Search

Supplementary Figures PDF file - 529K, Supplementary Figure 1: Application of the Gaussian mixture model to ESR1- and ERBB2-expression data Supplementary Figure 2: Normalization Supplementary Figure 3: Silhouette Analysis Supplementary Figure 4: Silhouette Analysis (P-Values) Supplementary Figure 5: Validation of Limma-corrected IBC/nIBC signature Supplementary Figure 6: Networks overrepresented in IBC Supplementary Figure 7: Networks overrepresented in nIBC Supplementary Figure 8: Prognostic value of the IBC/nIBC signature in nIBC

Supplementary Data File and Legend PDF file - 113K, Supplementary Data File and legend

Supplementary Tables XLS file - 236K, Supplementary Table 1: Comparison of ER- and ErbB2-status determined using gene expression profiling and immunohistochemistry Supplementary Table 2: Validation of the algorithm to identify transcription factor activation Supplementary Table 3: Description of the public nIBC data sets Supplementary Table 4: Molecular data of IBC and nIBC samples Supplementary Table 5: List of 491 probe sets differentially expressed between IBC and nIBC samples in a molecular subtype-independent manner Supplementary Table 6: IPA analysis of the list of 443 unique genes differentially expressed between IBC and nIBC Supplementary Table 7: Networks identified by IPA analysis of the list of 443 unique genes differentially expressed between IBC and nIBC Supplementary Table 8: Uni- and multivariate analyses for DMFS in the public pooled nIBC data set

Background: Inflammatory breast cancer (IBC) is a poorly characterized form of breast cancer. So far, the results of expression profiling in IBC are inconclusive due to various reasons including limited sample size. Here, we present the integration of three Affymetrix expression datasets collected through the World IBC Consortium allowing us to interrogate the molecular profile of IBC using the largest series of IBC samples ever reported.Experimental Design: Affymetrix profiles (HGU133-series) from 137 patients with IBC and 252 patients with non-IBC (nIBC) were analyzed using unsupervised and supervised techniques. Samples were classified according to the molecular subtypes using the PAM50-algorithm. Regression models were used to delineate IBC-specific and molecular subtype-independent changes in gene expression, pathway, and transcription factor activation.Results: Four robust IBC-sample clusters were identified, associated with the different molecular subtypes (P < 0.001), all of which were identified in IBC with a similar prevalence as in nIBC, except for the luminal A subtype (19% vs. 42%; P < 0.001) and the HER2-enriched subtype (22% vs. 9%; P < 0.001). Supervised analysis identified and validated an IBC-specific, molecular subtype-independent 79-gene signature, which held independent prognostic value in a series of 871 nIBCs. Functional analysis revealed attenuated TGF-β signaling in IBC.Conclusion: We show that IBC is transcriptionally heterogeneous and that all molecular subtypes described in nIBC are detectable in IBC, albeit with a different frequency. The molecular profile of IBC, bearing molecular traits of aggressive breast tumor biology, shows attenuation of TGF-β signaling, potentially explaining the metastatic potential of IBC tumor cells in an unexpected manner. Clin Cancer Res; 19(17); 4685–96. ©2013 AACR.

Purpose Inflammatory breast cancer (IBC) is an aggressive form of breast cancer with elevated metastatic potential, characterized by tumor emboli in dermal and parenchymal lymph vessels. This study has investigated the hypothesis that TGFβ signaling is implicated in the molecular biology of IBC. Methods TGFβ1-induced cell motility and gene expression patterns were investigated in three IBC and three non-IBC (nIBC) cell lines. Tissue samples from IBC and nIBC patients were investigated for the expression of nuclear SMAD2, SMAD3, and SMAD4. SMAD protein levels were related to gene expression data. Results TGFβ1-induced cell motility was strongly abrogated in IBC cells (P = 0.003). Genes differentially expressed between IBC and nIBC cells post TGFβ1 exposure revealed attenuated expression of SMAD3 transcriptional regulators, but overexpression of MYC target genes in IBC. IBC patient samples demonstrated a near absence of SMAD3 and -4 expression in the primary tumor compared to nIBC patient samples (P

Introduction: Inflammatory Breast Cancer (IBC) is an aggressive and highly metastatic form of breast cancer. Recent expression profiling studies revealed IBC-specific modulation of TGFβ-signalling. In the present study, we aim to validate these findings by evaluating nuclear SMAD staining patterns. Materials and Methods: Immunohistochemistry (IHC) was performed for SMAD2, -3, -4 and -6 on tissue sections from patients with (N=79) and without (N=133) IBC. Protocols for IHC were established on cell blocks from breast cancer cell lines with and without SMAD staining identified by western blotting on nuclear protein extracts. A staining score was assigned by multiplying the percentage of stained cancer cell nuclei by the staining intensity evaluated on a three-scale basis (1=weak, 2=intermediate, 3=strong). Expression data were subjected to unsupervised hierarchical cluster analysis (UHCA) and statistical assessment of staining differences was done using uni- and multivariate models. Additionally, we looked at the correlation between nuclear SMAD expression and gene expression in 11 patients. For each gene, we generated a spearman correlation coefficient, and selected only those genes that were positively correlated with nuclear SMAD expression in 11 patients (p Results: UHCA for SMAD2, -3 and -4 nuclear protein expression data identified two sample clusters downstream of the first bifurcation. IBC samples were significantly (OR=60.34, P Discussion: In line with the expectations, our data show that TGFβ-signalling indeed differs between samples from patients with and without IBC. In addition, this study does offer novel insights on IBC biology. First, SMAD2 nuclear staining is gained in IBC in the absence of its canonical binding partners SMAD3 and -4, suggesting non-canonical TGFβ-signalling in IBC. Second, it has been shown that SMAD2 induces an invasive response program in cancer cells without affecting their epithelial morphology, while SMAD3 accounts for invasion by induction of Epithelial-to-Mesenchymal transition (EMT). Thus our results indicate that IBC cell invasion occurs in a collective, sheet-like fashion instead of EMT, which is probably responsible for the invasion of non-IBC cells. Citation Format: Melike Marsan, Gert Van den Eynden, Patrick Neven, Ignace Vergote, Peter Vermeulen, Luc Dirix, Steven Van Laere. SMAD nuclear staining patterns in inflammatory breast cancer suggest non-canonical TGFβ-signalling establishing collective, sheet-like invasion SMAD nuclear staining patterns in inflammatory breast cancer suggest non-canonical TGFβ-signalling es [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P1-07-09.

Recent studies suggest that anaplastic lymphoma kinase (ALK) could serve as a therapeutic target for inflammatory breast cancer (IBC) and that strategies targeting ALK should be considered for evaluation in clinical trials. Reverse phase protein arays revealed activation of the ALK receptor tyrosine kinase and biochemically-linked downstream signalling molecules in pre-clinical models of IBC. ALK genetic abnormalities have been reported in 80% (20/25) of IBC patients with a high prevalence of ALK alterations in basal-like IBC (Robertson F. et al., 2013). However, ALK protein expression has not been studied in human IBC samples. Immunohistochemical detection of the ALK protein is increasingly being used as a screening tool to test samples for ALK rearrangements. The biological premise is that the genetic abnormality of the ALK gene leads to overexpression of the ALK protein and therefore overactivity of the ALK tyrosine kinase; since this kinase is the target of crizotinib, it makes sense to assess the drug target directly. Formalin fixed paraffin embedded tissue samples from pre-treatment surgical biopsies of 79 consecutive IBC patients were immunohistochemically (IHC) tested for ALK protein over-expression using a validated IHC test (NCL-ALK, clone 5A4) with a sensitivity of 93% and a specificity of 100% when the Vysis ALK-FISH break apart test is used as a gold standard. Stained tissue sections were evaluated using the IHC histoscore proposed by Thunnissen E. et al., 2012, assessing both the proportion of tumor cells showing cytoplasmic staining and the staining intensity. Cytoplasmic staining of appendiceal ganglion cells was used as an on-slide external positive control; weak to moderate cytoplasmic staining of macrophages was used as an internal positive control. In 75 of the 79 tissue samples none of the tumor cells showed any ALK immunoreactivity. One tumor showed moderate cytoplasmic staining in a few cells ( These results demonstrate that ALK protein overexpression is not a feature of IBC. Although genetic abnormalities of ALK without protein overexpression are not excluded, our results show that ALK IHC cannot be used to identify IBC patients eligible for enrollment in clinical trials evaluating ALK targeted therapeutics. Citation Format: Cecile Colpaert, Melike Marsan, Peter Vermeulen, Luc Dirix, Steven Van Laere, Inflammatory Breast Cancer International Consortium. Anaplastic lymphoma kinase (ALK) protein overexpression is not a feature of inflammatory breast cancer [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P3-06-39.

Introduction: Genome-wide expression profiling of samples from patients with and without Inflammatory Breast Cancer (IBC) has revealed novel insights into the biology of IBC. The present study was undertaken to compare these novel insights with data obtained from all available preclinical IBC models including 2 new models that we have recently developed that recapitulate the characteristics of IBC including retention of E-cadherin, formation of tumor emboli and encircling lymphoangiogenesis Materials and Methods: Five replicates of 7 preclinical IBC models (SUM149, SUM190, FC-IBC-01, FC-IBC-02, MDA-IBC-03, KPL-4, and Mary-X) were profiled using Affymetrix HGU133plus2 GeneChips. Using a nearest shrunken centroid algorithm, each expression profile was classified according to an IBC-specific signature identified in patient samples. Available expression profiles were further queried for expression patterns related to Epithelial-to-Mesenchymal Transition (EMT), TGFβ-signaling and IBC-specific patterns of transcription factor activation. Results: Application of our IBC-specific signature (posterior probabilities exceeded 0.50 in at least 4/5 replicates) revealed that out of 7 pre-clinical models of IBC, 3 of these robustly classified as IBC (FC-IBC-01, FC-IBC-02, and KPL-4). All preclinical IBC models were characterized by retention of E-Cadherin expression, absence of ZEB1 expression, attenuated expression of specific components of the TGFβ pathway (TGFβR2, SMAD3, SMAD7, and TGFβ1), and ambiguous activation patterns of several transcription factors involved in regulating cellular plasticity and cell fate decisions (Up in IBC: NR4A2, RARB/RXRA, PTX3, GSC2, and ZEB1; Down in IBC: SOX10, PAX5, and SMAD2). For each of the molecular alterations described above, Z-scores greater than 2 were achieved in at least 4/5 replicates. Conclusions: The observations that we have made using IBC patient tumor samples with regards to EMT, cell plasticity and TGFβ-signaling are corroborated in pre-clinical models of IBC using current analytic approaches, despite the fact that expression patterns of the majority of preclinical models of IBC deviate from the IBC-specific expression patterns observed in patient samples. Our data suggest that despite their highly invasive nature, IBC cancer cells retain an epithelial cell phenotype characterized by E-cadherin expression and loss of ZEB1 which appears to be mediated by, amongst others, attenuated TGFβ-signaling. This study strengthens our hypothesis that cancer cells from IBC exhibit cohesive invasion, and invade as a unit, possibly explaining the presence of florid tumour emboli which is a primary characteristic observed in IBC. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P2-05-04.

Background: Inflammatory breast cancer (IBC) is a poorly characterized form of breast cancer. So far, the results of expression profiling in IBC are inconclusive due to various reasons including limited sample size. Here, we present the integration of three Affymetrix expression datasets collected through the World IBC Consortium allowing us to interrogate the molecular profile of IBC using the largest series of IBC samples ever reported. Experimental Design: Affymetrix profiles (HGU133-series) from 137 patients with IBC and 252 patients with non-IBC (nIBC) were analyzed using unsupervised and supervised techniques. Samples were classified according to the molecular subtypes using the PAM50-algorithm. Regression models were used to delineate IBC-specific and molecular subtype-independent changes in gene expression, pathway, and transcription factor activation. Results: Four robust IBC-sample clusters were identified, associated with the different molecular subtypes (P < 0.001), all of which were identified in IBC with a similar prevalence as in nIBC, except for the luminal A subtype (19% vs. 42%; P < 0.001) and the HER2-enriched subtype (22% vs. 9%; P < 0.001). Supervised analysis identified and validated an IBC-specific, molecular subtype-independent 79-gene signature, which held independent prognostic value in a series of 871 nIBCs. Functional analysis revealed attenuated TGF-β signaling in IBC. Conclusion: We show that IBC is transcriptionally heterogeneous and that all molecular subtypes described in nIBC are detectable in IBC, albeit with a different frequency. The molecular profile of IBC, bearing molecular traits of aggressive breast tumor biology, shows attenuation of TGF-β signaling, potentially explaining the metastatic potential of IBC tumor cells in an unexpected manner. Clin Cancer Res; 19(17); 4685–96. ©2013 AACR.

Introduction: Metastases remain the primary cause of cancer-related death. The acquisition of invasive tumour cell behaviour is thought to be a cornerstone of the metastatic cascade. Therefore, gene signatures related to invasiveness could aid in stratifying patients according to their prognostic profile. In the present study we aimed at identifying an invasiveness gene signature and investigated its biological relevance in breast cancer. Methods & Results: We collected a set of published gene signatures related to cell motility and invasion. Using this collection, we identified 16 genes that were represented at a higher frequency than observed by coincidence, hereafter named the core invasiveness gene signature. Principal component analysis showed that these overrepresented genes were able to segregate invasive and non-invasive breast cancer cell lines, outperforming sets of 16 randomly selected genes (all P

Inflammatory breast cancer (IBC) is an aggressive disease. To date, no molecular feature reliably predicts either the response to chemotherapy (CT) or the survival. Using DNA microarrays, we searched for multigene predictors.The World IBC Consortium generated whole-genome expression profiles of 137 IBC and 252 non-IBC (nIBC) samples. We searched for transcriptional profiles associated with pathological complete response (pCR) to neoadjuvant anthracycline-based CT and distant metastasis-free survival (DMFS) in respective subsets of 87 and 106 informative IBC samples. Correlations were investigated with predictive and prognostic gene expression signatures published in nIBC (nIBC-GES). Supervised analyses tested genes and activation signatures of 19 biological pathways and 234 transcription factors.Three of five tested prognostic nIBC-GES and the two tested predictive nIBC-GES discriminated between IBC with and without pCR, as well as two interferon activation signatures. We identified a 107-gene signature enriched for immunity-related genes that distinguished between responders and nonresponders in IBC. Its robustness was demonstrated by external validation in three independent sets including two IBC sets and one nIBC set, with independent significant predictive value in IBC and nIBC validation sets in multivariate analysis. We found no robust signature associated with DMFS in patients with IBC, and neither of the tested prognostic GES, nor the molecular subtypes were informative, whereas they were in our nIBC series (220 stage I-III informative samples).Despite the relatively small sample size, we show that response to neoadjuvant CT in IBC is, as in nIBC, associated with immunity-related processes, suggesting that similar mechanisms responsible for pCR exist. Analysis of a larger IBC series is warranted regarding the correlation of gene expression profiles and DMFS.

e23008Background: Inflammatory breast cancer (IBC) is rare and aggressive type of breast cancer, characterized by its rapid invasion and spread. Recent evidence suggests that TGFβ signaling may be ...

Inflammatory breast cancer (IBC) is a rare and aggressive form of breast cancer with poor prognosis. Multimodal treatment including neo-adjuvant chemotherapy is the treatment of choice for IBC patients, with pathological complete response (pCR) being one of the most important prognostic factors. Tumor infiltrating lymphocytes (TILs) are an independent predictor of response to neo-adjuvant chemotherapy in breast cancer (Denkert C. et al., 2010). In a search for multigene predictors of pCR in IBC using DNA micro-arrays, a 107 gene signature was found that distinguished between responders and non-responders. This signature was enriched for immunity-related genes showing that in IBC, as in non-IBC, response to neo-adjuvant chemotherapy is associated with immunity related processes (Bertucci F. et al., 2014). Standard H&E stained sections of formalin-fixed paraffin-embedded pre-treatment tumor tissue of 77 IBC patients were used to evaluate the presence of TILs according to the Denkert method: mononuclear cell infiltrates were considered to be intratumoral TILs when making direct contact with the tumor cells, while stromal TILs are mononuclear cells being present in the tumor stroma without making direct contact with tumor cells. Intratumoral TILs were observed in 10,4% (8/77) of the patients. Stromal TILs (strTILs ) were present in all tumors: in 35% (27/77) strTILs occupied less than 5% of the tumor stroma, in 62 % (48/77) strTILs occupied between 5 and 50% of the tumor stroma and 2,6% (2/77) were "lymphocyte predominant breast cancers (LPBC)" with strTILS occupying more than 50% of the tumor stroma. One LPBC was HER2 positive, the other one was triple negative. There was a significant association of strTILs with response to neo-adjuvant chemotherapy: pCR was obtained in 13% of the patients with strTILs Tertiary lymphoid structures - lymphoid follicles with germinal centres, known to be associated with better clinical outcomes (Gu-Trantien C. et al., 2013) - were detected in 2.6% of the patients (2/77); both patients had a disease free survival of more than 3 years. This study shows that IBC tumors do not contain more TILs than non-IBC tumors and that, as in non-IBC, there is a positive association of stromal TILs with pCR in IBC. Further validation of these results and further study of lymphocyte subsets in IBC is warranted. Citation Format: Cecile Colpaert, Melike Marsan, Peter Vermeulen, Luc Dirix, Steven Van Laere, Inflammatory Breast Cancer International Consortium. Tumor infiltrating lymphocytes in inflammatory breast cancer [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P4-04-20.

Proceedings: AACR 103rd Annual Meeting 2012‐‐ Mar 31‐Apr 4, 2012; Chicago, IL Background: Patients with estrogen receptor (ER)-positive breast cancers are generally treated with endocrine therapy, but in a considerable fraction of patients resistance occurs. Previous research showed that nuclear factor kappa B (NFkB) may play a role in the development of resistance to hormonal therapy. In this study, we investigated the role of NFkB, and its interaction with ER in the development of endocrine resistance by studying breast cancer cell lines with variable ER and NFkB activity. Material and methods: Three gene expression data sets of breast cancer cell lines ([GSE12777][1]: Hoeflich et al, 2009; [GSE16795][2]: Hollestelle et al, 2009; E-TAM-157: Neve et al, 2005) were retrieved from Gene Expression Omnibus or ArrayExpress. The expression profiles were screened for ER and NFkB target genes to identify cell lines with strong and weak activation of ER and NFkB. Based on these activation profiles, 8 cell lines (MCF7, BT20, CAMA1, MDA-MB-361, MDA-MB-231, MDA-MB-134, MDA-MB-468 and MDA-MB-436) with high and low ER and NFkB activation profiles were selected. Using qRT-PCR on RNA from these cell lines, the expression of 6 biomarkers for endocrine resistance (ADAMDEC1, ITK, ABAT, CLEC7A, ETS1 and STC2) were profiled. 18S and beta-Actin were used as housekeeping genes. Relative expression levels were analyzed in function of ER and NFkB activation. Results: There is a significant lower expression of ADAMDEC1 (P=0.011), ITK (P=0.012) and ETS1 (P

Introduction. We identified a core invasiveness gene (CIG) signature that predicts the invasive properties of breast cancer cell lines. In the current study, we investigate the translational utility of the current gene signature in human breast cancer, including inflammatory breast cancer (IBC). The latter is a highly invasive and metastatic breast cancer subtype. Materials and methods. Six publicly available gene expression data sets and a data set on 137 IBC and 252 non-IBC samples were analyzed. Each sample was classified according to sets of stromal (N=2), prognostic (N=2), stem cell (N=3), epithelial-to-mesenchymal transition (EMT) (N=3) and pathway (N=20) gene signatures. In addition, the cell-of-origin subtype classifier and our own CIG-signature were applied. Associations between the CIG-signature and clinicopathological or molecular features were searched for using univariate and multivariate analysis. Results. We identified associations of the CIG signature (FDR Discussion. We show that the CIG signature in breast cancer is correlated with EMT, stem cell biology and EGFR-, TGFb- and STAT3-activation. Samples from patients with IBC demonstrate ambiguous EMT-patterns, suggest that tumor cells from patients with IBC are in a specific state of cell plasticity and EMT as such is not the main mode of invasion in IBC. On the contrary, lowered TGFb signaling in IBC suggest a form of collective invasion. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr 3427. doi:10.1158/1538-7445.AM2011-3427