M30 Neoepitope Expression in Epithelial Cancer: Quantification of Apoptosis in Circulating Tumor Cells by CellSearch Analysis

The finding of tumor cells in peripheral blood raises questions as to their metastatic potential. In fact, notwithstanding that a single tumor cell was proved to sustain metastasis in vivo (1), in humans the half-life of circulating tumor cells (CTC) in peripheral blood is estimated at between 1 and 2.4 hours, depending on the mathematical model of the extrapolated curve (2) and on the fact that apoptotic cells significantly contribute to the CTC fraction in breast (3) and prostate cancer (4) patients. On the other hand, a strict correlation was established between CTC count and prognosis (5), and elevated numbers of CTC at any time during therapy was reported to be an accurate indication of subsequent rapid progression and mortality (6). Nevertheless, the phenotypic and biological properties of the CTC that are necessary for the metastatic process are far from clear. Theoretically, CTC should be adapted to shed into peripheral blood and at least some of the CTC should be live cells. Moreover, although the metastatic phenotype may be later acquired as a result of selective pressure exerted at secondary sites, at least some of the CTC should be able to self-renew (7).

Addressing the role and mechanism of CTC in the development of metastasis, we investigated whether or not CTC are live cells, considering mainly when and how often the percentage of apoptotic CTC changes throughout disease course and treatment. For this purpose, we analyzed CTC in our patient cohort by the CellSearch system, an automated platform that permits serial testing with good sensitivity and reproducibility (5). CTC assay was integrated with a monoclonal antibody (mAb), anti-M30, for recognizing (8) a neoepitope in cytokeratin 18 (CK18) that becomes available at a caspase cleavage event during apoptosis and is not detectable in vital epithelial cells; the M30 neoepitope appears early in the apoptotic cascade, with Annexin V reactivity, and it is generally regarded as a stable biomarker, specific for epithelial cell apoptosis.

The results obtained in breast, renal, and colorectal cancer patients are presented, indicating that variable numbers of apoptotic CTC can be detected in all solid tumors. The M30-integrated assay seems to be a feasible tool for monitoring apoptotic CTC.

Peripheral blood was consecutively drawn from 34 breast cancer patients (33 female and 1 male, ages 39-83 years), 29 metastatic renal cancer (mRCC) patients (6 female and 23 male, ages 26-87 years), and 59 colorectal cancer patients (21 female and 38 male, ages 30-87 years) at baseline, before starting treatment. Breast and colorectal cancer patients were enrolled in this study regardless of type or line of therapy. Prior adjuvant treatment, treatment of metastatic disease, or both were permitted in the case of breast or colorectal metastatic cancer, whereas mRCC patients were consecutively enrolled in a pilot study, “Metastatic renal cancer: CTC determination in first-line Sunitinib treated patients,” conducted at Istituto Oncologico Veneto (IOV-IRCCS), Padova, Italy. Whole blood was also drawn from healthy control subjects (4 female and 4 male, ages 30-60 years) who had neither known illness at the time of sampling nor history of malignant disease. All enrolled patients and healthy subjects gave their informed consent for study inclusion and were enrolled using institutional review board–approved protocols. After baseline evaluation, serially monitored (1-10 months) reevaluations of disease status in the breast cancer patients were conducted depending on the type and schedule of treatment. Tumor measurements by appropriate scans were done using Response Evaluation Criteria in Solid Tumors guidelines without independent radiology review, with no knowledge of the levels of CTC.

The breast cancer cell line MCF7, the prostate cancer cell line PC3, and the colon cancer cell line LoVo were purchased from the American Type Culture Collection and were grown as described (9).

To quantitatively evaluate spontaneous and drug-induced apoptosis in the cancer cell lines, four different methodologies were compared: Annexin V apoptosis assay, anti-M30 immunostaining, terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) assay, and Western blood analysis (WB) for poly(ADP-ribose) polymerase (PARP) cleavage, according to the manufacturer's instruction as detailed in Supplementary Materials.

The enumeration of CTC in whole blood was done by the CellSearch System according to manufacturer's instruction as described (5).

To quantify the fraction of apoptotic CTC, M30-positive CTC were detected integrating CTC assay with a specific mAb, M30 CytoDEATH Fluorescein (ALX-804-590, Alexis Biochemicals), recognizing the M30 neoepitope of CK18, analyzed with the fourth filter of the CellSearch System; results were expressed as the total number of CTC and M30-positive CTC per 7.5 mL of blood.

Data were analyzed utilizing the StatGraphics software (version 2.6), as previously reported (10). Unless otherwise indicated, all results are expressed as mean values ± 1 SD, and mean values of three experiments are shown. The nonparametric Mann-Whitney test was used to compare quantitative variables. Frequencies were compared by Fisher' exact test (two tails) or χ² test with Yates' correction where appropriate.

Several apoptosis assays devised to detect different components of the apoptosis signaling cascade or specific apoptotic features, including DNA fragmentation, caspase activity, membrane alterations, and mitochondrial changes, are currently available (11). Using more than one method is strongly recommended because of the limited sensitivity and specificity of current assays; moreover, choosing the most appropriate apoptotic assay should also be based on sample type (tissues or cellular effusions). Therefore, we firstly evaluated the sensitivity and specificity of M30 immunostaining, comparing spontaneous and pharmacologically induced apoptosis in cancer cell lines by flow cytometry and WB.

The MCF7, LoVo, and PC3 cell lines were cultured in the presence or the absence of cisplatin for 24 hours, raising apoptotic events in drug-treated cells, as shown (Fig. 1A) by Annexin V immunostaining (15.25% ± 4.8 MCF7-, 15.3% ± 0.3 LoVo-, and 13.6% ± 2.3 PC3-positive cells, respectively) and M30 immunostaining (15.2% ± 8 MCF7-, 17.2% ± 8.3 LoVo-, and 12.6% ± 1.3 PC3-positive cells, respectively) and confirmed by WB, showing strong PARP cleavage in MCF7 (64.2% ± 0.5), in LoVo (55.2% ± 6.6), and in PC3 (34.4% ± 9.8).

To quantitatively compare M30 immunostaining at single-cell level, double fluorescence was done with TUNEL assay in the cell lines 36 hours after apoptosis induction with cisplatin. In the drug-treated cultures double staining discriminates the initial apoptotic phase before DNA fragmentation (M30-positive TUNEL-negative cells; Fig. 1B, top left quadrant in right plot) from the apoptotic cells, which are TUNEL-positive and M30-positive (Fig. 1B, top right quadrant in right plot). Again, a fraction of cells were M30-negative and TUNEL-positive (Fig. 1B, bottom right quadrant in right plot), representing the end phase of the process, when the cells become necrotic and the M30 epitope is lost. The data are consistent with previous observations that the exposure of the M30 neoepitope occurs at the initial phase of the apoptotic cascade, before the appearance of apoptotic features in the nuclei (8).

The M30 mAb was integrated into the CTC assay to specifically quantify apoptosis of spreading tumor cells. The integrated test was developed using the MCF7 cell line that was maintained in culture for 24 hours in the absence or the presence of paclitaxel, raising both early and late apoptotic events in drug-treated MCF7 cells, as shown by flow cytometry (Fig. 2A). Untreated controls and drug-conditioned cells were then spiked into whole blood samples, at numbers similar to those observed in vivo in cancer patients (200-1,000 cells/7.5 mL peripheral blood) to be finally processed by the CellSearch system.

The CTC assay revealed a significant increase of M30-positive CTC (71 ± 16, 11.6%) in the drug-conditioned MCF7-spiked samples compared with untreated controls (6 ± 4, 1%; P < 0.05; Fig. 2B). Because the values obtained by the CTC assay and flow cytometry are generated starting from different pools of events (only nucleated cells in the CTC assay versus both cells and nude nuclei in the flow cytometry), the fraction of apoptotic CTC by the integrated test differs from the flow cytometric results obtained before the spiking. Figure 2C shows an immunofluorescence image that stresses this point: M30 immunostaining clearly discriminates intact CTC, which are M30 negative, and early apoptotic CTC, which appeared M30 positive (Fig. 2C, left photo series); both these cells satisfy the morphologic features required to be defined as CTC (clear visible nucleus, cytoplasm/nuclear area overlying >50%, uniformly immunostaining of cytoplasm). Conversely, when the intact intermediate filament network is being progressively replaced by cytokeratin inclusions (Fig. 2C, central photo series), followed by chromatin condensation and loss, all characteristic of apoptosis (Fig. 2C, right photo series), the events, which still entered in flow cytometric analysis, cannot be assigned as CTC. Remarkably, the percentage of M30-positive CTC measured in the drug-conditioned MCF7-spiked samples (11.6%) closely resembles the quote of early apoptosis (14% in Fig. 1A) as determined by flow cytometry before the spiking into the whole blood.

Because CTC are considered “fragile” cells (12), the possibility that an event assigned as M30-positive CTC could be an artifact, due to apoptotic death occurring during the procedure of enrichment and immunostaining, was addressed. To this purpose, blood samples spiked with untreated MCF7 cells were treated with paclitaxel directly into the CellSave tube, 12 to 24 hours before the CellSearch processing. Although we acknowledge that lab-adapted MCF7 cells could be less fragile than CTC, we did not disclose in this case relevant changes of the M30-positive fraction (1.5 ± 1, 0.2%).

The M30-integrated CTC assay was fully developed in blood samples obtained from healthy donors and cancer patients. To use the test in follow-up studies, we generated an on-line staining procedure that is detailed in Supplementary Materials. The obvious advantage of this approach is that antibodies of interest were added and processed simultaneously with the CK-PE and CD45-APC antibodies, minimizing cell loss or disruption (13) during permeabilization and staining steps.

Compared with the MCF7-spiked samples, analysis of patient's samples showed that the integrated assay allows discrimination of the heterogeneous staining profile of the CTC: an irregular CK staining (Fig. 3, event 519) that is frequently observed in vivo and may closely resemble disruption of the filamentous network can be clearly distinguished from a M30-positive CTC (Fig. 3, event 524), prospectively minimizing discretionary interpretation of morphologic features.

To test whether apoptosis could be detected in CTC from patients with carcinoma, blood samples from 122 patients were tested (34 breast cancer, 59 colorectal cancer, 29 mRCC) before therapy. CTC were detected in 19 of 34 (56%) breast cancer patients; in 15 of these 19 patients (79%), M30-positive CTC were also detected. The number of CTC and M30-positive CTC ranged from 1 to 44 (median, 4) and 1 to 13 cells (median, 3), respectively. CTC were also detected in 26 of 59 (44%) colorectal cancer patients; in 24 of these 26 patients (92.3%), M30-positive CTC were detected. The number of CTC and M30-positive CTC ranged from 1 to 10 (median, 2) and 1 to 7 cells (median, 2), respectively. Finally, CTC were detected in 19 of 29 (66.5%) mRCC patients; in 17 of these 19 patients (89.5%), M30-positive CTC were detected. The number of CTC and M30-positive CTC ranged from 1 to 141 (median, 3) and 1 to 67 cells (median, 3), respectively.

The percentage of CTC-positive patients and total CTC numbers strictly resemble data previously reported in breast (5, 14, 15) and colorectal cancer patients (16). As summarized in Table 1, the presence of CTC and M30-positive CTC at diagnosis was not associated with any specific clinicopathologic feature in epithelial tumors, with remarkable exceptions. The presence of CTC at baseline seemed weakly associated with metastasis (P for trend = 0.075) in colorectal cancer (Table 1B). Moreover, the presence of CTC at baseline was associated with distant sites of metastasis (lung, mediastinal lymph node, liver, or bone, P for trend = 0.026) in mRCC; in this group, M30-positive CTC at baseline was weakly associated with clear cell tumor (P for trend = 0.08; Table 1C). In addition, M30-positive CTC were associated with elevated grading in breast cancer patients (Table 1A, P for trend = 0.018).

To investigate whether the integrated test may predict therapeutic response, CTC and M30-positive CTC were sequentially assessed in eight breast cancer patients. To this purpose, depending on their consensus to undergo multiple CTC tests, the patients were consecutively enrolled regardless of type or line of therapy and were monitored for a time in the range of 1 to 10 months. The results are summarized in Table 2.

Overall the number of total and M30-positive CTC decreased during treatment in six and increased in two of eight patients.

In the first group, in four cases, consisting of patients 43 [progressive disease (PD)], 55 (PD), 94 (PD), and 97 [stable disease/partial response (SD/PR)], the total CTC number switched from values greater than or equal to the threshold of poor prognosis (5 CTC/7.5 mL for MBC; ref. 5) to values less than the threshold at the end of observation time, indicating a pharmacodynamic response that was related to overall disease progression only in patient 97; the M30-positive CTC were very few in these patients and the relative percentage of M30-positive CTC fluctuated, creating a pattern of peaks and troughs that were difficult to evaluate. Only one or two cells were enumerated in patients 29 (PD) and 53 (PD), being too few CTC to discriminate treatment effect.

In the second group, patient 84 (PD) showed a major shedding of CTC that went from all alive to all dead (6 CTC at the time point 3); in patient 50 (SD/PR) the total CTC number increased over the follow-up period and essentially all cells were apoptotic (10 of 11 CTC at the time point 2). The switch from values <5 to values >5 CTC was related to overall disease progression only in patient 84.

In principle, both decreased total CTC numbers to value <5 CTC and increased fraction of apoptotic CTC may represent response-related markers. However, the fact that these are both rare events may preclude the possibility to accurately assess significant differences in the M30-negative/positive CTC numbers in any patient; only follow-up studies of adequate patient cohorts monitored for an appropriate time can address the predictive relevance of apoptotic CTC.

For this purpose the observed variations were expressed by a simpler parameter: the detected numbers of M30-negative and M30-positive CTC were separately plotted in relation to time, and the area under the curve (AUC) of longitudinal graphs was calculated (Fig. 4), following a procedure commonly adopted to evaluate cumulative changes of serologic tumor markers (17). The difference between live and apoptotic CTC concentration-time area was calculated in all patients according to the following formula:

Relative numbers were obtained in the following way:

• Positive ΔAUC value is the expression of extra live CTC over the follow-up period (e.g., patients 84 and 94; Fig. 4B);

• Negative ΔAUC value is the expression of extra apoptotic CTC over the follow-up period (e.g., patients 50 and 97; Fig. 4B);

• ΔAUC = 0 derives from balanced numbers of live and apoptotic CTC.

As shown in Table 2, positive ΔAUC value was associated with radiologic recurrence of disease (P for trend = 0.036), including cases where a switch under the threshold of 5 CTC was observed during therapy (patients 43, 55, and 94 in Table 2); conversely, negative ΔAUC was associated with SD/PR also in patient 50, whose total CTC number increased to value >5 CTC.

CTC can today be quantified in cancer patients, providing a robust predictor of treatment efficacy and survival throughout the continuum of the care (6, 18, 19). Data obtained in metastatic breast (5), colorectal (16, 20), and prostate cancer (4, 20, 21) by immunocytometric approach strongly support extending these observations to other solid tumor histotypes. Otherwise, it was recently published (22) that the threshold of 5 CTC/7.5 mL of peripheral blood, firstly set up by Cristofanilli (5), lacks prognostic significance in inflammatory metastatic breast cancer, where a value <5 CTC was not associated with better prognosis than ≥5 CTC. As suggested by the authors, the biological characterization of CTC should be addressed to discover unknown properties of these cells.

Furthermore, early clinical trials require validated pharmacodynamic biomarkers (hopefully blood-based) showing proof of mechanism (drug hits target) and/or proof of concept (tumor responds to drug). Apoptosis is often deregulated in cancer, and the induction of tumor cell death is a primary goal of many targeted therapies, directly or indirectly hinting molecular components of apoptosis regulatory pathways. Either way, apoptosis is regarded as a unique biomarker of treatment efficacy, and in the absence of tumor biopsies CTC may offer a surrogate sample.

In evaluating different components of the apoptotic cascade we focused on the M30 neoepitope for three reasons. First, the anti-M30 mAb defines an epitope of CK18 disclosing early phases of apoptosis. In this phase, despite caspase cleavage, CK18 is still retained in a filamentous network and tumor cells still satisfy the morphologic features of CTC. Second, a clinical correlate exists between serum levels of these CK fragment and tumor load and prognosis in breast (23) and colorectal cancer (24). Finally, both anti-M30 and Annexin V provided us consistent results in flow cytometry, altogether recommending including M30 in the new test.

The integrated assay was validated in 122 cancer patients before therapy at the first blood draw, disclosing that the M30 neoepitope is expressed on CTC at a very high frequency. Based on the presence of cytokeratin aggregates in their cytoplasm, apoptotic CTC in the late phase of the process were previously described in prostate (4, 25) and in lung cancer (26). M30-positive CTC were also documented in metastatic castration-resistant prostate cancer patients (27), but their enumeration was beyond the purposes of that study. To our knowledge, this is the first report on CellSearch technology applied to quantify apoptotic CTC. Larger studies are warranted to determine the prevalence of M30-positive CTC in vivo; however, although the detection of large numbers of these cells is counterintuitive, the data presented here are not surprising.

First, it is well known that higher grade and increased proliferation are often associated with tumor necrosis and apoptosis that may also be regarded as adverse biological features (28). Moreover, it was recently reported that both intact CTC and granular CTC (whose morphologic features strictly resemble early apoptotic cells) are inversely related to survival in castration-resistant prostate cancer (25). Indeed, in our breast cancer series, M30-positive CTC were associated with higher grading (P for trend = 0.018).

Second, apoptotic CTC were previously described in long-surviving breast cancer patients and have been considered a sign of occult niches of proliferating tumor cells, periodically shedding into the blood flow (2). Here we show that M30-positive CTC were detectable in the majority of cancer patients at different disease stages, possibly supporting that apoptotic dying is the mechanism because <0.1% of CTC released daily into the circulation successfully will settle in secondary organs (29).

Third, criticism was raised when, surprisingly, the median overall survival did not further decrease when >5 CTC were detected in 7.5 mL of blood (30). In challenging with rare cells, technical limits were evoked to account for a threshold of poor prognosis that is otherwise difficult to explain, for example increasing mistakes in assigning an event as a CTC when <5 CTC are detected or volume collected for the assay, but biological reasons cannot be excluded. Here we show that CTC frequently express a M30 neoepitope, which may offer a rationale to the argument that only the viable CTC cause the decreased chances for survival (30).

On the other hand, the high frequency of M30-positive CTC in radio-chemo-free patients raises doubts as to whether the integrated CTC test may be a useful tool to monitor drug-induced cell death. Our data show that CTC represent a heterogeneous cell population, among which both apoptotic cells and viable cells with possible metastatic potential exist; we show that the M30-integrated assay may be used to accurately quantify them during treatment. In this setting, preliminary data obtained by tracking a small case series of breast cancer patients indicate that changes in the M30-negative/positive CTC balance as expressed by ΔAUC may be used as a “dynamic” parameter disclosing an active disease, as documented by consistent radiologic findings. As in the case of the CTC absolute number, whether such an early assessment of response to treatment may result in an improved overall outcome or quality of life needs to be prospectively assessed in clinical trials designed to investigate this question. In ongoing clinical trials at IOV-IRCCS, we are now testing whether assaying the quote of apoptotic CTC provides a more sensitive marker for rating pharmacodynamic effects in patients compared with total CTC counts.

In conclusion, although apoptosis is thought to play a major role in anticancer therapy, the clinical relevance of induction of apoptosis remains uncertain, particularly in solid tumors. The proposed test might contribute to clarify this point, and possibly provides a secondary end point other than tumor size and tumor burden for evaluation of response in early phase trials.

No potential conflicts of interest were disclosed.

We thank Ms. Colette Case for editing the manuscript and P. Gallo for artwork preparation. The CellSearch platform was sponsored by the association “Il faro per lo IOV” of the ASCOM Padova.

Grant Support: Grants from the Italian Ministry of Health, (Oncology Program, Gender/Task 4 “Characterization of circulating tumor cells in breast and ovary cancer ”, R. Zamarchi); Banco Popolare di Verona (S. Indraccolo); Alleanza contro il cancro (ACC4; S. Indraccolo); Regione Veneto (Ricerca sanitaria finalizzata n.11/2008, A. Amadori); AIRC (A. Amadori).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.