[show abstract][hide abstract] ABSTRACT: First-generation molecular profiles for human breast cancers have enabled the identification of features that can predict therapeutic response; however, little is known about how the various data types can best be combined to yield optimal predictors. Collections of breast cancer cell lines mirror many aspects of breast cancer molecular pathobiology, and measurements of their omic and biological therapeutic responses are well-suited for development of strategies to identify the most predictive molecular feature sets.
We used least squares-support vector machines and random forest algorithms to identify molecular features associated with responses of a collection of 70 breast cancer cell lines to 90 experimental or approved therapeutic agents. The datasets analyzed included measurements of copy number aberrations, mutations, gene and isoform expression, promoter methylation and protein expression. Transcriptional subtype contributed strongly to response predictors for 25% of compounds, and adding other molecular data types improved prediction for 65%. No single molecular dataset consistently out-performed the others, suggesting that therapeutic response is mediated at multiple levels in the genome. Response predictors were developed and applied to TCGA data, and were found to be present in subsets of those patient samples.
These results suggest that matching patients to treatments based on transcriptional subtype will improve response rates, and inclusion of additional features from other profiling data types may provide additional benefit. Further, we suggest a systems biology strategy for guiding clinical trials so that patient cohorts most likely to respond to new therapies may be more efficiently identified.
[show abstract][hide abstract] ABSTRACT: Systemic chemotherapy in the adjuvant setting can cure breast cancer in some patients that would otherwise recur with incurable, metastatic disease. However, since only a fraction of patients would have recurrence after surgery alone, the challenge is to stratify high-risk patients (who stand to benefit from systemic chemotherapy) from low-risk patients (who can safely be spared treatment related toxicities and costs).
We focus here on risk stratification in node-negative, ER-positive, HER2-negative breast cancer. We use a large database of publicly available microarray datasets to build a random forests classifier and develop a robust multi-gene mRNA transcription-based predictor of relapse free survival at ten years, which we call the Random Forests Relapse Score (RFRS). Performance was assessed by internal cross-validation, multiple independent data sets, and comparison to existing algorithms using receiver operating characteristic and Kaplan-Meier survival analysis. Internal redundancy of features was determined using k-means clustering to define optimal signatures with smaller numbers of primary genes, each with multiple alternates.
Internal OOB cross-validation for the initial (full-gene-set) model on training data reported an ROC AUC of 0.704, which was comparable to or better than those reported previously or obtained by applying existing methods to our data set. Three risk groups with probability cut offs for low, intermediate and high-risk were defined. Survival analysis determined a highly significant difference in relapse rate between these risk groups. Validation of the models against independent test datasets showed highly similar results. Smaller 17-gene and 8-gene optimized models were also developed with minimal reduction in performance. Furthermore, the signature was shown to be almost equally effective on both hormone-treated and untreated patients.
RFRS allows flexibility in both the number and identity of genes utilized from thousands to as few as 17 or 8 genes, each with multiple alternatives. The RFRS reports a probability score strongly correlated with risk of relapse. This score could therefore be used to assign systemic chemotherapy specifically to those high-risk patients most likely to benefit from further treatment.
Genome Medicine 10/2013; 5(10):92. · 3.40 Impact Factor
[show abstract][hide abstract] ABSTRACT: Components of the plasminogen activation system (PAS) which are overexpressed in aggressive breast cancer subtypes offer appealing targets for development of new diagnostics and therapeutics. By comparing gene expression data in patient populations and cultured cell lines, we identified elevated levels of the urokinase plasminogen activation receptor (uPAR, PLAUR) in highly aggressive breast cancer subtypes and cell lines. Recombinant human anti-uPAR antagonistic antibodies exhibited potent binding in vitro to the surface of cancer cells expressing uPAR. In vivo these antibodies detected uPAR expression in triple negative breast cancer (TNBC) tumor xenografts using near infrared (NIR) imaging and (111)In single-photon emission computed tomography (SPECT). Antibody-based uPAR imaging probes accurately detected small disseminated lesions in a tumor metastasis model, complementing the current clinical imaging standard (18)F-fluorodeoxyglucose (FDG) at detecting non-glucose-avid metastatic lesions. A monotherapy study using the antagonistic antibodies resulted in a significant decrease in tumor growth in a TNBC xenograft model. Additionally, a radioimmunotherapy (RIT) study, using the anti-uPAR antibodies conjugated to the therapeutic radioisotope (177)Lu, found that they were effective at reducing tumor burden in vivo. Taken together, our results offer a preclinical proof-of-concept for uPAR targeting as a strategy for breast cancer diagnosis and therapy using this novel human antibody technology.
[show abstract][hide abstract] ABSTRACT: INTRODUCTION: Angiogenesis represents a potential therapeutic target in breast cancer. However, responses to targeted anti-angiogenic therapies have been reported to vary among patients. This suggests that the tumor vasculature may be heterogeneous and that an appropriate choice of treatment would require an understanding of these differences. METHODS: In order to investigate whether and how the breast tumor vasculature varies between individuals, we have isolated tumor-associated and matched normal vasculature from 17 breast carcinomas by laser capture microdissection, and generated gene expression profiles. Since microvessel density has previously been associated with disease course, tumors with low (n=9) or high (n=8) microvessel density were selected for analysis to maximize heterogeneity for this feature. RESULTS: We identify differences between tumor and normal vasculature, and describe two subtypes present within tumor vasculature. These subtypes exhibit distinct gene expression signatures that reflect features including hallmarks of vessel maturity. Potential therapeutic targets (MET, ITGAV and PDGFR) are differentially expressed between subtypes. Taking these subtypes into account has allowed us to derive a vascular signature associated with disease outcome. CONCLUSIONS: Our results further support a role for tumor microvasculature in determining disease progression. Overall, this study provides a deeper molecular understanding of the heterogeneity existing within the breast tumor vasculature, and opens new avenues towards the improved design and targeting of anti-angiogenic therapies.
Breast cancer research: BCR 08/2012; 14(4):R120. · 5.87 Impact Factor
[show abstract][hide abstract] ABSTRACT: Women older than 50 years account for 75% of new breast cancer diagnoses, and the majority of these tumors are of a luminal subtype. Although age-associated changes, including endocrine profiles and alterations within the breast microenvironment, increase cancer risk, an understanding of the cellular and molecular mechanisms that underlies these observations is lacking. In this study, we generated a large collection of normal human mammary epithelial cell strains from women ages 16 to 91 years, derived from primary tissues, to investigate the molecular changes that occur in aging breast cells. We found that in finite lifespan cultured and uncultured epithelial cells, aging is associated with a reduction of myoepithelial cells and an increase in luminal cells that express keratin 14 and integrin-α6, a phenotype that is usually expressed exclusively in myoepithelial cells in women younger than 30 years. Changes to the luminal lineage resulted from age-dependent expansion of defective multipotent progenitors that gave rise to incompletely differentiated luminal or myoepithelial cells. The aging process therefore results in both a shift in the balance of luminal/myoepithelial lineages and to changes in the functional spectrum of multipotent progenitors, which together increase the potential for malignant transformation. Together, our findings provide a cellular basis to explain the observed vulnerability to breast cancer that increases with age.
Cancer Research 05/2012; 72(14):3687-701. · 8.65 Impact Factor
[show abstract][hide abstract] ABSTRACT: Breast cancers are comprised of molecularly distinct subtypes that may respond differently to pathway-targeted therapies now under development. Collections of breast cancer cell lines mirror many of the molecular subtypes and pathways found in tumors, suggesting that treatment of cell lines with candidate therapeutic compounds can guide identification of associations between molecular subtypes, pathways, and drug response. In a test of 77 therapeutic compounds, nearly all drugs showed differential responses across these cell lines, and approximately one third showed subtype-, pathway-, and/or genomic aberration-specific responses. These observations suggest mechanisms of response and resistance and may inform efforts to develop molecular assays that predict clinical response.
Proceedings of the National Academy of Sciences 02/2012; 109(8):2724-9. · 9.74 Impact Factor
[show abstract][hide abstract] ABSTRACT: The liver represents the third most frequent site of metastasis in patients with breast cancer. We performed in vivo selection using 4T1 breast cancer cells to identify genes associated with the liver metastatic phenotype. Coincident with the loss of numerous tight-junctional proteins, we observe claudin-2 overexpression, specifically in liver-aggressive breast cancer cells. We further demonstrate that claudin-2 is both necessary and sufficient for the ability of 4T1 breast cancer cells to colonize and grow in the liver. The liver-aggressive breast cancer cells display a claudin-2-mediated increase in their ability to adhere to extracellular matrix (ECM) components, such as fibronectin and type IV collagen. Claudin-2 facilitates these cell/matrix interactions by increasing the cell surface expression of α(2)β(1)- and α(5)β(1)-integrin complexes in breast cancer cells. Indeed, claudin-2-mediated adhesion to fibronectin and type IV collagen can be blocked with neutralizing antibodies that target α(5)β(1) and α(2)β(1) complexes, respectively. Immunohistochemical analyses reveal that claudin-2, although weakly expressed in primary human breast cancers, is readily detected in all liver metastasis samples examined to date. Together, these results uncover novel roles for claudin-2 in promoting breast cancer adhesion to the ECM and define its importance during breast cancer metastasis to the liver.
[show abstract][hide abstract] ABSTRACT: Glycoprotein non-metastatic melanoma protein B (GPNMB)/Osteoactivin (OA) is a transmembrane protein expressed in approximately 40-75% of breast cancers. GPNMB/OA promotes the migration, invasion and metastasis of breast cancer cells; it is commonly expressed in basal/triple-negative breast tumors and is associated with shorter recurrence-free and overall survival times in patients with breast cancer. Thus, GPNMB/OA represents an attractive target for therapeutic intervention in breast cancer; however, little is known about the functions of GPNMB/OA within the primary tumor microenvironment.
We have employed mouse and human breast cancer cells to investigate the effects of GPNMB/OA on tumor growth and angiogenesis. GPNMB/OA-expressing tumors display elevated endothelial recruitment and reduced apoptosis when compared to vector control-derived tumors. Primary human breast cancers characterized by high vascular density also display elevated levels of GPNMB/OA when compared to those with low vascular density. Using immunoblot and ELISA assays, we demonstrate the GPNMB/OA ectodomain is shed from the surface of breast cancer cells. Transient siRNA-mediated knockdown studies of known sheddases identified ADAM10 as the protease responsible for GPNMB/OA processing. Finally, we demonstrate that the shed extracellular domain (ECD) of GPNMB/OA can promote endothelial migration in vitro.
GPNMB/OA expression promotes tumor growth, which is associated with enhanced endothelial recruitment. We identify ADAM10 as a sheddase capable of releasing the GPNMB/OA ectodomain from the surface of breast cancer cells, which induces endothelial cell migration. Thus, ectodomain shedding may serve as a novel mechanism by which GPNMB/OA promotes angiogenesis in breast cancer.
PLoS ONE 01/2010; 5(8):e12093. · 3.73 Impact Factor
[show abstract][hide abstract] ABSTRACT: The tumour stroma is believed to contribute to some of the most malignant characteristics of epithelial tumours. However, signalling between stromal and tumour cells is complex and remains poorly understood. Here we show that the genetic inactivation of Pten in stromal fibroblasts of mouse mammary glands accelerated the initiation, progression and malignant transformation of mammary epithelial tumours. This was associated with the massive remodelling of the extracellular matrix (ECM), innate immune cell infiltration and increased angiogenesis. Loss of Pten in stromal fibroblasts led to increased expression, phosphorylation (T72) and recruitment of Ets2 to target promoters known to be involved in these processes. Remarkably, Ets2 inactivation in Pten stroma-deleted tumours ameliorated disruption of the tumour microenvironment and was sufficient to decrease tumour growth and progression. Global gene expression profiling of mammary stromal cells identified a Pten-specific signature that was highly represented in the tumour stroma of patients with breast cancer. These findings identify the Pten-Ets2 axis as a critical stroma-specific signalling pathway that suppresses mammary epithelial tumours.
[show abstract][hide abstract] ABSTRACT: Although it is increasingly evident that cancer is influenced by signals emanating from tumor stroma, little is known regarding how changes in stromal gene expression affect epithelial tumor progression. We used laser capture microdissection to compare gene expression profiles of tumor stroma from 53 primary breast tumors and derived signatures strongly associated with clinical outcome. We present a new stroma-derived prognostic predictor (SDPP) that stratifies disease outcome independently of standard clinical prognostic factors and published expression-based predictors. The SDPP predicts outcome in several published whole tumor-derived expression data sets, identifies poor-outcome individuals from multiple clinical subtypes, including lymph node-negative tumors, and shows increased accuracy with respect to previously published predictors, especially for HER2-positive tumors. Prognostic power increases substantially when the predictor is combined with existing outcome predictors. Genes represented in the SDPP reveal the strong prognostic capacity of differential immune responses as well as angiogenic and hypoxic responses, highlighting the importance of stromal biology in tumor progression.
Nature medicine 06/2008; 14(5):518-27. · 27.14 Impact Factor
[show abstract][hide abstract] ABSTRACT: The skeleton is a preferred site of metastasis in patients with disseminated breast cancer. We have used 4T1 mouse mammary carcinoma cells, which metastasize to bone from the mammary fat pads of immunocompetent mice, to identify novel genes involved in this process. In vivo selection of parental cells resulted in the isolation of independent, aggressively bone metastatic breast cancer populations with reduced metastasis to the lung. Gene expression profiling identified osteoactivin as a candidate that is highly and selectively expressed in aggressively bone metastatic breast cancer cells. These cells displayed enhanced migratory and invasive characteristics in vitro, the latter requiring sustained osteoactivin expression. Osteoactivin depletion in these cells, by small interfering RNA, also lead to a loss of matrix metalloproteinase-3 expression, whereas forced osteoactivin expression in parental 4T1 cells was sufficient to elevate matrix metalloproteinase-3 levels, suggesting that this matrix metalloproteinase may be an important mediator of osteoactivin function. Overexpression of osteoactivin in an independent, weakly bone metastatic breast cancer cell model significantly enhanced the formation of osteolytic bone metastases in vivo. Finally, high levels of osteoactivin expression in primary human breast cancers correlate with estrogen receptor-negative status and increasing tumor grade. Thus, we have identified osteoactivin as a protein that is expressed in aggressive human breast cancers and is capable of promoting breast cancer metastasis to bone.
Molecular Cancer Research 11/2007; 5(10):1001-14. · 4.35 Impact Factor
[show abstract][hide abstract] ABSTRACT: The role of the cellular microenvironment in breast tumorigenesis has become an important research area. However, little is known about gene expression in histologically normal tissue adjacent to breast tumor, if this is influenced by the tumor, and how this compares with non-tumor-bearing breast tissue.
To address this, we have generated gene expression profiles of morphologically normal epithelial and stromal tissue, isolated using laser capture microdissection, from patients with breast cancer or undergoing breast reduction mammoplasty (n = 44).
Based on this data, we determined that morphologically normal epithelium and stroma exhibited distinct expression profiles, but molecular signatures that distinguished breast reduction tissue from tumor-adjacent normal tissue were absent. Stroma isolated from morphologically normal ducts adjacent to tumor tissue contained two distinct expression profiles that correlated with stromal cellularity, and shared similarities with soft tissue tumors with favorable outcome. Adjacent normal epithelium and stroma from breast cancer patients showed no significant association between expression profiles and standard clinical characteristics, but did cluster ER/PR/HER2-negative breast cancers with basal-like subtype expression profiles with poor prognosis.
Our data reveal that morphologically normal tissue adjacent to breast carcinomas has not undergone significant gene expression changes when compared to breast reduction tissue, and provide an important gene expression dataset for comparative studies of tumor expression profiles.
Breast cancer research: BCR 02/2006; 8(5):R58. · 5.87 Impact Factor
[show abstract][hide abstract] ABSTRACT: High throughput genomic/proteomic strategies, such as microarray studies, drug screens, and genetic screens, often produce a list of genes that are believed to be important for one or more reasons. Unfortunately it is often difficult to discern meaningful biological relationships from such lists. This study presents a new bioinformatic approach that can be used to identify regulatory subnetworks for lists of significant genes or proteins. We demonstrate the utility of this approach using an interaction network for yeast constructed from BIND, TRANSFAC, SCPD, and chromatin immunoprecipitation (ChIP)-Chip data bases and lists of genes from well known metabolic pathways or differential expression experiments. The approach accurately rediscovers known regulatory elements of the heat shock response as well as the gluconeogenesis, galactose, glycolysis, and glucose fermentation pathways in yeast. We also find evidence supporting a previous conjecture that approximately half of the enzymes in a metabolic pathway are transcriptionally co-regulated. Finally we demonstrate a previously unknown connection between GAL80 and the diauxic shift in yeast.
[show abstract][hide abstract] ABSTRACT: MOTIVATION: We introduce a development platform especially tailored to Bioinformatics research and software development. BIAS (Bioinformatics Integrated Application Software) provides the tools necessary for carrying out integrative Bioinformatics research requiring multiple datasets and analysis tools. It follows an object-relational strategy for providing persistent objects, allows third-party tools to be easily incorporated within the system and supports standards and data-exchange protocols common to Bioinformatics. AVAILABILITY: BIAS is an OpenSource project and is freely available to all interested users at http://www.mcb.mcgill.ca/~bias/. This website also contains a paper containing a more detailed description of BIAS and a sample implementation of a Bayesian network approach for the simultaneous prediction of gene regulation events and of mRNA expression from combinations of gene regulation events. CONTACT: email@example.com.
[show abstract][hide abstract] ABSTRACT: Abstract Motivation:,We introduce a development,platform,especially tailored to Bioinformatics research and software development.,BIAS (Bioinformatics Integrated Application Software) provides the tools necessary for carrying out integrative Bioinformatics research requiring multiple datasets and analysis tools. It follows an object-relational strategy for providing persistent objects, allows third-party tools to be easily incorporated within the system, and supports,standards,and data-exchange,protocols common,to Bioinformatics. Availability: BIAS is an OpenSource project and is freely available to all interested users at http://www.mcb.mcgill.ca/�bias. This web site also contains a paper containing a more,detailed description of BIAS and an example implementation,of a Bayesian network,approach,for the simultaneous prediction of gene regulation events and of mRNA expression from combinations of gene regulation events. Contact: firstname.lastname@example.org There now exists a wide range of ontologies, standards, databases, and applications available to the Bioinformatics community. However, the vast majority of these objects exist independently,and are not tightly integrated into a single system. This situation makes,it difficult for research groups to quickly and effectively use state-of-the-art Bioinformatics tools. Furthermore, it necessitates a great effort on their part to both integrate all the relevant objects together into a usable platform and to maintain this system over time. The slow progress towards,a development,framework for Bioinformatics is due to a complex,set of reasons including the heterogeneity and