-phase cell proliferation
• Validate our findings using CRC orthotopic mouse models treated with CB-5339 alone and in combination with
FOLFIRI and FOLFOX.
A deeper understanding of the role of specialised machinery (97-SPRTN-TEX264 complex) for repair of cytotoxic TOP1-ccs
in mediating the response to topoisomerase inhibitors will allow us to better predict the response to FOLFIRI in CRC
patients. Moreover, p97 inhibition is likely to synergize with irinotecan-based chemotherapy to further improve the
outcomes of patients treated with TOP1ccs-inducing drugs in many cancer types. This is a collaborative work between,
basic science (K. Ramadan), pre-clinical validation (S. Leedham) and industry partner (Cleave Therapeutics; a support letter
can be provided).
The PhD student will have the opportunity to get trained in basic science (molecular, cell biology and rodent work), pre-
clinical research (mouse work) and understand how to translate a clinical problem (chemotherapy resistance) to the bench
and return this knowledge back to the clinics.
References:
Mattiuzzi, C., F. Sanchis-Gomar, and G. Lippi, Concise update on colorectal cancer epidemiology. Ann Transl Med, 2019.
7(21): p. 609.
2. Tournigand, C., et al., FOLFIRI followed by FOLFOX6 or the reverse sequence in advanced colorectal cancer: a
randomized GERCOR study. J Clin Oncol, 2004. 22(2): p. 229-37.
3. Aggarwal, N., et al., Systematic review and meta-analysis of tumour microsatellite-instability status as a predictor
of response to fluorouracil-based adjuvant chemotherapy in colorectal cancer. Int J Colorectal Dis, 2022. 37(1): p. 35-46.
4. Thomas, A. and Y. Pommier, Targeting Topoisomerase I in the Era of Precision Medicine. Clin Cancer Res, 2019.
25(22): p. 6581-6589.
5. Fielden, J., M. Popovic, and K. Ramadan, TEX264 at the intersection of autophagy and DNA repair. Autophagy,
2022. 18(1): p. 40-49.
6. Fielden, J., et al., TEX264 coordinates p97- and SPRTN-mediated resolution of topoisomerase 1-DNA adducts. Nat
Commun, 2020. 11(1): p. 1274.
7. Lina Benajiba, et al., 2870 Trials in Progress: A Phase I Study to Evaluate the Safety and Pharmacokinetic Profiles
of CB-5339 in Participants with Relapsed/Refractory Acute Myeloid Leukemia or Relapsed/Refractory Intermediate or High-
Risk Myelodysplastic Syndrome. 62nd ASH Annual Meetig and Exposition, 2022.

Assessment of Oesophageal cancer patient responses to
immunotherapy via human tissue avatars.1,2,3 – Prof. Richard Owen
Primary Supervisor: Prof. Richard Owen
Second Supervisor: Prof. Eric O’Neil
Eligibility: Track 1,2 and 3 students only are eligible to apply for this project.
Abstract
Currently, the majority of in vitro therapeutic testing is carried out using patient-derived cell lines, xenografts (PDX) and
genetically engineered mouse models (GEMMs). While murine models provide valuable information about the systemic
effects of therapeutics, GEMMs fail to capture the genomic heterogeneity of native tumours and PDX models are
challenging to establish orthotopically and do not allow for the interaction of tumour with a functional immune system to
be studied (Kim et al., 2009). Introduction of patient- derived organoids has allowed in vitro analysis of treatment
interactions with three-dimensional tumour structures (Ooft et al., 2019) but, similarly to xenograft models, frequently
fails to capture the complexity of the TME (Larsen et al., 2021). Recent efforts move in the direction of incorporating
multiple different cell types and vascular mimics to reconstruct tumour complexities (Neal et al., 2018). Patient-derived
tumour slices provide a platform through which tumour, stroma and immune infiltrate can be studied in their native
architecture (Ghaderi et al., 2020; Kokkinos et al., 2021). Through this system therapeutics can be investigated for their
impact throughout the tumour, allowing analysis of intra-patient variation in a clinically relevant timeframe.
O’Neill lab has developed the use of live patient-derived tumour slices for dissection of pancreatic cancer
microenvironment and investigation of therapy responses. Methods have established to maintain superior cellular fitness
and preservation of tumour microenvironment compared to standard cultures, organoids or spheroids. Analysis of
transcriptomic changes induced by a combination of therapies aimed to target metabolic reprogramming treatment shows
the potential of the platform to interrogate treatment responses across all cellular compartments of the
microenvironment, in particular immune, in an unprecedented manner. Having demonstrated that organotypic tumour
slices can maintain viability and provide novel insights enhancing both novel therapeutic discovery and precision medicine
to improve current standard of care.
Oesophageal cancer is the sixth leading cause of cancer mortality worldwide (Bray et al., 2018). The predominant subtype
in the western world, oesophageal adenocarcinoma, is among the cancer types with the highest increase in incidence over
the past few decades (Devesa et al., 1998; Fitzgerald, 2004; Groulx et al., 2020; Lepage et al., 2008; Pennathur et al., 2013).
About 40% of oesophageal cancers present with distant metastases at diagnosis (Smyth et al., 2017) and for these
inoperable patients, median overall survival (OS) with conventional agents is less than one year (Cunningham et al., 2008;
Dijksterhuis et al., 2019; Janmaat et al., 2017; Jatoi et al., 2006; Waddell et al., 2013). Treatment regimens using αPD-1
with chemotherapy have been approved and an Oxford-based trial (LUD2015-005) recently performed comprehensive
clinical and molecular profiling throughout treatment using a combination of whole genome sequencing (WGS), single-cell
RNA-sequencing (scRNA- seq), and bulk RNA-sequencing (bulk RNA-seq) to identify patients that benefit. Treatment-
responsive molecular signatures were identified that effectively predict response and resistance to first-line αPD-1 and
also predicted long-term αPD-1 outcomes in other settings (Carroll et al. in press). Notably, high PD-L1 expression and
tumour mutational burden composed indicators to establish pre-treatment biomarkers that could improve prediction of
long-term outcomes of αPD-1 treatment.
Research objective
This project is aimed to develop live tissue patient avatars from oesophageal adenocarcinoma biopsies using the
technology validated for immune monitoring of pancreatic cancer avatars in the O’Neill lab. We aim to use engineer this
approach to screen for patients susceptible to αPD-1 therapy and a platform to assess further immune-therapies as
potential combinations for patients not served by the pre-treatment biomarkers I have found (Carroll et al. in press).

The T-cell receptor landscape of adult diffuse gliomas, a non-invasive
tool for tumour detection and classification? 1,2,3,4 – Assoc Prof. Olaf
Ansorge
Primary Supervisor: Assoc Prof. Olaf Ansorge
Second Supervisor: Dr. Bo Sun
Abstract
Adult diffuse gliomas are the most common malignant tumours affecting the central nervous system[1]. Glioblastoma
multiforme (GBM) is the most common and aggressive form. It affects up to 18 people per 100,000 and has a poor survival
time of approximately 12-18 months. Although diagnostic scans and treatments have advanced, there is still a significant
need for non-invasive and reliable methods to detect and monitor gliomas. Indeed, current imaging-based methods are
often insensitive to early tumour occurrence/recurrence[2,3], and biopsies carry significant neurologic risks. To address
this critical unmet need, our interdisciplinary study aims to evaluate the potential of T cell receptor (TCR) sequencing
approaches as a non-invasive biomarker for early detection and monitoring of adult diffuse gliomas.
T-cell receptor sequencing captures intra-tumoural T-cell responses

T-cell receptor (TCR) sequencing shows promise for early cancer detection and monitoring in various cancer types,
including GBM[4-8]. T-cells play a crucial role in the adaptive immune system and their involvement in cancer is increasingly
recognised. TCR sequencing of peripheral blood is a sensitive method for detecting the clonal evolution of intra-tumoural
T-cell populations, which has been associated with better prognosis[4-6]. In glioblastoma, recent studies have shown that
TCR repertoires in blood can serve as a sensitive biomarker of the intra-tumoural T-cell response, despite the brain's
perceived immune privilege[6-7].
Machine learning
demonstrates feasibility
in disease TCR repertoire
classification tasks
Our own re-analysis of a
published glioma TCR
sequencing dataset[8]
lends support to the
feasibility of the
proposed research by
revealing disease-specific
TCR signatures in
patients with adult
diffuse glioma. We
observed a greater clonal
overlap between
matched tumour/brain
and blood samples in
adult diffuse glioma
patients compared to
healthy donors (Figure
1A). Additionally, we
observed decreased Gini
coefficients and
increased Shannon's
diversity indices,
supporting tumour-specific disturbances of TCR clonality (Figure 1B, C). To demonstrate feasibility of TCR-guided disease
classification, we trained a deep learning model on one-hot encoded CDR3 amino acid sequences of TCRs, which achieved
an area under the receiver operating characteristic (AUC) curve score of 0.82 (Figure 1D).
110
Research objective
The proposed project aims to investigate the utility of TCR sequencing in early detection and monitoring of gliomas using
both established short-read and novel long-read sequencing approaches. The overall objective is to evaluate the ability of
TCR sequencing to detect and monitor immune responses in patients with adult diffuse gliomas. To achieve this, the study
has three primary objectives:
1. To perform cross-sectional short-read and long-read TCR sequencing of paired tumour and blood samples from
patients with adult diffuse glioma, and matched controls.
a. Paired tumour and blood samples will be derived from the established Tessa Jowell Brain Matrix
project (UK).
2. Evaluate and compare the reproducibility of long-read nanopore TCR sequencing to the previously established
short-read approach.
3. Use machine learning models, such as deep neural networks, support vector machine and variational autoencoders
to identify peripheral TCR repertoire patterns that predict the intra-tumoural T-cell response, and correlate these
to tumour mutation burden, and glioma molecular subtype.
b. Computational training and support will be provided by co-supervisor Dr Bo Sun in collaboration with
Associate Prof. Rachael Bashford-Rogers.
The proposed research has the potential to significantly impact patient care by providing a non-invasive and reliable
method for early detection and monitoring of glioblastoma and other diffuse gliomas. By identifying tumour-specific T-cell
receptor (TCR) repertoire patterns and developing machine learning models to predict the presence and molecular subtype
of cancer cells, this research may enable clinicians to detect tumours at an earlier stage and track tumour recurrence post-
surgical intervention. This is particularly pertinent in light of the active development of immune therapies in neuro-
oncology such as CAR T-cells. Furthermore, recent FDA authorisation of the T-detect COVID test[9], a TCR sequencing-
based diagnostic tool for identifying infected individuals, is evidence of the clinical feasibility of TCR-based approaches. Its
approval paves the way for the development of TCR-based approaches for use in clinical settings.
Where appropriate, the student will receive training in Good Clinical Practice principles and the conduct of observational
studies. From a laboratory perspective, flow cytometry, cell sorting, RT-PCR and library preparation of both short-read and
long-read sequencing modalities will be used in this project. Hands-on immunology experience and training in T cell
receptor enrichment and library preparations will be provided. For bioinformatics training, the student will be supported
to attend dedicated computational course offered under the wider network of the Medical Sciences Division; further,
direct supervision will be provided by Dr Bo Sun. The student will have full access to the facilities, expertise, and resources
available within NDCN, the department of Neuropathology and across the broader community at the University of Oxford.
By the end of the project, the candidate will be in a strong position to drive fundamental and translational clinical research
in neuro immuno-oncology and adaptive receptor repertoire sequencing.
References
[1] Komori, T. Grading of adult diffuse gliomas according to the 2021 WHO Classification of Tumors of the
Central Nervous System. Lab. Investig. 102, 126–133 (2022).
[2] Bernstock, J. D. et al. Standard clinical approaches and emerging modalities for glioblastoma imaging.
Neuro-Oncology Adv. 4, (2022).
[3] Müller Bark, J., Kulasinghe, A., Chua, B., Day, B. W. & Punyadeera, C. Circulating biomarkers in patients
with glioblastoma. Br. J. Cancer 122, 295–305 (2020).
[4] Valpione, S. et al. The T cell receptor repertoire of tumor infiltrating T cells is predictive and prognostic
for cancer survival. Nat. Commun. 12, (2021).
[5] Beshnova, D. et al. De novo prediction of cancer-associated T cell receptors for noninvasive cancer
detection. Sci. Transl. Med. 12, (2020).
[6] Melody, S. H. et al. TCR sequencing can identify and track glioma-infiltrating t cells after dc vaccination.
Cancer Immunol. Res. 4, 412–418 (2016).
[7] Platten, M. et al. A vaccine targeting mutant IDH1 in newly diagnosed glioma. Nature 592, 463–468
(2021).
[8] Sims, J. S. et al. Diversity and divergence of the glioma-infiltrating T-cell receptor repertoire. Proc. Natl.
Acad. Sci. 113, E3529–E3537 (2016).
[9] https://www.fda.gov/media/146481/download

111
Take epigenetic diagnostics “from bench to bedside”1,2,3,4 –Prof. Prof
Benjamin Schuster-Böckler
Primary Supervisor: Prof Benjamin Schuster-Böckler
Second Supervisor: Assoc Prof. Ben Fairfax
Abstract
Cancer initiation and progression is not solely driven by mutations: it is becoming increasingly clear that environmental
factors, differentiation dynamics and inflammation put cells in a state that is more (or less) susceptible to cancerous
transformation. Such dynamic cell states are maintained by cells through epigenetic marks on histones and DNA. We have
a particular interest in decoding the role of changes in the DNA methylation in cancer formation and progression.
Together with the group of Prof Song, we previously developed “TAPS” (TET Assisted Pyridine Borane Sequencing), a new
sequencing method for measuring methylated cytosine in DNA that produces much cleaner data from low-input samples
such as cell-free or single-cell DNA1. We recently used TAPS to generate a base-resolution atlas of tissue-specific
methylation and hydroxymethylation from over 20 different cell and tissue types, as well as from a range of different
cancer types. This map has revealed a whole range of differentially methylated regions that uniquely identify cell types
and show dysregulation in cancer.
We are involved in several clinical trials that collect whole-genome TAPS data from cell-free DNA of patients at risk of or
diagnosed with cancer. For example, we collaborate with Prof Lu and Dr Owen on a trial of checkpoint inhibition in
oesophageal adenocarcinoma2, as part of which we are collecting cell-free DNA methylation data of nearly 100 patients.
The goal of this DPhil project will be to leverage cell-type and tissue-specific epigenetic information to detect whether
patients have cancer and/or whether they will respond to immunotherapy 3.
Research objective
Objective 1: Predict cancer immune landscape from ctDNA
In our recent Cancer Cell paper2, we demonstrate that mutation burden and monocyte content of the tumour at diagnosis
are independent predictive biomarkers of immune-checkpoint therapy success in oesophageal adenocarcinoma. Immune
cells are the main contributors to cell-free DNA in healthy individuals as well as cancer patients. The first objective of the
DPhil project would be to determine whether the contribution of T-cell, macrophage and monocyte DNA varies between
healthy donors and cancer patients, and whether we can identify changes in the regulatory regions that control e.g.
monocyte specific marker genes that we identified in our previous RNA-seq analysis. The proposed outcomes of this work
will be a method to quantify the immune-cell contribution in cell-free DNA based on epigenetic marks.
Objective 2: Quantify mutation burden from ctDNA
We are currently working on methods to improve somatic variant prediction from cell-free TAPS data using machine-
learning to improve the signal-to-noise ratio4. Using such pre-processed data, we will test whether ctDNA derived mutation
burden correlates with mutation burden measurements derived from whole-genome sequencing of the same patient’s
tumour. The expected outcome of this work will be a method to quantify tumour mutation burden based on cell-free
DNA TAPS data, and an evaluation of the accuracy of these predictions.
Objective 3: develop an integrated predictor of treatment success
Finally, we hope to combine the results of objective 1 and 2 into an integrated classifier that combines mutation burden
and cell-composition estimated from cfDNA to predict patient outcome. We will evaluate the performance of this classifier
on an independent dataset of patients treated with neo-adjuvant immunotherapy. The expected outcome will be a
quantitative evaluation of prediction accuracy, which will be fundamental for securing further collaborations to
translate this work into a clinically useful tool.
112
This is a highly translational project by design. Any biomarkers and predictive algorithms developed as part of this project
will be applied first within existing clinical trials and, if promising, would form the basis for industry collaborations to test
these approaches in larger cohorts.
This project would be an ideal opportunity for a candidate with a passion for epigenetics and an interest in computational
biology. You will learn how to process and analyse DNA methylation and other epigenetic data using statistical tools. Some
experience with R or Python and basic linux command line usage would be helpful.
Initially, this project will be very “data science” heavy, ie identifying correlations and associations between linked datasets.
Further down the line, you will also learn to build predictive algorithms, apply them to data and evaluate their
performance.
References
1. Liu, Y. et al. Bisulfite-free direct detection of 5-methylcytosine and 5-hydroxymethylcytosine at base resolution.
Nature Biotechnology 37, 424-429 (2019).
2. Carroll, T.M. et al. Tumor monocyte content predicts immunochemotherapy outcomes in esophageal
adenocarcinoma. Cancer Cell (2023, in press)
3. Cabel, L. et al. Clinical potential of circulating tumour DNA in patients receiving anticancer immunotherapy. Nat Rev
Clin Oncol 15, 639-650 (2018).
4. Widman et al. Machine learning guided signal enrichment for ultrasensitive plasma tumor burden monitoring. BioRxiv
(2022, doi: https://doi.org/10.1101/2022.01.17.476508)

Spatial interrogation of low grade prostate cancer to identify genomic
events responsible for driving indolent not aggressive disease,1,2,3,4 – Dr.
Alistair Lamb
Primary Supervisor: Dr Alistair Lamb
Second Supervisor: Prof. Ian Mills
Abstract
Background: Over the past decade we have increasingly accepted that prostate cancer classified as “low risk” (Gleason
Grade Group 1 (Gleason Score 3+3 = 6), PSA <10, Stage T1-T2a) are indolent and do not progress.1 This has led to a
recommendation for most men with this type of prostate cancer to undergo active surveillance. 2 3 Indeed, opinion leaders
in prostate cancer are increasingly calling for Grade Group 1 prostate cancer to cease being called “carcinoma”.4 5 However,
we know little about the genetic composition of such low grade tumours, largely because research has focussed on higher
grade disease, but also because we have lacked the spatial genomic technologies to carefully interrogate discrete selected
areas of prostate tissue. Those technologies are now available6 7 and have been developed to a point where the clonal
copy number composition of epithelial regions can be defined,7 8 although they still lack spatial epigenomic analyses.
In our organscale study of a prostate removed at radical prostatectomy which had multifocal prostate cancer, we identified
an area of Gleason Grade Group 1 prostate cancer which lacked most of the defining somatic mutations of higher grade
disease (e.g. chr8q gain corresponding to well-known prostate cancer oncogene c-Myc9, or chr10p loss corresponding to
tumour suppressor gene PTEN10; Figure 1). This raises the possibility that low grade cancer is fundamentally distinct from
higher grade disease at a genetic level. We also found that areas of non-transformed ‘benign’ epithelia contained areas of
greater genomic risk than low grade cancer. Perhaps low grade cancer is therefore ‘safer’ than benign epithelia which still
retains the potential to undergo somatic mutations that develop aggressive disease.
Questions: Why do low grade Gleason pattern 3 prostate cancers seem not to progress? How can we better differentiate
“good” from “bad” pattern 4 disease? Can we drive “bad” pattern 4 disease to indolence?
Hypothesis: Low grade prostate cancer clones harbour genomic alterations which pre-determine indolence; when present
in higher grade clones, such alterations differentiate high grade disease with good prognosis.
Research objectives
1. Consolidate: To undertake clonal siCNV analysis7 on a selection of men with pure Gleason pattern 3 prostate cancer
(n=5; from archive as we no longer operate on these men) extending the findings of our recent paper in Nature,
Erickson et al7, to identify “indolence factors” when benchmarked against pattern 4 & 5 [tissue handling; spatial
transcriptomics; bioinformatics]
2. Extend: To track clonal composition of Gleason pattern 3 and 4 cancer in serially sampled MRI-targeted biopsies
from men on active surveillance for “low” and “intermediate risk” prostate cancer (n=5 each) [spatial
transcriptomics; DNA sequencing; epigenomics; bioinformatics]
3. Establish phenotype: To genetically modify cell-line models of aggressive PCa (e.g. PC3, DUCaP) to upregulate
indolence factors as a pre-clinical basis for pharmacological manipulation [cell culture; cloning; lentiviral
transduction; functional assays]
The data produced could provide the evidence our field is looking for to underpin a redefining of low grade prostate cancer,
as well as helping us understand what makes some pattern 4 prostate cancers so much worse than others. This could help
a proportion of men who currently undergo radical therapy (currently approx. 25,000 pa in UK) to avoid the side-effects
of such treatment.
114
The project incorporates a range of wet and dry lab techniques (see [] brackets above) and can be tailored, given our
overall lab skill sets, to someone with an interest in biology or computational work. We will also train the individual,
whatever their background, in our overall approach to translational surgical science with opportunities to work with
surgical oncologists, pathologists and radiologists as well as biologists and data scientists.
https://www.nds.ox.ac.uk/research/prostate-biology
References:
1. Ross HM, Kryvenko ON, Cowan JE, et al. Do adenocarcinomas of the prostate with Gleason score (GS) </=6 have the potential to metastasize to lymph
nodes? Am J Surg Pathol 2012;36(9):1346-52. doi: 10.1097/PAS.0b013e3182556dcd [published Online First: 2012/04/26]
2. Eastham JA, GBA, DAB, et al. Clinically Localized Prostate Cancer: AUA/ASTRO Guideline, Part I: Introduction, Risk Assessment, Staging and Risk-Based
Management. J Urol 2022 doi: doi.org/10.1097/JU.0000000000002757
3. Lowrance WT, Breau RH, Chou R, et al. Advanced Prostate Cancer: AUA/ASTRO/SUO Guideline PART I. J Urol 2021;205(1):14-21. doi:
10.1097/JU.0000000000001375 [published Online First: 2020/09/23]
4. Chappidi MR, Bell A, Cowan JE, et al. The Natural History of Untreated Biopsy Grade Group Progression and Delayed Definitive Treatment for Men on
Active Surveillance for Early-Stage Prostate Cancer. J Urol 2022;207(5):1001-09. doi: 10.1097/JU.0000000000002420 [published Online First:
2022/01/05]
5. Labbate CV, Paner GP, Eggener SE. Should Grade Group 1 (GG1) be called cancer? World J Urol 2022;40(1):15-19. doi: 10.1007/s00345-020-03583-4
[published Online First: 2021/01/13]
6. Stahl PL, Salmen F, Vickovic S, et al. Visualization and analysis of gene expression in tissue sections by spatial transcriptomics. Science
2016;353(6294):78-82. doi: 10.1126/science.aaf2403
7. Andrew Erickson EB, Mengxiao He et al, Alastair D Lamb, Joakim Lundeberg. Spatially resolved clonal copy number alterations in benign and malignant
tissue. Nature volume 608, pages 360–367 (2022) 2022 [https://www.nature.com/articles/s41586-022-05023-2]
8. Erickson A. Spatial iCNV 2021 [Available from: https://github.com/aerickso/SpatialInferCNV.
9. Ramos-Montoya A, Lamb AD, Russell R, et al. HES6 drives a critical AR transcriptional programme to induce castration-resistant prostate cancer through
activation of an E2F1-mediated cell cycle network. EMBO molecular medicine 2014;6(5):651-61. doi: 10.1002/emmm.201303581
10. Jurmeister S, Ramos-Montoya A, Sandi C, et al. Identification of potential therapeutic targets in prostate cancer through a cross-species approach.
EMBO molecular medicine 2018;10(3) doi: 10.15252/emmm.201708274 [published Online First: 2018/02/14]

New Immune Therapies for Acute Myeloid Leukaemia (AML) And
Myeloid Blood Cancers 1,2,3,4 – Prof. Paresh Vyas
Primary Supervisor: Prof. Paresh Vyas
Second Supervisor: Dr, Ricardo Fernandes, Prof. Omer Dushek and Professor Gillespie.
Abstract
Immunotherapy is transforming cancer outcomes. This project focuses on developing a mechanistic understanding for
developing new T cell receptor (TCR) based treatments for AML, the most common aggressive adult blood cancer. We
have identified peptides bound to HLA class II and T cell receptors that recognise these peptides by studying AML patients
cured by allogeneic stem cell transplantation. Though many groups are working on developing HLA-I restricted therapies,
very little is known about how to harness HLA-II therapies in cancer. To address this important knowledge gap, the project
will study the mechanisms that govern binding of peptide to HLA-II and the binding of TCR to peptide-HLA-II (pHLA) antigen.
This is a pre-requisite to developing effective HLA-II directed TCR-based therapies. The applicant will work in a multi-
disciplinary group of immunologists, AML clinical academics and structural biologists to transform immune therapy for
AML and related myeloid cancers. There are three discrete packages of work. First, experimental and computational
approaches will be employed to study HLA-II peptide binding. Second, experimental approaches will study the binding and
crossreactivity of TCRs to pHLA-II antigens. Finally, the applicant will collaborate with a structural biology group to study
the structural basis of p-HLA-II-TCR binding.
Research objectives
Background: Over the last few years, we have been working to understand the mechanistic basis of cure mediated by the
most established curative cellular immunotherapy, allogeneic stem and immune cell transplantation (allo-SCT). Allo-SCT
involves transfer of blood stem and immune cells from a healthy person (donor) to a patient (recipient) (Fig. 1a). Allo-SCT
has been in routine clinical practice since the 1960s and around 20,000 allo-SCTs are performed worldwide annually. The
most common disease treated by allo-SCT is AML. Allo-SCT is curative because some of the donor immune cells, called T
cells, attack and eradicate the patient's cancer cells. This is called Graft-versus-Leukaemia (GvL). However, donor immune
cells can also attack the patient's normal healthy tissue, which can cause great harm. This is known as Graft-versus-Host
Disease (GvHD) (Fig 1b).
Fig.1. Alloreactive antigen-specific T

cells mediate GvL and GVHD in
allo-SCT. a) Donor (left) stem and
immune cells are infused into a
patient after chemotherapy. b)
Alloreactive antigens mainly arise
from coding germline SNPs that
differ from patient and donor.
These are known as minor
histocompatibility antigens
(mHAG). Some mHAGs are
presented only on AML cells
(left) and elicit GvL T cell responses.
Other mHAGs are expressed only on
normal tissue and elicit GvHD (right). Finally, some mHAGs are expressed on both AML and normal cells and elicit both GvL
and GvHD
Remarkably, until our recent work, the field has limited insight into the antigenic specificity of GvL, or how to distinguish
T cells that cause beneficial GvL from those that cause harmful GvHD. To address this knowledge deficit, we used an
unbiased systematic reverse immunology approach to study elite responders who had been cured by allo-SCT, with
minimal GvHD. We identified 22 peptide antigens targeted by alloreactive T cells due to single amino acid differences
116
between patient and donor encoded by germline single nucleotide polymorphisms (SNPs) that differed between patient
and donor, rather than somatic mutations in the patient’s tumour (Fig. 2).
Fig. 2. Alloreactive antigens. In principle

these could be either mHAGs (due to
coding germline SNPs) or neoantigens
(due to somatic AML-specific mutations.
Our data shows the antigens identified
were mHAGs consistent with the low
mutational burden in AML.
1. Characterisation of HLA-II binding to alloreactive peptide Using molecular biology and protein engineering methods the
applicant will express soluble pHLA-II with varying peptide lengths and study the biophysical properties of peptide-HLA-II
binding and compare this to computationally derived prediction of peptide-HLA-II binding. By mutating peptide resides
the applicant can identify which peptide residues are critical for binding. This work will be combined with structural studies
in Aim 3.
2. Characterisation of TCR binding and crossreactivity to pHLA-II antigen. Using molecular biology and protein engineering
methods the applicant will express soluble TCR and test the biophysical properties of binding to pHLA-II to TCR. By mutating
TCR resides the applicant can identify which TCR residues are critical for binding. If time permits, the applicant will generate
a large, unbiased, peptide-MHC library for yeast display to determine the TCR crossreactivity of wild-type and engineered
TCRs. Finally, for select TCRs that bind pHLA-II the mode of T cell activation and need for co-stimulation will be studied.
This work will be combined with structural studies in Aim 3.
3. Structural studies of TCR-PHLA-II binding The data in Aims 2 and 3 will be greatly strengthened by structural studies of
pHLA-TCR interaction. The applicant will take select soluble p-HLA-II antigens and cognate TCRs from aims 2 and 3
collaborate with a structural biology group to enable them to make crystals for structural studies using a range of structural
resolution methods.
The combination of Aims 1-3 will provide detailed insight into the mechanisms of p-HLA-II binding and the binding of pHLA-
II antigens with TCRs.
This project aims to identify the optimal pHLA antigens and TCRs for either TCR engager therapy or TCR T-cell therapies
for AML and myeloid cancers.
The DPhil student will be trained in: (i) fundamental aspects of immunology and specifically pHLA interactions and the
binding of pHLA to TCR; (2) molecular biology, protein engineering and biophysical measurements of protein-protein
interaction: (3) computational biology (4) structural biology. The training will be focussed on specific skill sets that are
critical for developing immune therapies.
References:
s1. Polonen, P., et al.,. Cancer Res, 2019. 79(10): p. 2466-2479. 2. Christopher, M.J., et al.,. N Engl J Med, 2018. 379(24): p.
2330-2341. 3. Toffalori, C., et al. Nat Med, 2019. 25(4): p. 603-611. 4. Hernandez-Malmierca, P., et al.Cell Stem Cell, 2022.
29(5): p. 760-775 e10.

117
Mapping the ovarian cancer ascites ecosytem for discovering novel
therapeutic approaches 1,2,3 – Dr. Nicola Ternette
Primary Supervisor: Dr. Nicola Ternette
Second Supervisor: Eleni Adamopoulou and Prof Tim Elliott
Abstract
Dynamic interactions between cancer, immune and stromal cells of the tumour microenvironment are critical for tumour
development and contribute to treatment resistance. However, a systematic molecular and cellular understanding of how
these interactions change during tumour progression, treatment, and their association with clinical prognosis is still
lacking. For the most part, this is due to limited sampling and cell availability of primary tumour tissues for extensive
studies. Ovarian cancer (OC) is the most lethal of gynaecological cancers largely due to its late diagnosis, high metastatic
potential, and resistance to chemotherapy. Almost all advanced stage OC patients develop malignant ascites 1. Cancer-
associated ascites is enriched for tumour, immune, and stromal cells, creating a unique local tumour-immune
microenvironment (TIME) that could be used as a surrogate resource for immune cell monitoring. However, the extent
that ascites reflects the primary tumour tissue microenvironment remains unknown. The aim of this project is to
investigate how the TIME in the ascites of OC shape tumour progression over time and dissect the effect of treatment on
immune and tumour cells associated with better prognosis. This will be conducted by integrating deep cell phenotyping,
multiplex spatial profiling and mass spectrometry (MS)-based immunopeptidomics for uncovering unique and common
features of the TIME of primary tumours and ascites during disease progression and response to therapy, thereby
determining predictors of good prognosis. Leveraging network-based analysis, this approach has broad potential for
informing the design of more effective therapeutic strategies.
Research objective
1.High-dimensional phenotyping of the TIME in ascites from advanced-stage OC patients. To comprehensively measure
the trajectory of interactions between immune and tumour cells prior to and during chemotherapy, a multiplexed
immunofluorescence approach for cells isolated from ascites will be used. This will be complemented by in-vitro functional
assays in cross-sectional studies. In collaboration with key experts we have developed and validated 35-colour spectral
flow cytometry panels for high-throughput protein-based screening of tumour, stromal, T-, B-, NK- and myeloid cells.
Preliminary interrogation of ascites pre- and post-chemotherapy revealed complex immune cell profiles with a good
representation of cell subsets compared to those observed in peripheral blood mononuclear cells from healthy individuals.
Chemotherapy results in redistribution of T-cell subsets, increasing PD-1+ non-cytotoxic CD8+ T cells. Given the role of
tumour-specific CD8+ cytotoxic cells in clearing tumour cells, this suggests a potential clinical advantage of using immune
checkpoint blockade therapies to rescue this crucial cell subset. However, immune checkpoint blockade in ovarian cancer
patients has to date not proven beneficial to disease outcomes2. Our study aims to identify factors driving this lack of
response and potential new targets and/or combinations. Furthermore, tumor cells are phenotypically heterogeneic with
specific subsets dominating the ascites ecosystem after extensive chemotherapy. This heterogeneity can have important
implications for various aspects of cancer, including tumor growth, response to treatment, and the development of
resistance. We aim to gain a deeper view of the different subsets of tumour cells with unique characteristics and
properties. Our preliminary findings support the hypothesis that the composition of ascites represents the TIME and
supports its use as a relatively non-invasive source of patient material that could yield new predictive and prognostic
biomarkers. We aim to profile a large number of ascites samples from well-stratified OC patients and (a) identify
correlations between ascitic immune-tumour profiles and treatment outcomes with a view to establishing tractable
mechanistic hypotheses using deep immunophenotyping and (b) compare the composition of ascites to that of the primary
tumour microenvironment by employing multiplex spatial profiling in a subset of patients where matched primary tissue
samples are available, extending our multiplex panels to include other immune and stromal cells 3. Outcomes: These data
will help to define the functional immune-stromal-tumour cell interactions occurring in ascites contributing to tumour
clearance and chemoresistance. It will enable the development of a Tumour Immune Cell Atlas compendium in ascites of
OC that will serve as a reference model for studies in other cancer types like colorectal and pancreatic and aid in the
development of innovative therapeutic approaches.
2. Exploring the tumour cell immunopeptidome. Characterisation of tumour antigens is essential for making the link
between T cell specificity and therapeutic efficacy, and for the design of safe and effective cancer vaccines. Mass
118
spectrometry (MS)-based immunopeptidomics is currently the only method that enables the direct identification of MHC-
bound peptide antigens but is limited by the volume of cells required and also unable to assign candidate targets to cell
subtypes within the tumour mass. Ascites represents a great alternative source for tumour cells in sufficient numbers for
deep immunopeptidome profiling and in a state (suspension) that is amenable to bulk sorting prior to Immunopeptidomics.
Leveraging our synergistic expertise in antigen processing/presentation, cutting edge MS-based immunopeptidomics and
antigen discovery 4-6, we will employ deep immunopeptidome analysis of flow-sorted tumour cell subsets from ascites
samples collected pre- and post-chemotherapy. Tumour cell signatures identified in Objective 1 will be used to select the
most appropriate surface markers for the isolation of representative tumour cell populations for novel antigen target
discovery. Outcomes: These data will provide a comprehensive characterization of the antigenic landscape of OC, enabling
better-informed immunotherapies and facilitating the development of next-generation cancer vaccines.
In-depth patient sample profiling has direct translational potential. Ascitic fluid is an underutilised but unique biospecimen
for research due to easy accessibility, successive sampling, and capacity for capturing changes in the cell populations that
compose the TIME. Furthermore, deep knowledge of the tumour cell immunopeptidome could better inform
immunotherapies and help develop the next-generation cancer vaccines. This study will support further translational
studies to investigate diagnostic, therapeutic, and prognostic markers.
Training on cell culture, spectral flow cytometry, MS-based immunepeptidomics, multiplex immunofluorescence, digital
spatial profiling, bioinformatics techniques for data analysis will be provided.
References:
1) Huang, H. et al. Neoplasma 60, 546–552 (2013).

2) Pujade-Lauraine, E. et al. Lancet Oncol. 22, 1034–1046 (2021).
3) Izar, B. et al. Nat. Med. 26, 1271–1279 (2020).
4). Adamopoulou E et al. Nat Commun. 4:2039 (2013).
5) Purcell AW et al., Mass spectrometry-based identification of MHC-bound peptides for immunopeptidomics. Nat Protoc.
6:1687 (2019 ).
6) Peng X et al.,. Novel canonical and non-canonical viral antigens extend current targets for immunotherapy of HPV-driven
cervical cancer. iScience. 3:106101(2023)

119
Elucidating the role of trans-lesion synthesis DNA polymerases in
mutational processes and therapy resistance,1,2,3,4 - Marketa Tomkova
Primary Supervisor: Marketa Tomkova
Second Supervisor: Prof. Ian Tomlinson
Abstract
Trans-lesion synthesis polymerases (TLS) enable cells to replicate damaged DNA that would otherwise lead to replication fork
collapse and cell death. However, TLS polymerases are inherently error-prone and introduce new mutations into the DNA,
potentially contributing to the development of cancer. Understanding the sources and mechanisms of cancer-causing
mutagenesis is critical for identifying effective preventive strategies, predicting personalised response to therapy, and designing
novel therapeutics. One of the open questions in cancer mutagenesis is what proportion of the cancer-causing mutations are
due to errors made by DNA polymerases during DNA replication. 1,2 The aim of this project is to elucidate the role of the error-
prone TLS polymerases in mutational processes, using computational genomics combined with wet-lab approaches. Second, we
aim to develop computational tools predictive of response to therapy, as TLS polymerases contribute to resistance to
chemotherapy by bypassing replication-blocking lesions induced by chemotherapy such as cisplatin3–5.
Research objective
Aim 1: Computational genomics approach to identify mutational signatures of TLS polymerases

Carcinogens and mutagenic processes leave distinct footprints in DNA, detectable using the computational approach of
mutational signatures6. Remarkably, aetiology of nearly one-third of the mutational signatures in cancer patients is unknown,
and there are open questions about the exact molecular mechanisms in many of the described mutational signatures. 7
Understanding these mechanisms is important for prevention (e.g., to know how we can change our lifestyle to avoid cancer),
predicting risk and personalised therapy (e.g., using the signatures as biomarkers), and designing novel therapeutics (e.g., based
on synthetic lethality).
The first aim of this project is to identify the contribution of TLS polymerases to the previously detected mutational signatures
and to develop refined TLS signatures using novel computational approaches by utilising additional genomic features and other
data (including TLS gene expression, locations of regions where TLS polymerases get recruited, and specialised datasets of
samples deficient in one TLS polymerase and compensated by other TLS polymerases). Candidate mutational signatures of TLS
polymerases will be identified, comparing traditional ways of de novo signature extraction, with novel approaches, such as
deep-learning-based methods.
Aim 2: Direct in vitro and in vivo measurement of error-signatures of TLS polymerases

One of the challenges in studying DNA polymerase errors is that they are very hard to measure. We have recently developed a
specialised technique called Polymerase Error Sequencing (PER-seq) to detect the errors made by DNA polymerases in single
molecules in vitro (cell-free) in unprecedented detail. Here, we will apply PER-seq to selected TLS polymerases to obtain direct
measurements of their error signatures, unobscured by DNA repair or other complex cellular processes. We will then
complement this with sequencing of TLS-mutant/overexpression mouse and/or cellular models (Tomlinson lab) and analysis of
sequencing data from other previously published resources.
Aim 3: Prediction of resistance to therapy due to TLS polymerases

Finally, we will evaluate the potential of these signatures to predict survival and resistance/response to treatment using data
from cell-lines9, recently cleaned and curated TCGA Resource10, Genomics England, Hartwig Medical Foundation, ICGC and
focussed datasets such as the SCOT clinical trial, and GDSC. Selected candidate predictions may be validated experimentally.
The expected outcomes of this project include (a) mutational signatures of TLS polymerases with support in human cancer data,
in vitro, and in vivo models, (b) novel computational methods for signature detection, (c) mechanistic understanding of TLS role
in mutagenesis, and (d) biomarkers of TLS-based therapy resistance.
TLS polymerases enable bypass of chemotherapy-induced DNA damage, leading to therapy resistance. TLS polymerases thus
represent an attractive target for sensitizing cancer cells to genotoxic therapies. Indeed, inhibitors of TLS or their protein-protein
interactions show promising synergy with therapies such as cisplatin, temozolomide, PARP inhibitors, and others 3–5,11,12. It is
120
thus of increasing importance to understand the mechanisms and extent of TLS contribution to chemoresistance and to develop
biomarkers of resistance due to TLS polymerases. The signatures of individual TLS polymerases will elucidate which TLS
polymerases are involved in resistance to different therapies, and will help to predict which patients would benefit from TLS
inhibitor-based treatment. Finally, the aims 1 and 2 are also expected to elucidate the mutagenic role of TLS polymerases in
genesis of different cancer types, with potential implications for cancer prevention.
The student will have the opportunity to learn transferable skills, including big data analysis, data visualization, machine learning
and potentially deep learning, statistics, high-throughput computing, bioinformatics, and computational genomics, including
integration of large sequencing genomic, epigenomic, transcriptomic, and other data set. The interdisciplinary nature of the
project will provide opportunity to also gain laboratory skills in a range of molecular biology techniques, PER-seq, and other
methods. Support will be provided to develop soft skills in presenting, writing, critical thinking, experimental design, and
networking within the Oxford scientific community and at conferences.
References:
1. Tomasetti & Vogelstein. Variation in cancer risk among tissues can be explained by the number of stem cell divisions. Science
347, 78–81 (2015).
2. Abascal et al. Somatic mutation landscapes at single-molecule resolution. Nature 593, 405–410 (2021).
3. Taniguchi. REV1-POL ζ Inhibition and Cancer Therapy. Mol Cell vol. 75 419–420 (2019).
4. Wilson, Duncton, Chang, Lee Luo, Georgiadis, Pellicena, Deacon, Gao & Das. Early Drug Discovery and Development of Novel
Cancer Therapeutics Targeting DNA Polymerase Eta (POLH). Front Oncol 11, 4776 (2021).
5. Yamanaka, Chatterjee, Hemann & Walker. Inhibition of mutagenic translesion synthesis: A possible strategy for improving
chemotherapy? PLoS Genet 13, e1006842 (2017).
6. Alexandrov et al. Signatures of mutational processes in human cancer. Nature 500, 415–21 (2013).
7. Hu, Xu & De. Characteristics of mutational signatures of unknown etiology. NAR Cancer 2, (2020).
8. Davies et al. HRDetect is a predictor of BRCA1 and BRCA2 deficiency based on mutational signatures. Nat Med 23, 517–525
(2017).
9. Petljak et al. Characterizing Mutational Signatures in Human Cancer Cell Lines Reveals Episodic APOBEC Mutagenesis. Cell
176, 1282-1294.e20 (2019).
10. Liu et al. An Integrated TCGA Pan-Cancer Clinical Data Resource to Drive High-Quality Survival Outcome Analytics. Cell 173,
400-416.e11 (2018).
11. Wojtaszek et al. A Small Molecule Targeting Mutagenic Translesion Synthesis Improves Chemotherapy. Cell 178, 152-
159.e11 (2019).
12. Ler & Carty. DNA Damage Tolerance Pathways in Human Cells: A Potential Therapeutic Target. Front Oncol vol. 11 (2022).
13. Vaziri, Rogozin, Gu, Wu & Day. Unravelling roles of error-prone DNA polymerases in shaping cancer genomes. Oncogene
2021 1–17 (2021).
121
Multi-cancer detection testing in clinical practice. 1,2,3,4 - Dr. Brian
Nicholson
Primary Supervisor: Dr. Brian Nicholson
Second Supervisor:

Non-invasive MCED tests presents a new opportunity to improve early cancer detection by optimising patient selection
for targeted cancer testing. Whilst MCED technologies are designed to detect a cancer signal across multiple cancer
sites, their performance varies by cancer site and cancer stage. These technologies will not be used in isolation by
clinicians in primary or secondary care: they will be used in people with a prior risk of cancer based on their risk factors
(most importantly age), symptoms, signs, and test results, who are referred into clinical pathways for definitive testing.
Care will be needed to select at-risk populations that complement the performance characteristics of the test to balance
the likelihood of missed cancers and unnecessary referrals for invasive or expensive investigation. With a rapidly
increasing number of MCED technologies in development, their performance characteristics are likely to improve.
However, understanding the performance characteristics of MCED technologies alone will not be sufficient to guarantee
the success of their implementation. Many promising innovations fail to reach clinical adoption as little attention has
been given to the drivers of uptake in clinical practice. Successful clinical implementation of MCEDs in clinical practice
is critically dependent upon intimate understanding of the patient, clinician and system-level factors that influence
uptake. The successful candidate would join an exciting multidisciplinary programme of work investigating the accuracy,
utility, and implementation of MCED testing in NHS clinical practice.
Research objectives
There is scope for the successful candidate to develop research objectives within the broad framework of the MCED
focussed CRUK Oxford Cancer Centre’s Early Detection theme. The Early Detection theme focusses on patient selection
for MCED testing, MCED test development, and MCED test evaluation in clinical practice. The successful candidate will
be supported to develop and lead research into MCED testing using methods that suit their intended career path.
Examples of areas for development could be to:
• compare the performance of existing risk algorithms and clinical guidance to identify populations most at risk
of cancers (combined and individually) who could be offered MCED testing by using existing health records
data or by developing studies to collect new cohort data. These multi-parametric algorithms could take
patterns of a patient’s symptoms, signs, test results, consultation patterns, medical history and risk factors to
calculate their individual risk of cancer diagnosis to be updated as MCED tests are completed.
• Utilise the Rapid Diagnostic Centre Digital Research Platform (RDC-DRP) curated to include clinical, research,
and biobank data derived from the expanded Suspected CANcer (SCAN) pathway and biobank. The RDC-DRP
could support fundamental and basic science researchers seeking to study early-stage disease and enhance
risk factor and symptom data capture, clinical epidemiologists interested in the MCED signatures in patients
with non-specific symptoms, and health services researchers hoping to use an online secure patient survey
portal to collect patient data prior to and following their appointment.
• develop community-based prospective MCED cohorts and trials engaging patients across to promote diversity
and inclusivity with the team who delivered the SYMPLIFY study. Together with a focus on assessing the
accuracy and placement of MCED technologies within NHS clinical workflow key implementation questions
could be asked using qualitative methodologies to understand the public, patient, clinician and system-level
factors that influence MCED uptake and impact.
122
In order for the NHS to maximise the benefit of MCEDs for patients in clinical practice research is required to
understand how MCEDs complement existing diagnostic pathways, if they replace commonly used diagnostic
tests, and how patients and practitioners will use them. As MCEDs develop, with improved or different analytical
performance, the candidate’s research findings will be required to understand where to best place MCED in the
diagnostic pathway. Oxford is uniquely placed to investigate MCED technologies as the supervisory team are
involved in the development of MCED technologies and NHS evaluations of MCEDs in clinical practice.
In addition to the training provided by the Cancer Science DPhil programme, NDPCHS offers broad methodological
expertise in applied health services research and evidence-based health care with training available to support the
approach chosen by the candidate under guidance from their supervisory team. For example, the Medical Statistics
group specialises in quantitative diagnostic, prognostic, monitoring, and prediction methodologies, the Medical
Sociology and Health Experiences Research Group specialises in social science informed, qualitative and mixed
methods implementation studies of health and illness, and the Primary Care Clinical Trials Unit delivers world class
clinical trials in the community. In addition, the Oxford-led NIHR Community Healthcare MedTech and In vitro
Diagnostics Co-operative (CH-MIC) works upstream and downstream of the CE-marking process to both influence
the development of novel technologies and the evaluation of clinic-ready products.

123
Understanding STING regulation in cancer and the crucial role of
ubiquitination in the ER – 1,3 Prof. John Christianson
Primary Supervisor: Prof. John Christianson

Second Supervisor: Assoc Prof. Eileen Parkes
Eligibility: Track 1 and 3 students are eligible to apply for this project
Abstract
Cancers interact with their surrounding environment (the tumour microenvironment) by remodelling it to contain cells
promoting tumour invasion and spread, and resistance to anti-cancer therapies. Innate immune pathways, typically
used to defend cells from infection by viral and bacterial pathogens, are hijacked in cancer. The mechanisms by which
cancer cells modify innate immunity are currently not well understood. A key pathway is the cGAS-STING pathway – the
cytoplasmic sensor cGAS recognises non-self or mislocalised DNA and activates STING (the STimulator of Interferon
Genes). STING is embedded in the endoplasmic reticulum (ER) – activation of the STING-mediated interferon response
requires oligomerisation and efflux from the ER (1). Fine tuning of this response is paramount, and ubiquitination of
STING has emerged as an important post-translational modification capable of modulating these signalling events.
Importantly, evidence is emerging of important interferon-independent effects of cGAS-STING signalling which may
drive tumour progression. Establishing how ubiquitination and its conjugating machinery impact the cGAS-STING
pathway is key to understanding how cancers subvert this pathway to their own ends.
This DPhil project will biochemically and functionally characterise ER-resident ubiquitination machinery that modulates
STING signalling in order to delineate its regulation of the interferon response.
Research Objectives and Outcomes

Recently, our lab identified a multi-subunit complex organised around ER-resident ubiquitin ligase (E3) RNF26, whose
constituents modulate signalling through STING to scale the magnitude of the interferon response (2). We are now
investigating how each component of this RNF26 complex impacts STING to contribute to the response, focusing on
defining protein-protein interactions, key functional domains, ubiquitin linkages, complex assembly, and its synergy (or
competition) with other ubiquitin ligases. This is crucial as understanding STING regulation will identify mechanisms of
resistance to immune targeting agents (immune checkpoint blockade and STING agonists) in advanced cancers.
Objective (1): Molecular dissection of ubiquitin conjugating machinery competing to modify STING in the ER. Genomic
editing, gene silencing and dominant negatives will establish the individual and combinatorial contributions of ER-
resident E3s (RNF26, RNF5, gp78, Hrd1) to STING properties including its; stability/degradation, ubiquitination profile,
oligomerisation, trafficking, and activation of the downstream interferon response, in model cell lines. The diversity and
dynamic nature of ubiquitin chain linkages modifying STING will then be explored using both mass spectrometry and
sensitivity to linkage-specific deubiquitinases.
Outcome: Establishment of key ubiquitination events governing STING in the ER and consequently the magnitude of its
downstream signalling cascade.
Objective (2): Defining how cofactors contribute to ER-E3 recognition and/or ubiquitination of STING. Potentially
important regulatory domains of E3 complex components identified bioinformatically will be evaluated functionally
using truncations and site-directed mutagenesis. This will be complemented by proximity-labeling strategies coupled
with proteomics to define the spatiotemporal organisation of E3 complexes and their interaction/s with STING. STING
agonists and antagonists (currently being developed for clinical applications) will be used to pharmacologically probe
for changes in E3-STING interaction.
Outcome: An understanding of the how ubiquitin multifaceted regulation of STING at the ER influences response to
activating treatments.
124
Objective (3): Preclinical validation of STING modulating factors. Identified STING regulating factors will be modified
using gene editing and CRISPR-cas9 approaches using an ex vivo platform (i.e. culturing cells from patient samples).
These samples will be used to generate organoids with/without fibroblasts and patient-matched immune cells. This
near-patient system will be used to determine the effect of modulation of STING regulating factors on the tumour
microenvironment. Using this platform, immunotherapeutic stimulants can be added to determine the role of STING-
modulating E3s and co-factors in response to existing immunomodulating treatment will be investigated. Flow
cytometry and T cell activity assays will be employed to measure the impact of novel targets on immune response in
this near-patient model.
Outcome: Characterisation of targetable mechanisms of STING suppression determining response to cancer

immunotherapy.
Collectively this research will develop insights into the fundamental cellular controls of immune signalling. Along with
ongoing work in the lab, it will form part of our broad effort to explore and define ubiquitination events and mechanisms
at the ER responsible for essential cellular homeostatic functions
Translational relevance of the project

This project will address important fundamental and clinical questions relevant to personalising immunotherapy
treatment in cancer. Tailoring immune targeting approaches and understanding resistance mechanisms (such as STING
repression) has potential to improving clinical responses. In this study novel STING regulating mechanisms will be
characterised as potential biomarkers and/or targets for further clinical study. Moreover, this proposal uses patient
samples for 3D modelling, further supporting translation of this work to the clinical setting.
There will be multiple training opportunities available during the project including; advanced cell biology and
biochemistry, proteomic sample preparation and analysis, flow cytometry, and 2D and 3D cell culture modelling. There
will also be opportunities to present findings at local, national and international conferences.
References
1. Hopfner K and Hornung V (2020) Molecular mechanisms and cellular functions of cGAS–STING signalling.
Nature Reviews Molecular Cell Biology. 197: 1-21 (LINK)
2. Fenech EJ, Lari F, Charles PD, Fischer R, Thezenas ML, Bagola K, Paton AW, Paton JC, Gyrd-Hansen M, Kessler
BM, Christianson JC (2020) Interaction mapping of endoplasmic reticulum ubiquitin ligases identifies
modulators of innate immune signaling. eLife 2020;9:e57306 DOI: 10.7554/eLife.573 (LINK)
3. Parkes EE, Walker SM, Taggart LE, et al. Activation of STING-Dependent Innate Immune Signaling By S-
Phase-Specific DNA Damage in Breast Cancer. J Natl Cancer Inst . 2017;109(1). doi:10.1093/jnci/djw199.

125
Data Driven approaches for addressing inequalities in Cancer
Outcomes - 1,2,3,4 Prof Eva Morris
Primary Supervisor: Prof. Eva Morris
Second Supervisor: Prof. Sasha Shepperd and Dr Ben Lacey
Eligibility: All tracks are eligible to apply for this project
Abstract
There is growing concern that there is significant variation in the quality of cancer care and outcomes across the UK. 1
Although the NHS seeks to ensure that everyone receives the best treatment available and attains the best outcomes
possible, significant variation between patient groups persists. The first steps to tackling such inequities are quantifying
their size, investigating why they arise and understanding what impact they have on patient outcomes. 2 Data are key to
enabling this but, at present, there is a lack of available information to fully explore, and hence understand, cancer
inequalities and so generate evidence to tackle them.
For some social groups there are good data to categorise people and quantify differences in outcomes but less
information about what is driving those differences.3 For example, it is possible to group people into age categories and
demonstrate differences in care and outcomes, but understanding whether these differences are driven by associated
factors such as levels of frailty or comorbidity, or indicative of true inequities, is challenging. In contrast, for other
potentially excluded groups there is extremely limited information to even identify members, let alone investigate
patterns and outcomes of care. For example, identifying individuals in the LBGTQ+ community, or even ethnicity, in
routine administrative data used to investigate epidemiological trends is extremely challenging.
This project will investigate the administrative data available to quantify cancer-related inequalities in the UK, as well
as to use these data to understand what drives them and generate the evidence needed to tackle any inequities found.

The project’s research objectives are:
1. Systematically review the existing evidence base on cancer inequalities and map the data available to investigate
them
2. Use the evidence gathered to identify and fulfil novel opportunities to quantify inequalities and undertake such
studies. For example, using the cancer patient experience survey to investigate issues relevant to the LGBTQ+
3. Use the evidence gathered to identify and fulfil novel opportunities to better distinguish between unavoidable
inequalities and true inequities, and understand what is driving any variation seen. For example, investigate to
what extent lower treatment rates in older people are a result of age-discrimination, or ineligibility for treatment
due to concomitant disease or frailty.

The translation potential of the project is substantial as there is good evidence of substantial variation in cancer care
across the UK, and the outcomes obtained lag behind many of our economic neighbours with similar health systems.
Eliminating cancer inequalities is a priority for cancer charities and the National Health Service but there is limited
evidence to inform interventions. This study will help provide the information required.
Summarise the training opportunities

The student will work within the rich academic environment of the Nuffield Department of Population Health and
affiliated institutions, gaining research experience and skills training in epidemiology and health data science. Training
will also include information governance and research ethics, and patient/public engagement to ensure appropriate and
effective use patient level administrative data. Translational/implementation skills will also be developed to help enable
research findings to inform cancer policy to improve health outcomes. Candidates will receive professional mentorship
through regular supervisory meetings, and acquire research skills by attending seminars and workshops. Candidates will
126
work closely with other team members, and will have the opportunity to communicate their findings in international
conferences and with the public.
References
1. Martins T, Abel G, Ukoumunne OC, et al. Ethnic inequalities in routes to diagnosis of cancer: a population-based UK
cohort study. Br J Cancer 127(5): 863-71, 2022
2. S. Vaccarella et al., ‘Reducing Social Inequalities in Cancer: Setting Priorities for Research’, CA Cancer J Clin, vol. 68,
no. 5, pp. 324–326, Sep. 2018
3. M. Marmot, S. Friel, R. Bell, T. A. J. Houweling, and S. Taylor, ‘Closing the gap in a generation: health equity through
action on the social determinants of health’, The Lancet, vol. 372, no. 9650, pp. 1661–1669, 2008

127
Molecular enhancement of vaccine-induced immune responses for
cancer therapy - 1,2,3 Prof. Paul Klenerman
Primary Supervisor: Prof. Paul Klenerman
Second Supervisor: Nicholas Provine
Eligibility: Tracks 1,2 & 3 are eligible to apply for this project
Abstract
The role of immune responses in cancer therapy has been highlighted in recent years by the success checkpoint inhibitor
therapy. However, not all cancers respond to such therapies and other methods to initiate or boost anti-cancer
responses are being explored. This includes the development of new cancer vaccines. Leading molecular platforms for
such vaccines – as for infectious diseases - are mRNA and Adenovirus approaches. By introducing relevant target
antigens into these vectors, it is possible to generate T and also B cell responses to tumours. However, the efficacy of
such approaches currently is very variable. To be most effective in a cancer setting, the strongest possible immune
response must be first induced and then boosted and sustained: the molecular rules which govern these processes for
Adenoviral vectors and mRNA are not yet known, particularly the latter. There are a number of factors which can
influence this – in particular the local cell-cell signals which occur at the interface of an antigen presenting cell (such as
a dendritic cell) and a T cell, which can be through surface proteins and secreted cytokines at the immunologic
“synapse”.
In this project we are partnering with our colleagues at the University of Naples, led by Alfredo Nicosia, who was
instrumental in the development of simian adenoviral vectors as used in Covid-19. The team have systematically
developed a set of vectors which can be co-delivered with mRNA and/or Adenoviral vaccines to enhance their
immunogenicity and hence their efficacy. These vectors include a range of surface proteins, soluble triggers and soluble
blocking molecules which can influence the priming environment of the T and B cells and impact of immune memory.
We aim to explore the functional impact of these new molecular “adjuvants” in vivo following vaccination, using
functional assays, spatial and genomic approaches to define the quality of immune memory induced. By careful analysis
of the impact of a range of such new tools we will help both select the optimal approach for each platform, and also
understand much better the rules which govern different types of immune priming – thus applicable across many
platforms.
This project overall represents an exciting opportunity to work in a highly translational project which will not only
develop new therapeutic tools for cancer but also use novel techniques to address fundamental questions about
vaccines and immunology.
Fig | Molecular adjuvant-expressing Ad vector enhances tumour control. Ad-9D9 is a MA-expressing vector that enhances
control of CT26 tumor growth in vivo in mice when combined with a cancer vaccine expressing T cell epitopes (GAd). D’Alise
… Nicosia, Front Immunol, 2023

In this programme of work, a series of mRNA and Adenoviral constructs (molecular adjuvants, MAs) are currently being
developed by our collaborators and screened for their ability to enhance T cell responses in vivo following vaccination
with an established cancer vaccine. The impact of the MAs is being tested for their ability to boost immunogenicity using
128
ex vivo T cell assays, and to provide in vivo protection using a standardised challenge with the MC38 tumour cell line.
We have already had experience in our lab with an early product of this pipeline. Tools to assess the immunogenicity
include FACS staining using MHC-peptide tetramer tools to evaluate individual tumour-specific responses, a range of
functional assays to assess cytokine release and killing by such T cells, single cell RNAseq approaches to evaluate the
underlying regulation of these cells and their T cell receptor usage and multicolour staining and spatial transcriptomic
approaches to understand the local niche where priming and boosting occurs. These are essential tools in modern
molecular immunology. We will also assess a second model developed in the lab for SARS-CoV-2, which includes
evaluation of B cell responses in molecular detail. We aim to assess not only the impact on the antigen-specific cells,
but also on the cells in the priming microenvironment, and address the question as to where and how the antigen is
presented and what the additional impact of each MA is. The effects will differ between mRNA and adenoviral vectors
and between boost and prime, adding further depth to the project. There will be the opportunity to test a series of
hypotheses related to the specific action of each co-stimulatory and co-inhibitory pathway.
The project is an important new collaboration between the University of Oxford and University of Naples – we have
already had one visit from a PhD student from Naples so we are sure this connection is good and the MA pipeline is
flowing already. It also represents a collaboration between the newly established Provine lab, with a focus on vaccine
immunobiology, and the Klenerman lab, with a focus on T cell function.

New cancer vaccines based on mRNA or adenoviral vectors are already under development. These tools will strongly
enhance such approaches, including the development of strong cytotoxicity and the maintenance of immunity in the
face of exhaustion within tumours. Our labs have a track record of moving such constructs into human trials,
collaborating on the first studies brining a range of new adenoviral vectors to the clinic and establishing the mechanisms
of immunogenicity. These approaches can be applied to a range of targets, especially cancers of the GI tract, which are
generally less susceptible to checkpoint inhibition, and where new approaches to induce immunity are a key focus for
many groups.
As mentioned above, the students would be trained in flow cytometry, cell culture, functional B and T cell assays, in vivo
immunology techniques, as well as molecular approaches including scRNASeq, TCR sequencing, spatial transcriptomic
studies, and the accompanying bioinformatic approaches (mainly based in R studio). As well as testing the panel of new
molecular tools generated by our collaborators, there will also be opportunities to develop new adenoviral vector and
mRNA tools.
References
Barnes, E., … A. Nicosia, R. Cortese, and P. Klenerman, Novel adenovirus-based vaccines induce broad and sustained T
cell responses to HCV in man. Sci Transl Med, 2012. 4(115): p. 115ra1.
Provine, N.M., A. Amini, L.C. Garner,…and P. Klenerman, MAIT cell activation augments adenovirus vector vaccine
immunogenicity. Science, 2021. 371(6528): p. 521-526.
Garner, L.C., …N.M. Provine*, and P. Klenerman*, Single-cell analysis of human MAIT cell transcriptional, functional and
clonal diversity. Nature Immunology, `(in press) September 2023.
Cupovic, J….N.M. Provine… P. Klenerman*, and B. Ludewig, Adenovirus vector vaccination reprograms pulmonary
fibroblastic niches to support protective inflating memory CD8(+) T cells. Nat Immunol, 2021. 22(8): p. 1042-1051.
D’Alise, A.M. … A. Nicosia, Adenovirus Encoded Adjuvant (AdEnA) anti-CTLA-4, a novel strategy to improve Adenovirus
based vaccines against infectious diseases and cancer. Front Immunol 14, 1156714 (2023).

Peptide specific natural killer cell receptors in cancer – 2,3 Malcom
Sim
Primary Supervisor: Malcom Sim
Second Supervisor: Prof. Tim Elliott
Eligibility: Tracks 2 & 3 are eligible to apply for this project
Abstract
Natural killer (NK) cells are important innate immune effector cells that can directly kill tumours and produce
proinflammatory cytokines. NK cell function is tightly regulated by an array of germline encoded activating and inhibitory
receptors. The killer-cell immunoglobulin-like receptors (KIR) are a family of activating and inhibitory NK cell receptors
that bind class I human leukocyte antigens (HLA-I). HLA-I molecules present short peptide antigens to T cells but the
contribution of HLA-I bound peptides to KIR binding is poorly understood. Recent work by the Sim lab demonstrated
that KIR binding can be highly peptide dependent, with activating KIR displaying high peptide-specificity, analogous to
recognition by CD8+ T cells [1, 2]. The goal of this project will be to develop high-throughput systems to identify novel
ligands for activating KIR across multiple HLA-I allotypes. Tools that can predict KIR binding peptides from sequence will
be developed using machine learning and other computational approaches. Using large tumour genome datasets (such
as TCGA), predicted neoantigens will be analysed for their impact on predicted KIR binding and correlated with
prognosis. Further work will explore the translational potential of delivering activating KIR binding peptide to tumours
for the purpose of redirecting NK cells towards. This work will advance our basic understanding of peptide-specific NK
cell receptors, their role in NK cell recognition of tumours and explore their therapeutic potential.

The KIR are critical regulators of NK cell function. Disease association studies link specific KIR-HLA-I ligands with multiple
disease processes including cancer, infectious disease, autoimmunity and disorders of pregnancy [3]. Understanding the
unique features of KIR-HLA-I interactions is necessary to understand these disease associations and the contribution of
KIR+ NK cells. Furthermore, defining the peptide-specificity of KIR is required to potentially exploit this property in
therapeutic settings. The award will allow the student to absorb the technical and scientific knowledge of multiple
collaborators, while adapting them to the problem of determining KIR specificity. The student will translate this
knowledge to understand the contribution of KIR binding peptides to NK cell recognition of cancer.
Objective 1. Develop high throughput screens to determine KIR peptide-specificity

Comprehensively determining the sequences of KIR binding peptides involves multiple layers of complexity. Firstly,
there are many KIRs. Our lab focuses on the HLA-C binding KIRs of which there are 3 inhibitory receptors (KIR2DL1,
KIR2DL2 and KIR2DL3) and four activating receptors (KIR2DS1, KIR2DS2, KIR2DS4 and KIR2DS5). Secondly, HLA-C is highly
polymorphic, with thousands of different allotypes. Uncovering bone fide ligands for KIR bind is fundamental to
understanding their contribution to NK cell function and diseases linked to specific KIR-HLA-I combinations. By
collaborating with Dr. Ricardo Fernandes (CAMS-COI), we will exploit the multiple technologies (such as yeast display)
previously developed for determining the specificity of T cell receptors (TCRs) [4]. We will develop cell-based platforms
to express large libraries of DNA encoded HLA-I bound peptides for screening via binding to soluble KIR.
Objective 2. Utilize KIR binding data into tools that can predict KIR binding from sequence
KIR binding screens (Objective 1) will generate large datasets of KIR binding and non-binding peptides. The next goal will
be to integrate this data into tools that can predict KIR binding peptides from sequence. Predicting KIR binding peptides
is necessary to understanding how changes in HLA-I bound peptides can impact KIR binding and NK cell function.
Developing these tools will be in collaboration with Prof. Hasheem Koohy (WIMM), an expert in machine learning whose
team is developing tools to predict ligands for TCRs.
Objective 3. Investigate the contribution of peptide dependent changes in KIR binding to NK cell recognition of tumours
With the ability to predict KIR binding from sequence (Objective 2), the contribution of KIR binding peptides to NK
recognition of tumours will be investigated. Firstly, the predicted impact of neoantigens on KIR binding will be assessed
in large patient datasets with known tumour genome sequencing and patient prognosis, such as the TCGA. Predicting
neoantigens from TCGA is feasible and has been published by other groups [5]. Patients will be stratified by whether
neoantigens are predicted no have no impact on KIR binding, increase binding to activating receptors or increase binding
to inhibitory receptors. Our second approach will investigate how delivering peptide ligands for activating KIR to
tumours may render them susceptible to NK cell killing. Using soluble peptide first, followed by antibody-based
130
technology [6], we will test whether changes in the HLA-I bound peptides on tumours can modify their susceptibility to
NK cell lysis.
With the ability to predict KIR binding from sequence (Objective 2), the contribution of KIR binding peptides to NK
recognition of tumours will be investigated. Firstly, the predicted impact of neoantigens on KIR binding will be assessed
in large patient datasets with known tumour genome sequencing and patient prognosis, such as the TCGA. Predicting
neoantigens from TCGA is feasible and has been published by other groups [5]. Patients will be stratified by whether
neoantigens are predicted no have no impact on KIR binding, increase binding to activating receptors or increase binding
to inhibitory receptors. Our second approach will investigate how delivering peptide ligands for activating KIR to
tumours may render them susceptible to NK cell killing. Using soluble peptide first, followed by antibody-based
technology [6], we will test whether changes in the HLA-I bound peptides on tumours can modify their susceptibility to
NK cell lysis.
Translational potential.
The translational potential of this project is two-fold and described largely in Objective 3. Firstly, patient
prognosis/response to therapy will be analysed in the context of predicted changes in KIR binding due predicted
neoantigens. This may lead to discoveries of patient subgroups more amenably to NK based therapeutics. Secondly,
methods to deliver peptides ligands for activating KIR will be developed with the goal of redirecting tumour associated
NK cells towards the tumour. In addition, this work will provide a deeper understanding of KIR-HLA-I interactions
facilitating the development of novel therapeutic opportunities in diseases associated with KIR.
Training opportunities.
The student will deploy a multitude of wet and dry lab techniques as part of this project. Prominent wet lab techniques
will be molecular biology (cloning, library preparation etc), cell biology (transfection, transduction, sterile culture) and
flow cytometry. Dry lab techniques will include standard lab tools such as FlowJo and Prism, some basic coding and
database searches will also be employed.
References
1. Sim, M.J.W., et al., Innate receptors with high specificity for HLA class I-peptide complexes. Sci Immunol,
2023. 8(87): p. eadh1781.
2. Sim, M.J.W., et al., Human NK cell receptor KIR2DS4 detects a conserved bacterial epitope presented by
HLA-C. Proc Natl Acad Sci U S A, 2019. 116(26): p. 12964-12973.
3. Parham, P. and A. Moffett, Variable NK cell receptors and their MHC class I ligands in immunity,
reproduction and human evolution. Nat Rev Immunol, 2013. 13(2): p. 133-44.
4. Joglekar, A.V. and G. Li, T cell antigen discovery. Nat Methods, 2021. 18(8): p. 873-880.
5. Wu, J., et al., TSNAdb v2.0: The Updated Version of Tumor-specific Neoantigen Database. Genomics
Proteomics Bioinformatics, 2022.
6. Millar, D.G., et al., Antibody-mediated delivery of viral epitopes to tumors harnesses CMV-specific T cells for
cancer therapy. Nat Biotechnol, 2020. 38(4): p. 420-425.

Modulating inhibitory receptor signalling to enhance the T-Cell
response in Cancer – 2,3,4 Ricardo Fernandes
Primary Supervisor: Ricardo Fernandes
Second Supervisor: Prof. Omer Dushek
Eligibility: Tracks 2, 3 and 4 are eligible to apply for this project
Abstract
The T cell response against pathogens, tumours, and self relies on integrating signals from diverse receptors, including
the T cell receptor, co-stimulatory receptors, and inhibitory receptors (IRs). Inhibitory receptors counterbalance positive
signals, suppressing T cell responses. A feature of IRs is the presence of signalling motifs in their cytoplasmic tail, which
attenuate T cell responses. While antibody blockade, which limits IR/ligand interactions, represents a significant
advancement in antitumour therapy, only a small fraction of patients benefits from this approach. Moreover, the
mechanistic basis whereby IRs damped T cell responses is still poorly understood, and a systematic comparison of the
potency, mediators, and targets of various clinically relevant IRs has not yet been undertaken. This project aims to
dissect ligand-driven and ligand-independent IR signalling processes to gain insights into the IR signalling mechanism.
This will be achieved by determining the potency of IR signalling for a group of receptors, identifying direct mediators
of IR signalling, and developing a phosphatase-mediated approach to inhibit IR signalling. This proposal seeks to deepen
our understanding of IR signalling and generate novel insights to potentiate T cell responses in cancer.

Regulation of T cell signalling by immune checkpoints such as PD-1 and CTLA-4 has been at the centre of recent
breakthroughs in cancer immunotherapy. Signalling by PD-1 and CTLA-4 reduces T cell activity and contributes to an
“exhausted” phenotype, severely compromising antitumor
responses. In the case of PD-1, binding to PD-L1/2 triggers
the tyrosine phosphorylation of signalling motifs and results
in the recruitment of cytosolic phosphatases such as
SHP1/2, which in turn reduces TCR and CD28 signalling.
Strikingly, signalling by several immune receptors relies on
the Tyr phosphorylation of signalling motifs. We
hypothesise that tonic receptor phosphorylation and
sustained signalling by ‘ligand-experienced’ receptors
impact T cell function and resist classic antibody blockade.
To address this issue, we engineered a bi-specific molecule
to recruit CD45, an abundant and promiscuous receptor
tyrosine phosphatase, within close proximity of PD-1 (Fig.
1)1. In this approach, the phosphatase domain of CD45 acts
intracellularly, in cis, on the p-Tyr residues of the PD-1
signalling motif, thus inhibiting sustained signalling. We Fig. 1 | Receptor Inhibition by Phosphatase
have shown that Receptor Inhibition by Phosphatase Recruitment. RIPR combines checkpoint blockade
Recruitment (RIPR), potentiates T cell activity beyond that with CD45 recruitment. CD45 recruitment
seen with PD-1/PD-L1 antagonist antibodies, both in the dephosphorylates receptors such as PD-1.
presence and absence of PD-1 ligand-binding in vitro, and
reduces tumour growth in mouse models1.
Aim 1: Determine the relative signalling potency of IRs in T cells

IRs have a distinct organisation of inhibitory motifs in their cytoplasmic domain, ranging from non-conventional single
tyrosine (Tyr) motifs to a combination of two or more “classic” inhibitory motifs like ITIMs. These distinct domains are
likely to affect the IR signalling potency. To test this, we will determine the relative IR potencies by establishing an IR
signalling platform in primary CD4+ and CD8+ T cells. We will perform systematic screens to determine the ability to
suppress T cell function by various IRs, including PD-1, TIGIT, BTLA, TIM-3, CTLA-4, LAIR-1, ILT-2 and ILT-4. In addition,
we will compare IR signalling potencies in the presence and absence of IR-ligand binding using surrogate receptor-ligand
pairs.
132
Aim 2: Identification ofxc early and late mediators of IR signalling
We will determine the direct signalling mediators to IR signalling using targeted protein pull-down strategies
complemented by mass spectrometry (MS)-based proteomics (in collaboration with Dr Adan Pinto-Fernandez’s Group)
and epigenetic profiling (in collaboration with Dr. Chunxiao Song’s Group). Moreover, we will map the phosphorylation
status of the inhibitory motifs and protein-complex composition by western blot and optimised affinity purification.
Next, we will determine the contribution of specific mediators and transcription factors by deleting a single or a
combination of targets using CRISPR/Cas9. Collectively, this approach aims to identify early and late signalling mediators
of various IRs and their role in suppressing T cell responses.
Aim 3: Development of RIPR-based molecules to inhibit inhibitory receptors

Bispecific diabodies that recruit CD45 phosphatase to IRs, such as PD-1 and CTLA-4, were found to potentiate T cell
responses. We aim to extend this concept to target additional IRs, including BTLA, TIGIT and TIM-3, which have been
implicated in suppressing antitumour responses. Moreover, we will systematically test the potency of newly generated
RIPR proteins using various anti-CD45 nanobodies. Newly generated molecules will be characterised in biophysical and
in vitro stimulation assays. Binding on-rate, off-rate and affinity will be determined by surface plasmon resonance. After
characterising the binding properties of the RIPR molecules, their ability to potentiate T cell cytotoxic functions will be
determined in vitro using co-culture assays with T cells and target cells. Markers of T cell activation will be quantified
longitudinally using flow cytometry, western blotting and ELISAs. This comprehensive approach is expected to identify
determinants of RIPR activity for various checkpoint receptors. This information will be used to guide the design of
future antagonists of checkpoint receptor signalling with strong potential for therapeutic applications.
Summary of milestones and expected outcomes

(i) Obtaining a comparative list of signalling potencies for all tested IRs;
(iii) identification of early and late signalling mediators of IRs;
(iv) Providing proof-of-concept for inhibiting IR signalling by phosphatase recruitment;

We expect that the described approach will establish a rapid and facile method to systematically probe the contribution
of inhibitory receptors in suppressing T cell effector functions. This information will enable the identification of new
targets and guide the development of IR therapeutics. We anticipate the next stage of immunotherapy development to
include new molecules that exploit specific aspects of the mechanisms involved in receptor signalling. The RIPR approach
may offer a new avenue to directly target receptor phosphorylation and shut down inhibitory receptor signalling with a
strong potential for being used to target various surface receptors found in distinct immune cell populations.
The candidate will receive training in molecular biology, protein design and expression, biophysical characterisation of
protein interactions and various cellular assays. Moreover, the candidate will be trained in protein engineering, library
design and selection using yeast display. T cell signaling assays will be used to determine the activity of newly generated
RIPR molecules, providing an opportunity for training in flow cytometry and MS, among others. This training will allow
the candidate to drive fundamental and applied research in academia and industry. The candidate will have full access
to the facilities and resources available within the Department and the broader community at the University of Oxford.
Key publications
1. Fernandes RA, Su L, Nishiga Y, Ren J, Bhuiyan AM, Ali LR, Majzner R, Ohtsuki S, Rietberg SP, Yang X, Picton L, Savvides
CS, Mackall, CL, Sage J, Dougan M, Garcia KC. Immune receptor inhibition through enforced phosphatase
recruitment. (2020) Nature, Oct;586(7831):779-784
2. Fernandes RA*, Ganzinger KA*, Tzou J, Jonsson P, Lee SF, Palayret M, Santos AM, Chang VT, Macleod C, Lagerholm
BC, Lindsay AE, Dushek O, Tilevik A, Davis SD, Klenerman D. A cell-topography based mechanism for ligand
discrimination by the T-cell receptor. (2019) Proc Natl Acad Sci U S A. Jul; 116(28), 14002-14010
3. Chang VT*, Fernandes RA*, Ganzinger KA*, Lee SF*, Siebold C, McColl J, Jönsson P, Palayret M, Harlos K, Coles CH,
Jones EY, Lui Y, Huang E, Gilbert RJ, Klenerman D, Aricescu AR, Davis SJ. Initiation of T cell signaling by CD45
segregation at 'close contacts'. (2016) Nat Immunol. May;17(5):574-82

133
Engineered oncolytic herpes virus therapy for neuroblastoma –
1,2,3,4 Dr Alistair Easton
Primary Supervisor: Dr Alistair Easton
Second Supervisor: Dr Kerry Fisher
Abstract
Neuroblastoma (NB) is a peripheral neuroectodermal malignancy that accounts for 15% of paediatric cancer related
mortality and is the most common fatal extra-cranial malignancy in children. Current treatment for high-risk cases is
multimodal, involving surgery, chemoradiotherapy and bone marrow ablation. This is a gruelling treatment regime with
significant toxicity and high risk of relapse.
Immunotherapy, in the form of a mAb targeting ganglioside GD2, has resulted in a step-change in survival. However,
treatment is often sub-optimal and limited due to on-target, off-tumour binding to peripheral nerves, which results in
intractable neuropathic pain. Despite significant improvements in outcomes, the majority of high-risk patients still die
from their disease.
Neuroblastoma is a model immune-cold tumour, that downregulates MHCI and exhibits low mutational burden. As
such, patients do not respond to immune check-point inhibitors1. However, complete spontaneous regression of widely
metastatic, stage 4s, NB is a well-recognised phenomenon that correlates with evidence of immune activation 2,
including autoimmunity. In addition, pre-treatment immune infiltrate is highly predictive of survival. These observations
indicate that NB will be amenable, and highly sensitive, to immune therapy, if an effective protocol can be developed.
This research project explores a novel approach that leverages attenuated herpes viruses to enhance the precision and
efficacy of immune therapy against NB. We will collaborate with NB specialists at The University of Southampton, who
have developed and characterised a number of spontaneous and allograft NB mouse models, as well as patient derived
NB cell lines3. These are ideal for testing our viral constructs, which are capable of infecting murine and human cells,
and delivering immunostimulating molecules.
The research involves a multi-faceted approach, combining virology, immunology, and cancer biology. We will use single
cell multiplex imaging and advanced image analysis software to investigate the ability of our viral constructs to infiltrate
tumour tissue, trigger localized immune responses, and induce tumour cell death. Furthermore, the project seeks to
understand the potential synergistic effects of genetically modified herpes viruses, expressing existing
immunotherapeutic modalities, such as anti-GD2 mAb, cytokines and immune checkpoint inhibitors. We will examine
the impact of this approach on localised and overall systemic immunity, ensuring that the therapy elicits a potent and
specific anti-tumour immune response while minimizing adverse effects.
Neuroblastoma is an area of unmet need and this combinatorial approach may unlock new avenues for enhanced
tumour eradication and durable responses in cancer patients.
Objectives:
• Create attenuated oncolytic herpes virus constructs that express NB targeting mAb and immune stimulating
molecules.
• Use murine models and human cell lines to characterise virus and immune infiltration, focusing on localised
and systemic immune responses to NB.
• Optimise the combination of immune modulating modalities while minimising side-effects.
Proposed Outcomes:
• Using pre-clinical models, demonstrate enhanced immune infiltration and systemic activation to generate
durable anti-NB immune responses.
• Show localised and systemic tumour eradication and/or sensitisation to current therapies.
134
Academic value
Neuroblastoma is an ideal model immune-cold tumour that we know is amenable to spontaneous immune eradication
by pathways that we do not fully understand. This project is an opportunity to study and manipulate the anti-tumour
immune response through precision targeting of the tumour and its microenvironment.
Collaborations
This collaboration involves the Translational Histopathology Lab (THL), led by academic clinical pathologist Dr Alistair
Easton specialising in multiplex single cell imaging, and Prof Len Seymour and Dr Kerry Fisher’s gene therapy group,
specialising in precision immune therapy using viral delivery vectors. We will collaborate with Dr Juliet Gray’s
Neuroblastoma research group at the University of Southampton, who specialise in immunotherapies and preclinical
modelling of NB. This award will allow a doctoral student gain experience generating therapeutic viral vectors, applying
these to preclinical mouse models of disease and imaging/topological analysis of tumours.
As described earlier, NB is an area of unmet need. Current treatments are inadequate, resulting in high morbidity and
mortality in a paediatric population. The dire prognosis of this disease means it is a prime target for advanced precision
therapeutics, such as mAb. Whilst immune therapy has shown transformative potential, it is currently hampered by
toxicity, partly due to a lack of precision. This project addresses this issue by allowing the delivery of small molecules
directly into the tumour site. There are well established global clinical trial networks for this disease, some led by our
collaborators, meaning that any improvements in treatment will have a rapid path to clinic.
Training Opportunities:
This project will involve tissue culture, virology, immunology, oncology, molecular biology, genetic engineering, animal
handling, histology, multiplex fluorescence immunohistochemistry, image analysis, using machine learning software and
statistics.
References
1. Anderson, J., Majzner, R. G. & Sondel, P. M. Immunotherapy of Neuroblastoma: Facts and Hopes. Clinical Cancer
Research 28, 3196 (2022).
2. Hammouche, D., Easton, A., Lopez, M. & Gray, J. Combining multiplex immunohistochemistry with computer aided
topographical and single-cell analysis to investigate the mechanism of spontaneous regression in a case of metastatic
congenital neuroblastoma. (2020) doi:10.47496/NL.CRIO.2020.01.
3. Webb, E. R. et al. Immune characterization of pre-clinical murine models of neuroblastoma. Sci Rep 10, (2020).

Pre-Clinical modelling of oncolytic virus therapy for hepatic
Colorectal Cancer metastases using ex-vivo, perfused,
hepatectomies 1,2,3,4 Dr Kerry Fisher
Primary Supervisor: Dr. Kerry Fisher
Second Supervisor: Dr. Alistair Easton
Abstract
Colorectal cancer (CRC) is the fourth most common cancer in the UK, affecting 43,000 people annually. Hepatic
metastases occur in 25-30% of patients. If there are multiple deposits, treatment options are limited and 5-year survival
is 11%. In order to improve upon this, we need to better understand the complex vascular, immune and stromal
microenvironment in liver metastases, so that we can achieve better penetration of therapeutics and counter resistance
mechanisms.
Conventional cell culture systems fail to recapitulate the complexity of the vascular and stromal microenvironment. To
address this, we have developed an ex-vivo liver explant model, using tissue obtained from patients undergoing partial
hepatectomy for CRC metastases. These serve as an invaluable resource for studying real tumour responses within a
physiologically relevant context.
Initially, this project will determine the physiological effects of our perfusion model on the liver and tumour, using serial
histological biopsies and metabolic / hypoxia markers. Further to this, single-cell imaging/multiplex, will determine the
baseline immune and stromal microenvironment, and any changes due to the explant model. Establishing baseline
parameters will allow this model to be used to study chemotherapeutic, radiological and biological agents.
The second half of this project will involve perfusing our model with oncolytic viruses, engineered to deliver
immunomodulatory molecules directly into tumour cells. We will study the distribution and localised effects of virus
infection. In addition, we will introduce leukocytes in order to study immune-cell trafficking and in-situ immune
responses. Ultimately, we will optimise the immuno-modulatory agents to enhance tumour killing and provoke durable
anti-tumour immune responses. Dynamic sampling will allow a temporal analysis to capture real-time changes in gene
expression, cellular behavior, and therapeutic resistance. This comprehensive approach will provide unprecedented
granularity in understanding how metastatic tumours in this microenvironment respond to treatment
Objectives:
1. Characterise the baseline physiological and metabolic changes that occur in liver and tumour in our ex-vivo,
liver, explant model.
2. Deep phenotype and topologically analyse the baseline immune and stroma components in CRC hepatic
metastases.
3. Topologically map penetration, infection, spread and tissue responses (including immune responses) to
oncolytic virus infusion in our ex-vivo model.
4. Characterise and optimise tumour killing and in-situ immune responses following perfusion with armed
oncolytic viruses, with the aim of generating durable anti-tumour immunity with minimal bystander pathology.
Academic value
Our ex-vivo liver explant model is an invaluable resource for pre-clinical modelling of any liver primary, CRC and non-
CRC metastatic malignancy where tissue is available from partial hepatectomies. The model can used for conventional
chemotherapy, radiotherapy and embolization. This project addresses the need for a baseline understanding of the
physiological changes that occur due to the ex-vivo perfusion process. The dynamic sampling and imaging protocols,
that will optimised as part of this project, will be easily transferable into alternative therapeutic models.
Collaborations
This will be a collaboration between Prof Len Seymour’s and Dr Kerry Fisher’s molecular therapy group, the Translational
Histopathology Lab (THL), led by Dr Alistair Easton, and the Tumour Evolution and Cell Identity group, including Mr Alex
Gordon-Weeks. It will combine expertise in precision gene therapy, histopathology, imaging and hepatobiliary surgery.
This award will facilitate this new collaboration between these groups.
136
Translational Potential:
I. This will result in a robust and novel pre-clinical modelling tool that can be used in a variety of diseases and
with a variety of therapeutics.
II. Oncolytic viruses are a novel technique for precision gene therapy and immune stimulation. This project
will improve our understanding of the biodistribution and effects of these viruses, and allow us to test novel
constructs to achieve optimal responses.
Training Opportunities:
This project will provide opportunities/training in tissue culture, ex-vivo tissue preps/tissue handling, histological
analysis, multiplex imaging, molecular biology, virology, immunology and oncology.

137
Building Patient specific digital twins for cancer drug development
and cancer treatment - 1,2,3,4 Assoc Prof. Rachael Bashford Rogers
Primary Supervisor: Assoc Prof. Rachael Bashford Rogers
Second Supervisor: Dr. Isabela Pedroza-Pacheco
Abstract
Cancer poses a highly multi-factorial problem, where no single dataset captures the full complexity of the tumour, the
tumour microenvironment and accompanying host genetics, vasculature, lymph node and co-morbidity components
that have been shown to impact patient survival and response to treatments, resulting in a highly complex problem. By
capturing real-time data, the digital twin understands the current state, simulates the future state and is a basis for
optimisation. It enables to detect problems early, can be used as a basis for in silico testing and is an opportunity to
develop novel therapeutics. This project proposes to construct high-quality multi-scale cancer datasets, virtual cohorts
of patients and patient digital twins which can be used for improved screening, diagnosis, clinical decision making,
disease management and drug development. We will assess the short- and long-term predictive and prognostic power
of such digital twins and quantify the explainable and interpretable key features underlying these predictions. Through
the development of virtual clinical trials, we aim to accelerate the selection of drug combination and prioritisation in
patient therapies. Overall, this may lay the foundation for more accurate and mechanistically-driven clinical decision
support systems and accelerate therapeutic design in cancer.

Objective 1. Develop high-quality multi-scale cancer datasets and virtual patient cohorts. Cancer poses a highly multi-
factorial problem, where no single dataset captures the full complexity of cancer. In this project, you will:
1. Bring together cancer datasets containing all -omic data types describing key features, including key
genetics, transcriptomic signals, cellular neighbourhoods, cellular interactions, acellular signals, and wide
organ system or whole-body signals across tumour types and corresponding healthy tissue.
2. Train different multi-layered machine learning models, with others developed through the hub, to
characterise the short- and long-term predictive power, predict optimal patient treatment, provide
interpretable biological learning, and identify suitable lead targets and molecules in the drug discovery
process.
3. Generate virtual cohorts of patients or tumours (i.e. cohorts of human virtual twins) based on biological
data. These realistic populations of virtual patients will be generated with the desired clinical
characteristics in terms of clinical stage, grade, molecular alterations and ethnicity, sex, age and co-
morbidities. Metrics will objectively assess the performance of each model in making correct predictions.
4. Leverage virtual twins together with augmented synchronous data and partial longitudinal data to
reconstruct disease trajectories.
This work will be done collaboratively with experts in UCL and Sheffield.
Objective 2. Provide biologically interpretable outputs from ex vivo experimentally-informed virtual twins. We aim to
develop global models of tumour microenvironments (TME) which could be easily streamlined into a clinical workflow
using clinical data and routinely collected clinical samples. Such models will consist of five components:
1. To identify biologically-informed feature selection from routinely collected clinical datasets learnt from
higher resolution reference datasets. This may be applied to cancers where biopsies are routinely taken.
2. Clinical history data needs to be encoded and features selected, including tumour genotype.
3. Drug sensitivity experimental work for a subset of patient tumour and healthy tissue will be developed.
4. These three components could be fed into a model of the TME. This would then inform patient outcome
model in which key features could be pulled out.
5. Finally, these ex vivo experimentally-informed virtual twins will be projected onto the models in Objective
1 to inform and update in silico experiments which will be performed in the virtual clinical trials. Virtual
clinical trial simulations will be parameterised to reproduce the outcome of several real trials related to
cancer treatment, where responders, non-responders and partial responders may be observed.
138
Objective 3. Accelerating drug discovery for targeted patient groups. We will consider the different cancer drug classes
together for accelerated drug discovery for targeted patient groups together with key novelties in each area prioritising
specificity and reduction of off-target effects:
1. Chemotherapeutic agents targeting key tumour genes will be accelerated by coupling machine-learning
(ML) with physics-based (PB) methods. Data generated from PB methods will be fed into ML models which
are iteratively refined to generate potentially better structures. The most promising compounds will be
synthesised and their thermodynamic and pharmacokinetic properties be examined, and validated in in
vitro cell line experiments in collaboration with UCL and Sheffield.
2. Immunotherapeutic targets will be predicted and prioritised using perturbation network modelling of cell-
cell communication networks derived from single cell multi-omics data. This process will be run over
patient-specific tumour and germline mutational profiles, and biomarkers will be predicted for patient
stratification for each prioritised molecular target.
Comparisons of the effect of the different drug approaches and combinations will be tested using the patient slide-
perfusion system (outlined in objective 2) on relevant patient sets (collectively decided by clinicians, biologists and data
scientists). Our virtual twins can be used to predict the treatment efficiency in each patient. Finally, we will apply cross-
work package cutting approaches to gain insights into off-target effects of prioritised drugs/targets.
This study will provide a unique platform to understand the relationship between different tumour-associated features
across scales (molecular, cellular, acellular, organ systems, systemic), with the overall aim of defining improved
therapeutic options and patient outcomes on a personal basis. Furthermore, the methods developed here will not just
be broadly applicable to cancer, but will have wider applications in biotechnology and in health services worldwide. This
will be achieved through the development and application of novel experimental and computational approaches,
working in partnership with a global network of clinicians, immunologists and sample/data cohorts.
Training opportunities: The DPhil will gain experience and training in laboratory molecular biology, immunology, cancer
biology, and bioinformatics. These include:
• Genomic, bulk and single-cell transcriptomic, and other ‘omic analyses across large patient cohorts.
• Development and/or implemmentation of novel computational pipelines for the integration of multi-scale
longitudinal data with clinical covariates.
• Model building and generation of digital twins.
• The project will work in partnership with a global network of clinicians, immunologists, and computational
experts.
The Bashford-Rogers laboratory has a strong track record of collaboration over the last 15 years and established systems
for co-supervision.
References:
1. Single-cell immune multi-omics and repertoire analyses in pancreatic ductal adenocarcinoma reveal differential
immunosuppressive mechanisms within different tumour microenvironments (2023)
(https://www.biorxiv.org/content/10.1101/2023.08.31.555730v1) Shivan Sivakumar, Ashwin Jainarayanan, Edward
Arbe-Barnes, et al.
2. Predicting risk of pancreatic cancer in individuals with new-onset type-2 diabetes in primary care: protocol for the
development and validation of a clinical prediction model (QPancreasD, 2021)
(https://www.medrxiv.org/content/10.1101/2021.12.22.21268161v1) Pui San Tan, Ashley Kieran Clift, Weiqi Liao, et al.
3. Analysis of the B cell receptor repertoire in six immune-mediated diseases. (Nature, 2019) RJM Bashford-Rogers, L
Bergamaschi, EF McKinney, et al.
4. Opportunities and challenges for digital twins in biomedical research: Proceedings of a workshop in brief (The National
Academies Press, 2023). Linda Casola, et al.
139
The Oxford Cancer Vaccine Project – Move from neoantigens to
state of the art mRNA-Lipid nanoparticles (LNP) - 1,2,3,4 Cesar Lopez-
Camacho & Lennard Lee
Primary Supervisor: Cesar Lopez-Camacho & Lennard Lee
Second Supervisor:
Abstract
Oxford University is renowned for its pioneering contributions to vaccine development and immune-oncology. Over the
past two years, the Jenner Institute has focused on developing and refining mRNA-based vaccine development,
leveraging new manufacturing capabilities to increase efficacy and improve dosing regimes.
Simultaneously, under the leadership of Tim Elliot at the Oxford Centre for Immuno-oncology in conjunction with
Oxford’s University Math department, we have been successfully honing pipelines for the prediction of cancer antigens.
The latest Oxford model, POEM, has aligned mechanisitic insight with statistical models to create an analytical pipeline
that has been shown to significantly outperform and better predict peptide immunogeneity than existing models.
This PhD project will provide a student and Oxford University the opportunity to bring together and synergise these two
research projects and rapidly advance our capabilities to develop bespoke cancer vaccines by our small research
manufacturing facility. The project aims to develop a cancer vaccine to treat 10 types of cancer for patients with
metastatic disease to provide a pivotal boost to immunotherapy responses.
The project will work in conjunction with our wider strategy to bring in new rsources, and facilitate the development of
a highly skilled student who will serve as the linchpin in bridging these two distinguished institutions. The overarching
goal is to expedite the transition from neoantigen discovery to the clinical applications of a state-of-the-art mRNA-LNP
(messenger RNA lipid nanoparticle) vaccine platform.
By amalgamating the expertise of the Jenner Institute and the Oxford Centre for Immuno-oncology, the translation of
cutting-edge cancer immunotherapy research into tangible solutions will be rapidly accelerated for the benefit of
patients, repositioning Oxford on a new trajectory at the forefront of mRNA-based neoantigen vaccine development.
Research objectives
1. Integration of Expertise: The project will enable the expertise creating POEMs immunogeneity algorithm
to be utilised widely. It will take the critical groundwork laid by Tim Elliot and Miles Weatherseed and
ensure iterative improvements to generate an improved state of the art multi-epitope prediction model.
The project will enable validation of the POEMs tools using datasets such as the caner genome atlas.
2. Experimental Validation: The project will take the Oxford University technology in developing mRNA-LNPs,
and refocus these at cancer. Upon POEM-besed neoantigens identification, we will utilise techniques such
as gene synthesis, mRNA In vitro transcription (IVT) production, encapsulation of mRNA into LNPs. Such
technologies have been used for several infectious disease vaccines in Lopez-Camacho’s group, such as
Malaria, Tickborne Pathogens, Chikungunya Virus, etc. Pre-clinical validation for the neoantigents mRNA-
based vaccines will be performed at Lopez-Camacho's laboratory using in vitro experiments and mouse
models (Animal Project Licence in place, granted to Dr. Lopez-Camacho) to assess vaccine transduction
efficiency and immunogenicity.
3. Strategic Alignment: The project will show Oxford capabilities and will involve new opportunities for the
project to be embedded in a wider strategic agenda. Dr. Lennard Lee will provide expertise towards
aligning the project on a national clinical flagship project footing, and the Cancer Vaccine Launchpad,
using his experience in working with the Office for Life Sciences and NHSE.

The translational potential of this project is significant, with benefits for both patients and the life sciences sector.
Building upon the successful model of the Oxford-AstraZeneca coronavirus vaccine, this project holds the promise of
delivering cutting-edge cancer immunotherapy solutions rapidly. This project aspires that similar to Oxford’s last vaccine
140
product, we do this at a not-for-profit basis to the NHS, and additionally extend its impact to resource-limited third-
world countries.
Patients will gain significantly from the outcomes of this research, as it aims to expedite the development of state-of-
the-art mRNA-LNP vaccines tailored to individual cancer profiles. This personalised approach by our small research
manufacturing facility that holds the potential to revolutionize cancer treatment, offering enhanced efficacy and fewer
side effects.
Furthermore, the translational success of this project aligns seamlessly with the broader life sciences sector. By iterating
each stage of a cancer vaccine, we will set ourselves on a course to have optimised all aspects, from genomics,
neoantigen selection to manufacturing, reinforcing the UK's position as a global leader in medical innovation and our
national cancer vaccine advance.
Finally, this project not only provides a clear path to improve patient outcomes but also embodies the ethos of
technology acceleration, accessibility and affordability, following in the footsteps of the Oxford-AstraZeneca coronavirus
vaccine's global distribution model. Its potential to deliver high-impact, low-cost cancer immunotherapies positions it
as a project that can best bring our best scientists together, for those in need, both within the UK and in underserved
regions around the world.
Summarise the training opportunities
1. mRNA-LNP Vaccine Manufacturing Skills- small research manufacturing facility: The manufacturing of
mRNA-LNP vaccines is a critical component of this project. Aspiring scientists and engineers can gain
hands-on experience in the production of cutting-edge vaccines, expanding their knowledge in
biomanufacturing and contributing to the development of life-saving therapies.
2. Maths and Bioinformatics Expertise: This project presents a unique training opportunity for individuals
interested in the intersection of mathematics and bioinformatics. Aspiring researchers can deepen their
understanding of computational techniques for neoantigen prediction, equipping them with highly
sought-after skills in the field.
3. Diverse Experimental Models: For those interested in cellular and molecular biology, this project offers the
opportunity to work with a range of in vitro and mouse models. Researchers can hone their skills in
experimental design, cell culture techniques, and animal model studies, contributing to the advancement
of cancer immunotherapy research.
Additionally, this project strategically positions trainees within the broader landscape of cancer research and healthcare
priorities together with state of the art genetic-based technologies. By aligning with the national cancer advance and
the objectives of government, charities, patient groups, and industry partners, trainees can gain exposure to the
intricate dynamics of healthcare policy and industry collaboration. This holistic approach to training not only equips
individuals with technical expertise but also cultivates a deep understanding of the strategic and ethical considerations
that drive innovation in cancer immunotherapy.

141
Influence of viral based delivery methods on distinct pathways of
antigen uptake and presentation: implications for precision vaccine
development - 1,2,3,4 Prof. Sarah Rowland-Jones
Primary Supervisor: Prof. Sarah Rowland-Jones
Second Supervisor: Dr Ester Gea-Mallorquí
Abstract
Myeloid-cell-based immunotherapies are becoming increasingly popular, as they are able to target the early stages of
the immune response involving dendritic cells (DCs). Most such immunotherapies use lentiviral (LV)-based vaccines to
induce antigen expression in DCs. To optimise these therapies, it is important to gain a better understanding of the
biology of LV-DC interactions.
This project aims to decode the very fundamentals of antigen presentation. We aim to understand what factors
influence the route of antigen uptake, based on our previous observations of antigen delivery using different lentiviral-
based vectors. We have detected marked differences in the way two closely related human retroviruses interact with
DCs, and plan to use this knowledge to develop vaccine strategies that optimize the generation of potent T-cell
responses against immunogens from persistent viruses and tumours.
Dendritic cells (DCs) play a pivotal role in priming adaptive immune responses. As professional antigen-presenting cells,
DCs instruct the development of the adaptive response, mainly through distinct routes of antigen presentation. How an
antigen is delivered to DCs can have a profound impact on its processing, loading into the antigen presenting pathway
and subsequent presentation to the adaptive immune system.
Typically, intracellular antigens will be processed by the proteasome and loaded into major histocompatibility complex
(MHC) class I molecules; while internalised external antigens will be digested and loaded into MHC class II in late
endosomes before being transported to the plasma membrane. However, external antigens can also be presented via
MHCI in a process known as cross-presentation, which has been shown to be particularly powerful in inducing effective
CD8+ T cell responses against tumours.
LV are based on the HIV-1 backbone. However, HIV-2 is a genetically and structurally similar virus that more rarely
causes AIDS as it is able to induce an immune control on the infection. We have previously shown that HIV-2 viral control
correlates with the potent induction of polyfunctional and high avidity cytotoxic T lymphocytes (CTLs)1–3. The generation
of similarly potent polyfunctional T-cell responses against tumour antigens is a key goal for therapeutic cancer vaccines.
Pathogenic HIV-1 and immune-controlled HIV-2 interact differently with DCs4–5. Our initial observations show that HIV-
1 is more likely to be captured on the surface of DCs, whereas very little HIV-2 is captured, so the virus largely causes
DC infection. We have data to show which viral proteins are responsible for these distinct interactions. We hypothesise
that the different interactions between HIV-1 and HIV-2 with DCs will translate into distinct pathways of antigen
delivery, presentation and development of an adaptive response with distinct qualities, that will ultimately correlate
with long-term immune control. In this project we will investigate how the initial priming of the adaptive response by
HIV-1 and HIV-2 dictates the induction of a potent CD8+ T-cell response, and how specific viral proteins can be used to
divert the immune response into particular pathways. These studies should lead to an understanding of how we can
improve LV platforms to develop effective cancer vaccines.
This project offers the candidate an excellent opportunity for training in a multi-disciplinary range of skills and state-of-
the-art laboratory techniques, including cellular immunology (flow cytometry, dendritic cell culture and functional
assays of T-cell activity), virology, proteomics analysis, confocal microscopy and antigen processing studies. In addition,
analysis of laboratory data in relation to clinical data will be performed to understand the potential translational impact
of the laboratory studies.
142
The student will have access to additional training on scientific writing and presentation of research findings: he/she
will be encouraged to present his/her work during regular weekly lab and unit meetings, take part in regular journal
clubs, and attend national/international conferences, as well as preparing the first draft of manuscripts and review
articles in close collaboration with the supervisors.
References
1. Duvall MG, Precopio ML, Ambrozak DA, Jaye A, McMichael AJ, Whittle HC, Roederer M, Rowland-Jones SL, Koup
RA. Polyfunctional T cell responses are a hallmark of HIV-2 infection. Eur J Immunol. 2008;38(2):350-363.
doi:10.1002/eji.200737768
2. de Silva TI, Peng Y, Leligdowicz A, Zaidi I, Li L, Griffin H, Blais M-E, Vincent T, Saraiva M, Yindom L-M, van Tienen C,
Easterbrook P, Jaye A, Whittle H, Dong T, Rowland-Jones SL. Correlates of T-cell-mediated viral control and
phenotype of CD8(+) T cells in HIV-2, a naturally contained human retroviral infection. Blood. 2013;121(21):4330-
4339. doi:10.1182/blood-2012-12-472787
3. Leligdowicz A, Onyango C, Yindom L-M, Peng Y, Cotten M, Jaye A, McMichael A, Whittle H, Dong T, Rowland-Jones
S. Highly avid, oligoclonal, early-differentiated antigen-specific CD8+ T cells in chronic HIV-2 infection. Eur J
Immunol. 2010;40(7):1963-1972. doi:10.1002/eji.200940295
4. Kijewski SDG, Akiyama H, Feizpour A, Miller CM, Ramirez N-GP, Reinhard BM, Gummuluru S. Access of HIV-2 to
CD169-dependent dendritic cell-mediated trans infection pathway is attenuated. Virology. 2016;497:328-336.
doi:10.1016/j.virol.2016.07.029
5. Duvall MG, Lore K, Blaak H, Ambrozak DA, Adams WC, Santos K, Geldmacher C, Mascola JR, McMichael AJ, Jaye A,
Whittle HC, Rowland-Jones SL, Koup RA. Dendritic Cells Are Less Susceptible to Human Immunodeficiency Virus
Type 2 (HIV-2) Infection than to HIV-1 Infection. J Virol. 2007;81(24):13486-13498. doi:10.1128/jvi.00976-07
6. Gea-Mallorqui, E. and Rowland-Jones, S.L. DC-targeting Lentivectors for Cancer Immunotherapy Immunotherapy
Advances 2023 (in press)

Bioinformatics and machine learning approaches for early cancer
detection in a resource restricted health system using whole
genome nanopore sequencing of cell-free DNA. - 1,2,3,4 Prof. Anna
Schuh
Primary Supervisor: Prof. Anna Schuh
Second Supervisor: Dr. Dimitris Vavoulis
Abstract
Cancer diagnosis from circulating tumour DNA (ctDNA) obtained from a simple blood draw (liquid biopsy) could
transform cancer care, if they can reliably diagnose cancer at a stage when it is curable with surgery alone. However,
the huge commercial efforts underway to evaluate “liquid biopsies” as screening tests in asymptomatic individuals in
the US/Europe have revealed major concerns for healthcare systems, starting with high false positive rates[1,2]. These
limitations are of even greater relevance in resource-restricted regions where any interventions must be aimed at
reducing, not increasing the burden on overstretched diagnostic services. With this in mind, we propose that in this
setting liquid biopsies should be used instead to triage patients presenting to primary care with non-specific symptoms
of cancer such as weight loss, night sweats, fatigue or anaemia. About one-third of these patients have early-stage
disease that would make them eligible for surgery with curative intent. Our observational data from the Muhimbili
National Hospital (MNH) in Tanzania indicates that only between 10-20% of patients over the age of 45 with non-specific
symptoms of cancer eventually turn out to have cancer. However, most patients with non-specific symptoms are not
referred to secondary care for investigations as a cancer diagnosis is deemed unlikely. Those who do enter the cancer
diagnostic pathway, compete for the limited access to invasive procedures and often undergo a diagnostic odyssey that
ruins entire families even without the cost of subsequent treatment.

The primary objectives of this proposal are: 1) to develop a predictive machine learning model for the detection of
cancer using shallow (1x) whole-genome nanopore sequencing of circulating cell-free DNA (cfDNA) collected from the
plasma of patients presenting with non-specific symptoms in a low-resource clinical environment, and 2) to perform a
case-control study in said environment to establish the diagnostic accuracy of the test. The major outcome of this study
will be an analytical pipeline, along with a user-friendly web interface for facilitating easy access to the pipeline by
clinicians. The pipeline will take as input multiple genomic data modalities (copy number aberrations, methylation
modifications, infection-related DNA detection, and fragmentomics) from a single patient, as well as the patient’s clinical
data (including sex, age, serum FLC ratio, MBL levels, B-cell count and others), and will generate at its output the
probability that the patient has cancer, as well as an estimate of the tissue(s) of origin of the detected ctDNA.
The proposed project builds directly upon the capacity and capabilities developed at the Oxford Molecular Diagnostics
Centre (OMDC) in the context of various projects, including the £3.9M NIHR-funded “Aggressive Infection-Related East
Africa Lymphoma” (AI-REAL) study, which concludes in June 2024. AI-REAL successfully developed and validated a liquid
biopsy test to aid in the rapid diagnosis of Burkitt’s Lymphoma (BL), the most common and aggressive form of childhood
cancer world-wide. The existing OMDC liquid-biopsy test for early cancer detection and disease monitoring [3] detects
traces of ctDNA in the plasma based on paired-end deep (80x or higher) whole-genome sequencing (Figs. 1 and 2). In
this proposal, the test will be adapted to work with shallow (1x) nanopore sequencing data for affordability and
feasibility in a low-resource clinical environment (MNH) and its diagnostic accuracy will be estimated in a retrospective
case-control study involving 400 patients (200 cases and 200 controls) with non-specific symptoms recruited at MNH
(see primary objectives).
Given its multidisciplinary nature, the project is founded on a tight collaboration between the Oxford Mo-lecular
Diagnostics Centre at the Department of Oncology, the Muhimbili University of Health and Allied Sciences (MUHAS) and
the Rapid Diagnostic Clinic of the Muhimbili National Hospital in Dar es-Salaam, Tanzania. This research collaboration
was initiated in the context of the AI-REAL study, and it will be further facilitated by this award through sharing of
sequencing data and collaborative data analysis, which are made possible via the joint use of OMDC’s HIPAA-compliant
and secure AWS cloud environment by Oxford- and Tanzania-based researchers.
144
In sub-Saharan Africa, while other causes of death are holding steady or declining, annual cancer deaths are projected
to nearly double by 2030 to 984,000. Although detecting cancer and starting intervention at earlier stages of disease is
known to improve patient outcomes and overall cancer-related survival, in Tanzania cancer is typically diagnosed late
and at a younger age. The fundamental barrier to improving cancer diagnosis in the region, is a lack of diagnostic capacity
due to a critical skill shortage, which cannot be ad-dressed in the foreseeable future by more training at a national level.
Early-stage cancer diagnosis is especially challenging as healthcare providers in Tanzania have a low index of suspicion
for cancer. Patients presenting with non-specific symptoms such as fatigue, weight-loss or night sweats are initially
investigated for parasites or infectious diseases like TB and HIV, leading to long delays in referral to cancer specialists.
Even if a patient has specific cancer symptoms such as a breast lump or vaginal bleeding, traditional belief systems about
the causes of cancer, the invasive nature of diagnostic procedures, the stigma around cancer and the cost and logistics
of a long journey promote delayed presentation to specialist cancer clinics. Therefore, an affordable non-invasive triage
tool that effectively rules cancer in/out, as we propose here, could reduce diagnostic delays, facilitate fewer, less costly
visits for patients, A
relieve pressure
on existing
diagnostic services
and allow them to
be used more
efficiently and
lead to
substantially Bi Bii C
improved patient
outcomes.
Figure1: Integration of genomic data modalities for ctDNA detection. A) Multimodal scores for the quantification of plasma ctDNA generated from the integration of copy
number aberrations, somatic SNVs and INDELs, and methylation signals in each plasma sample. A red circle indicates a higher ctDNA burden in comparison to the healthy
controls. 52 out of 61 cancer plasma samples were correctly identified as such, which corresponds to 85.2% sensitivity. This is higher than the sensitivity of any of the three
data modalities. B) Multimodal scores against cancer stage and type (Bi) and correlation between median multimodal scores and stage (Bii). C) In silico validation of
multimodal analysis at increasing ctDNA fractions. At each ctDNA fraction, we simulated 1000 healthy and 1000 cancer plasma samples using actual healthy and cancer
plasma samples as templates. The area under the receiver operating characteristic (ROC) curve (AUC) is ≥80% at ctDNA fractions ≥0.7%. CTRL: controls; CRC: colorectal; OES:
oesophageal; PNCR: pancreatic; RNL: renal; OVR: ovarian; BST: breast
A Bi
Bii Ci Cii
RR event / death
● Yes No
(post-Tx)
Detected 4 1
ctDNA
Not
0 4
detected
145
Figure 2: Multimodal ctDNA detection for post-operative MRD and adjuvant therapy response tracking in colorectal cancer without matched tumour. A) Tracking post-
operative MRD in case GEL193. ctDNA was detectable in the plasma 1 year after surgery and this correlated with inoperable metastatic rectal cancer and a possible lung
adenocarcinoma suggested through radiological examination, both of which were recorded approximately 3 years after the post-operative plasma sample was collected. B)
Tracking response to adjuvant therapy following surgery. Case GEL282 (Bi) did not have detectable ctDNA immediately after the last cycle of treatment. However, low ctDNA
burden was detected 5 months later, which correlated with the presence of tubular adenomas with low grade dysplasia in the sigmoid colon at around the same time. Case
GEL432 (Bii) did not have detectable ctDNA shortly after the last cycle of treatment and was still alive approximately 6 years after the last plasma sample was collected. C)
Confusion matrix (Ci) and event-free (i.e., no recurrence, metastasis, or precancerous adenomas present) survival (Cii) in 9 patients with colorectal cancer. In 8 out 9 patients,
ctDNA burden after the end of surgery and/or adjuvant therapy correlated with the presence/absence of clinical events, such as recurrence or pre-cancerous adenomas (Ci).
Absence of ctDNA detection after the end of surgery/adjuvant treatment correlated with longer survival times (Cii). Dx: diagnosis; Sx: surgery; Tx0: first cycle of adjuvant
therapy; Tx1: last cycle of adjuvant therapy; RR: recurrence; LGD: low-grade dysplasia
The DPhil candidate for this project will be working in a multidisciplinary environment of bioinformaticians,
statisticians, cancer geneticists and clinicians. This will provide ideal conditions for developing technical skills in
state-of-the-art methods for data analysis and software engineering, which are transferable to domains other than
genomics, as well as making substantial scientific contributions to a cutting-edge research field. In particular, the
student is expected to acquire expertise in the following domains: a) cancer genomics, particularly as applied in
early detection, b) bioinformatics approaches for WGS data analysis, c) Bayesian approaches to data modelling
and analysis, and d) supervised and unsupervised machine learning techniques, including deep learning, for the
integration of heterogeneous data modalities, feature selection, prediction, classification, and clustering. The
student is also expected to hone essential soft skills, including time management, presentations skills and task
planning, which will allow them to communicate and operate effectively in a vibrant and diverse team of clinical
and computational scientists.
References
[1] Liu MC et al. Ann. Oncol. 31, 745–759 (2020) [2] Nicholson BD et al. Lancet Oncol. 24, 733–743 (2023) [3]
Vavoulis DV et al. medRxiv (2023) [4] Choy LYL et al. Clin. Chem. 68, 1151–1163 (2022)

Spatial analysis of the bone marrow in myeloproliferative
neoplasms DNA. -4 Prof. Helen Byrne
Primary Supervisor: Prof. Helen Byrne
Second Supervisor: Dan Royston, Jens Rittscher, Ros Cooper
Eligibility: Track 4 applicants are eligible to apply for this project
Abstract
Myeloproliferative neoplasia (MPN) is a blood cancer characterised by overproduction of myeloid blood cells typically
driven by well-defined driver mutations in JAK2, CALR and MPL. However, genotype-phenotype correlation is
unpredictable and risk of disease progression is variable. MPN patients are at increased risk of venous thromboembolic
events, and some patients develop myelofibrosis or experience acute leukaemic transformation, events associated with
poor clinical prognosis. Recent single-cell sequencing approaches in murine models have confirmed the role of the bone
marrow microenvironment in driving fibrosis progression in MPN, with critical reprogramming of stromal populations
in the context of inflammation and neoplastic haematopoietic stem and progenitor cells (HSPCs) populations. However,
little is known about the spatiotemporal relationships between these cell populations in MPN. We hypothesise that in-
situ spatial approaches will provide key insights into how stromal and immune populations in the bone marrow niche
drive fibrosis development. Opportunities afforded by new highly multiplexed spatial transcriptomic approaches will
allow in-situ phenotyping of these populations at single cell resolution. We will use multiple transcriptomic-based
approaches to map the bone marrow niche in patients with MPN and those with healthy/reactive marrows. We will use
novel computational and mathematical approaches to analyse and integrate these datasets. We will cross-validate key
biomarkers using protein-based approaches including multiplex immunofluorescence and immunohistochemistry
across larger clinical cohorts. This has the potential to identify markers of early-stage disease progression and guide the
search for novel therapeutic targets.

Background: The chronic myeloproliferative neoplasms (MPN) are largely underdiagnosed blood cancers characterised
by the overproduction of blood cells derived from haematopoietic stem and progenitor cells (HSPCs) that harbour
mutations resulting in cytokine-independent or hypersensitive proliferative signals. In over 90% of MPN, a driver
mutation in the genes encoding JAK2, CALR or MPL results in constitutive activation of the MPL-JAK-STAT signalling
pathway. Despite this commonality, disease phenotype and risk of progression in MPN are highly variable.
The origin and nature of the earliest cellular and extracellular abnormalities in MPN have recently been elucidated using
RNA sequencing and genetic fate-tracing experiments in mouse and human. Key insights include recognition of early
disease initiation in the endosteal niche adjacent to the bone surface, and the perivascular niche within the central area
of the bone marrow. Reprogramming and abnormal differentiation of stromal cells in these regions, fuelled by mutation-
bearing HSPCs and inflammation, appear fundamental to the emergence and evolution of MPN within various murine
MPN models. However, the significance of these pathological processes within the bone marrow remains unclear.
Further, the spatiotemporal relationships of key pathological changes occurring in bone marrow biopsies from newly
diagnosed and progressing MPN patients are not well understood. New spatial transcriptomic platforms allow for in-
situ analysis of both cellular phenotype and pathway activation within tissue microenvironments with up to single-cell
resolution. The complexity of these highly multiplexed, high-resolution datasets and challenges associated with their
interpretation are becoming increasingly apparent.
Using a range of complementary approaches, including automated morphological analysis, multiplexed immunostaining
(IF) and whole slide / regional spatial transcriptomics, we will generate powerful, multi-modal spatial datasets and
develop novel computational and mathematical approaches to comprehensively characterise the cellular and
extracellular matrix of patient biopsies. This will enable us to determine the key spatial changes in the tissue
microenvironment at MPN initiation and to track subsequent changes that correspond to disease progression, including
myelofibrosis. Using complementary spatial protein and gene-expression data employing different technologies and
different commercial platforms, we will build robust and validated computational models of the marrow in health and
disease. The multi-platform approach will also enable us to investigate the extent to which shared and overlapping
probes improve the accuracy and robustness of inference about protein expression levels drawn from spatial
transcriptomic data. Such insights will have short-term implications for current diagnostic approaches to MPN and will
inform the search and evaluation of novel therapeutics. In this way, we aim to achieve the following aims and objectives:
147
Figure 6: Overview of project workflow
Aims and objectives:

(1) To develop novel computational and mathematical approaches for the analysis of highly multiplexed
spatially-resolved datasets to facilitate whole slide / regional computational modelling of the bone marrow
microenvironment in normal and MPN diagnostic samples
(2) To use these approaches to analyse spatial transcriptomic data generated with existing commercial
platforms (Merscope & Nanostring GeoMx/CosMx)
(3) To cross validate transcriptomic data by comparative analysis of shared/overlapping probe-sets from
competing commercial platforms employed under (1)
(4) To validate protein expression analysis using sample-matched multiplexed immunofluorescence and
immunohistochemistry sample staining
(5) To create a spatial expression atlas of normal bone marrow (and MPN) from whole slide bone marrow
samples
Academic value of the research and funding justification: Funding for a DPhil student will enable us to transform
our understanding of healthy bone marrow to include detailed information about its spatial structure and how
this changes in MPN and subsequent disease progression. Realising this goal will requires A multidisciplinary team
who work closely together is needed to address the new challenges associated with realising this goal. The
requested funds will cover the costs of a graduate student who will carry out research in the Mathematical
Institute.
Collaborations: The project will initiate a new, multidisciplinary collaboration between Byrne, Rittscher, Royston
and Cooper, building on an established, successful collaboration between Royston and Rittscher. The student will
make frequent visits to the labs of Royston and Cooper where they will interact with pathologists and clinicians
investigating MPN and other diseases. This will enable them to learn the relevant MPN biology and how to use
different imaging platforms. Byrne and Rittscher will meet weekly with the student to provide guidance on the
image analysis methods, while the full team will meet monthly.
Translational potential: This project will enable us to reclassify the bone marrow of healthy and MPN patients
based on spatial cell interactions. Identifying MPN-specific cell interactions associated with disease progression
will improve diagnosis and assist with the identification of signalling pathways that could be targeted to normalise
aberrant cell-cell communication. The project will facilitate spatial biomarker development and identification of
novel therapeutics for MPN. It will also generate new methodologies for analysing multi-modal data and
strengthen expertise in multidisciplinary approaches to tackling cancer. As such, it will contribute to the scientific
themes “Cancer big data” and “Early cancer detection”.
148
Training opportunities: The student will learn how to use and develop software for the extraction, analysis and
interpretation of spatial information from multiplexed and spatial transcriptomics images. The project would suit
a mathematician (or similarly numerate graduate) interested in applying statistical, mathematical and
computational techniques to biological imaging data or a pathologist interested in the digital image analysis and
quantification.
References
6. H Ryou et al (2023). Continuous Indexing of Fibrosis (CIF): improving the assessment and classification of
MPN patients. Leukemia. 37(2): 348-358.
7. K Sirinukunwattana et al (2020). Artificial intelligence-based morphological fingerprinting of
megakaryocytes: a new tool for assessing disease in MPN patients. Blood Adv. 4 (14): 3284-3294.
8. JA Bull and HM Byrne (2023). Quantification of spatial and phenotypic heterogeneity in an agent-based
model of tumour-macrophage interactions. PLOS Computational Biology 19: e1010994.
9. O Vipond et al (2021). Multiparameter persistent homology landscapes identify immune cell spatial
patterns in tumours. PNAS 118 (41): e2102166118.

149

DPhilin Cancer Science 2024 Project Booklet V6

Uploaded by

Copyright:

Available Formats

DPhilin Cancer Science 2024 Project Booklet V6

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

DPhilin Cancer Science 2024 Project Booklet V6

Uploaded by

Copyright:

Available Formats

DPhil in Cancer Science

• a choice of interdisciplinary cutting-edge cancer research projects.

At the end of their DPhil course, students should:

• commitment and passion to a career in cancer research

• evidence of motivation for and understanding of the proposed area of study

• commitment to the subject, beyond the requirements of the degree course

• preliminary knowledge of relevant research techniques

• capacity for sustained and intense work

• reasoning ability and academic curiosity.

Abstract of the project

Research objectives and proposed outcomes

Summarise the training opportunities.

Return to Projects list

Abstract of the project

Return to Projects list

Work package 1: Apply scCOLOR-seq to Clear Cell Sarcoma tumours.

12 Return to Projects list

Research objectives and proposed outcomes

Return to Projects list

The China Kadoorie Biobank (CKB) included 0.5 million

Based on this CKB cohort, we are conducting a large case-

Return to Projects list

18 Return to Projects list

Return to Projects list

Crucially, however, there is a “Goldilocks zone” of ROS for

The central aim of this project is to exploit the ROS-

Return to Projects list

24 Return to Projects list

26 Return to Projects list

28 Return to Projects list

1. Bryant et al (2005) Specific killing of BRCA2-deficient tumours with inhibitors of poly(ADP-ribose)

30 Return to Projects list

Primary Supervisor: Prof. Tim Elliot

Figure 1. Ultrasound-activated nanoparticles for delivery of drugs.

(A) Schematic diagram for ultrasound-guided delivery of nanoparticle encapsulated

Work Package 1: Development of ultrasound-activated nanoparticles for targeted drug delivery

33 Return to Projects list

Primary Supervisor: Prof. Clive Wilson

Return to Projects list

Objective1. Produce a highly sensitive synthetic antigen receptor targeting CD19.

Return to Projects list

Co-culture experiments combining CRLM

3. Identify replacement-dependent signalling mechanisms by comparing PDO-fibroblast co-cultures to identify

2. Mass spectrometry, proteomics, metabolomics, data analytics/integration (Roman Fisher)1,2

3. Multiplexed IHC performance and image analysis (Mark Coles)

Translational potential of the project

Return to Projects list

44 Return to Projects list

46 Return to Projects list

Research objectives and proposed outcomes

• Investigation of the timing of TLS formation alongside lung cancer development

• Characterisation of tumour-controlling TLS to understand factors driving effective anti-cancer immunity

52 Return to Projects list

Primary Supervisor: Assoc Prof. Eileen Parkes

• Characterisation of fibroblast phenotype in human breast cancer using spatial transcriptomics

54 Return to Projects list

56 Return to Projects list

Translational potential of the project

Return to Projects list