B I O P R O C E S S TECHNICAL

Assays for Controlling

Host-Cell Impurities in
Biopharmaceuticals
Tanja Wolter and Andreas Richter

O
ne critical task during Other references suggest that higher
the manufacture of levels may be acceptable (2). The 100-
biopharmaceuticals pg limit requires a purification process
is purification of the that is very effective and robust in
drug substance. The removing DNA — as well as analytical
downstream process must remove methods that are extremely sensitive
all contaminants, including host and reliable to prove it. Host-cell
cell material such as DNA and proteins in the drug substance should
cellular protein. Trace amounts of be “below detectable levels using a
host-cell DNA and proteins can highly sensitive analytical method”
be copurified along with the drug (1). As a rule, that amount should
substance. Such contaminants are not exceed a level of about 100 ppm.
obviously undesirable because But no exact limit is set for proteins;
of possible consequences if they therefore, the specification for proteins
are injected into patients along must be determined case by case.
with it. They could potentially According to the European
cause allergic reactions (proteins) Agency for the Evaluation of WWW.PHOTOS.COM

or even transfection of cells Medicinal Products, EMEA (3),

product. Routine testing determines
(DNA) resulting in tumors. two different strategies can be used
host-cell DNA and proteins in each
Because of those potential negative to ensure that a drug substance
product batch as part of the lot-
effects, regulatory authorities have is within the allowable limits of
release data.
released several guidance documents contamination with DNA or host-
No matter which approach is used,
about what levels of impurities are cell proteins. You can either validate
the assays and methods involved in
acceptable. Residual DNA in final bulk that a process removes sufficient
determination of residual DNA and
products should be generally lower amounts of such contaminants
proteins must fulfill ICH validation
than 100 pg per therapeutic dose (1). or perform routine final product
requirements (4, 5). Obviously
testing to determine whether they
those techniques must be sensitive
PRODUCT FOCUS: PRODUCTS OF CELL are present.
enough to determine very low levels
CULTURE AND FERMENTATION With the validation approach,
of contamination (e.g., in the ppm
known amounts of DNA are spiked
PROCESS FOCUS: DOWNSTREAM range). Here we describe some
into a downstream process. The
(SEPARATION/PURIFICATION) common analytical methods used.
various steps within that process
WHO SHOULD READ: ANALYTICAL are then examined to determine
QUANTIFICATION
AND PROCESS SCIENTISTS, QA/QC their capability to remove the
OF HOST-CELL DNA
contamination. This approach is
KEYWORDS: QUALITY CONTROL, The limit of 100 pg DNA per dose
most commonly used for DNA.
DNA AND PROTEIN DETECTION set by regulatory authorities roughly
Robustness and consistency
METHODS, HOST-CELL IMPURITIES, ASSAY equals the amount of DNA from
in DNA removal also must be
DEVELOPMENT, ANALYTICAL TESTING less than 17 diploid Chinese hamster
shown by measuring the amounts
ovary (CHO) cells. To determine
LEVEL: INTERMEDIATE contaminating several batches of
such small amounts of DNA, an
40 BioProcess International FEBRUARY 2005
analytical method must be extremely Figure 1: Dot-blot hybridization assay for quantifying residual genomic E. coli DNA.
sensitive and robust. In principle, Samples are analyzed in duplicate with and without addition of reference DNA (DNA
three techniques have the required spike). A genomic E. coli DNA calibration row is applied from 1600 pg to 0.8 pg. In this
case, the typical quantification range is 3.1−800 pg. Calibration row, controls, and test
sensitivity: hybridization, methods samples are applied to the filter by a vacuum-based blotting procedure. DNA is cross-
based on DNA-binding protein linked by ultraviolet (UV) light and then incubated with a 32P-labeled, random-primed
(such as the Threshold assay from DNA probe. After it is washed, the filter is evaluated by a phosphor imaging instrument
Molecular Devices, (BAS Reader from Fuji of Saitama City, Japan, www.fujimed.com). The DNA content of
www.moleculardevices.com) (6), the samples is calculated using the calibration row signals. For each test sample, an
individual recovery is calculated with the values of the spiked sample, spike controls,
and quantitative PCR (q-PCR) (7).
� and the negative controls.
The 2004 European Pharmacopoeia
clearly advises that residual DNA
should be determined using
sequence-independent techniques —
and goes on to specify that methods
such as hybridization or DNA-
binding protein assays be used.
� �
The basic principle of a ����������������
hybridization assay is binding
of DNA probes to immobilized ������������������������ �
and denatured host-cell DNA.
Demands of the guidelines suggest �� �� �� �� �
that those probes be manufactured �
independent of the DNA sequence. ��������������������� ��������
That can be achieved, for example, ��������
by a “random priming” procedure. �� � �� �� ����� �
Probes are usually labeled with ��������
radioactive tags or fluorescent dyes, ��������������� �
�
resulting in a signal proportional to
the amount of DNA immobilized
Table 1: Intraassay, interassay, and biological variability coefficients of variability (CVs)
on a filter. Signal detection is
performed by phosphor- or Hybridization Threshold PCR
fluorescence-imaging systems.
By contrast, the Threshold assay Specific for Random sequence, Single-stranded Specific for target
species-specific DNA, not species- sequence
is based on sequence-independent specific
binding of two proteins specific to
Estimated minimal detection 50 600 150
single-stranded DNA (ssDNA). One length of DNA (base pairs)
binding protein is an antibody, and Robustness to interfering ++ + +
the other is called single-stranded substances
binding (SSB) protein. First, a Time required (hours) 48 6 2
reaction complex is formed when Sensitivity 6 pg 3 pg <1 pg
the biotinylated SSB protein and the
anti-ssDNA antibody (conjugated
requires matrix-specific pretreatment. from a few base pairs (bp) up to
to urease) bind to single-stranded
All assays differ in sensitivity to thousands of base pairs in length. It
host-cell DNA. A filtration stage
residuals such as organic solvents, is specific for the source DNA but
follows, during which the strong
detergents, high salt concentrations, not the sequence.
affinity of streptavidin for biotin is
ethanol, or residual proteins. Several DNA-specific binding during
used to capture and concentrate
DNA purification methods — such Threshold assays is performed by
those reaction complexes onto a
as protease treatment, organic two DNA-binding proteins, both
biotinylated membrane. For detection,
extraction, phenol-free “Wako” of which must bind to each ssDNA
the membrane is placed into a reader
extraction (Wako Chemicals, fragment. So this assay requires
that contains the substrate urea.
www.wako-chem.co.jp/english), and DNA fragments longer than about
Inside, urea is hydrolyzed by urease
precipitation by ethanol or glycogen 600 bp (6) to form a reaction
to produce a pH change, which is
— should be checked for their complex that will produce a signal
relative to the amount of host-cell
suitability. in the reader. And this assay is not
DNA in the sample.
A few differences between DNA specific to the DNA source.
Initial sample treatment is often
testing procedures should be taken PCR-based assays are specific
the most critical factor in successful
into account when interpreting to their target sequence, and the
analysis, regardless which analytical
quantitative data. The hybridization amount of total DNA is derived
technique is used. Each sample
assay randomly measures total DNA, from measured target copy numbers.
FEBRUARY 2005 BioProcess International 41
Figure 2. Development steps of a quantitative host cell protein assay and assay validation. Assay
development, regardless of which
����������� ������������
������� ������������ ������������ ����� format is used, requires about 9–12
�������������� ���������� ������� ���������� months. That timeframe covers
purification of the antigen to be
� ������� ����������
��������������
����������������
����������������������
used, immunization of animals,
�������������������� ������������������������ and antisera collection, through
����������������
������������ to the validation of the assay. The
������������������� �����������
���������� ��������������� actual time required for antibody
��������������� ����� ����������� production can vary greatly
�������������������������
������������������������ � depending on the antigenicity of
the proteins used for immunization.
As a rule, you should allow about
Figure 3: Selection of the antigen for a host cell protein assay
six months’ time for antibody
������ �����
������������ ������������ ������� ������������ ���������� production alone.
���� ����� The first step in development
of an HCP assay is generation and
purification of an appropriate antigen
�������������������������������������������
������������������������������������������������� to be used for immunization. This
��������������������������������������������������� requires fermentation of the native
cell line without the gene sequence
��������������������� ���������������������
����������������������������������������������������������� of the biopharmaceutical inserted
������������������������������������������� (8), commonly termed a “mock
fermentation run.” Supernatant
from this mock run is purified using
The amplified sequence must be significant concentrations following � exactly the same purification scheme
demonstrated to represent the total several purification steps — or even as for the product itself. The result is
amount of DNA. be copurified and concentrated a protein-containing solution without
All DNA quantification assays along with the drug substance itself. the product that can be used as an
need sample pretreatment and Many purification steps may be antigen for production of polyclonal
therefore should be designed as needed within a downstream process antibodies for the HCP assay.
spike-recovery assays to control to remove HCPs. Each step has the Because there are normally
the loss of material during sample potential for loss of product during multiple steps in a downstream
preparation. Figure 1 illustrates a purification. So during process process, however, it is necessary to
spike-recovery control strategy for a development, a suitable assay must determine which step will produce
hybridization assay. DNA standards be available for determining both that antigen. A broad spectrum of
for the host cell and vector in use product concentration and the HCPs would be detectable with
are usually commercially unavailable. levels of HCP contamination. the results of the first purification
Therefore, in-house standards must The type of assay required to step, but that does not represent the
be produced and quantified by determine HCP concentration potential HCP in your final product.
UV absorption. Those reference depends on the phase of clinical If you use the results of the final
materials should be additionally studies being performed with purification step as your antigen, the
quantified and qualified by agarose the material produced. In earlier HCP may better represent that in
gel electrophoresis to confirm the clinical phases, a generic assay may a final product but be too specific
UV data and control degradation be sufficient. However, by the and thus demonstrate insufficient
of the DNA. Assay robustness is time a biopharmaceutical is going HCP removal. Decisions must be
different for each assay, especially into phase 3 studies, regulatory made case by case. Look at the
with respect to interfering authorities normally require a spectrum of proteins present (using
substances (Table 1). Overall, product-specific HCP assay. The Western blotting, for example), and
hybridization assays are more robust rest of this article describes steps base your decision on that protein
than Threshold and PCR assays. involved in producing, optimizing, pattern. Figure 3 shows one possible
and validating a product-specific approach to choosing antigens from
QUANTIFICATION quantitative HCP assay (Figure 2). a mock fermentation run.
OF HOST-CELL PROTEIN Antigen and Antibody Production: Another important aspect to
Those designing downstream Development of a specific HCP consider during antigen preparation
processes for biopharmaceutical assay must begin early in the is the presence or absence of serum
drug products must also consider development of a biopharmaceutical antigens. Serum proteins can be
removal of host-cell proteins product because of the time very antigenic and thus “mask” the
(HCPs). These can be present in required for antibody production antibodies to the HCPs. Here is
42 BioProcess International FEBRUARY 2005
yet another reason why serum-free total proteins. Figure 4: Separation of E. coli host cell
media should be used whenever Quality Control: The most proteins on a two-dimensional gel using
possible. important technique for quality a pH gradient from 4 to 7
���
Various immunization schemes control of polyclonal antibodies is ���� ����

can be used to increase polyclonal two-dimensional gel electrophoresis ���

antibody production. A cascading followed by Western blotting (9).

��
immunization scheme (9, 10) is Together they allow separation and
commonly used because it allows identification of the antigen protein
��
less antigenic proteins to produce an mixture used. The electrophoresis
immune response. That broadens the step separates individual proteins,
spectrum of polyclonals and prevents and the Western blot controls the ��
focusing on the most strongly specificity of antibodies to those
antigenic proteins within the protein individual proteins. Signals from
��
mixture presented to the animal. the blots and silver-stained gels
Affinity chromatography is the prepared in parallel are compared
ideal method of purifying those using software developed for Figure 5: Western blot of proteins
antibodies. The antigen used for this purpose.The results of that separated according to Figure 4 on a
immunization is immobilized on comparison can be used to predict nylon membrane. Proteins were
column chromatography media, and the suitability of the antibodies for detected by a polyclonal goat serum
then animal sera is passed through use in an HCP assay. Figure 4 shows against E. coli host cell proteins
(antibodies affinity purified). Detection
that column to separate and purify the proteins of an Escherichia coli was performed by a rabbit antigoat anti-
the antibodies specific to it. The HCP mixture separated using two- body coupled to horseradish
results tend to have a higher HCP dimensional gel electrophoresis, peroxidase.
affinity than those purified by a and Figure 5 is the corresponding ���� ���� ���
���
different method. Some antibodies Western blot.
��
may be lost during affinity The quality, specificity, and
��
purification, however, so they sensitivity of any antibody-based
should be quantified using two- assay used for detection of HCP is ��
dimensional gel electrophoresis and related directly to the quality of the
Western blotting to ensure sufficient antibodies themselves. They are not ��

amounts. likely to detect all the HCP proteins

Antigens and antibodies that will present. Those that were low in ��

be used as future reference material concentration or that cause only

must be stored properly. Limited a weak antigenic response will go
amounts should be purified from undetected. The goal of the assay
Threshold system. Table 2 compares
the sera: The shelf life of purified is, however, to detect a variety of
the two assay formats.
polyclonals is limited, whereas raw different proteins that represent the
Validation of Analytical Methods:
sera can be stored for over 10 years HCP spectrum (11).
Any assay used for biopharmaceutical
at –80 °C. The antigen should be Developing a Quantitative HCP
lot-release testing, including the
divided into aliquots before storage Assay: Based on either the enzyme-
quantitative HCP assay, should be
to avoid unnecessary freeze–thaw linked immunosorbent (ELISA)
validated as described in the ICH
cycles. Control the quality of or immunoligand (ILA) assay
guidelines (4, 5). Parameters to
your reference material using format, quantitative assay can be
validate include precision, accuracy,
electrophoresis and determination of developed using qualified antisera
linearity, and specificity. During
for the marker and the antigen
validation, the quantitation limit is
as the reference standard. The
defined, which thereby determines
sensitivity of this assay is of great
assay sensitivity. Table 2 compares
importance. As a general rule, 100
validation data for a generic HCP
ppm can be considered the upper
tested with the ILA and the ELISA
As a general rule, limit for acceptable levels of HCP
methods. The same polyclonal
100 ppm can be (11). So the assay used to determine
antibodies were used in both assays.
HCP concentration must be more
considered the Generic and Specific HCP
sensitive than that. Both ELISAs
Assays: Depending on the
UPPER limit for and the ILAs can be developed with
stage of development for the
acceptable levels of significantly >100-ppm sensitivities.
biopharmaceutical in question,
Our experience has shown, however,
host-cell proteins. either a generic HCP assay or a
that the ILA format is much more
specific HCP assay can be useful.
sensitive than the ELISA method.
The generic assay should detect
But it requires the use of the
44 BioProcess International FEBRUARY 2005
 Harmonisation of Technical Requirements
Table 2: Validation data of host-cell protein assay (E. coli) for the Registration of Pharmaceuticals for
Human Use, March 1995; www.fda.gov/
Validation Parameter ELISA ILA cder/guidance/ichq2a.pdf.
5 ICH. Q2B Validation of Analytical
Intermediate precision 9% 10%
Procedures: Methodology (Guidance for
Accuracy 91% 101% Industry). International Conference on
Lower quantitation limit 12.5 ng/mL 5.0 ng/mL Harmonisation of Technical Requirements
for the Registration of Pharmaceuticals for
Working range between 12.5 ng/mL between 5.0 ng/mL
Human Use, November 1996; www.fda.gov/
and 200 ng/mL and 40 ng/mL
cder/guidance/1320fnl.pdf.
6 Kung TV, et al. Picogram
any possible HCP contamination. with cellular-based contamination
Quantitation of Total DNA Using DNA-
Normally, cell lysates are used as have come up yet, it is an issue that Binding Proteins in Silicon Sensor-Based
antigens when developing antisera will continue to need addressing. System. Analytical Biochem. 187, 1989:
for such assays. For a product- Levels of allowable contamination 220–227.
specific assay, proteins derived from may change, but the purity 7 Lovatt A. Applications of Quantitative
the mock fermentation–purification of biopharmaceuticals must be PCR in the Biosafety and Genetic Stability
Assessment of Biotechnology Products. Rev.
run are used to develop antisera. ensured. Regulatory authorities will
Mol. Biotechnol. 82, 2002: 279–300.
A generic assay is useful continue to expect development of
8 Hoffman K. Strategies for Host Cell
in developing a downstream product-specific HCP assays — not Protein Analysis, BioPharm 13(6) June 2000:
manufacturing process. HCP levels only for cellular contaminants, but 38–45.
can be controlled following any also for any potential contaminants 9 Anicetti VR, et al. Immunization
change in that process at relatively that could come from media Procedures for E. coli Proteins. Appl.
low cost. The generic assay also can components (for example). Assays Biochem. Biotechnol. 22, 1989: 151–168
be used for lot-release testing in will also need to show that those are 10 Briggs J, Panfili PR. Quantitation
of DNA and Protein Impurities in
early clinical studies. However, once not present in the final product.
Biopharmaceuticals. Analytical Chem. 63,
manufacturing process development Assay development for possible 1991: 850–859.
has been finalized, product-specific contaminants should start as early 11 Eaton LC. Host Cell Contaminant
assay development should be in the drug development process Protein Assay Development for Recombinant
initiated. It should be started early as possible. Once developed, these
1995: 105–114J. 
Biopharmaceuticals. J. Chromatogr. A 705,
because it will take up to 12 months assays are useful in developing
to finish — and because the generic a production process as well as
assay is normally unacceptable for in determining product purity.
lot-release testing material used Because many are product specific,
in later clinical trials. Such studies you cannot underestimate the
normally require the use of a time involved not only in their
product-specific HCP assay. development, but also for their
validation as required by the
THE SOONER THE BETTER authorities.
Early concerns over negative
influences from cellular-based REFERENCES
contamination in biopharmaceuticals 1 Center for Biologics Evaluation
and Research. Points to Consider in the
have proven unjustified. Residual
Manufacture and Testing of Monoclonal
DNA, for example, has caused Antibody Products for Human Use. US Food
no cellular transformations (e.g. and Drug Administration, 28 February 1997;
cancers), nor have there been www.fda.gov/cber/gdlns/ptc_mab.pdf.
serious immunological problems due 2 WHO Expert Committee on
to HCPs. Perhaps the danger was Biological Standardization. Forty-Seventh
Report, Technical Report Series No. 878.
not as great as originally thought WWW.PHOTOS.COM
World Health Organisation, 1998; www.who.
— or maybe the precautions int/bookorders/anglais/detart1.jsp?sesslan=1 Corresponding author Tanja Wolter is
taken during manufacturing to &codlan=1&codcol=10&codcch=878. leader of the protein analysis group
reduce those potential dangers 3 CPMP Position Statement on DNA at NewLab BioQuality AG NewLab
have been effective. Special and Host Cell Proteins (HCP): Impurities,
BioQuality AG, Max-Planck-Str. 15a,
chromatographic media can be Routine Testing Versus Validation Studies.
European Agency for the Evaluation of D-40699 Erkrath, Germany; 49-
used to reduce both residual DNA
Medicinal Products. June, 1997; www.emea. 211-9255300, fax 49-211-9255344;
and residual HCP. Sensitive assays eu.int/pdfs/human/press/pos/038297en. [email protected]. Andreas Richter is
have been developed to detect pdf. director of operations, richter@newlab.
possible residual contamination 4 ICH. Q2A Text on Validation de.
at extremely low levels. of Analytical Procedures (Guidance for
Although no serious problems Industry). International Conference on

46 BioProcess International FEBRUARY 2005