Evaluation of a multi-marker immunomagnetic enrichment assay for the quantification of circulating melanoma cells

A2058 metastatic melanoma cells and human embryonic kidney 293 (HEK293) cells were cultured in DMEM supplemented with 10% foetal bovine serum (FBS).

Mouse anti-human monoclonal antibodies to MCAM (clone P1H12), MCSP (clone 9.2.27), CD271 (clone C40-1457) and CD45 (clone HI30) were purchased from BD Biosciences (San Jose, CA, USA). Rabbit anti-human polyclonal antibody to ABCB5 (clone RB16781) was purchased from Abgent (San Diego, CA, USA). Secondary antibodies were anti-mouse and anti-rabbit conjugated to Alexa-Fluor 488 (A488) and biotinylated anti-mouse, which was used with streptavidin-A488.

Patients, recruited from 3 clinics in Perth, Australia, were diagnosed and staged according to guidelines of the American Joint Committee on Cancer (AJCC). All patients recruited between August 2011 and March 2012 were included in the study. Peripheral blood samples were obtained from 10 non-metastatic (stage I and II) and 13 metastatic (stage III and IV) patients. Additionally, 15 control blood samples were obtained from healthy volunteers. Blood was drawn by phlebotomists into EDTA tubes, after the first few millilitres was discarded to avoid epithelial contamination, and refrigerated until use. For 21 of 23 patients, blood was collected into five EDTA tubes; only one tube was collected for the remaining two patients and all controls. Patient samples ranged from 4 to 10 ml and control samples were 9 ml. All results were therefore calculated per ml of whole blood. Participants signed informed consent with the clinician in accordance with protocols safeguarding patient rights. All procedures have been accepted by the Human Research Ethics Committees at ECU (No. 2932) and SCGH (No. 2007–123).

Expression of MCAM, MCSP, ABCB5, CD271 and CD45 was tested in A2058 and HEK293 cells by immunofluorescence. Approximately 60,000 cells were seeded per coverslip and allowed to adhere overnight. Attached cells were fixed with 4% paraformaldehyde for 10 minutes at room temperature and stained with primary antibody, diluted in PBS with 3% BSA for one hour at room temperature. MCAM, MCSP and CD45 antibodies were diluted 1/500, ABCB5 1/250 and CD271 1/750. Cells were then stained with secondary antibody, diluted 1/500, for 20 minutes at room temperature. Cells were washed three times with PBS between each step. Coverslips were mounted onto microscope slides with ProLong Gold anti-fade mounting medium containing DAPI for nuclear staining (Invitrogen, Carlsbad, CA, USA) before analysis with an Olympus BX51 microscope equipped with an Olympus DP71 camera.

MCSP, MCAM, ABCB5 and CD271 antibodies were covalently bound to magnetic beads using the Dynabeads Antibody Coupling Kit (Invitrogen), as per the manufacturer’s instructions. For efficient antibody-bead coupling, 10 μg of antibody was used per mg of Dynabeads.

A2058 cells were harvested in DMEM containing 5 mM EDTA and analysed with the Vi-Cell-XR Cell Viability Analyser (Beckman Coulter, Brea, CA, USA). 0.5 × 10⁶ A2058 cells were added to six tubes. 0.5 μl of MCSP beads was added to three tubes and 0.5 μl of MCAM beads was added to the remaining three tubes. Tubes were incubated for one hour at 4°C with rotation and subsequently placed on a DynaMag-2 magnet (Invitrogen). Enriched cells were washed three times with PBS, resuspended in 500 μl of PBS and counted with the Vi-Cell-XR Cell Viability Analyser. 0.5 μl of MCSP beads bound a mean of 1.42 × 10⁵ cells and 0.5 μl of MCAM beads bound a mean of 0.98 × 10⁵ cells.

A2058 cells were harvested as previously described. Only cells with greater than 90% viability were used. Cells were spiked into DMEM/10% FBS or white blood cells (WBCs) resuspended in 1 ml incubation buffer (0.5% BSA, 2 mM EDTA in PBS, pH 7.2), following lysis of control blood with lysis buffer (140 mM NH4Cl, 17 mM Tris, pH 7.65). Spikes of 1, 10 and 20 cells were performed by pippetting under a microscope, while 50 and 100 cell spikes were performed by dilution. Cells were incubated with 0.5 μl of MCSP beads or a combination of MCSP, MCAM, ABCB5 and CD271 beads (with a combined volume of 0.5 μl) for one hour at 4°C with rotation. Enriched cells were placed on a DynaMag-2 magnet (Invitrogen), washed three times with PBS and mounted on microscope slides as previously described. Before mounting, cells enriched from control blood were fixed with 4% paraformaldehyde for 10 minutes at room temperature, stained with CD45 antibody (diluted 1/500 in PBS containing 3% BSA) for one hour at room temperature, followed by biotinylated anti-mouse and streptavidin-A488, both for 10 minutes at room temperature. Two PBS washes were performed between each antibody incubation. Enriched cells were quantified by microscopy where WBCs were defined as CD45 positive and melanoma cells as CD45 negative.

Control and patient blood was processed within 24 hours of collection. Blood was lysed as previously described. For controls and patients, one blood sample was incubated with a combination of MCSP, MCAM, ABCB5 and CD271 beads (combined volume 0.5 μl). The four remaining patient samples were incubated with 0.5 μl of one of the four individual bead types. Samples were incubated, washed and stained as previously described and quantified by microscopy.

Enriched cells attached to magnetic beads were lysed for whole genome amplification using the Repli-g kit (Qiagen, Germantown, MD, USA). In summary, after two PBS washes, all buffer was removed and cells attached to the magnetic beads were resuspended in 3.5 μl of cell lysis buffer containing DTT, incubated 10 minutes at room temperature then stop solution was added. Whole genome amplification was carried out as per the manufacturer’s instructions.

The V600E mutation was detected using the qBiomarker Somatic Mutation PCR Assay (c.1799 T > A) relative to a reference PCR for the B-raf gene (SABiosciences, Qiagen). Quantitative PCR was performed in a ViiA 7 Real-Time System (Applied Biosystems, Foster City, CA, USA).

Statistical tests were performed using SPSS Statistics 19 (IBM). The Mann–Whitney U Test was used to compare the number of CMCs between patients and controls and between metastatic and non-metastatic patients. The Wilcoxon Signed Rank test was used to compare the number of CMCs enriched with a combination of beads versus each individual bead type. P < 0.05 was considered statistically significant. The two patients with missing data were excluded from the appropriate analyses.

Expression of MCAM, MCSP, ABCB5, CD271 and CD45 was demonstrated in A2058 melanoma cells by immunofluorescence. HEK293 cells, used as negative controls, were 100% negative when stained with each antibody (not shown). A2058 cells were 100% positive for MCSP and MCAM (Figure 1a and Figure 1b respectively), while ABCB5 and CD271 were expressed on <1% of A2058 cells (Figure 1c and Figure 1d respectively), suggesting the presence of rare subpopulations within the cell line. A2058 cells were 100% negative when stained with CD45 (Figure 1e) and secondary antibody only (Figure 1f). This result supports the use of MCSP and MCAM to enrich the general CMC population. ABCB5 and CD271 appear useful for isolating rare CMC subsets.

To compare the sensitivity and specificity of enrichment targeting a single marker versus multiple markers (Figure 2) an increasing number of A2058 cells (1, 10, 20, 50 and 100) were spiked into media (DMEM/10% FBS) or control blood in triplicate and enriched with MCSP beads or a combination of all beads.

Percentage retrieval of cells spiked into media was high (mean = 76.7%). However, retrieval of cells from control blood was much less sensitive (mean = 35.8%), presumably due to the low specificity of the beads. This is supported by the observation of a large number of CD45 positive cells in the enriched fraction. Increasing the volume of beads increased the amount of non-specific binding but did not improve the recovery of spiked A2058 cells.

The highest percentage recovery was observed in single cell spikes (Figure 2), however it should be noted that CD45 negative cells were counted as recovered A2058 cells in blood. As we occasionally observed single CD45 negative cells in control blood (Figure 3b), a CD45 negative cell enriched from a single cell spike may not always be an A2058 cell and may account for the increased percentage recovery of single A2058 cells from blood.

There was little difference between recoveries of spiked cells using MCSP beads alone versus a combination of beads. However, given our observation that 100% of A2058 cells are MCSP positive, the benefit of a combination of beads may not be apparent in the recovery of spiked cells as much as it may improve the recovery of CMCs from patient blood, where CMC phenotype is unknown.

In order to improve the sensitivity and specificity of the assay we tested the indirect method of immunomagnetic enrichment, where cells are first labelled with primary antibody and subsequently enriched with antibody-specific beads. We enriched spiked A2058 cells from control blood using the Dynabeads CELLection Pan Mouse IgG Kit (Invitrogen) according to the manufacturer’s instructions. While non-specific enrichment was decreased, the percentage of spiked cells recovered was also reduced (data not shown).

CMCs were enriched from 15 control and 23 patient blood samples with a combination of MCSP, MCAM, ABCB5 and CD271 beads. In order to exclude non-specifically bound blood cells, enriched cells were stained for CD45, which is expressed on all human leukocytes [37]. Positively enriched CD45 negative cells were considered CMCs (Figure 4). A maximum of one CMC, or 0.11 CMCs per ml of whole blood, was detected in control blood samples. More than one CMC was detected in 73.9% of patients and ranged from 0 to 2.5 CMCs per ml (Figure 3b). There were significantly more CMCs per ml of whole blood in patients than controls (p < 0.001, Mann–Whitney U-Test). It is important to note that a maximum of one CMC was detected in 9 ml of blood in 4 of the 15 control samples. By contrast, increased numbers of CMCs were detected in patient samples even when lower volumes, usually 4 or 6 ml, were analysed.

Although positive enrichment has been shown to result in higher purity than negative enrichment [26], we still observed low purity of CMCs (data not shown). We trialled a CD45 depletion step before positive enrichment to increase the purity of enriched CMCs, however, our results showed that CD45 depletion did not reduce the number of non-specifically enriched leukocytes (data not shown).

An increasing number of A2058 cells (1, 5, 10 and 100) were spiked into control blood. Unspiked blood was used as a negative control in all experiments. Enriched cells attached to the beads were lysed and whole genome amplified. Amplified DNA was used to detect the B-raf V600E mutation by real-time PCR. A2058 cells are heterozygous for the V600E mutation [38]. We were consistently able to detect the mutation when 10 and 100 cells were spiked. However, PCR positivity was only observed in 30% of cases when 1 or 5 cells were spiked. This is in line with the percentage retrieval demonstrated above (Figure 2). It is possible that allele drop-out, which is a common problem in single-cell PCR [39], might have occurred in the samples where only one cell was present, reducing the number of tests positive for the B-raf V600E mutation in 1 and 5 cell spikes. Nevertheless, our results demonstrated that we were able to detect the V600E mutation in one cell amid a background of non-specifically bound leukocytes. The V600E mutation was further detected in CMCs enriched with a combination of magnetic beads in 6 of 8 patients tested, who had confirmed V600E-positive tumours. This result confirmed the enriched fraction contained tumour derived cells.

CMCs were enriched with antibody-coupled beads individually and in combination to see whether targeting multiple markers resulted in enrichment of more CMCs. Significantly more CMCs were enriched per ml whole blood when beads were used in combination rather than individually (MCAM, p = 0.028; MCSP, p = 0.002; ABCB5, p < 0.001; or CD271, p = 0.001; Wilcoxon Signed-Rank Test) (Figure 3a).

To determine whether there was a difference in the number of CMCs between early and late stage patients, CMCs were enriched in 10 non-metastatic and 13 metastatic patients. Significantly more CMCs were enriched from metastatic patients when a combination of beads was used (p = 0.007, Mann–Whitney U-Test) (Figure 3b). There was however no statistically significant difference when any individual marker was targeted.