Fathi Elloumi

ACToR - Aggregated Computational Toxicology Resource

ACToR (Aggregated Computational Toxicology Resource) is a database and set of software applications that bring into one central location many types and sources of data on environmental chemicals. Currently, the ACToR chemical database contains information on chemical structure, in vitro bioassays and in vivo toxicology assays derived from more than 150 sources including the U.S. Environmental Protection Agency (EPA), Centers for Disease Control (CDC), U.S. Food and Drug Administration (FDA), National Institutes of Health (NIH), state agencies, corresponding government agencies in Canada, Europe and Japan, universities, the World Health Organization (WHO) and non-governmental organizations (NGOs). At the EPA National Center for Computational Toxicology, ACToR helps manage large data sets being used in a high-throughput environmental chemical screening and prioritization program called ToxCast.

Interactions with Fibroblasts are Distinct in Basal-like and Luminal Breast Cancers

Basal-like breast cancers have several well-characterized distinguishing molecular features, but most of these are features of the cancer cells themselves. The unique stromal–epithelial interactions, and more generally, microenvironmental features of basal-like breast cancers have not been well characterized. To identify characteristic microenvironment features of basal-like breast cancer, we performed cocultures of several basal-like breast cancer cell lines with fibroblasts and compared these with cocultures of luminal breast cancer cell lines with fibroblasts. Interactions between basal-like cancer cells and fibroblasts induced expression of numerous interleukins and chemokines, including IL-6, IL-8, CXCL1, CXCL3, and TGFβ. Under the influence of fibroblasts, basal-like breast cancer cell lines also showed increased migration in vitro. Migration was less pronounced for luminal lines; but, these lines were more likely to have altered proliferation. These differences were relevant to tumor biology in vivo, as the gene set that distinguished luminal and basal-like stromal interactions in coculture also distinguishes basal-like from luminal tumors with 98% accuracy in 10-fold cross-validation and 100% accuracy in an independent test set. However, comparisons between cocultures where cells were in direct contact and cocultures where interaction was solely through soluble factors suggest that there is an important impact of direct cell-to-cell contact. The phenotypes and gene expression changes invoked by cancer cell interactions with fibroblasts support the microenvironment and cell–cell interactions as intrinsic features of breast cancer subtypes. Mol Cancer Res; 9(1); 3–13 ©2010 AACR.

A comparison of machine learning algorithms for chemical toxicity classification using a simulated multi-scale data model

BackgroundBioactivity profiling using high-throughput in vitro assays can reduce the cost and time required for toxicological screening of environmental chemicals and can also reduce the need for animal testing. Several public efforts are aimed at discovering patterns or classifiers in high-dimensional bioactivity space that predict tissue, organ or whole animal toxicological endpoints. Supervised machine learning is a powerful approach to discover combinatorial relationships in complex in vitro/in vivo datasets. We present a novel model to simulate complex chemical-toxicology data sets and use this model to evaluate the relative performance of different machine learning (ML) methods.ResultsThe classification performance of Artificial Neural Networks (ANN), K-Nearest Neighbors (KNN), Linear Discriminant Analysis (LDA), Naïve Bayes (NB), Recursive Partitioning and Regression Trees (RPART), and Support Vector Machines (SVM) in the presence and absence of filter-based feature selection was analyzed using K-way cross-validation testing and independent validation on simulated in vitro assay data sets with varying levels of model complexity, number of irrelevant features and measurement noise. While the prediction accuracy of all ML methods decreased as non-causal (irrelevant) features were added, some ML methods performed better than others. In the limit of using a large number of features, ANN and SVM were always in the top performing set of methods while RPART and KNN (k = 5) were always in the poorest performing set. The addition of measurement noise and irrelevant features decreased the classification accuracy of all ML methods, with LDA suffering the greatest performance degradation. LDA performance is especially sensitive to the use of feature selection. Filter-based feature selection generally improved performance, most strikingly for LDA.ConclusionWe have developed a novel simulation model to evaluate machine learning methods for the analysis of data sets in which in vitro bioassay data is being used to predict in vivo chemical toxicology. From our analysis, we can recommend that several ML methods, most notably SVM and ANN, are good candidates for use in real world applications in this area.

NFκB Promotes Ovarian Tumorigenesis via Classical Pathways That Support Proliferative Cancer Cells and Alternative Pathways That Support ALDH+ Cancer Stem–like Cells

Understanding the mechanisms supporting tumor-initiating cells (TIC) is vital to combat advanced-stage recurrent cancers. Here, we show that in advanced ovarian cancers NFκB signaling via the RelB transcription factor supports TIC populations by directly regulating the cancer stem-like associated enzyme aldehyde dehydrogenase (ALDH). Loss of RelB significantly inhibited spheroid formation, ALDH expression and activity, chemoresistance, and tumorigenesis in subcutaneous and intrabursal mouse xenograft models of human ovarian cancer. RelB also affected expression of the ALDH gene ALDH1A2 Interestingly, classical NFκB signaling through the RelA transcription factor was equally important for tumorigenesis in the intrabursal model, but had no effect on ALDH. In this case, classical signaling via RelA was essential for proliferating cells, whereas the alternative signaling pathway was not. Our results show how NFκB sustains diverse cancer phenotypes via distinct classical and alternative signaling pathways, with implications for improved understanding of disease recurrence and therapeutic response. Cancer Res; 77(24); 6927-40. ©2017 AACR.

Abstract 2213: The normal breast study: Understanding the normal tissue adjacent to breast cancers. Epithelial-to-Mesenchymal transition (EMT) signatures

Proceedings: AACR 102nd Annual Meeting 2011‐‐ Apr 2‐6, 2011; Orlando, FL Significance: Breast Cancers (BC) evolve and acquire adaptive changes while in active communication with the surrounding host normal tissue. Understanding how host tissue interacts with cancers during breast cancer progression could lead to novel biomarkers or targeted therapies. Innovation: We hypothesized that novel molecular subtypes of microenvironment (ME) can be identified and have prognostic value. Stromal signatures are poorly understood and the independent prognostic value of stroma and/or ME response has not been widely studied. Approach: We used gene expression data 1) to identify molecular subtypes of microenvironment, 2) to study the distribution and prevalence of microenvironment subtypes in an ethnically diverse group of BC cases and 3) to test the value of ME in predicting breast cancer progression. In 2009, our team initiated the UNC NORMAL BREAST STUDY (NBS), a unique epidemiologic study of normal tissue from ethnically diverse patients at UNC Hospitals. The NBS has recruited over 200 patients undergoing breast surgery at UNC Hospitals, including cosmetic surgeries, excisional diagnostic breast biopsies, lumpectomies and mastectomies. All participants donate snap frozen and paraffin-embedded normal breast tissue. For all patients who have been diagnosed with breast cancer, the gross distance between the tumor and the normal breast tissue specimen is measured. In addition all study participants are asked to submit a blood sample and complete a telephone interview regarding demographics and environmental exposures. Medical record abstraction is performed to obtain treatment data and anthropometry. Each sample is carefully analyzed for histopathology and whole genome gene expression data is collected. Results: Among the breast cancer patients profiled in our initial studies, unsupervised clustering resulted in two groups of patients, one of which showed expression features suggestive of an activated mesenchyme. We then generated an EMT signature using cell line models and observed that this signature is enriched in one cluster or subgroup of these patients. In the approximately 40% of patients where the EMT signature is enriched, survival was decreased, and survival was significantly decreased among patients with ER positive disease. Conclusion: The NBS allows a unique opportunity to make an impact in discovering novel biology of human breast cancer microenvironment. The information gained from this translational study will establish whether microenvironment subtypes are associated with recurrence risk and will elucidate how variation in host biology contributes to BC disparities. 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 2213. doi:10.1158/1538-7445.AM2011-2213