Differential Network Analysis Applied to Preoperative Breast Cancer Chemotherapy Response

• Published in: PloS one Vol. 8; no. 12; p. e81784
• Main Authors: Warsow, Gregor, Struckmann, Stephan, Kerkhoff, Claus, Reimer, Toralf, Engel, Nadja, Fuellen, Georg
• Format: Journal Article
• Language: English
• Published: United States Public Library of Science 09.12.2013; Public Library of Science (PLoS)
• Subjects: Adjuvant chemotherapy; Aging; Anthracyclines; Apoptosis; Biomarkers; Biomarkers, Pharmacological - metabolism; Biotechnology; Breast cancer; Breast Neoplasms - drug therapy; Breast Neoplasms - genetics; Breast Neoplasms - pathology; Cancer; Cancer therapies; Cell Cycle Proteins - genetics; Cell Cycle Proteins - metabolism; Cell Line, Tumor; Chemotherapy; Clinical decision making; Data processing; Decision making; Doxorubicin; Doxorubicin - administration & dosage; Drug Synergism; Female; Gene expression; Gene Expression Regulation, Neoplastic - drug effects; Gene Regulatory Networks; Genes; Genomes; Hepatocellular carcinoma; Humans; Hypotheses; Informatics; Kinases; Medical prognosis; Medical schools; Medicine; Methods; Molecular modelling; Neoadjuvant Therapy; Network analysis; Paclitaxel; Paclitaxel - administration & dosage; Protein Interaction Mapping; Protein-Serine-Threonine Kinases - genetics; Protein-Serine-Threonine Kinases - metabolism; Proteins; Software; Software development tools; Synergism; Synergistic effect; Trans-Activators - genetics; Trans-Activators - metabolism; Tumor cell lines; Germany
• ISSN: 1932-6203; 1932-6203
• DOI: 10.1371/journal.pone.0081784

In silico approaches are increasingly considered to improve breast cancer treatment. One of these treatments, neoadjuvant TFAC chemotherapy, is used in cases where application of preoperative systemic therapy is indicated. Estimating response to treatment allows or improves clinical decision-making and this, in turn, may be based on a good understanding of the underlying molecular mechanisms. Ever increasing amounts of high throughput data become available for integration into functional networks. In this study, we applied our software tool ExprEssence to identify specific mechanisms relevant for TFAC therapy response, from a gene/protein interaction network. We contrasted the resulting active subnetwork to the subnetworks of two other such methods, OptDis and KeyPathwayMiner. We could show that the ExprEssence subnetwork is more related to the mechanistic functional principles of TFAC therapy than the subnetworks of the other two methods despite the simplicity of ExprEssence. We were able to validate our method by recovering known mechanisms and as an application example of our method, we identified a mechanism that may further explain the synergism between paclitaxel and doxorubicin in TFAC treatment: Paclitaxel may attenuate MELK gene expression, resulting in lower levels of its target MYBL2, already associated with doxorubicin synergism in hepatocellular carcinoma cell lines. We tested our hypothesis in three breast cancer cell lines, confirming it in part. In particular, the predicted effect on MYBL2 could be validated, and a synergistic effect of paclitaxel and doxorubicin could be demonstrated in the breast cancer cell lines SKBR3 and MCF-7.