Transcripts of circulating tumor cells detected by a breast cancer-specific platform correlate with clinical stage in bladder cancer patients

Abstract

Purpose

There is increasing interest in circulating tumor cells (CTCs) as a biomarker in bladder cancer (BC). In the present pilot study, we used a platform originally developed for detection of breast cancer CTCs to assess breast cancer-associated transcripts in CTCs of patients with different stages of BC. Moreover, transcripts specific for cancer stem cells and epithelial mesenchymal transition (EMT) were assessed.

Methods

We prospectively enrolled 83 BC patients and 29 controls. The AdnaTest^® system was used to enrich epithelial cells in peripheral blood and to detect breast cancer-associated, stem cell-specific or EMT-specific transcripts. Test results were correlated with clinical and pathological stage.

Results

A positive AdnaTest^® BreastCancerDetect was present in 6.9 % of controls (group A), 6.7, 15.0 and 18.7 % of patients with non-muscle-invasive BC (B), cM0 muscle-invasive BC (C) and metastatic BC (D) (p = 0.13). Stem cell-specific transcripts in group A, B, C and D were detected in 10.3, 10.0, 22.5 and 31.3 % (p = 0.03). EMT-associated transcripts were present in 3.5, 3.3, 15.0 and 18.7 % (p = 0.03). In group C, epithelial and stem-like transcripts correlated with tumor stage (p = 0.01 and 0.04).

Conclusions

CTCs with expression of breast cancer-associated transcripts are present in a considerable proportion of patients with BC. EMT and stem cell-specific transcripts of CTCs correlate with clinical stage and can be detected in patients negative for epithelial transcripts. The prognostic relevance of AdnaTest^® results in BC patients and potential implications for therapy decisions remain to be determined in prospective studies.

Introduction

In a significant proportion of patients with advanced tumors, circulating tumor cells (CTCs) can be detected in the peripheral blood. The presence of CTCs in patients with solid malignancies such as breast, prostate or colorectal cancer is associated with impaired prognosis (Pantel and Alix-Panabieres 2012). Therefore, CTCs have emerged as an important tool for risk stratification and treatment monitoring. In contrast to prostate and breast cancer, there is only sparse data on the presence, prognostic and predictive role of CTCs in bladder cancer (BC) patients (Msaouel and Koutsilieris 2011). Several studies using the CellSearch^® system demonstrated that depending on the clinical stage, 0–57 % of patients with invasive bladder cancer have circulating tumor cells in their blood (Rink et al. 2012; Flaig et al. 2011). The presence of these CTCs appears to be associated with impaired prognosis after cystectomy (Rink et al. 2012). Several groups have used polymerase chain reaction (PCR)-based approaches to detect epithelial transcripts in whole blood (Gazzaniga et al. 2005). However, data on detection of CTCs by alternative standardized CTC detection platforms are lacking. The AdnaTest^® enables an analysis of cancer-specific transcripts by PCR after immunomagnetic enrichment of epithelial cells. Therefore, the system provides broad opportunities for analysis of mRNA-based biomarkers. A predictive significance of mRNA expression profiles in CTCs has been shown recently by Antonarakis et al. who demonstrated that the presence of androgen receptor splice variant 7 detected by the AdnaTest^® was associated with resistance to enzalutamide and abiraterone (Antonarakis et al. 2014). So far, the AdnaTest^® system has been evaluated and commercialized for colon, prostate (Todenhofer et al. 2012) and breast cancer (Kasimir-Bauer et al. 2012).

Recently, several studies have shown a significant molecular overlap of breast and bladder cancer (Choi et al. 2014b; Damrauer et al. 2014). Moreover, the transcripts used in the AdnaTest^® BreastCancerDetect kit have been shown to be overexpressed also in BC (Hayashi et al. 2015a; Stojnev et al. 2014). It is therefore likely that breast cancer-associated transcripts (epithelial cell adhesion molecule (EPCAM), MUC1 and human epidermal growth factor receptor 2 (HER2)) can be detected in BC CTCs. To enrich BC CTCs, we used the AdnaTest^® EMT2/SC kit originally developed for analysis of CTCs in EMT or displaying tumor stemness phenotype (Phosphatidylinositol 3-kinase α (PI3Kα), twist-related protein 1 (TWIST1), proteinkinase B (AKT2) and aldehyde dehydrogenase isoform 1 (ALDH1)). The aim of the present study was to evaluate the use of the AdnaTest^® in patients with different stages of BC and to correlate results with clinical and pathologic parameters.

Patients and methods

Patients

We prospectively enrolled four different consecutive groups of patients: (A) 26 patients undergoing transurethral resection of the prostate or ureterorenoscopy without any evidence of malignancy (Control group), (B) 30 patients with non-muscle-invasive BC undergoing urethrocystoscopy (NMIBC group), (C) 40 patients undergoing radical cystectomy for muscle-invasive BC (MIBC group) and (D) 16 patients with metastatic BC (M1 group). Written informed consent was obtained from all patients prior to inclusion in the study. The study was approved by the ethics committee of the University of Tuebingen (Approval Number 295/2015BO2).

CTC detection and criteria for positive BreastCancerDetect test

7.5 ml of blood was collected prior to surgery or, in case of patients with metastatic disease, before or during chemotherapy. Blood samples were collected in AdnaCollect tubes (Qiagen Hannover, Langenhagen, Germany) allowing storage for 24 h at 4 °C. Within 24 h after collection, analysis of CTCs using the AdnaTest^® EMT2/SC (Qiagen, Langenhagen, Germany) platform was performed according to the manufacturer’s protocol. The platform enables an enrichment of cells displaying EPCAM, epidermal growth factor receptor (EGFR) or HER2 by an antibody-mix (anti-EPCAM, anti-HER2, anti-EGFR) linked to magnetic particles. mRNA is isolated using Oligo-dT-coupled magnetic beads. The BreastCancerDetect system combines reverse transcription of the isolated mRNA into cDNA and a multiplex PCR performed for the analysis of tumor-associated gene expression (HER2, MUC1, EPCAM). Visualization and quantification of DNA was performed using an Agilent Bioanalyzer 2100 (Agilent, Böblingen, Germany). The AdnaTest^® result was considered to be positive if the fragment concentration after PCR of at least one of the markers was above 0.1 ng/µl.

Detection of CTC with EMT or stem cell properties

Using the cDNA resulting in the AdnaTest^® EMT2/SC procedure, the overexpression of EMT-related markers (PI3Kα, TWIST1 and AKT2) was analyzed by PCR as per manufacturer’s instruction as well as the detection of the tumor stem cell-related overexpression of ALDH1. The resulting fragment concentrations resulting from PCR were quantified using the Agilent Bioanalyzer. The results were considered positive if the cut-off values as per the manufacturer’s instructions of one of the included markers were exceeded (namely 0.2 ng/µl for AKT2, 0.15 ng/µl for TWIST1 and ALDH1, 0.25 ng/µl for PI3Kα).

Association with clinical characteristics

Dichotomized results of the respective marker kits (BreastCancerDetect, EMT, StemCellDetect) were correlated with the assigned groups (Controls, NMIBC, MIBC, BC cM1) using Cochrane armitage test for trend. In patients with clinically localized MIBC undergoing cystectomy, results of the respective marker kits were compared to T-stage and N-stage using Wilcoxon-Mann–Whitney test.

Results

Patients’ characteristics

Patients’ characteristics of Control, NMIBC, MIBC and M1 groups are summarized in Tables 1, 2, 3 and 4. In the group of patients with cM1 BC, 62.5 % of patients underwent blood collection before having received any chemotherapy.

Table 1.

Patients’ characteristics of control group

General
Age, years (range)	60 (33–85)
Gender, male/female	19/10
Disease conditions
Stone disease, n (%)	16 (55.2 %)
Benign prostate syndrome, n (%)	13 (44.8 %)

Table 2.

Patients’ characteristics of patients undergoing TUR-BT (group B)

General
Age, years (range)	73 (39–89)
Gender, male/female	28/2
Tumor stage
pTa, n (%)	24 (80.0 %)
pT1, n (%)	4 (13.3 %)
Grading
Low grade, n (%)	19 (63.3 %)
High grade, n (%)	9 (30.0 %)
Carcinoma in situ (Cis)
Pure Cis, n (%)	2 (6.7 %)
Concomitant Cis, n (%)	2 (6.7 %)

Table 3.

Patients’ characteristics of patients undergoing radical cystectomy (group C)

General
Age, years (range)	68 (44–92)
Gender, male/female	29/11
Tumor stage
pT2, n (%)	15 (37.5 %)
pT3, n (%)	19 (47.5 %)
pT4, n (%)	6 (15.0 %)
Grading
Low grade, n (%)	0
High grade, n (%)	40 (100.0 %)
CIS
Concomitant, n (%)	16 (40 %)
Lymph node status
N+, n (%)	14 (35.0 %)
N0, n (%)	26 (65.0 %)
Distant metastases
M0, n (%)	40 (100 %)
M1, n (%), n (%)	–
Surgical resection margin status
R0, n (%)	33 (82.5 %)
R1, n (%)	4 (10.0 %)
Rx, n (%)	3 (7.5 %)
Incidental prostate cancer
n (%)	10 (34.5 %)

Table 4.

Patients’ characteristics of patients with metastatic urothelial carcinoma (group D)

General
Age, years (range)	64 (54–81)
Gender, male/female	14/2
Metastases location
Lymph nodes, n (%)	9 (56.3 %)
Visceral metastases, n (%)	11 (68.8 %)
Bone, n (%)	6 (37.5 %)
Timing of blood collection
No chemotherapy before blood collection, n (%)	10 (62.5 %)
Chemotherapy before blood collection, n (%)	6 (37.5 %)

Results of AdnaTest^®BreastCancerDetect

The association of test results with clinical stage is shown in Fig. 1a (p = 0.13). In the subgroup of M1 patients, who had no previous chemotherapy (n = 10), 30.0 % had this test positive whereas none of the patients with metastatic BC, who underwent CTC testing during chemotherapy, showed a positive AdnaTest^® BreastCancerDetect.

Fig. 1.

Results of AdnaTest^® BreastCancerDetect (a, b), StemCellDetect (c, d), EMT (e, f) and the combination of all three tests (g, h) in patients with different stages of bladder cancer (BC). The combination of the tests was considered positive if at least one of the tests included (BreastCancerDetect, StemCellDetect or EMT) was positive. b, d, f and h show results of tests in patients with muscle-invasive BC (MIBC) according to T-stage and lymph node involvement

In the group of patients with cM0 MIBC, patients with locally advanced disease were more likely to be positive (p = 0.01) whereas no significant association with nodal involvement was observed (p = 0.41) (Fig. 1b).

Results of single transcripts of the AdnaTest^® BreastCancerDetect (HER2, MUC1, EPCAM) are summarized in Table 5.

Table 5.

Detection rates of different transcripts in different stages of bladder cancer

Transcript	Controls (group A) n = 29	NMIBC (B) n = 30	MIBC (C) n = 40	cM1 (D) n = 16	p value groups A versus B versus C versus Da	cM1 D* without previous chemotherapy, n = 10	p value groups A versus B versus C versus D*b
HER2	1 (3.5 %)	2 (6.7 %)	5 (12.5 %)	3 (18.8 %)	0.06	3 (30.0 %)	0.02
MUC1	1 (3.5 %)	0 (0.0 %)	4 (10.0 %)	3 (18.8 %)	0.02	3 (30.0 %)	0.01
EPCAM	1 (3.5 %)	0 (0.0 %)	1 (2.5 %)	0 (0.0 %)	0.59	0 (0.0 %)	0.68
TWIST	1 (3.5 %)	0 (0.0 %)	0 (0.0 %)	2 (12.5 %)	0.27	2 (20.0 %)	0.17
AKT2	1 (3.5 %)	0 (0.0 %)	2 (5.0 %)	2 (12.5 %)	0.15	1 (10.0 %)	0.30
PI3Kα	1 (3.5 %)	1 (3.3 %)	5 (12.5 %)	2 (12.5 %)	0.11	1 (10.0 %)	0.13
ALDH1	3 (10.3 %)	3 (10 %)	9 (22.5 %)	5 (31.3 %)	0.03	5 (50.0 %)	0.007

Detection rates of transcripts from AdnaTest^® BreastCancerDetect (Human epidermal growth factor receptor 2 (HER2), MUC1 and Epithelial cell adhesion molecule (EPCAM)), AdnaTest^® EMT (TWIST, AKT2, Phosphatidylinositide 3-kinase alpha (PI3Kα)) and AdnaTest^® StemCellDetect (Acetaldehyde dehydrogenase 1 (ALDH1))

^a p value from Cochrane Armitage test for trend

^bCochrane Armitage test for trend group A versus B versus C versus D* (patients with previous chemotherapy excluded)

Results of AdnaTest^® StemCellDetect

Transcripts for ALDH1 significantly correlated with clinical stage (p = 0.03) (Fig. 1c). In the metastatic group, patients without previous chemotherapy were positive in 50.0 % of cases whereas versus 0.0 % of the group of metastatic patients with ongoing chemotherapy.

In non-metastatic MIBC, presence of ALDH1 transcripts significantly correlated with pT-stage (p = 0.04) but not with lymph node involvement (p = 0.90) (Fig. 1d).

Results of AdnaTest^®EMT

Results of the AdnaTest^® EMT significantly correlated with clinical stage (p = 0.03) (Fig. 1e). In patients with cM1 BC without previous chemotherapy, two patients (20.0 %) had a positive test versus 16.7 % of patients with ongoing chemotherapy.

Results of the AdnaTest^® EMT did not correlate with T-stage (p = 0.49) or N-stage (p = 0.92) (Fig. 1f).

Table 5 shows results of single transcripts.

Combined test results

The proportions of patients showing at least one of the tests AdnaTest^® BreastCancerDetect, AdnaTest^® StemCellDetect or AdnaTest^® EMT positive correlated with clinical stage (p = 0.04) (Fig. 1g). In cM1 BC patients without previous chemotherapy, five patients (50.0 %) were positive versus one patient (16.7 %) in the group of patients with ongoing chemotherapy.

In the group of patients with cM0 MIBC, results were positive in 36.0 % of patients with locally advanced disease (pT > 2) versus 20.0 % in patients with organ-confined disease (p = 0.27). Patients with lymph node-positive disease were positive in 35.7 % of cases versus 26.9 % in patients with nodal-negative disease (p = 0.56) (Fig. 1h).

Discussion

The detection of circulating tumor cells has become an important tool in the monitoring of patients with advanced solid malignancies. The number of studies addressing CTCs in BC is limited (Msaouel and Koutsilieris 2011). The majority of studies on CTCs in BC use the CellSearch^® platform (Rink et al. 2011, 2012; Flaig et al. 2011). Various studies used PCR to detect epithelial, potentially tumor-associated antigens in whole blood (Ribal et al. 2006; Gazzaniga et al. 2005). However, these methods did not include an enrichment of CTCs leading to detection of significant background expression of peripheral blood mononuclear cells (PBMCs). The present pilot study therefore aimed to assess the value-standardized PCR-based assay for detection of immunomagnetically enriched CTCs in BC. The platform used in our study has been developed and validated for patients with advanced breast cancer (Muller et al. 2012).

Our results show that HER2 and MUC1 expression are frequent events in circulating epithelial cells, presumably bona fide CTCs, in patients with advanced BC. In patients with metastatic BC without previous chemotherapy, these transcripts can be found in circulating epithelial cells of 30 % of patients. In non-metastatic MIBC, these transcripts significantly correlated with T-stage. Rink et al. (2012) already assessed the clinical and prognostic value of HER2-positive circulating tumor cells in a cohort of 100 patients undergoing radical cystectomy. Using the CellSearch^® platform, they observed that 23 % of patients had CTCs detectable at time of radical cystectomy. Of these patients, three (14 %) showed a strong expression of HER2 on CTCs. In contrast to our study, they did not observe a association of presence of CTCs or expression of HER2 with clinicopathologic parameters. It is noteworthy that in contrast to our study they used FISH for HER2 gene amplification status. Previous studies have shown that HER2 status is strongly depending on the method of detection (Tvrdik et al. 2012).

There are ongoing controversies on a potential stem cell-like phenotype of circulating tumor cells (Tinhofer et al. 2014). Moreover, CTCs have been shown to undergo EMT promoting migration and invasion of tumor cells (Liu et al. 2015). These features may prevent systems that are based on detection of epithelial markers from detection of all subpopulations of CTCs. In the current study, we used the AdnaTest^® Stem cell and EMT kits to assess the presence of transcripts associated with features of stem cell-like cells and EMT.

Interestingly, the majority of CTCs detected were positive for EMT and/or stem cell properties but not for EPCAM as an epithelial marker, indicating that a switch to EMT and stem cell phenotype might be a common phenomenon in BC. Prior studies in breast cancer have also shown that this system detects features of stem cells and EMT in a significant proportion of patients with breast and colorectal cancer (Kasimir-Bauer et al. 2012; Ning et al. 2015). In our study, ALDH1 as a marker of stem cells was more frequently detected than the epithelial markers HER2, MUC1 or EPCAM. This may indicate that ALDH1-positive subpopulations of CTCs exist that cannot be detected by PCR for epithelial markers. Interestingly, the presence of transcripts not only correlated with clinical stage but also with T-stage, indicating a higher frequency of these CTC subtypes in patients with locally advanced tumors. It has not been clarified yet whether these ALDH1-positive cells show a different biologic behavior compared to other subtypes of CTCs. However, the fact that more than 10 % of healthy control patients also showed transcripts of ALDH1 indicates that these transcripts are partially derived from circulating benign cells. Therefore, the presence of these transcripts has to be interpreted with caution.

The presence of EMT-related transcripts during chemotherapy in the absence of epithelial or stem cell-associated transcripts may be in accordance with prior studies showing that cells with an EMT-like phenotype show increased chemoresistance in solid tumors (Fischer et al. 2015). To address this issue, a prospective trial correlating presence and characteristics of CTCs during chemotherapy with clinical response in patients with urothelial carcinoma is urgently needed.

The assessment of multiple different transcripts by the AdnaTest^® may improve non-invasive molecular characterization of BC. Non-invasive methods for assessing changes on RNA and DNA level (e.g., cell free DNA analysis) in metastatic tumors have drawn considerable attention in the last years.

Recently, several groups have introduced the concept of molecular subtypes of MIBC (Choi et al. 2014a; Damrauer et al. 2014) differing in prognosis and responsiveness to neoadjuvant chemotherapy. It remains to be determined whether a transcript profile derived from analysis of CTCs with the AdnaTest^® indeed may enable a differentiation of select molecular subtypes. Previous studies in breast cancer indicate that ALDH1 and HER2 are differentially expressed in distinct molecular subtypes of breast cancer (Ricardo et al. 2011). To validate the correlation of these markers with subtypes of BC, RNA gene expression profiling of primary tumors (in case of non-metastatic BC) and metastatic lesions (in case of metastatic BC) would be required, which was beyond the scope of the present study.

The transcripts assessed in this study may also provide important information on molecular alterations of the tumor amendable to targeted therapeutics. HER2 overexpression has been shown to be a frequent event in BC and is associated with response to HER2 targeting drugs in preclinical studies (Hayashi et al. 2015b). A recent phase II study showed no clear benefit when adding HER2 targeting drugs to chemotherapy but was limited by a low number of HER2-positive patients (Oudard et al. 2015). The platform used in our study might complement other techniques helping to identify patients that may benefit from HER2 inhibition in the framework of clinical trials (Hussain et al. 2007). The knowledge about EMT and tumor stemness-related molecular profiles in BC may also have the potential to trigger alternative targeted therapeutic strategies.

In BC, CTCs have been shown to correlate with prognosis in localized MIBC, whereas no data are available on prognostic relevance of CTCs in metastatic BC (Rink et al. 2012). For the present pilot study, no follow-up data of the patients had yet been collected. Previous studies have shown that the method used for CTC detection has a significant impact on the prognostic relevance (Muller et al. 2012). In breast cancer, Cellsearch^® as well as the AdnaTest^® BreastCancerDetect and the AdnaTest^® EMT/SC were able to show prognostic value in metastatic disease (Muller et al. 2012; Andreopoulou et al. 2012; Tewes et al. 2009; Aktas et al. 2009). However, the prognostic value of the AdnaTest^® in BC has still to be determined prospectively in different subgroups of BC patients.

CTC platforms that solely enable enumeration of circulating tumor cells cannot exploit the full potential of these cells. The AdnaTest^® provides important advantages by allowing to assess information on gene expression of CTCs. However, the system cannot provide information on protein expression and genomic changes that might correlate with therapy response or prognosis (Alva et al. 2015).

Methods enabling multiple downstream analyses and in vitro culture of the cells are urgently needed as they may provide an important step to personalized medicine (Yu et al. 2014). Ideally, these methods should be able to enrich CTCs regardless of an expression of epithelial surface antigens (Karabacak et al. 2014).

The present study has inherent limitations: First, it is a pilot study with a limited number of patients included (especially in the group of metastatic patients). Second, no results of CellSearch^® were available for the patients. A general limitation of the AdnaTest^® is that the cell morphology of CTCs cannot be evaluated. Therefore, it cannot be guaranteed that the transcripts detected by this platform are ultimately derived from tumor cells. To date, no marker has been discovered both on protein and mRNA level that has been shown to be specific for BC cells. Therefore, it is currently not possible to provide proof by PCR or immunocytochemical methods that these cells are indeed BC cells. Moreover, current CTC platforms including the AdnaTest do not allow to differentiate CTCs with metastatic potential from tumor cells that will undergo dormancy or apoptosis. Both in patients with non-metastatic BC and metastatic BC, the rate of CTC-positive patients using the AdnaTest^® BreastCancerDetect is lower compared to other studies using the CellSearch^® system (Rink et al. 2012; Flaig et al. 2011). However, when combining the results of the AdnaTest^® BreastCancerDetect and the AdnaTest^® StemCellDetect and EMT, rates of positive patients are similar to prior studies. A combined use of the AdnaTest^® BreastCancerDetect and AdnaTest^® StemCell Detect/EMT might therefore not only provide additional information on gene expression profiles of CTCs but may also improve the sensitivity of the system. Although previous studies have shown low proportions of CTC-positive patients with localized prostate cancer, we cannot rule out that in patients with MIBC and incidental prostate cancer at time of cystectomy CTCs with prostatic origin were detected.

The present pilot study focused on the analysis of CTCs in different stages of BC. We did not assess the influence of surgical therapy on CTCs. To date, only sparse data are available on how CTCs and their numbers change during and after transurethral resection of a bladder tumor (TUR-BT) and cystectomy. Recent evidence exists showing that CTCs can be found in central vein blood during TUR-BT (Engilbertsson et al. 2015). Whether these CTCs contribute to dissemination of disease remains to be elucidated. Moreover, the influence of chemotherapy on circulating tumor cells in patients with BC and a potential prediction of therapy response by monitoring of CTCs is an important issue, which should be addressed by further studies.

Conclusions

The present study provides evidence that a platform originally designed for detection of circulating tumor cells in breast cancer patients and optimized for the enrichment of such cells detects tumor-associated transcripts of epithelial, EMT-like and tumor stem cell-like cells in a considerable proportion of patients with muscle-invasive and metastatic bladder cancer. The detection of stem cell- and EMT-associated transcripts in patients with missing epithelial transcripts indicates the presence of a subgroup of CTCs in these patients that might be missed by methods using detection of epithelial markers. Future studies should focus on the use of this system for correlating the presence of specific transcripts with molecular subtypes and response to treatment.

Compliance with ethical standards

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. The study was approved by the ethics committee of the University of Tuebingen (Approval Number 295/2015BO2).

Conflict of interest

DS, NF and SH are employees of Qiagen Hannover.