14716barriere.qxd:protipiar

ANTICANCER RESEARCH 32: 3363-3370 (2012) Mesenchymal Characterization: Alternative to Simple
CTC Detection in Two Clinical Trials
GUISLAINE BARRIERE1, ALAIN RIOUALLON2, JOËL RENAUDIE2, 1Astralab Clinical Laboratory, Limoges, France; 2Department of Gynecology and Surgery. Clinique du Colombier, Limoges, France Abstract. Background: Breast cancer is one of the most
enriched for CTCs by EpCAM-coated immunomagnetic common malignancies in women. Approximately 25% of beads. CTCs are then stained for cytokeratins (CK8, CK18 patients with early-stage disease will develop metastatic and CK19), for CD45 to eliminate white cells, and with 4’,6- recurrence. Two clinical trials were undertaken in order to Diamidino-2-Phenylindole (DAPI) nuclear counterstain for detect circulating tumor cells (CTCs) in primary breast cancer. essessment of their viability. Finally, they are enumerated by Patients and Methods: Four-hundred patients with early breast at least two pathologists. This method was cleared by the US cancer were enrolled in the trial. After enrichment from their Food and Drug Administration (FDA) but not endorsed by the peripheral blood, their CTCs were characterized by gene American Society of Clinical Oncology (ASCO). Due to the expression of cancer cell markers. Results: CTCs had a use of a sole antibody, CTCs can escape the enrichment and predominant epithelial phenotype in 8.75% of patients and de- detection phases, as EpCAM expression is decreased in differentiated characteristics (mesenchymal, stem phenotypes mesenchymal cells (3). In one-third of patients, CTCs are alone or both) in 37.6%. Conclusion: Tumor epithelial cells undetected by conventional methods (4). This impairment undergoing epithelial–mesenchymal transition give rise to prompted the development of methods where CTCs are cells with mesenchymal aggressive phenotype. Detection of qualified by molecular biology techniques, revealing mesenchymal and cancer stem cells, which are tumor- expression of specific genes (5). AdnaGen technology is one initiating cells, is more relevant than simple counting of CTCs of those techniques (6). It is based on a combined mixture of to assess their presence in the blood of patients with breast antibodies (EpCAM and two epitopes of MUC1) to trap cancer. This study will be the basis for future evaluation of the CTCs which are defined by expression of genes involved in outcome of the disease and the prognostic value of early- epithelial–mesenchymal transition (EMT). Most CTC detection trials were dedicated to metastatic breast cancer (7).
CTCs were detected, depending on the stage of the disease Detection of circulating tumor cells (CTCs) has been recently and the undertaken methodology, in 8% to 95% breast cancer developed and can be considered as a prognostic tool (1). A patients (8-12). Due to the range of results, conclusive limited number of markers are currently used for the isolation relevance is still pending. However, these studies have shown (cell surface antigens) or detection (antigens or mRNA) of the prognostic potential of CTCs in primary and metastatic CTCs. One such marker is the epithelial cell adhesion breast cancer (13). Our preliminary results on primary breast molecule (EpCAM) (2); others are more cancer-specific, such cancer and CTCs arising from EMT were presented in Prague as human epidural growth factor receptor 2 (HER2) and (June 13-17, 2011) at the TATAA Biocenter qPCR mucin 1 (MUC1) for breast carcinoma. CTCs are generally symposium (14). These results were confirmed by Kasimir- counted by using the CellSearch system (Veridex, Warren, Bauer et al. (15). Today, the term CTCs encompasses all New Jersey, USA). In this technique, blood sample is types of cells which are considered as foreign entities in theblood, exhibiting cancerous characteristics. This term doesnot describe the diversity of CTC subpopulations. Amongthem, cancer stem and mesenchymal cells have to be taken Correspondence to: Guislaine Barrière, Astralab Clinical into account. Their features are invasiveness linked to Laboratory, 7-11 Avenue de Lattre de Tassigny, 87000 Limoges, increased cell motility and their capability to avoid apoptosis, France. Tel: +33 555302930, e-mail: [email protected] anoikis, and general immune defense (16-19). Upon hypoxia, Key Words: Circulating tumor cells, de-differentiated circulating epithelial cancer cells can modify their phenotype. This tumor cells, epithelial–mesenchymal transition, breast cancer.
phenomenon depends on pleitropic cytokines such as ANTICANCER RESEARCH 32: 3363-3370 (2012) transforming growth factor β (TGF-β) (20). The latter is were used: AdnaTest Breast Select and Detect, AdnaTest EMT- induced by hypoxia-inducible factor 1α (HIF1α) (21). The 1/Stem Cell Select and Detect, according to the suppliers signaling pathway induces loss of the epithelial polarity of instructions (AdnaGen AG, Langenhagen, Germany). Steps arebriefly described. For semi-quantitative reverse transcription and cells, which undergo cytoskeletal remodeling. Expression of polymerase chain reaction (RT-PCR), Sensiscript and HotStarTaq some proteins which are involved in adhesion structures are from Qiagen GmbH (Hilden, Germany) were used and the inhibited. Numerous experiments on this transition have housekeeping gene was β actin. Thermal profiles are those demonstrated a decrease of E-cadherin and an increase of N- recommended by the supplier. Visualization of PCR fragments was cadherin and vimentin (22-24). Thus, cells acquire a carried out with a bioanalyser (2100 Bioanalyser; DNA 1000 migratory phenotype and exhibit mesenchymal features.
LabChip; Agilent Technologies; Santa Clara; CA). Peaks were Epithelial cells and mesenchymal cells can exist considered to be positive when concentrations were ≥0.15 ng/μl(according to AdnaGen indications). simultaneously, both types being CTCs. Hence, a clinical trial Protocol 1: Tumor cell enrichment was realized by using was developed to establish the percentage of patients carrying AdnaTest Breast Select. The immunomagnetic beads were coated these cells at the time of early diagnosis of breast cancer and with three antibodies, one to EpCAM and two others to MUC1.
before and after surgical eradication of the primary tumor. In Then, tumor-associated mRNA expressions were analyzed by this protocol, 400 patients were included and CTCs were AdnaTest Breast Detect kit. The following transcripts were separated evidenced by three markers: HER2, MUC1, epithelial by capillary electrophoresis: HER2, MUC1 and GA733-2 with base glycoprotein 40 (GA733-2). Moreover, we analyzed 130 pairs of size 270, 293 and 395 bp, respectively. Protocol 2: AdnaTest EMT-1/Stem Cell Select avoided an excess clinical samples belonging to this cohort to detect contaminating leukocytes by a special washing buffer procedure.
mesenchymal and stem cells by using more specific EMT Expressions of tumor-associated mRNA were depicted by AdnaTest markers: transcription factor TWIST1, protein kinase B EMT-1/Stem Cell Detect kit. The following transcripts: TWIST1, (AKT2), phosphatidylinositol 3-kinase A (PI3KA) and AKT2, PI3KA and ALDH1 (203 bp, 306 bp, 595 bp, 165 bp, respectively) were revealed by capillary electrophoresis. To thesupplied method, we added CD44 and the polycomb group protein Materials and Methods
BMI1 (BMI1) as subsidiary markers of stemness. For each one, theRT sample, previously obtained, was amplified by singleplex PCR.
Sequences of primers were CD44 forward: GCCCAATGCCT Patients. Four-hundred patients (T1/T3, N–/N+, M0) were enrolled TTGATGGACC and reverse: GCAGGGATTCTGTCTGTGCTG; and between 2009-2011. Patients were selected provided they fulfilled BMI1 forward: CATTGTCTTTTCCGCCCGC and reverse: CAAAG the following criteria: age 40-75 years; breast cancer diagnosis CACACACATCAGGTGGG. The thermal profile used for CD44 PCR confirmed by a pathologist’s analysis of the primary tumor; absence was as follows: after 15 min denaturation at 95˚C, 33 cycles of PCR of bone, visceral, cerebral metastasis (controlateral breast were carried out by denaturation at 94˚C for 30 s, annealing/extension mammography, liver ultrasonography and entire body-bone at 59˚C for 30 s, and elongation for 72˚C for 30 s. Termination of the scanning). Axillary lymph node invasion was assessed. These PCR reaction was subsequently carried out at 72˚C for 5 min patients were treated by tumorectomy or mammectomy; axillary followed by storage of the sample at 10˚C. The procedure for BMI1 clearance was added when required. Characteristics of the primary was as follows: after 15 min denaturation at 95˚C, 36 cycles of PCR tumor at the time of diagnosis are shown in Table I. All blood were carried out by denaturation at 94˚C for 30 s, annealing/extension samples were obtained after informed consent. The protocol was at 59.7˚C for 30 s, and elongation for 72˚C for 30 s. Termination of conducted at the Department of Gynecological Surgery (Private the PCR reaction was subsequently carried out at 72˚C for 5 min Hospital Clinique le Colombier) and Astralab Laboratory followed by storage of the sample at 10˚C. The primers generate (Department of Specialized Clinical Analyses) in Limoges France.
fragments of the following sizes: (CD44, 257 bp and BMI1, 132 bp).
The study was performed with approval of an appropriate Local CD44 and BMI1 expressions were considered positive when the Ethics Committee (Comité de protection des personnes Sud-Ouest et transcript concentration was above 0.50 ng/μl. Blood collected from Outre-mer IV. France) and was in compliance with the Helsinki 20 healthy donors was investigated to determine this cut-off value. Declaration. This cohort enabled us to realize two trials: the totalityof patients (400) was incorporated in a first protocol and the final Statistical analysis. Statistical analyses were performed using 130 patients were also included in a second protocol. XLStat2011 software (Addinsoft; Paris; France). The Chi-squaretest was used to establish a relationship between CTC detection and Sampling of biological material. Blood samples of 7 ml were collected tumor characteristics and lymphatic invasion. for cellular enrichment with AdnaCollect tubes (AdnaGen AG,Langenhagen, Germany) before surgery and at least 3 weeks aftersurgery. Blood collections were performed before any drug therapy.
Two tubes of 7 ml were necessary when patients were enrolled intothe second trial. Samples were stored, shipped in the dark at 4-8˚C and Protocol 1: Patients’ characteristics. A total of 400 patients were analyzed within 24 h. Blood collection tubes contain EDTA and (aged from 40 to 75 years, average 66 years of age) were a chemical agent to prevent illegitimate RNA expression. included in the trial. They were studied at the time of newly- CTC selection and detection. Selection and detection of CTCs were diagnosed breast cancer and patients were enrolled after performed as described elsewhere (5). The following AdnaGen kits elimination of bone visceral and cerebral metastasis, whether Table I. Tumor characteristics for 400 patients included in protocol 1. CTC-positivity is indicated for each category of breast cancer and before orafter surgery. they had axillary lymphatic node invasion or not. Among the patients failed to clear cancer cells from their blood despite cohort, 87% of patients had T1 or T2 tumors. Most of the suppression of the primary tumor. Out of the 20 patients patients had no positive axillary lymph nodes (67%). The most found to be CTC-positive after surgery, 12 were cases of new common molecular characteristic phenotype, assessed on detection. These data are presented in Table I. There was a estrogen receptors (ER), progesterone receptors (PR) and discrepancy between the HER2 phenotype of CTCs and that HER2 expression of the primary tumor, was hormone receptor of the primary tumor. Effectively, out of four patients with positivity. Luminal A type breast cancer was observed in 77%, HER2-positive primary tumor, only two had CTCs bearing luminal B in 7%, triple negative (TN) in 8% and HER2 was the HER2 receptor. In seven cases of HER2-negative tumor, overexpressed in 4% of breast cancer cases (Table I). the corresponding CTCs were positive for HER2. This resultshowed a high incidence of conversion for the EGF receptor.
Incidence of CTCs. In 35 patients, CTCs were detected at CTC positivity was distributed as follows: 25 luminal A, 3 least once (either before and/or after surgery). Thus 8.75% of luminal B, 1 HER2-overexpressing and 2 TN. Four breast the total cohort expressed at least one marker. Out of the 400 tumors were not molecularly characterized. CTCs are often patients, 359 were submitted to the two series of analyses detected in luminal phenotypes. However, the imbalance before and after surgery. The analysis of markers for the 35 between the different types of tumors in the studied cohort is CTC-positive patients gave the following positivity rates: not sufficient to support this assertion. Statistical studies GA733-2, 42%; MUC1, 65%; HER2, 26%. In eight patients, indicated that CTC detection is independent of the primary CTCs were detected both pre- and post-operatively. Out of tumor characteristics. This non-correlation between CTCs and the 23 patients found to be CTC-positive before surgery, eight parameters of the tumor was demonstrated by p-values, which remained positive. Thus, 23% of the initially CTC-positive ANTICANCER RESEARCH 32: 3363-3370 (2012) Table II. Tumor characteristics for 130 patients included in protocol 2. Positivity of mesenchymal or/and stem cells is indicated for each categoryof breast cancer before surgery. EMT: Epithelial–mesenchymal transition; ddCTC: de-differentiated circulating tumor cell; ND: not determined.
Protocol 2: Blood samples from 130 patients, submitted to applied a transcript cut-off value of 0.50 ng/μl (specificity of the CTC analyses described above, were tested for CTC the cut-off is more than 90%, as confirmed in 20 healthy presence by using EMT and stemness markers. All patients donor samples). Thus, ddCTCs detected by mesenchymal were studied at the time of newly-diagnosed breast cancer.
markers and/or ALDHI were 63% and 40% positive for BMI1 Table II describes clinical and pathological characteristics of and CD44, respectively. Among 84 N– patients, 32 had the tumors. We distinguish these cells by calling them de- ddCTCs, and of 37 N+ patients, only 11 had ddCTCs.
differentiated CTCs (ddCTCs), as markers used for their Statistical studies indicated that ddCTC detection is an detection are implicated in their mesenchymal or stemness independent factor of the other pronostic factors based on the pathologist’s analysis. No p-value was less than 0.05. In aprevious report, we noticed a correlation between the EMT-stemness detection. Among 130 patients, 49 had presence of ddCTCs and lymph node positivity with p<0.05 ddCTCs expressing at least one EMT marker or ALDH1 (5). This value is at the limit of significance, indicating a mRNA. Thus 37.6% of samples were positive for ddCTCs trend, and this correlation disappeared as the number of (Table II). Moreover among this population, 36.7%, 34.6% and 28.6% exhibited EMT cell, stem cell or both EMT andstem cell markers respectively. The analysis of each marker Discussion
for the 49 ddCTC-positive patients gave the followingpositivity rates: TWIST1, 8.2%; PI3KΑ, 61.2%; AKT2, In the first protocol, detection of CTCs led to discovery of a 10.2%; and ALDH1, 63.2%. When analyses detected at least small rate of CTC positivity at the time of early-breast one EMT marker or ALDH1, the sample was further cancer diagnosis, when cancer cells were detected on the examined for BMI1 and CD44. For these two markers, we basis of GA733-2, HER2, MUC1 marker expression (8.75%). These results seem low when compared to those implicated in regulating the EMT status, such as TWIST1.
previously published, which could be due to the use of Thus these markers are relevant for the detection of different technologies. However, Molloy et al. and Banys et mesenchymal cells. By using these, we detected 49 patients al. who conducted trials comparable to ours, described CTC out of 130 who had ddCTCs in the blood at the time of early positivity with 7% and 12% rates, respectively for their diagnosis of primary breast cancer. These ddCTCs can be patients (8, 9). To explain such a low percentage of patients distinguished into three phenotypes: mesenchymal, stem, or with CTCs, one must bear in mind that CTC isolation is mixed status (EMT and stemness). Even patients with node- based on epithelial characterized antibodies of cells (25).
negative invasion had ddCTCs: 32 out of 84 cases. This Moreover, at the detection step, the markers are limited to positivity is independant of tumor clinicopathological some of the total CTC populations. It can be expected that features. It would be of interest to test blood of these patients detected cells are epithelial cancer cells and that those after surgery and during progression-free survival to loosing the epithelial marker (EpCAM) are missed. Cancer determine whether these CTCs are implicated in the cells in the blood arise from the primary tumor and two types recurrence or metastases occurring in 25% of women at 5 of processes could be implicated: passive or active years. The stemness characteristics of these cells, particularly delamination. In the first phenomenon, escape of cancer cells dormancy, would be at the origin of the “waking-up” of the occurs as clusters and cells retain their epithelial disease. We cannot emphasize enough that the selected characteristics. Collective epithelial cell migration is an method is crucial in order to truly detect cancer cells in the integral part of development and cancer progression.
blood. Most of the conventional methods are not able to Apicobasal polarity is maintained when cell strands or sheets delineate the subpopulations issued from EMT. In the cohort leave the primary tumor (26, 27). The lifespan of such cells of 400 patients, we only found positivity for 8.75% of is limited by anoikis and solely their presence in the blood is samples, whereas 37.6% were positive when we examined an indicator of the existence of a primary tumor. Their EMT characteristics. Thus it seems evident that the second evidence does not necessarily indicate that they will survive method is more appropriate to analyse CTCs in breast cancer and grow but they can testify to the existence of hidden cell at the time of early diagnosis. Thus an EMT signature of subpopulations not detected by the method, coming from an CTCs may help to stratify early-stage breast cancer.
active delamination process during EMT (28). The data However, additional studies are required to determine the reported here also demonstrated that out of the 23 CTC- prognostic potential of EMT associated with CTCs.
positive patients, eight had persistence of cancer cells in the Consequently, a follow-up of at least five years may be blood even three weeks after removal of the primary tumor.
requiered to include mature outcome data.
Moreover, regarding CTC-negative patients before surgery,12 became CTC-positive four weeks after removal of the Conclusion
primary tumor. This positivity, demonstrating that cancercells are not always cleared by surgery, can be explained by The aim of the study was to assess CTCs expressing the presence of a secondary cancer site which releases CTCs mesenchymal markers in patients with primary breast cancer.
into the blood. Many questions still remain about this Due to frequent loss of epithelial antigens by CTCs, the most positive recurrence. The major concern is invasive cell populations are hidden. Detection has to be micrometastases exist a long time before they are revealed improved to identify CTCs which have undergone EMT. The by their growth. 12-37% of small breast tumors (<1 cm), two protocols demonstrated that initiating tumor cells which have already metastasized at diagnosis (29, 30) and some are subpopulations of ddCTCs are often missed. Such a result data suggest that systemic dissemination of tumor cells is of importance because it is these cells which support occurs at early stages of tumor development (31). Based on invasion and metastases, and potentially support the genesis these considerations, we started the second protocol to detect of residual disease. It has been demonstrated that ddCTCs are cells arising from EMT and which are not evidenced by the resistant to chemotherapy and radiotherapy. Neoadjuvent therapy to reduce the volume of tumor before surgery could An oncological type of EMT is able to convert concentrate these cells in the primary tumor. Thus it might be differentiated epithelial cancer cells into migratory necessary to establish a follow-up of ddCTCs to monitor how mesenchymal cancer cells. EMT is associated with poor they are cleared from the blood. Methodologies to identify clinical outcome in breast cancer (32, 33). PI3K signaling ddCTCs should allow for an accurate screening of new plays a key role in inducing and maintaining EMT. The most molecules which would be able to target or reverse EMT. important downstream effector of PI3KΑ is AKT. Moreover,migratory mesenchymal cells acquire some stem cell Competing Interests
characteristics and can be detected by using ALDH1, BMI1and CD44. Numerous transcription factors have been The Authors declare that they have no competing interests.
ANTICANCER RESEARCH 32: 3363-3370 (2012) Authors’ Contributions
11 Giuliano M, Giordano A, Jackson S, Hess KR, De Giorgi U, Mego M, Handy BC, Ueno NT, Alvarez RH, De Laurentiis M, De All contributors participated equally to this study. All authors read Placido S, Valero V, Hortobagyi GN, Reuben JM and Cristofanilli M: Circulating tumor cells as prognostic and predictive markersin metastatic breast cancer patients receiving first-line systemictreatment. Breast Cancer Res 13: R67-75, 2011.
Acknowledgements
12 Krag DN, Ashikaga T, Moss TJ, Kusminsky RE, Feldman S, Carp NZ, Moffat FL, Beitsch PD, Frazier TG, Gaskin TA, Shook Our work was financially supported by URCAM- Limousin (Union JW, Harlow SP and Weaver DL: Breast Cancer Cells in the Régionale des Caisses d’Assurance Maladie). Blood: A Pilot Study. Breast J 5: 354-358, 1999.
13 Cristofanilli M, Budd GT, Ellis MJ, Stopeck A, Matera J, Miller References
MC, Reuben JM, Doyle GV, Allard WJ, Terstappen LW and HayesDF: Circulating tumor cells, disease progression, and survival in 1 Liu MC, Shields PG, Warren RD, Cohen P, Wilkinson M, metastatic breast cancer. N Engl J Med 351: 781-791, 2004.
Ottaviano YL, Rao SB, Eng-Wong J, Seillier-Moiseiwitsch F, 14 Barrière G, Riouallon A, Renaudie J, Tartary M, Pelpel M and Noone AM and Isaacs C: Circulating tumor cells: a useful Rigaud M: Breast cancer, what do CTCs mean? EMT–Stemness predictor of treatment efficacy in metastatic breast cancer. J Clin invasivenes, metastasis and circulating tumor cells. Symposium Praga (13 June 2011): “Developments in Real-time PCR – From 2 Maetzel D, Denzel S, Mack B, Canis M, Went P, Benk M, Kieu Preanalytics to Molecular Diagnostics” Circulating Tumor Cells C, Papior P, Baeuerle PA, Munz M and Gires O: Nuclear signalling by tumour-associated antigen EpCAM. Nat Cell Biol 15 Kasimir-Bauer S, Hoffmann O, Wallwiener D, Kimmig R and Fehm T: Expression of stem cell and epithelial–mesenchymal 3 Sieuwerts AM, Kraan J, Bolt J, van der Spoel P, Elstrodt F, transition markers in primary breast cancer patients with Schutte M, Martens JW, Gratama JW, Sleijfer S and Foekens JA: circulating tumor cells. Breast Cancer Res 14: R15-23, 2012.
Anti-epithelial cell adhesion molecule antibodies and the 16 Morel AP, Lièvre M, Thomas C, Hinkal G, Ansieau S and detection of circulating normal-like breast tumor cells. J Natl Puisieux A: Generation of breast cancer stem cells through Cancer Inst 101: 61-66, 2009.
epithelial–mesenchymal transition. PLoS One 3: e2888, 2008.
4 Gradilone A, Raimondi C, Nicolazzo C, Petracca A, Gandini O, 17 Takebe N, Warren RQ and Ivy SP: Breast cancer growth and Vincenzi B, Naso G, Aglianò AM, Cortesi E and Gazzaniga P: metastasis: interplay between cancer stem cells, embryonic Circulating tumour cells lacking cytokeratin in breast cancer: signaling pathways and epithelial-to-mesenchymal transition.
The importance of being mesenchymal. J Cell Mol Med 15: Breast Cancer Res 13: 211-221, 2011.
18 Huber MA, Kraut N, and Beug H: Molecular requirements for 5 Barriere G, Riouallon A, Renaudie J, Tartary M and Rigaud M: epithelial–mesenchymal transition during tumor progression.
Mesenchymal and stemness circulating tumor cells in early Curr Opin Cell Biol 17: 548-558, 2005.
breast cancer diagnosis. BMC Cancer 12: 114, 2012. 19 Oliveras-Ferraros C, Cufí S, Vazquez-Martin A, Menendez OJ, 6 Barriere G, Tartary M and Rigaud M: Epithelial Mesenchymal Bosch-Barrera J, Martin-Castillo B, Joven J and Menendez JA: Transition: A New Insight into the Detection of Circulating Metformin rescues cell surface major histocompatibility complex Tumor Cells. IRSN Oncology (in press), 2012.
class I (MHC-I) deficiency caused by oncogenic transformation.
7 Fehm T, Hoffmann O, Aktas B, Becker S, Solomayer EF, Cell Cycle 11: 865-870, 2012.
Wallwiener D, Kimmig R and Kasimir-Bauer S: Detection and 20 Bi WR, Yang CQ and Shi Q: Transforming Growth Factor-β1 characterization of circulating tumor cells in blood of primary Induced Epithelial – Mesenchymal Transition in Hepatic breast cancer patients by RT-PCR and comparison to status of Fibrosis. Hepatogastroenterology 59, in press, 2012.
bone marrow disseminated cells. Breast Cancer Res 11: R59-67, 21 Zhou G, Dada LA, Wu M, Kelly A, Trejo H, Zhou Q, Varga J and Sznajder JI: Hypoxia-induced alveolar epithelial–mesenchymal 8 Molloy TJ, Bosma AJ, Baumbusch LO, Synnestvedt M, Borgen transition requires mitochondrial ROS and hypoxia-inducible E, Russnes HG, Schlichting E, van’t Veer LJ and Naume B: The factor 1. Am J Physiol Lung Cell Mol Physiol 297: L1120-1130, prognostic significance of tumour cell detection in the peripheral blood versus the bone marrow in 733 early-stage breast cancer 22 Thiery JP and Sleeman JP: Complex networks orchestrate patients. Breast Cancer Res 13: R61-71, 2011.
epithelial–mesenchymal transitions. Nat Rev Mol Cell Biol 7: 9 Banys M, Krawczyk N, Becker S, Jakubowska J, Staebler A, Wallwiener D, Fehm T and Rothmund R: The influence of 23 Yang J and Weinberg RA: Epithelial–mesenchymal transition: at removal of primary tumor on incidence and phenotype of the crossroads of development and tumor metastasis. Dev Cell circulating tumor cells in primary breast cancer. Breast Cancer 24 Kang Y and Massagué J: Epithelial-mesenchymal transitions: 10 Bidard FC, Hajage D, Bachelot T, Delaloge S, Brain E, twist in development and metastasis. Cell 118: 277-279, 2004.
Campone M, Cottu P, Beuzeboc P, Rolland E, Mathiot C and 25 Criscitiello C, Sotiriou C and Ignatiadis M: Circulating tumor Pierga JY: Assessment of circulating tumor cells and serum cells and emerging blood biomarkers in breast cancer. Curr Opin markers for progression-free survival prediction in metastatic breast cancer: A prospective observational study. Breast Cancer 26 Wood JM and Olson MF: Collective migration: spatial tension relief. Curr Biol 22: R125-127, 2012.
27 Scott RW, Crighton D and Olson MF: Modeling and imaging 3- 32 Liu R, Wang X, Chen GY, Dalerba P, Gurney A, Hoey T, dimensional collective cell invasion. J Vis Exp 58: 3525-3528.
Sherlock G, Lewicki J, Shedden K and Clarke MF: The prognostic role of a gene signature from tumorigenic breast 28 Wicha MS and Hayes DF: Circulating tumor cells: not all cancer cells. N Engl J Med 356: 217-226, 2007.
detected cells are bad and not all bad cells are detected. J Clin 33 Sheridan C, Kishimoto H, Fuchs RK, Mehrotra S, Bhat- Nakshatri P, Turner CH, Goulet R JR, Badve S and Nakshatri H: 29 Wilhelm MC, Edge SB, Cole DD, deParedes E and Frierson HF CD44+/CD24– breast cancer cells exhibit enhanced invasive JR: Nonpalpable invasive breast cancer. Ann Surg 213: 600-605, properties: an early step necessary for metastasis. Breast Cancer 30 Chadha M, Chabon AB, Friedmann P and Vikram B: Predictors of axillary lymph node metastases in patients with T1 breastcancer. A multivariate analysis. Cancer 73: 350-353, 1994.
31 Hüsemann Y, Geigl JB, Schubert F, Musiani P, Meyer M, Burghart E, Forni G, Eils R, Fehm T, Riethmüller G and Klein CA: Systemic spread is an early step in breast cancer. Cancer

Source: http://www.astralab.fr/IMG/pdf/anticancer.pdf