Isolation of circulating tumor cells using a

microvortex-generating herringbone-chip
Shannon L. Stotta,b,c,1, Chia-Hsien Hsua,b,c,1,3, Dina I. Tsukrova, Min Yud, David T. Miyamotod,e, Belinda A. Waltmand,
S. Michael Rothenbergd,f, Ajay M. Shaha, Malgorzata E. Smasd, George K. Korira, Frederick P. Floyd, Jr.a, Anna J. Gilmand,
Jenna B. Lordd, Daniel Winokurd, Simeon Springerd, Daniel Irimiaa,b,c, Sunitha Nagratha,b,c, Lecia V. Sequistd,g,
Richard J. Leed,g, Kurt J. Isselbacherd,2, Shyamala Maheswaranc,d, Daniel A. Haberd,f,g, and Mehmet Tonera,b,c
a
Center for Engineering in Medicine, Massachusetts General Hospital, Harvard Medical School, Boston, MA 02114; bShriners Hospital for Children,
Harvard Medical School, Boston, MA 02114; cDepartment of Surgery, Massachusetts General Hospital, Harvard Medical School, Boston, MA 02114;
d
Massachusetts General Hospital Cancer Center, Harvard Medical School, Boston, MA 02114; eDepartment of Radiation Oncology, Massachusetts General
Hospital, Harvard Medical School, Boston, MA 02114; gDepartment of Medicine, Massachusetts General Hospital, Harvard Medical School, Boston,
MA 02114; and fHoward Hughes Medical Institute, Chevy Chase, MD 20815

Contributed by Kurt J. Isselbacher, August 24, 2010 (sent for review July 14, 2010)

Rare circulating tumor cells (CTCs) present in the bloodstream of present in microfluidic devices, cells follow streamlines and dis-
patients with cancer provide a potentially accessible source for play minimal molecular diffusion across flow channels. This lack
detection, characterization, and monitoring of nonhematological of mixing results in a limited number of interactions with the anti-
cancers. We previously demonstrated the effectiveness of a micro- body-coated surface of the device, which is critical for target cell
fluidic device, the CTC-Chip, in capturing these epithelial cell adhe- capture. In our first generation device called “the CTC-Chip,”
sion molecule (EpCAM)-expressing cells using antibody-coated 78,000 antibody-functionalized microposts were used to break
microposts. Here, we describe a high-throughput microfluidic mix- up the streamlines and enhance cell-surface interactions. The
ing device, the herringbone-chip, or “HB-Chip,” which provides an CTC-Chip provided improved yield and purity of captured CTCs
enhanced platform for CTC isolation. The HB-Chip design applies
(17), enabling noninvasive serial genotyping of lung cancers
passive mixing of blood cells through the generation of microvor-
carrying epidermal growth factor receptor (EGFR) mutations
tices to significantly increase the number of interactions between
during the course of therapy with targeted kinase inhibitors
target CTCs and the antibody-coated chip surface. Efficient cell cap-
ture was validated using defined numbers of cancer cells spiked
(18) and monitoring of patients with either localized or metastatic
into control blood, and clinical utility was demonstrated in speci- prostate cancer (19). While effective at isolating CTCs from a
mens from patients with prostate cancer. CTCs were detected cohort of patients with metastatic epithelial cancers (17), the
in 14 of 15 (93%) patients with metastatic disease (median ¼ CTC-Chip relied on laminar flow limiting the interactions of
63 CTCs∕mL, mean ¼ 386
238 CTCs∕mL), and the tumor-specific target cells with surfaces, and the complex micropost structure
TMPRSS2-ERG translocation was readily identified following RNA proved challenging to scale up for high-throughput production
isolation and RT-PCR analysis. The use of transparent materials and larger-scale clinical applications. The alternative strategy
allowed for imaging of the captured CTCs using standard clinical developed here involves the use of surface ridges or herringbones
histopathological stains, in addition to immunofluorescence-conju- in the wall of the device to disrupt streamlines, maximizing
gated antibodies. In a subset of patient samples, the low shear collisions between target cells and the antibody-coated walls
design of the HB-Chip revealed microclusters of CTCs, previously themselves, thus not requiring the construction and functionali-
unappreciated tumor cell aggregates that may contribute to the zation of a complex micropost geometry.
hematogenous dissemination of cancer. A strategy for mixing of distinct solutions using such transverse
flows within microchannels to induce fluidic microvortices was
microfluidics ∣ ctcs ∣ prostate cancer ∣ clusters ∣ point-of-care originally presented by Stroock et al. (20). In extending this prin-
ciple to the capture of CTCs from whole blood, we developed a

C irculating tumor cells (CTCs) provide a potential source

of cells derived from primary or metastatic sites that may
be amenable for analysis in patients with epithelial cancers. How-
high-throughput microvortex mixing device that ensures effective
contacts of cells with antibody-coated surfaces, while designed
with simple geometry that is amenable for large-scale manufac-
ever, isolating these cells is technically challenging due to their turing. The Herringbone (HB)-Chip effectively captured CTCs
rare numbers (1 in 109 blood cells) and their low recovery follow- from patients with metastatic prostate cancer, and its low shear
ing traditional batch purification techniques. Multiple batch ap-
flow properties revealed the presence of previously unappre-
proaches have been employed to detect CTCs, ranging from cell
ciated microclusters of CTCs.
size based separation (1, 2) and fast scanning cytometry (3) to
commercially available technology based on the use of immuno-
magnetic beads (4–10). To date, the low number of CTCs isolated Author contributions: S.L.S., C.-H.H., D.I.T., M.Y., D.T.M., B.A.W., S.M.R., A.M.S., G.K.K., D.I.,
from most patients (median ≤ 1 CTC∕mL in ∼half of patients S.N., L.V.S., R.J.L., K.J.I., S.M., D.A.H., and M.T. designed research; S.L.S., C.-H.H., D.I.T., M.Y.,
D.T.M., B.A.W., S.M.R., A.M.S., M.E.S., G.K.K., F.P.F., A.J.G., J.B.L., D.W., and S.S. performed
with known metastatic cancer) has allowed demonstration of research; S.L.S., C.-H.H., D.I.T., M.Y., D.T.M., B.A.W., S.M.R., M.E.S., G.K.K., F.P.F., J.B.L., S.N.,
their correlation with poor overall survival in metastatic prostate, R.J.L., K.J.I., S.M., D.A.H., and M.T. analyzed data; and S.L.S., C.-H.H., D.I.T., K.J.I., S.M.,
breast, and colon cancers (6, 8, 11–13). However, higher CTC D.A.H., and M.T. wrote the paper.
recovery rates are required to enable detailed molecular analyses The authors declare no conflict of interest.
and to provide reliable information for clinical monitoring of Freely available online through the PNAS open access option.
individual patients undergoing therapy. 1
S.L.S. and C-H.H. contributed equally to this work.
Microfluidic devices provide important advantages over batch 2
To whom correspondence should be addressed. E-mail: [email protected].
purification techniques in that they enable highly efficient proces- 3
Present address: Division of Medical Engineering Research, National Health Research
sing of complex cellular fluids, with minimal damage to sensitive Institutes, No. 35, Keyan Road, Zhunan Town, Miaoli County, 35053, Taiwan.
cell populations due to shear forces or need for cell fixation This article contains supporting information online at www.pnas.org/lookup/suppl/
(14–16). However, under the laminar, uniaxial flow conditions doi:10.1073/pnas.1012539107/-/DCSupplemental.

18392–18397 ∣ PNAS ∣ October 26, 2010 ∣ vol. 107 ∣ no. 43 www.pnas.org/cgi/doi/10.1073/pnas.1012539107

Results (Fig. 1C), increasing the number of cell-surface interactions in
Design of the Herringbone-Chip and Flow Visualization Studies. The the antibody-coated device. Flow visualization studies demon-
HB-Chip consists of a 1” × 3” glass slide bonded to a polydi- strate this concept, showing in small scale versions of the flat
methylsiloxane (PDMS) structure, containing eight microchan- and herringbone devices that two separate streams of equal visc-
nels with patterned chevrons or herringbones on their upper osity (one fluorescently labeled green, one clear) fail to mix in the
surface (Fig. 1A). The internal walls of the device are made former but mix rapidly in the latter (Fig. S1 A, B). Three-dimen-
chemically active by in-line functionalization and coating with sional projections of confocal images illustrate the high degree of
antibodies against EpCAM (see Materials and Methods). The mixing within one cycle of herringbone grooves, with the two
geometric design of the device is based upon the general strategy streams of fluid folding over one another (Fig. 1 E, F) (see
for inducing chaotic mixing at low Re numbers (20), adapted for Movie S1, Movie S2, Movie S3, and Movie S4 for complete
isolation of rare cells from whole blood by varying the ratio of visualization of the flow fields; Fig. S2A–D). As predicted
height of the grooves to that of the channel, chevron dimensions, (20), flow rate does not impact the degree of mixing for the range
and periodicity. The herringbone grooves are staggered periodi- of Re that is relevant in this system (Fig. S1 A, B). Thus, any in-
cally, with each mixing cycle defined by two sequential regions of terpatient variances in the rheological properties of the blood will
ten chevrons shifted asymmetrically (Fig. 1B). The final dimen- not impact the extent of mixing in the herringbone CTC-Chip.
sions of the device were selected based on the optimization
studies: overall height of channel (h) 50 μm, with the ratio of HB-Chip Optimization and Validation by Capture of Cancer Cell Lines.
the height of the grooves to that of the channel (α) set to 0.8; To demonstrate the impact of herringbone-mediated mixing, the
angle between the herringbones and the axis of the channel HB-Chip was first compared with a traditional smooth wall
(θ) 45°, and principal wave vector, q ¼ 2π∕100 μm (20, 21). A microfluidic channel of similar dimensions (Fig. 2A, and
branching inlet header was designed to feed into eight individual Fig. S3 A, B). We used a single channel of either the HB-Chip
channels to promote mechanical integrity and uniform flow dis- or the smooth chamber device, connected to a microfluidic waste
tribution across the device (Fig. 1 A). collection chamber (Fig. S4). This small footprint serial chamber
In comparison to a traditional flat-walled microfluidic device setup allowed for accurate counting of target PC3 prostate cancer
(Fig. 1D), the herringbone-induced microvortices disrupt the cells captured in the devices, as well as enumeration of non-
laminar flow streamlines that cells travel, causing them to “shift” captured cells flowing into a microfluidic waste chamber, thus

APPLIED BIOLOGICAL
SCIENCES
ENGINEERING

Fig. 1. (A) The HB-Chip consists of a microfluidic array of channels with a single inlet and exit. Inset illustrates the uniform blood flow through the device. (B) A
micrograph of the grooved surface illustrates the asymmetry and periodicity of the herringbone grooves. Cartoon illustrating the cell-surface interactions in (C)
the HB-Chip, and (D) a traditional flat-walled microfluidic device. Flow visualization studies using two paired streams of the same viscosity (one stream is green,
the other clear) demonstrate (E) the chaotic microvortices generated by the herringbone grooves, and the lack of mixing in (F) traditional flat-walled devices.

Stott et al. PNAS ∣ October 26, 2010 ∣ vol. 107 ∣ no. 43 ∣ 18393
 For clinical samples, a larger version of the HB-Chip was
created (Fig. 1A). Repeating the spiked cell experiments at
1.2 mL∕hr (Fig. 2A), the large HB-Chip yielded an average
capture efficiency of 91.8%
5.2% (n ¼ 6) for PC3 cells spiked
into whole blood. In comparing the HB-Chip and the micro-
post-based CTC-Chip, operating at a flow rate that favors high
performance in both devices (1 mL∕hr), the HB-Chip yielded a
significantly improved capture efficiency in comparison with the
CTC-Chip (26.3% improvement, p ¼ 0.0001, Fig. 2B). The purity
of spiked cancer cells captured among contaminating leukoctyes
was 14.0%
0.1% for the HB-Chip, compared with 9.2%
0.1%
for the CTC-Chip (p ¼ 0.033). The kinetics of flow through
the HB-Chip allowed an increased inner chamber height to
100 μm (corresponding to a twofold increase in blood volume
throughput) without a significant drop in capture efficiency
(Fig. 2C, p > 0.05). High data reproducibility (average variance ¼
5%
1.5%, n ¼ 12)) was also evident with the HB-Chip when
analyzing duplicate samples at different cell spiking densities
(Fig. 2D).
The viability of cancer cells spiked into whole blood and cap-
tured on the HB-Chip was high (95%
0.6%, n ¼ 4, Fig. 2E), and
these cells could be cultured in vitro following separation of the
top and bottom surfaces of the device (Fig. 2F). The relative sim-
plicity of the HB-Chip design allows for manufacture using trans-
parent, chemically stable materials, enabling high magnification
imaging and assays such as FISH. Androgen receptor (AR) gene
Fig. 2. (A) Device proof-of-principle studies were conducted using PC3 cells copy number was demonstrated using on-chip FISH in the
spiked into whole blood at 1;000 cells∕mL and processed with a small version LNCaP prostate cancer cell line, following spiking into blood
of the HB-Chip (Fig. S4, channel height 70 μm, α ¼ 0.43) and a flat-walled and capture on the device (Fig. 2G). Two copies of the AR gene
device (channel height 100 μm). Capture efficiency is shown for both devices were detected in LNCaP cells, consistent with the known X chro-
in addition to IgG controls. (B) Relative capture efficiency for head to head mosome copy number in this cell line (23). Imaging CTCs cap-
comparisons run between the full-sized HB-Chip (channel height 50 μm, tured on a transparent device also allows for the use of standard
α ¼ 0.8) and the silicon CTC-Chip (17). For the data in (B), a single preparation
cytology stains, which require immunohistochemical analysis
of blood was spiked with PC3 cells at a concentration of 500 cells∕mL, divided
into equal parts and processed on both platforms. Due to slight errors in spik-
using transmitted light. Hematoxylin and eosin staining was
ing concentration, the HB-Chip capture efficiency was determined to be adapted for flow-through microfluidic conditions (see Materials
greater than 100% (112.4%
1.8%, n ¼ 4). Consequently, the comparison and Methods), producing characteristic images of captured cancer
data is normalized against the HB-Chip. (C) Variations in device chamber cells (Fig. 2H).
height showed a minimal impact on capture efficiency, until the chamber
height was increased threefold. (D) Variability in device performance is Capture of CTCs from Patients with Metastatic Prostate Cancer. Blood
shown for four independent experiments, each with different spiking con- samples from 15 patients at various stages of treatment for meta-
centrations of PC3 cells in blood. (E) Micrograph of spiked cancer cells cap-
static prostate cancer were processed with the HB-Chip. We also
tured on the HB-Chip, representative of capture cell viability (LIVE/DEAD).
(Scale bar: 40 μm). (F) Micrograph of spiked cells captured on the herringbone
processed blood from healthy individuals as controls (n ¼ 10). All
chip and subsequently cultured on the device for 21 d. (G) On-chip FISH of cells captured on the HB-Chip were stained for DNA content
captured LNCaP cells with nuclei stained with DAPI, CEPX (green) and AR (4,6-diamidino-2-phenylindole or DAPI), prostate-specific anti-
gene locus (red). (H) Micrograph of a fluorescently labeled PC3 cell captured gen (PSA, a highly specific marker for prostate epithelial cells),
on the HB-Chip (top) and subsequent micrograph taken of the same cell and a leukocyte marker (CD45). All devices were imaged in a
stained with H and E. (Scale bar: 10 μm). blinded fashion, using our automated imaging algorithm (19).
Rare “false positive” events were observed in healthy donors
enabling a true mass-balance for cell capture efficiency (see Ma- (median ¼ 1 CTC∕mL, mean ¼ 3
1 CTCs∕mL, range 0 to
terials and Methods) (Fig. 2A). Compared with other epithelial 8 CTCs∕mL), allowing us to set a detection threshold for signif-
cancer cell lines, PC3 cells express relatively low levels of cell- icance in patients at ≥10 CTCs∕mL. Among men with prostate
surface EpCAM (51,667 EpCAM molecules per cell), making cancer, CTCs were captured in 14 of 15 cases (93%), with counts
them an appropriate choice for optimization of cell capture plat- ranging from 12 to 3;167 CTCs∕mL (median ¼ 63 CTCs∕mL,
forms (22). As predicted, the HB-Chip proved highly efficient at mean ¼ 386
238 CTCs∕mL) (Fig. 3A, see Table S1 for clinical
capturing PC3 cells (Fig. 2A), whereas significantly less capture characteristics). We have recently reported capture of CTCs in
was measured with the flat chamber device (p < 0.05 for all flow metastatic prostate cancer in 23 of 36 (64%) patients using
rates tested). At very low flow rates (0.12 mL∕hr, translating to the micropost CTC-Chip, under similar capture and enumeration
1.2 mL∕hr in the large-scale HB-Chip), the overall capture effi- conditions (19). Thus, in this patient population the HB-Chip
ciencies in the flat chamber (29%
4.3% (n ¼ 4)) and HB-Chip demonstrates equal or improved capture efficiency. To validate
(79%
4.5% (n ¼ 8)) reflect the settling of the PC3 cells towards the HB-Chip for molecular characterization of CTCs, we under-
the bottom surface of the device. However, as the flow rate is took on-chip lysis of captured cells and isolated RNA of sufficient
increased, the average capture efficiency for the flat chamber quality for RT-PCR assays, identifying the known TMPRSS2-
device drops to <8%, whereas the HB-Chip maintains a target ERG chimeric transcript, confirmed by nucleotide sequencing,
cell capture efficiency >40% for all flow rates up to 0.48 mL∕hr in a patient whose primary tumor harbored this translocation,
(or 4.8 mL∕hr in the large HB-Chip). The capability to run at but not in a patient whose tumor did not have the chromosome
higher flow rates is an additional benefit in comparison to the fusion (Fig. 3B).
CTC-Chip, previously shown to have a significant drop in capture The transparent materials used in the manufacture of the
efficiency at flow rates above 2–3 mL∕hr (17). HB-Chip made it possible to complement immunofluorescence

18394 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1012539107 Stott et al.

 Fig. 3. Blood from metastatic prostate cancer pa-
tients and healthy donors was processed using the
HB-Chip. (A) Healthy donors (n ¼ 10, red circles) were
used to select the signal intensity threshold in our
automated imaging system (19). The metastatic pros-
tate cancer patient data (n ¼ 15, green circles) had
significantly higher results, with 14∶15 patient sam-
ples presenting CTCs above-threshold levels. Addi-
tionally, (B) the TMPRSS2-ERG fusion transcript was
detected from RNA isolated from the HB-Chip, using
RT-PCR. Sequencing of the gel band identified the
rare gene fusion of TMPRSS2 exon 1 and ERG exon
5. (C and D) Micrographs of CTCs isolated from pa-
tients with metastatic prostate cancer costained
using antibodies against PSA (green), a leukocyte
marker, CD45 (red), and a nuclear stain (DAPI). Cor-
responding micrographs of the same cells are shown
to the right of the fluorescent images, demonstrating
their appearance after subsequent staining with H
and E. (Scale bars: 10 μm).

staining of CTCs with stains used in standard pathology labora- suggesting that CTC isolation strategies relying solely on size-
tories to identify tumor cells using light microscopy. Thus, a CTC based filtration may exclude a significant portion of the CTC
identified using a specific fluorescence-conjugated antibody can population.
be reevaluated under high magnification using transmitted light
to gain additional insights into cytomorphologic and nuclear Identification of CTC Microclusters in the Bloodstream of Patients with
features (Fig. 3 C, D and Fig. S5C–F). Characteristics of tumor Cancer. While most of the CTCs captured on the HB-Chip were
cells, such as large nuclei or visible nucleoli, are evident in some single cells, we observed rare clusters of 4–12 cells (Fig. 4 A, D).
circulating cells identified by expression of PSA (Fig. 3D), pros- Such clumps of tumor cells were never identified using the CTC-
tate-specific membrane antigen (PSMA) (Fig. 4C), or cytokeratin Chip, where interpost distances are designed to allow flow of
(Fig. 4D and Fig. S5E–F). single cells (17), nor have they been reported after the batch im-
The ability to combine immunofluorescence with light micro- munomagnetic bead purification process (4–9). Microclusters
scopic analysis provides insight into a common challenge in CTC were identified in specimens from 3 out of 19 patients with either
isolation, namely the identification of cells that stain both for the metastatic prostate or lung cancer. These were analyzed using
epithelial and the leukocyte marker. While these “double posi- immunofluorescence (PSMA or keratin), as well as hematoxylin
tive” cells are typically excluded from CTC enumeration, they and eosin stains (Figs. 4 C–D). Microclusters have never been
are more prevalent than CTCs, and their identity has been the identified following spiking of cancer cell lines into control blood
subject of much debate: coating of CTCs with leukocyte-derived specimens in our experiments, suggesting that they are not simply
proteins may occur during their circulation in the bloodstream, cell aggregations resulting from physical properties of the
while some monocytes may ingest tumor-derived markers. To HB-Chip. Three-dimensional imaging of the microclusters also
begin to address this challenge, which has been described with demonstrates that they have preserved shape and orientation,
all CTC isolation platforms, we identified double positive cells and are not groupings of cells flattened onto the walls or grooves
(PSAþ, CD45þ) captured by the HB-Chip from four patients of the device (rotating three-dimensional projection Movie S5).
with metastatic prostate cancer. In the majority of cases, hema- Discussion
toxylin and eosin staining revealed the characteristic morphology The application of microfluidic technology to the isolation of
of leukocytes (Fig. 4B), suggesting that these cells are appropri- CTCs provides an opportunity for improved capture yield for
ately excluded from CTC enumeration. However, there are also a these rare cells and for their preserved viability, allowing for
few cells with more ambiguous staining patterns in this popula- detailed molecular and eventually functional characterization.
tion (Fig. 4A), whose size and staining patterns are more consis- However, the constraints inherent in large-scale production
tent with CTCs. To maintain a conservative approach in our CTC and chemical modification of a complex microfluidic device, com-
APPLIED BIOLOGICAL

enumeration schema, these cells are still excluded from overall bined with the need to image rare cells captured in a nontranspar-
counts. It is of interest that double positive cells are observed ent three-dimensional array of microposts provide a challenge for
SCIENCES

at a significantly lower frequency in healthy blood donors, sug- “point-of-care” application of our previously described CTC-
gesting that these are not simply artifacts of postcapture staining. Chip (24). Here we demonstrate the application of a microfluidic
Further studies will be required to identify specific properties of vortex generator, the HB-Chip, which uses herringbone grooves
these cells to explain their costaining for epithelial markers in within the transparent wall of the device to create microvortices
patients with metastatic cancers. that direct target cells toward capture on the antibody-coated
In addition to patients with metastatic prostate cancer, we walls. This “second generation” device allows for higher blood
isolated CTCs from patients with lung cancer (n ¼ 4), using volume throughput, and increased CTC capture efficiency and
ENGINEERING

antiEpCAM capture followed by staining for DAPI, CK7/8, purity. In addition to the imaging benefits provided by its trans-
and CD45 (Fig. S5). As for prostate cancer patients, the variable parent walls, the less complex three-dimensional structures of the
size (8–16 μm in diameter) and the morphological heterogeneity HB-Chip greatly facilitates scaled up device production, and will
of CTCs is well illustrated (Fig. S5 A–D). Of note, the lower size thus enable the initiation of larger-scale clinical studies. In this
limit of CTCs is comparable to the median size of leukocytes, initial proof of principle, the HB-Chip successfully isolated CTCs

Stott et al. PNAS ∣ October 26, 2010 ∣ vol. 107 ∣ no. 43 ∣ 18395
Fig. 4. (A and B) Micrographs of cells isolated on the HB-Chip from a metastatic prostate cancer patient that were not classified as CTCs due to their dual-
positivity for both PSA (green) and CD45 (red). Corresponding H and E images are shown. (C) Micrographs of a CTC cluster isolated from a metastatic prostate
cancer patient on the HB-Chip; immunofluoresence staining (DNA (blue), prostate-specific membrane antigen (green), and CD45 (red)) and subsequent
immunohistochemical staining (H and E) are shown. A three-dimensional projection of this cluster can be viewed in SI Text. (D) A CTC cluster isolated from
a metastatic lung cancer patient on the HB-Chip; immunofluoresence staining (DNA (blue), cytokeratins 7∕8 (green), and CD45 (red)) and subsequent immu-
nohistochemical staining (H and E) are shown. All scale bars represent 10 μm.

from 14 of 15 (93%) patients with metastatic prostate cancer captured CTCs, the combined use of immunofluorescence
(median 63 CTCs∕mL) and allowed RNA-based detection of the and light microscopy-based stains also provides insight into the
TMPRSS2-ERG fusion transcript, showing a further improve- identity of the so-called “double positive” cells, positive for both
ment over our previous generation CTC-Chip (19). epithelial and hematopoietic markers. While these cells may
The application of standard histopathology imaging platforms greatly outnumber the “single positive” epithelial cells that are
and metrics to the study of CTCs is essential to their broad typically scored as CTCs, their identity and significance has been
application in cancer diagnostics and to realize the potential the subject of debate. The demonstration that most of these cells
of a “blood biopsy” for epithelial cancers. To date, the very have the morphology of mononuclear leukocytes confirms their
low numbers of CTCs isolated by commercial platforms, together appropriate exclusion from CTC enumeration.
with the reliance on relatively low resolution immunofluores- The most unexpected finding of this study was the isolation
cence-based imaging has limited such applications. As such, using the HB-Chip of clusters of CTCs from the blood of patients
the ability to collect significant numbers of CTCs on a transparent with metastatic cancer. In some cases, the size of these clusters
slide is likely to facilitate the use of standardized criteria for ana- was quite large, consisting of up to 14 CTCs, with cells within the
lysis of tumor cells, as commonly used in the cytologic examina- cluster having highly variable cytomorphologic features. These
tion of ascites or pleural effusions. Previous studies using straight clusters have not been reported using the commercial immuno-
deposition of blood onto glass slides postlysis of red blood cells magnetic bead platform, presumably because they do not with-
have shown that some Wright-Giemsa-stained CTCs may share stand the multiple steps associated with batch purification, and
morphological appearance with cells within the primary tumor they were also not evident using the CTC-Chip, since the tightly
biopsy (25, 26). As shown here, significant heterogeneity in he- packed spacing of microposts may have prevented their passage.
matoxylin and eosin staining intensity is evident among CTCs, While we cannot formally exclude the possibility that in vitro
despite uniform staining of contaminating leukocytes. Consistent clustering of CTCs might result from HB-Chip-induced flow
with standard cytology observations, cells with reduced staining patterns, this is unlikely, since we have never identified such clus-
intensity are likely to be at early stages of autolysis, a fate that ters in experiments with cancer cells spiked into control blood
would be expected for most CTCs in the circulation. In addition specimens, nor is the presence of CTC clusters correlated with
to the detailed morphological analysis now possible for some the absolute number of CTCs in clinical specimens. The existence

18396 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1012539107 Stott et al.

of clusters of CTCs in the blood of patients with cancer may pro- CTC Staining and Enumeration. After processing of blood on the HB-Chip, cells
vide insight into the process of metastasis in human cancer (27). were fixed with 4% paraformaldehyde, and subsequently permeabilized
with 0.2% Triton X-100 (Sigma-Aldrich) in PBS. Cells were immunostained
Given technological limitations to date, the mechanisms under-
with primary antibodies, either a rabbit polyclonal antibody to PSA (DAKO),
lying cancer metastasis have been studied exclusively in mouse monoclonal antibody against cytokeratin 7/8 (CAM 5.2 clone, BD Biosciences),
models. These have emphasized the altered motility of single cells or a monoclonal mouse IgG2b antibody to PSMA (Santa Cruz Biotechnology).
intravasating from the primary tumor into the circulation, fol- Following, the appropriately matched secondary AlexaFluor 488-conjugated
lowed by their active extravasation at distant sites of metastasis. antibody (Invitrogen) was used to identify epithelial cells. Nuclei were stained
Circulating CTC clusters could either constitute microemboli with DAPI. All samples were counterstained with mouse IgG1 antihuman
breaking off from a tumor mass or, alternatively, they could result anti-CD45 (Clone H130, BD Biosciences), followed by Alexa Fluor 594 fluor-
ophore (Invitrogen) to identify any bound leukocytes. After completion of
from intravascular proliferation of tumor cells (28). By whatever staining, all devices were washed PBS and stored at 4 °C. All devices were
mechanism they arise, CTC clusters have the potential to lodge in imaged using our previously described imaging system (19). See SI Text for
distal capillary beds, thereby initiating metastatic lesions. Further additional details.
characterization of these clusters in additional cancer types and at
various stages of cancer progression will be required to ascertain Statistical Analysis. Data are reported as mean
standard error of the mean
their potential contribution to the metastatic spread of cancer. as noted. If groups had a normal distribution and homogenous variances, the
group means were compared by an independent t-test. Differences were
Materials and Methods considered significant at the 95% confidence level (p < 0.05).
Device Fabrication and Surface Modification. Microfluidic devices were made See SI Text for details on the hematoxylin and eosin staining, on-chip FISH,
of PDMS bonded to glass substrates using soft lithography techniques (29). RT-PCR analysis and sequencing, cell line preparation, flow visualization, and
The microfluidic devices were functionalized with epithelial cell adhesion cancer cell line experiments.
molecule (EpCAM, R and D Systems) antibody using avidin-biotin chemistry.
The immobilization of antibody on the glass and PDMS surfaces was achieved ACKNOWLEDGMENTS. We express our gratitude to the all patients who
using a previously described method (30). See SI Text for additional details. participated in this study and the healthy volunteers who contributed blood
samples. We are also grateful to Octavio Hurtado, Jochen Lennerz, Douglas
Rubinson, Vijay Ambravaneswaran, Salil Desai, Ravi Kapur, John Walsh,
Study Subjects and Blood Processing. Patients with advanced prostate and Tom Barber, Justin Wong, Zev Nakamura, Matthew Ulman, Suchismita Paul,
lung cancer were recruited according to a protocol approved by the institu- Brian Brannigan, Alessandra Moore, Maria Kempner, Brooke Nentwig, and
tional review board (IRB). Blood specimens from healthy volunteers were col- Laura Libby for expert technical support. This work was supported by Dream
lected under a separate IRB-approved protocol. A total of 19 cancer patients Team Awards from the Prostate Cancer Foundation and Stand Up To Cancer
(15 prostate, 4 lung), who were treated at the Massachusetts General Hos- (to D.A.H. and M.T.), a Quantum Grant from the National Institute for
Biomedical Imaging and Bioengineering (to M.T.), National Institutes of
pital Cancer Center, donated 10–20 mL of blood on one or more occasions Health/National Cancer Institute (NIH)/(NCI) Grant CA89138 (to S.M.), and
for analysis on the HB-Chip. All specimens were collected into vacutainer generous donations from the Ellison Foundation, AstraZeneca, the Martell
(Becton-Dickinson) tubes containing the anticoagulant EDTA and were pro- Foundation, Alex and Sonja Spier, the Monell Foundation, and institutional
cessed through the HB-Chip within 6 h of blood draw. Samples were run on support from Massachusetts General Hospital. S.L.S. is supported through an
the previously described microfluidic processing machine (17). Briefly, a 5 mL American Cancer Society New England Division Research Grant, S.N. is
supported through an NIH Director’s New Innovator Award, R.J.L. is
aliquot of blood was placed in an air-tight conical tube on a rocker assembly,
supported by a Physician Research Training Award (Department of Defense,
and ∼4 mL of blood were pneumatically driven through the chip at a flow Prostate Cancer Research Program) and a Career Development Award
rate of 1.5–2.5 mL∕hr. Following, the HB-Chip was flushed with 2.5 mL of PBS (The American Society of Clinical Oncology (ASCO) Cancer Foundation),
at 2.5 mL∕hr to remove any nonspecifically bound cells. and D.A.H. is supported by the Howard Hughes Medical Institute.

1. Vona G, et al. (2000) Isolation by size of epithelial tumor cells: a new method for the 15. Adams AA, et al. (2008) Highly efficient circulating tumor cell isolation from
immunomorphological and molecular characterization of circulating tumor cells. Am J whole blood and label-free enumeration using polymer-based microfluidics with
Pathol 156:57–63. an integrated conductivity sensor. J Am Chem Soc 130:8633–8641.
2. De Giorgi V, et al. (2010) Application of a filtration- and isolation-by-size technique for 16. Phillips JA, Xu Y, Xia Z, Fan ZH, Tan W (2009) Enrichment of cancer cells using aptamers
the detection of circulating tumor cells in cutaneous melanoma. J Invest Dermatol immobilized on a microfluidic channel. Anal Chem 81:1033–1039.
130:2440–2447. 17. Nagrath S, et al. (2007) Isolation of rare circulating tumor cells in cancer patients by
3. Krivacic RT, et al. (2004) A rare-cell detector for cancer. Proc Natl Acad Sci USA microchip technology. Nature 450:1235–1239.
101:10501–10504. 18. Maheswaran S, et al. (2008) Detection of mutations in EGFR in circulating lung-cancer
4. Fehm T, et al. (2002) Cytogenetic evidence that circulating epithelial cells in patients cells. N Engl J Med 359:366–377.
with carcinoma are malignant. Clin Cancer Res 8:2073–2084. 19. Stott SL, et al. (2010) Isolation and characterization of circulating tumor cells from
5. Moreno JG, et al. (2001) Changes in circulating carcinoma cells in patients with patients with localized and metastatic prostate cancer. Sci Transl Med 2:25ra23.
metastatic prostate cancer correlate with disease status. Urology 58:386–392. 20. Stroock AD, et al. (2002) Chaotic mixer for microchannels. Science 295:647–651.
6. Moreno JG, et al. (2005) Circulating tumor cells predict survival in patients with 21. Hsu CH, Di Carlo D, Chen C, Irimia D, Toner M (2008) Microvortex for focusing, guiding,
metastatic prostate cancer. Urology 65:713–718.
and sorting of particles. Lab Chip 8:2128–2134.
7. Fizazi K, et al. (2007) High detection rate of circulating tumor cells in blood of patients
22. Rao CG, et al. (2005) Expression of epithelial cell adhesion molecule in carcinoma
with prostate cancer using telomerase activity. Ann Oncol 18:518–521.
cells present in blood and primary and metastatic tumors. Int J Oncol 27:49–57.
8. Danila DC, et al. (2007) Circulating tumor cell number and prognosis in progressive
23. Pan Y, et al. (1999) Characterization of chromosomal abnormalities in prostate cancer
castration-resistant prostate cancer. Clin Cancer Res 13:7053–7058.
cell lines by spectral karyotyping. Cytogenet Cell Genet 87:225–232.
9. Wang ZP, et al. (2000) Identification and characterization of circulating prostate
24. Schmidt C (2010) Assays that predict outcomes make slow progress toward prime
carcinoma cells. Cancer 88:2787–2795.
APPLIED BIOLOGICAL

time. J Natl Cancer Inst 102:677–679.

10. Talasaz AH, et al. (2009) Isolating highly enriched populations of circulating epithelial
25. Marrinucci D, et al. (2007) Case study of the morphologic variation of circulating tumor
cells and other rare cells from blood using a magnetic sweeper device. Proc Natl Acad
SCIENCES

Sci USA 106:3970–3975. cells. Hum Pathol 38:514–519.

11. de Bono JS, et al. (2008) Circulating tumor cells predict survival benefit from treatment 26. Marrinucci D, et al. (2009) Circulating tumor cells from well-differentiated lung
in metastatic castration-resistant prostate cancer. Clin Cancer Res 14:6302–6309. adenocarcinoma retain cytomorphologic features of primary tumor type. Arch Pathol
12. Olmos D, et al. (2009) Circulating tumor cell (CTC) counts as intermediate end points in Lab Med 133:1468–1471.
castration-resistant prostate cancer (CRPC): a single-center experience. Ann Oncol 27. Chiang AC, Massague J (2008) Molecular basis of metastasis. N Engl J Med
20:27–33. 359:2814–2823.
13. Scher HI, et al. (2009) Circulating tumor cells as prognostic markers in progressive, 28. Al-Mehdi AB, et al. (2000) Intravascular origin of metastasis from the proliferation of
castration-resistant prostate cancer: a reanalysis of IMMC38 trial data. Lancet Oncol endothelium-attached tumor cells: a new model for metastasis. Nat Med 6:100–102.
10:233–239. 29. Xia YN, Whitesides GM (1998) Soft lithography. Ann Rev Mater Sci 28:153–184.
ENGINEERING

14. Dharmasiri U, et al. (2009) Highly efficient capture and enumeration of low abundance 30. Murthy SK, Sin A, Tompkins RG, Toner M (2004) Effect of flow and surface
prostate cancer cells using prostate-specific membrane antigen aptamers immobilized conditions on human lymphocyte isolation using microfluidic chambers. Langmuir
to a polymeric microfluidic device. Electrophoresis 30:3289–3300. 20:11649–11655.

Stott et al. PNAS ∣ October 26, 2010 ∣ vol. 107 ∣ no. 43 ∣ 18397