Stem-like Subtype Research Articles

Abstract 706: Inflammatory and stem-like colorectal cell lines show differential response to MEK-162 and neratinib in combination

Abstract Background: Improvement of colorectal cancer (CRC) treatment depends on finding new effective agents and associated biomarkers that can predict patient response to drugs. Recent cell culture experiments show that a combination of MEK and pan ERBB inhibitors may be efficacious in treating CRC (Sun et al 2014). The purpose of this study was to determine whether a cell line's response to targeted agents would correlate with its mutation profile and/or gene expression subtype. Methods: K-RAS mutation status of 7 cell lines were identified and classified into stem-like or inflammatory subtypes using gene expression data from the cancer cell line encyclopedia and an algorithm described by Sadanandam et al. We tested 5 stem-like cell lines (SW480, SW620, HCT116, C2BBE1, HS675T) and 2 inflammatory cell lines (NCI-H747, SW837) for response to inhibitors of the RAS/RAF and PI3K pathways. MEK-162 and neratinib were tested as single agents and in combination, using cell viability and clonogenic assays. Results: Cell lines varied in response to the drugs when used as single agents. All cell lines were resistant to the EGFR inhibitor (Gefitinib) at concentrations ≥5 μM. Two pan ERBB inhibitors (neratinib and afatinib) showed IC50 values at or >4 μM and 2 MEK inhibitors (Selumetinib and MEK162) had IC50 values at or >1μM. The inflammatory cell lines, SW837 and NCI-H747, had greater sensitivity to the MEK inhibitors (IC50 value 1μM) while the KRAS mutant stem-like cell lines were relatively more resistant (SW480, SW620) with IC50 values ≥4-5 μM. Stem-like cell lines with WT-KRAS (C2BBE1, HS675T) were even more resistant; IC50s were not achieved at concentrations of 5 μM. Similar results were obtained in clonogenic growth assay. The MEK and pan ERBB inhibitor combinations were more effective at decreasing cell viability and inducing apoptosis in inflammatory cell lines than in stem-like cell lines. This combination effectively inhibited 60-70% of cell viability in inflammatory cell lines at a dose of 0.5 μM and 0.062 μM for neratinib and MEK-162 respectively. The lethality of this combination was the same even if the doses were reversed (0.5 μM MEK-162 and 0.062 μM neratinib). In contrast, we observed that this combination was not nearly as effective against stem-like cell lines irrespective of KRAS mutation status. Profiling of the phosphorylation status of proteins in the RAF-MEK-ERK signaling pathway after exposure to the MEK-162 plus neratinib combination showed a substantial inhibition of ERK phosphorylation in inflammatory cell lines but not in 4 of the 5 stem-like cell lines. Conclusion: This study demonstrates that colon cancer cell lines classified as the inflammatory subtype are sensitive to the combination of MEK-162 plus neratinib at concentrations at which cell lines of the stem-like subtype are resistant. The PA DOH disclaims responsibility for analysis, interpretations, or conclusions. Citation Format: Rekha Pal, Nan Song, Ashok Srinivasan, Samuel A. Jacobs, Soonmyung Paik, Katherine L. Pogue-Geile. Inflammatory and stem-like colorectal cell lines show differential response to MEK-162 and neratinib in combination. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 706. doi:10.1158/1538-7445.AM2015-706

Abstract 399: Nanoparticle silencing of β1 and β3 integrins augments Dasatinib sensitivity

Abstract Triple negative breast cancers (TNBCs) are an aggressive subcategory of breast cancers that currently lack FDA-approved targeted therapies. Previous studies have i) further subcategorized TNBCs into 5 distinct subtypes, each with a preferential response to specific chemotherapies and ii) identified beta1 and beta3 integrins to be key mediators of breast cancer EMT and metastasis. We surveyed beta1 and beta3 integrin expression in the 5 subtypes of TNBC with and without treatment of their proposed chemotherapies by western blot analysis. Using nanoparticles, we delivered beta1 and beta3 integrin-specific siRNAs to silence the expression of both integrins either with or without treatment of the proposed chemotherapies, and evaluated the effect of these treatment regimens on EMT, invasion, and 3D outgrowth. We demonstrated that beta1 and beta3 integrins are robustly expressed in the Mesenchymal Stem-Like, Mesenchymal, and Basal-Like2 subtypes of TNBCs. Furthermore, we observed elevated expression of beta1 integrin upon Dasatinib treatment of the Mesenchymal Stem-Like subtype. Lastly, we determined that delivery of beta1 and beta3 integrin siRNAs sensitized breast cancer cells to Dasatinib treatment. Collectively, our data suggests that delivery of beta1 and beta3 integrin-specific siRNAs in combination with Dasatinib treatment efficiently eliminates the aggressive characteristics of Mesenchymal Stem-Like breast cancer cell lines. Citation Format: Jenny G. Parvani, Vinson Chu, Sarah Roelle, Zheng-Rong Lu, William Schiemann. Nanoparticle silencing of β1 and β3 integrins augments Dasatinib sensitivity. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 399. doi:10.1158/1538-7445.AM2015-399

Abstract 574: CD133 and LGR5 characterize stem-cell like subtype of colorectal cancer

Abstract Background: Colorectal cancer (CRC) has recently been divided into six subtypes with cluster mRNA expression analysis. The subtype characterized as stem-like subtype exhibit the highest degree of stemness with poorest prognosis. Prominin-1 (CD133) and leucine-rich-repeat-containing G protein-coupled receptor (LGR5) mark cancer stem cells. High CD133 expression is associated with higher rate of metastases and disease progression. In mouse models, tumor cells expressing CD133 are enriched in chemotherapeutic-resistant colorectal cancer. LGR5+ cells characterize fast-cycling columnar-based stem cells of colon which are instrumental in neoplasia formation. Further, the LGR5+ cells are associated with metastasis formation. Material and methods: A prospective randomized multicenter randomized trial in Norway recruited from December 1992 to October 1996. 412 colorectal cancer patients stage II and III were treated with surgery alone or 5-fluoruracil (5-FU) and levamisole (Lev) after surgical resection. The patients were followed up by routine clinical examination until five years after surgery. We constructed tissue microarrays from 409 patient samples. CD133 protein expression was assessed by mouse anti-CD133 ([clone AC133], Miltenyi Biotec, Bergisch Gladbach, Germany) at a dilution 1:10. LGR5 mRNA expression was assessed by RNAscope 2.0 HD Detection Kit-Brown (Advanced Cell Diagnostics, Hayward, CA, USA). Results: Colon cancer stage III patients negative for CD133 had better five-year disease-free survival (DFS) when they received 5-FU/Lev (P = .016). In contrast, the patients did not respond to adjuvant therapy when CD133 was expressed (P = .500). LGR5 was a prognostic marker for colon cancer stage II (P = .045). In colon stage III, mRNA expression of LGR5 was associated with better DFS (P = .046) and patients with positive LGR5 responded to chemotherapy (P = .025). In stage II, LGR5 mRNA expression was associated with lower tumor diameter (P = .005), while in stage III there was no statistically significant association. In conclusion, CD133 expression was associated with chemotherapy resistance in the colon stage III cohort. LGR5 was associated with a metastatic phenotype in CRC stage II and colon cancer stage III. In addition, LGR5 expression predicted response to chemotherapy in colon cancer stage III. Citation Format: Luka Stanisavljevic, Mette Pernille Myklebust, Olav Dahl. CD133 and LGR5 characterize stem-cell like subtype of colorectal cancer. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 574. doi:10.1158/1538-7445.AM2015-574

Abstract 2925: WAVE3 is associated with chemoresistance in TNBC by activating tumor angiogenesis downstream of the STAT1/Hif-1α/VEGF-A signaling axis

Abstract Chemoresistance is the main cause of disease recurrence and eventual death of breast cancer patients, particularly those harboring triple-negative breast cancer (TNBC) tumors for which effective FDA-approved drugs currently do not exist. However, a recent study determined that TNBCs can be segregated into 6 genetically distinct subtypes that do in fact exhibit differential rates of pathological complete response (pCR) to standard-of-care chemotherapies. Interestingly, the mesenchymal and mesenchymal stem-like subtypes of TNBCs exhibit the lowest rates of pCR when treated with standard-of-care chemotherapies. WAVE3 is an actin-cytoskeleton remodeling protein, and recent studies have highlighted a potential role for WAVE3 in promoting tumor progression and metastasis in TNBC. However, whether WAVE3 activity is involved in the development of chemoresistance in TNBCs, remains unclear. Here we show that loss of WAVE3 expression resensitizes human TNBC cells to Doxorubicin and Docetaxel, as measured by increased apoptosis and cell death. We also show that WAVE3 knockdown in the chemotherapy-treated TNBC cells results in inhibition of STAT1 phosphorylation, as well as a significant decrease in expression levels of its downstream effector Hif-1α. Since Hif-1α is a major activator of VEGF-A production, and therefore, a stimulator of tumor angiogenesis, loss of Hif-1α in the WAVE3-knocdown cells resulted in the inhibition the chemotherapy-mediated VEGF-A secretion and the downstream activation of angiogenesis, a phenomenon that often accompanies chemoresistance. Our data identify a critical role of WAVE3 in sensitizing TNBC to chemotherapy by inhibiting the STAT1⌠Hif-1α⌠VEGF-A signaling axis, and support the possibility of WAVE3 inhibition as a promising approach for TNBC cancer therapy. Citation Format: Gangarao Davuluri, Edward F. Plow, William P. Schiemann, Khalid Sossey-Alaoui. WAVE3 is associated with chemoresistance in TNBC by activating tumor angiogenesis downstream of the STAT1/Hif-1α/VEGF-A signaling axis. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 2925. doi:10.1158/1538-7445.AM2015-2925

Abstract 1446: Induction of epithelial-to-mesenchymal transition (EMT) by SL-15 knockdown contributes to anoikis resistance in human colorectal cancer metastases

Abstract Purpose of the study: Metastasis or systemic dissemination of cancer cells into secondary organs is responsible for 90% of cancer-related fatalities. Our team previously screened out SL-15 as therapeutic target of gastric cancer through the analysis of gene expression profiling. In this study, we investigated the roles of human SL-15 in colorectal cancer (CRC) metastasis. Experimental procedures: SL-15 levels of metastatic liver CRC tissues and of primary CRC tissues as well as of adjacent nontumour tissues were compared using quantitative RT-PCR. Transwell and wound healing methods were used to determine the migration/invasion abilities of CRC cells. Flow cytometery analysis was performed to measure the apoptotic ratio as well as the levels of E-cadherin/p53/SL-15. Confocal observation and Western blot analysis were used for studying the underlying mechanisms of small interfering RNA against SL-15 (siSL-15)-induced CRC metastasis. Results: The level of SL-15 was higher in the order of metastatic liver CRC tissues, primary CRC tissues and then the adjacent nontumor tissues. siSL-15 significantly induced the migration/invasion of CRC cells that was almost entirely suppressed by the co-expression of SL-15. Importantly, the application of siSL-15 contributes to anoikis resistance of CRC cells via the upregulation of zinc finger E-box-binding homeobox 1 (ZEB1) in stem-like subtype CRC cells and intestinal trefoil factor (ITF) in goblet- like subtype CRC cells by the inhibition of glycogen synthase kinase 3 β (GSK3β) and of β-catenin degradation, thereby leading to downregulation of E-cadherin as well as p53 degradation. Importantly, the interactions among SL-15/p53/GSK3β that were dissociated upon the application of siSL-15 were determined. Conclusions: SL-15 is more highly up-regulated in metastatic liver CRC cancer than in primary CRC cancer and the adjacent nontumor tissue: SL-15 knockdown contributes to anoikis resistance of CRC cells and CRC metastasis via inducing epithelial-to-mesenchymal transition as well as p53 degradation due to the dissociation of SL-15 with GSK3β and p53. Citation Format: Soo Young Jun, Hyun-Soo Cho, Jeong-Ju Lee, Jun-Ho Ahn, Ji-Yong Yoon, Jae-Hye Lee, Min Hyuk Choi, Nam-Soon Kim. Induction of epithelial-to-mesenchymal transition (EMT) by SL-15 knockdown contributes to anoikis resistance in human colorectal cancer metastases. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 1446. doi:10.1158/1538-7445.AM2015-1446

Expression of epithelial-mesenchymal transition-related markers in triple-negative breast cancer: ZEB1 as a potential biomarker for poor clinical outcome.

Triple-negative breast cancer (TNBC) is a heterogeneous group of disease with a well-known association with epithelial-mesenchymal transition (EMT) and breast cancer stem cell phenotype. Recent studies have shown that TNBC can be classified into 6 subtypes, including basal-like, mesenchymal-like, and mesenchymal stem-like subtypes. However, clinical significance of the EMT in TNBC remains unclear. We analyzed immunohistochemical expression of EMT-related markers, including EMT markers (expression of vimentin, smooth muscle actin, osteonectin, and N-cadherin; loss of E-cadherin), EMT inducers (ZEB1 and CD146), and breast cancer stem cell markers (CD44(+)/CD24(-) and aldehyde dehydrogenase 1) in 173 TNBCs and correlated their expression with clinicopathological features of the tumors, including clinical outcome. Expressions of vimentin, CD44(+)/CD24(-), and CD146 were more frequent in basal-like TNBCs than non-basal-like TNBCs. Whereas CD146 expression was closely associated with the expression of various EMT markers and CD44(+)/CD24(-) phenotype, ZEB1 expression correlated only with the expression of smooth muscle actin. Expressions of vimentin, smooth muscle actin, osteonectin, and ZEB1 and loss of E-cadherin were more frequently found in metaplastic carcinomas than in other histologic subtypes. In survival analyses, EMT markers were not associated with patients' clinical outcomes. However, ZEB1 expression was found to be an independent prognostic factor for poor disease-free survival. These findings indicate that expression of EMT-related markers in TNBCs can be a signature of a certain subgroup of TNBC, which is associated with metaplastic carcinoma, and ZEB1 expression can serve as a potential biomarker to define a subgroup of TNBC associated with poor clinical outcomes.

Dual inhibition of EGFR and MET induces synthetic lethality in triple-negative breast cancer cells through downregulation of ribosomal protein S6

Triple-negative breast cancer (TNBC) exhibits innate resistance to the EGFR inhibition despite high level expression of EGFR. Recently, we found that the proliferation of basal-like (BL) subtype TNBC cells is synergistically inhibited by combination of EGFR and PI3K/AKT inhibitors. On the contrary, TNBC cells of mesenchymal stem-like (MSL) subtype are resistant to these combinations. To identify potential synthetic lethal interaction of compounds for treatment of MSL subtype TNBC cells, we performed MTT screening of MDA-MB-231 cells with a small library of receptor tyrosine kinase inhibitors (RTKIs) in the presence of gefitinib, an EGFR inhibitor. We identified MET inhibitors as potent RTKIs that caused synthetic lethality in combination with gefitinib in MDA-MB-231 cells. We demonstrated that combination of a MET inhibitor SU11274 with various EGFR inhibitors resulted in synergistic suppression of cell viability (in MTT assay) and cell survival (in colony formation assay) of MSL subtype TNBC cells. We further demonstrated that SU11274 alone induced G2 arrest and gefitinib/SU11274 combination sustained the SU11274-induced G2 arrest in these cells. In addition, SU11274/gefitinib combination synergistically reduced the level of ribosomal protein S6 (RPS6) in MSL subtype TNBC cells. In addition, knockdown of RPS6 itself, in both HS578T and MDA-MB-231, markedly reduced the proliferation of these cells. Taken together, our data suggest that dual targeting of EGFR and MET inhibits the proliferation of MSL subtype TNBC cells through down-regulation of RPS6.

Metastatic breast carcinomas display genomic and transcriptomic heterogeneity

Metaplastic breast carcinoma is a rare and aggressive histologic type of breast cancer, preferentially displaying a triple-negative phenotype. We sought to define the transcriptomic heterogeneity of metaplastic breast cancers on the basis of current gene expression microarray-based classifiers, and to determine whether these tumors display gene copy number profiles consistent with those of BRCA1-associated breast cancers. Twenty-eight consecutive triple-negative metaplastic breast carcinomas were reviewed, and the metaplastic component present in each frozen specimen was defined (ie, spindle cell, squamous, chondroid metaplasia). RNA and DNA extracted from frozen sections with tumor cell content >60% were subjected to gene expression (Illumina HumanHT-12 v4) and copy number profiling (Affymetrix SNP 6.0), respectively. Using the best practice PAM50/claudin-low microarray-based classifier, all metaplastic breast carcinomas with spindle cell metaplasia were of claudin-low subtype, whereas those with squamous or chondroid metaplasia were preferentially of basal-like subtype. Triple-negative breast cancer subtyping using a dedicated website (http://cbc.mc.vanderbilt.edu/tnbc/) revealed that all metaplastic breast carcinomas with chondroid metaplasia were of mesenchymal-like subtype, spindle cell carcinomas preferentially of unstable or mesenchymal stem-like subtype, and those with squamous metaplasia were of multiple subtypes. None of the cases was classified as immunomodulatory or luminal androgen receptor subtype. Integrative clustering, combining gene expression and gene copy number data, revealed that metaplastic breast carcinomas with spindle cell and chondroid metaplasia were preferentially classified as of integrative clusters 4 and 9, respectively, whereas those with squamous metaplasia were classified into six different clusters. Eight of the 26 metaplastic breast cancers subjected to SNP6 analysis were classified as BRCA1-like. The diversity of histologic features of metaplastic breast carcinomas is reflected at the transcriptomic level, and an association between molecular subtypes and histology was observed. BRCA1-like genomic profiles were found only in a subset (31%) of metaplastic breast cancers, and were not associated with a specific molecular or histologic subtype.

194 Pancreatic ductal organoids as a new platform for drug discovery

stroma content and vascularization. At the transcriptomic level, we identified several PDX clusters similar to Goblet-like, Enterocyte, Transit-Amplifying, Inflammatory and Stem-like subtypes recently reported in CRC patients. In addition, well and moderately differentiated PDX showed a different gene expression profile than poorly and undifferentiated models. At the genomic level, the colon PDX models exhibited alteration profiles comparable to those reported for patient tumors regarding signatures of mutational processes (patient-age related or associated to mismatch repair gene deficiencies), chromosomal rearrangement and mutated genes (e.g. APC, TP53, KRAS and TP53). Moreover, we identified 6 PDX models with microsatellite instability (MSI), 5 of them being hyper-mutant with a loss of MLH1 expression. Interestingly, colon PDX displaying the MSI phenotype showed gene expression patterns of the Globlet-like or Inflammatory subtypes. To demonstrate the potential of our approach, we searched for molecular determinants of cetuximab sensitivity in colon PDX by combining in vivo drug sensitivity and molecular data. The study identified the Transit Amplifying subtype, AREG/EREG expression and KRAS mutation among the best predictors of response. Similarly, predictive biomarkers for oxaliplatin and irinotecan were found, and their associations with particular PDX subtypes are under investigation. The accuracy in predicting patient outcome following treatment with FOLFOX and FOLFIRI will be evaluated by assessing publicly available datasets. The comprehensive characterization of our colon PDX models confirmed their similarities with patient tumors and their inter-tumoral diversity. The combined use of PDX molecular and drug sensitivity data represents a promising approach for biomarker identification and tumor response prediction at a preclinical stage.

193 The identification of new therapies for ependymoma subgroups

stroma content and vascularization. At the transcriptomic level, we identified several PDX clusters similar to Goblet-like, Enterocyte, Transit-Amplifying, Inflammatory and Stem-like subtypes recently reported in CRC patients. In addition, well and moderately differentiated PDX showed a different gene expression profile than poorly and undifferentiated models. At the genomic level, the colon PDX models exhibited alteration profiles comparable to those reported for patient tumors regarding signatures of mutational processes (patient-age related or associated to mismatch repair gene deficiencies), chromosomal rearrangement and mutated genes (e.g. APC, TP53, KRAS and TP53). Moreover, we identified 6 PDX models with microsatellite instability (MSI), 5 of them being hyper-mutant with a loss of MLH1 expression. Interestingly, colon PDX displaying the MSI phenotype showed gene expression patterns of the Globlet-like or Inflammatory subtypes. To demonstrate the potential of our approach, we searched for molecular determinants of cetuximab sensitivity in colon PDX by combining in vivo drug sensitivity and molecular data. The study identified the Transit Amplifying subtype, AREG/EREG expression and KRAS mutation among the best predictors of response. Similarly, predictive biomarkers for oxaliplatin and irinotecan were found, and their associations with particular PDX subtypes are under investigation. The accuracy in predicting patient outcome following treatment with FOLFOX and FOLFIRI will be evaluated by assessing publicly available datasets. The comprehensive characterization of our colon PDX models confirmed their similarities with patient tumors and their inter-tumoral diversity. The combined use of PDX molecular and drug sensitivity data represents a promising approach for biomarker identification and tumor response prediction at a preclinical stage.

Loss of WAVE3 sensitizes triple-negative breast cancers to chemotherapeutics by inhibiting the STAT-HIF-1α-mediated angiogenesis.

Chemoresistance allows for disease to recur and ultimately causes the death of most breast cancer patients. This scenario is particularly relevant in patients harboring triple-negative breast cancer (TNBC) tumors for which there are no effective FDA-approved drugs. However, a recent study determined that TNBCs can be segregated into 6 genetically distinct subtypes that do in fact exhibit differential rates of pathological complete response (pCR) to standard-of-care chemotherapies. Of these, the mesenchymal and mesenchymal stem-like subtypes of TNBCs exhibit the lowest rates of pCR when treated with standard-of-care chemotherapies. WAVE3 is an actin-cytoskeleton remodeling protein, and recent studies have highlighted a potential role for WAVE3 in promoting tumor progression and metastasis in TNBC. However, whether WAVE3 activity is involved in the development of chemoresistance in TNBCs remains unclear. Here we show that loss of WAVE3 expression resensitizes human TNBC cells to doxorubicin and docetaxel, as measured by increased apoptosis and cell death. We also show that WAVE3 knockdown in the chemotherapy-treated TNBC cells results in inhibition of STAT1 phosphorylation, as well as a significant decrease in expression levels of its downstream effector HIF-1α. Since HIF-1α is a major activator of VEGF-A production, and therefore a stimulator of tumor angiogenesis, loss of HIF-1α in the WAVE3-knockdown cells resulted in the inhibition the chemotherapy-mediated VEGF-A secretion and the downstream activation of angiogenesis, a phenomenon that often accompanies chemoresistance. Our data identify a critical role of WAVE3 in sensitizing TNBC to chemotherapy by inhibiting the STAT1→HIF-1α→VEGF-A signaling axis, and support the possibility that WAVE3 inhibition may be a promising target for TNBC cancer therapy.

Abstract 5559: Reciprocal expression of miR-125b and miR-192/miR-200 families define clinically relevant sub-types of colorectal cancer with differential sensitivity to EGFR and MEK targeted agents

Abstract The aims of this study were to determine if miRNA profiles could enhance prediction of pharmacological response compared to mutation alone. miRNAs are readily detectable in serum and plasma thus could be used in combination with mutations in liquid biopsies to support patient selection. Unsupervised hierarchical clustering of a 50 cell line panel global miRNA profile revealed a subset of miRNAs with a striking reciprocal expression pattern. Expression of miR-192+ve (3 miRs), miR-125b+ve (3miRs) and miR-200-ve (3miRs) defined 3 main cell sub-types which were further divided into 5 subtypes by expression of miR31+ve. Significantly, the miRNA sub-types are closely related to the five mRNA subtypes we had previously identified using iNMF (Schlicker et al, 2012) and to the subtypes identified in clinical CRC samples by Sadanandam et al 2013, Loboda et al, 2012 and Oh et al 2012. The miR192+ve subtype is an epithelial sub-type (iNMF sub-type 2 / Oh A-type) which has a gene expression pattern found in cells with transit amplifying phenotype, whereas the miR200 negative sub-type relates to the mesenchymal iNMF sub-type 1.1 and has a gene expression pattern similar to that found in CRC stem-like cells. Expression of miR125b is found in the stem-like subtype but also in mesenchymal sub-types with gene expression related to an inflammatory phenotype (iNMF sub-type 1.2, Oh B). Analysis of differentially expressed genes between the sub-types revealed the miR-192+ve positive sub-type was discriminated from the miR-125b+ve by higher expression of several genes specifically associated with intestinal differentiation including HNF1a (TCF1) (3-fold) and HNF4a (TCF14) (13-fold) which drive expression of miR192 and CDX1 (71-fold), HNF4G (6-fold) ISX (12-fold), and IHH (24-fold). TCGA clinical samples were also classified by the four mRNA methods and miRNAs differentially expressed between the sub-types analysed; all of the miRNAs identified in cell lines were statistically significantly differentially expressed in TCGA sub-types, however, there were additional miRNAs which clearly correlate with sub-types in clinical samples that have very low expression in cells. Importantly, the miR-192+ve sub-type exhibited significantly greater sensitivity to EGFR TKIs, cetuximab and MEK targeted agents compared to the miR-200-ve and miR-125b+ve sub-types. Taken together, the miR data enables segregation by differentiation status, this data enhances prediction of drug activity compared to mutation data alone and is of significance because miR expression and differentiation status are modulated by prior chemotherapy regimes. Citation Format: Christine M. Chresta, Andreas Schlicker, Garry Beran, Georgia Cerillo, Lodewyk F. A. Wessels, George Orphanides. Reciprocal expression of miR-125b and miR-192/miR-200 families define clinically relevant sub-types of colorectal cancer with differential sensitivity to EGFR and MEK targeted agents. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 5559. doi:10.1158/1538-7445.AM2014-5559

A3.34 The global microrna profile of skin in systemic sclerosis

BackgroundThe underlying pathogenesis of systemic sclerosis (SSc) remains poorly understood contributing to limited efficacy of therapeutic options.MicroRNAs (miRNAs) are small non-coding RNAs that play an important role in post-transcriptional gene...

Abstract P1-08-17: Impact of claudin-3 and basal-markers expressions on pathological complete response to neoadjuvant chemotherapy and prognosis in triple negative breast cancer

Abstract Background Triple negative breast cancer (TNBC) has responded well in the neoadjuvant setting, with rates of pathologic complete response (pCR) commonly higher than for other breast tumor types. However, more than half of TNBC patients do not achieve a pCR and have a very poor prognosis. Cluster analysis identified 6 TNBC subtypes displaying unique gene expression, including basal-like and a mesenchymal stem-like (claudin-low) subtype. The object of this study was to evaluate the clinical utility of claudin-3 and basal-markers expressions in the prediction of pCR and prognosis in a cohort of breast cancer patients receiving the combination of taxane and anthracycline as neoadjuvant chemotherapy. Materials and Methods A prospective data base of 412 Japanese women with stage I/II/III breast cancer who received neoadjuvant chemotherapy was analyzed. The regimens were 6 cycles of concurrent administration of doxorubicin (50 mg/m2) and docetaxel (60 mg/m2) or 4 cycles of epirubicin (100 mg/m2), 5-fluorouracil (500 mg/m2), and cyclophosphamide (500 mg/m2) followed by 4 cycles of docetaxel (100 mg/m2). We performed immunohistochemical staining for ER, PgR, HER2, epidermal growth factor receptor (EGFR), cytokeratin 5/6 and claudin-3 using specimens obtained at diagnosis by core needle biopsy. Based on immunoreactivity, TNBC tumors were classified into basal/claudin-high (BCH), basal/claudin-low (BCL), non-basal/claudin-high (NBCH) and non-basal/claudin-low (NBCL). Results The overall pCR of this patient cohort was 19%. pCR was observed 30 cases of 84 TNBC (36%) and 50 cases of 328 non-TNBC (15%)(p<0.0001). Among the TNBC tumors, pCR rate was significantly different between subgroups. BCH 50%, BCL 23%, NBCH 83% and NBCL 43% (p = 0.039). The 5-year overall survival rate was significantly different between TNBC and non-TNBC. TNBC 81% and non-TNBC 94% (p = 0.0017). The 5-year overall survival rate was different between triple-negative subgroups. BCH 92%, BCL 78%, NBCH 83% and NBCL 46%. Conclusions This study indicated that claudin-3 and basal-markers might be predictive biomarkers of pCR and prognosis in triple negative breast cancer patients treated with neoadjuvant chemotherapy. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P1-08-17.

Abstract P4-04-05: Molecular subtyping reveals the heterogeneity of metaplastic breast cancers

Abstract Introduction: Metaplastic breast carcinoma (MBC) is a rare (1%) and aggressive histological special type of breast cancer, which often (>90%) displays a triple-negative (TN) phenotype. Microarray-based gene expression classifications of breast cancer have been described, including the ‘intrinsic subtypes’ (i.e. luminal A, luminal B, basal-like, HER2-enriched, normal breast-like and claudin-low) and the six TN molecular subtypes with therapeutic implications in preclinical models (i.e. basal-like I, basal-like II, mesenchymal-like, mesenchymal stem-like, immunomodulatory and luminal androgen receptor). In addition, single nucleotide polymorphism (SNP) array analyses have led to the development of classifiers to define whether TN breast cancers would have genomic aberrations consistent with those found in tumors arising in BRCA1 germline mutation carriers. Here we sought to define i) the heterogeneity of MBCs using distinct microarray-based classifiers and ii) whether MBCs display gene copy number profiles consistent with those of BRCA1 breast cancers. Material and Methods: Thirty consecutive MBCs were retrieved from the tumor banks of the authors’ institutions, reviewed by two pathologists and classified into three groups: carcinomas with spindle cell metaplasia (n = 12), with squamous metaplasia (n = 10), and with heterologous elements (n = 8). RNA and DNA were extracted from representative frozen sections containing >50% of cancer cells from each tumor and subjected to gene expression profiling using the HumanHT-12 v4 platform (Illumina) and gene copy number analysis using SNP 6.0 arrays (Affymetrix), respectively. ‘Intrinsic subtyping’ was performed according to the UNC guidelines, and subtyping into the six TN types using a dedicated website (https://cbc.mc.vanderbilt.edu/tnbc/). Classification of MBCs into BRCA1-like or non-BRCA1-like was performed using an algorithm to identify and quantify large-scale state transitions. Results: PAM50/claudin-low subtyping of MBCs revealed that all spindle cell carcinomas (n = 12) were consistently of claudin-low subtype, whereas MBCs with heterologous elements were classified as of basal-like (n = 6) and normal breast-like (n = 2), and squamous cell carcinomas as of basal-like (n = 5), claudin-low (n = 4) and normal-like (n = 1). Using the six TN subtypes, spindle cell MBCs were classified as mesenchymal stem-like (n = 4), immunomodulatory (n = 3), basal-like 2 (n = 3), or mesenchymal (n = 2) subtype. Carcinomas with heterologous elements were classified as basal-like 1 (n = 5) or mesenchymal (n = 3) subtype, whereas carcinomas with squamous metaplasia were of immunomodulatory (n = 5), basal-like 1 (n = 2), basal-like 2 (n = 2) or mesenchymal (n = 1) subtype. Out of the 26 samples where SNP 6.0 arrays were successfully performed, only 9 (35%) were classified as BRCA1-like, including two spindle cell carcinomas, two carcinomas with heterologous elements and five carcinomas with squamous metaplasia. Conclusion: MBCs constitute a heterogeneous group of tumors and the histological subclassification of these cancers is of importance, given that tumors with distinct metaplastic elements are classified differently according to current molecular subtyping methods. Only 35% of MBCs display BRCA1-like patterns of gene copy number aberrations. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P4-04-05.

Abstract A153: Isolation and characterization of high-purity invasive CTCs from men with metastatic castration-resistant prostate cancer (mCRPC).

Abstract Introduction: As metastatic biopsies are often difficult and impractical to obtain; isolation, characterization, and genomic profiling of circulating tumor cells (CTCs) from men with metastatic castration resistant prostate cancer (mCRPC) is critical for understanding tumor progression and mechanisms of resistance to treatment. The Vitatex cell-adhesion matrix (CAM) platform captures invasive CTCs in both a cell surface marker- and size-independent manner based on the ability of CTCs to invade a fluorescently-labeled CAM. This platform allows for isolation and characterization of viable CTC sub-populations from individual patients with mCRPC. In this study, we sought to optimize CTC isolation and purification for downstream genomics and to explore epithelial-mesenchymal and stem-like CTC subtypes, all in the context of treatment resistance to Abiraterone, a novel inhibitor of androgen synthesis. Methods: Forty ml of blood was collected from 23 patients with mCRPC. The CTCs were isolated using the VitaAssay platform and paired metastatic biopsies were obtained for a subset of men. Fluorescence-activated cell sorting (FACS) was coupled with the VitaAssay platform to increase cell purity. Invasive CTCs having ingested the fluorescent CAM were sorted based on CAM+/CD45-/CD14-/DAPI- staining. To complement the FACS data and for quality control, a subset of cases (N=11) were also enumerated using Immunocytochemistry (ICC), where CTCs were identified as CAM+/CD45-/CD14-/DAPI+. Healthy donor blood was used as a control. FACS-sorted stem-like CTCs were defined as CD44+/CAM+/CD45-/CD14-/DAPI- and mesenchymal CTCs were identified with ICC as vimentin+/CAM+/CD45-/DAPI+. Array comparative genomic hybridization (aCGH) of FACS-sorted CAM-avid amplified CTCs using Agilent microarrays was performed to obtain the CTC and matched biopsy genomic profiles. Results: Enumeration of CTCs by VitaAssay coupled with FACS yielded an average of 72 CTCs per 7.5 ml blood (range 1 to 338) for the 23 patients. Only 4 of the 23 patients had ≤ 2 CTCs per 7.5 ml blood. No CTCs were detected in the healthy control subject after cell sorting. The CTC purity increased post-FACS to ∼90% from ∼1% pre-FACS. Enumeration along with genomic analysis (copy number aberrations) of CTCs and their subpopulations, comparison with matched biopsies and clinical data are currently being investigated. Conclusions: Combining the VitaAssay platform with FACS allows for recovery of high purity mCRPC CTC subpopulations suitable for downstream molecular and genomic analysis. Ultimately, studies incorporating these optimized CTC methods will allow for better understanding of mechanisms of treatment resistance, disease progression and potentially identify new drug targets for CRPC. Citation Information: Mol Cancer Ther 2013;12(11 Suppl):A153. Citation Format: Gayatri Premasekharan, Terence W. Friedlander, Vy T. Ngo, Elizabeth Gilbert, Priya Sathaye, Anna Harris, Charles J. Ryan, Pamela L. Paris. Isolation and characterization of high-purity invasive CTCs from men with metastatic castration-resistant prostate cancer (mCRPC). [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference: Molecular Targets and Cancer Therapeutics; 2013 Oct 19-23; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2013;12(11 Suppl):Abstract nr A153.

Abstract 230: Connectivity-map analyses identify FDA-approved drugs targeting breast cancer stem cells.

Abstract Recently developed genomics-based tools are allowing repositioning of established drugs as cancer treatments. This approach has not been employed extensively to identify drugs that may target cancer stem cells (CSCs) of solid tumors. CD44+/CD24- cancer cells are the first described CSCs for breast cancer, which are enriched in mesenchymal stem-like (MSL) and mesenchymal-like (ML) subtypes of triple negative (TNBCs) and in anti-estrogen resistant breast cancers. We subjected gene expression array datasets containing list of genes differentially expressed in CD44+/CD24- CSCs compared to non-CSCs as well as in cells enriched for other markers of CSCs including CD271 and GD3S (six datasets in total) to connectivity map (CMAP) to identify drugs that may reverse gene expression patterns unique to CSCs. All trans retinoic acid (ATRA) was negatively connected with gene expression in CD44+/CD24-, CD271+, and GD3S+ CSCs as well as in ML and MSL subtypes of TNBCs. Independently, ATRA was found to modulate the expression of several genes in the recently described CSC-enriched Yes-Associated Protein (YAP) signature. ATRA-inducible genes of the YAP signature were associated with good prognosis, whereas ATRA-repressible genes were associated with poor prognosis, depending on the subtype of breast cancer. ATRA reduced mammosphere forming ability of breast cancer cells including acquired anti-estrogen resistant cells, which correlated with ATRA-mediated reduction of the critical stemness-associated gene SOX2 and induction of SOX21, which functions as a differentiation factor depending on the cell type. CD44+/CD24- breast cancer cell line with K-ras mutation was resistant to ATRA, which was reversed by MAP kinase inhibitors. From these results, we propose ATRA as a potential therapy for specific subtypes of breast cancers enriched for CSCs. A biomarker-driven patient selection may help to identify patients who may benefit from ATRA therapy. Citation Format: Poornima Bhat-Nakshatri, Chirayu Goswami, Sunil Badve, George W. Sledge, Harikrishna Nakshatri. Connectivity-map analyses identify FDA-approved drugs targeting breast cancer stem cells. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 230. doi:10.1158/1538-7445.AM2013-230

Stem Cell-Like Gene Expression in Ovarian Cancer Predicts Type II Subtype and Prognosis

Although ovarian cancer is often initially chemotherapy-sensitive, the vast majority of tumors eventually relapse and patients die of increasingly aggressive disease. Cancer stem cells are believed to have properties that allow them to survive therapy and may drive recurrent tumor growth. Cancer stem cells or cancer-initiating cells are a rare cell population and difficult to isolate experimentally. Genes that are expressed by stem cells may characterize a subset of less differentiated tumors and aid in prognostic classification of ovarian cancer. The purpose of this study was the genomic identification and characterization of a subtype of ovarian cancer that has stem cell-like gene expression. Using human and mouse gene signatures of embryonic, adult, or cancer stem cells, we performed an unsupervised bipartition class discovery on expression profiles from 145 serous ovarian tumors to identify a stem-like and more differentiated subgroup. Subtypes were reproducible and were further characterized in four independent, heterogeneous ovarian cancer datasets. We identified a stem-like subtype characterized by a 51-gene signature, which is significantly enriched in tumors with properties of Type II ovarian cancer; high grade, serous tumors, and poor survival. Conversely, the differentiated tumors share properties with Type I, including lower grade and mixed histological subtypes. The stem cell-like signature was prognostic within high-stage serous ovarian cancer, classifying a small subset of high-stage tumors with better prognosis, in the differentiated subtype. In multivariate models that adjusted for common clinical factors (including grade, stage, age), the subtype classification was still a significant predictor of relapse. The prognostic stem-like gene signature yields new insights into prognostic differences in ovarian cancer, provides a genomic context for defining Type I/II subtypes, and potential gene targets which following further validation may be valuable in the clinical management or treatment of ovarian cancer.

Abstract P3-06-14: Identification of Prognosis-Relevant Subgroups in Patients with Chemoresistant Triple Negative Breast Cancer

Abstract Purpose: Triple-negative breast cancer (TNBC) is a highly heterogeneous disease. TNBC patients with pathologic complete response (pCR) have excellent survival, but those with residual disease after neoadjuvant chemotherapy have significantly worse outcome. However, some patients having extensive residual cancer burden after neoadjuvant chemotherapy do not relapse, and we hypothesize that there may be subgroups with diverse prognosis among these chemoresistant TNBC patients. Methods: Forty-nine cases with residual cancer from 111 TNBC patients treated with neoadjuvant chemotherapy (in M.D. Anderson Cancer Center, 2000–2006) constituted the discovery cohort. Twenty-five chemoresistant samples from 47 neoadjuvant chemotherapy-treated TNBC (in Baylor College of Medicine and Methodist Hospital, 2002–2006) were chosen for validation. Extended validation was performed in 269 operable TNBC predicted to be chemoresistant (using a JAMA-published genomic predictor) from public databases. Results: By comparing the gene expression data from cases in relapse with those from un-relapsed cases, we established a 7-gene prognostic signature (including AR, ESR2, GATA3, GBX2, KRT16, MMP28, and WNT11) using dChip and gene enrichment analyses. In the discovery cohort, the signature showed positive predictive value (PPV; i.e., cumulative relapse rate of patients predicted to relapse in 3 years) of 95.4% and negative predictive values (NPV; i.e., relapse-free survival of patients predicted not to relapse in 3 years) of 100%. In the validation cohort, the classifier predicted correctly with PPV of 75.0% and NPV of 76.9% at 3 years. Compared with patients predicted not to relapse, those predicted to relapse had a hazard ratio of 3.37 (95% CI, 1.15–9.85) for disease recurrence or death in 3 years. In an extended validation cohort of 269 patients, our signature discriminated chemoresistant TNBC in overall cohort (PPV, 52.4%; NPV, 77.7%; log rank P < 0.0001), or each subset (e.g., log rank p = 0.001 for Rotterdam set; p = 0.013 for Frankfurt set), with adjusted overall hazard ratio of 2.07 (95% CI, 1.26–3.39). This signature was the only marker that could effectively predict the relapse in patients with chemoresistant TNBC. Of note, the subgroup predicted not to relapse was characterized by high expression of luminal-like genes (AR, GATA3), while the subgroup predicted to have high possibility of relapse was characterized by high expression of cancer stem cell and epithelial-to-mesenchymal transition associated genes (WNT11, MMP28). The former corresponded to the luminal AR subtype and the latter to the mesenchymal stem-like subtype, according to Pietenpol's TNBC subtype classification. Conclusion: We developed a clinically useful prognostic signature for chemoresistant TNBC. For these chemoresistant TNBC patients, new therapeutic strategies targeting AR-activation or cancer stem cells need to be developed. Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr P3-06-14.

DLK1-DIO3 Genomic Imprinted MicroRNA Cluster at 14q32.2 Defines a Stemlike Subtype of Hepatocellular Carcinoma Associated with Poor Survival

Hepatocellular carcinoma (HCC) is a heterogeneous and highly aggressive malignancy, for which there are no effective cures. Identification of a malignant stemlike subtype of HCC may offer patients with a dismal prognosis a potential targeted therapy using c-MET and Wnt pathway inhibitors. MicroRNAs (miRNAs) show promise as diagnostic and prognostic tools for cancer detection and stratification. Using a TRE-c-Met-driven transgenic HCC mouse model, we identified a cluster of 23 miRNAs that is encoded within the Dlk1-Gtl2 imprinted region on chromosome 12qF1 overexpressed in all of the isolated liver tumors. Interestingly, this region is conserved among mammalian species and maps to the human DLK1-DIO3 region on chromosome 14q32.2. We thus examined the expression of the DLK1-DIO3 miRNA cluster in a cohort of 97 hepatitis B virus-associated HCC patients and identified a subgroup (n = 18) of patients showing strong coordinate overexpression of miRNAs in this cluster but not in other cancer types (breast, lung, kidney, stomach, and colon) that were tested. Expression levels of imprinted gene transcripts from neighboring loci in this 14q32.2 region and from a subset of other imprinted sites were concomitantly elevated in human HCC. Interestingly, overexpression of the DLK1-DIO3 miRNA cluster was positively correlated with HCC stem cell markers (CD133, CD90, EpCAM, Nestin) and associated with a high level of serum α-fetoprotein, a conventional biomarker for liver cancer, and poor survival rate in HCC patients. In conclusion, our findings suggest that coordinate up-regulation of the DLK1-DIO3 miRNA cluster at 14q32.2 may define a novel molecular (stem cell-like) subtype of HCC associated with poor prognosis.