Bioinformatics Analysis Of An Ad-Vegf Flank Angiogenesis Model Identifies Vessel Subtype Gene Signatures: Implications For Anti-Vegf Therapy

Proceedings: AACR Annual Meeting 2014; April 5-9, 2014; San Diego, CA Vascular Endothelial Growth Factor A (VEGF-A) is the predominant growth factor expressed by tumor cells to drive angiogenesis and solid tumor growth. Antiangiogenesis therapies have been clinically demonstrated to be effective, however, there are still not effective means to preselect those patients most likely to derive optimal benefit. To identify blood vessel markers that may differentiate patients responsive to VEGF pathway blockade, we used a well-established in vivo model to generate pathological angiogenesis following the delivery of adenovirus engineered to express VEGF-A164 (Ad-VEGF) to a localized subcutaneous site within the flank of a nude mouse. This model provides a temporally controlled induction of angiogenesis and subsequent remodeling and maturation of the vasculature. This enabled us to study discrete subclasses of vessel which have been previously shown to have differential sensitivities to anti-angiogenic therapy. Here we report gene expression profiling on a time course of the Ad-VEGF flank model which was used to characterize molecular and genetic changes associated with these distinct vessel subtypes. Using a nearest centroid-based classification algorithm we identified unique endothelial cell specific gene signatures representing ‘early’, ‘middle' and ‘late' stages of vessel development and maturation. To evaluate the utility of these gene signatures in predicting sensitivity to anti-VEGF therapy, we profiled a publically available gene array dataset generated from a mouse xenograft model following anti-VEGF treatment with our vessel subtype gene signatures. A majority of the early and middle vessel genes were significantly reduced in bevacizumab-treated animals, demonstrating they respond to VEGF pathway inhibition. To further evaluate the potential value of these gene signatures as predictive biomarkers we profiled a dataset of 248 gastric tumors using hierarchical clustering. Our analysis shows the middle vessel signature was highly expressed in a mesenchymal molecular subtype of gastric cancer. Interestingly, the majority of the mesenchymal gastric tumors were ‘diffuse' subtype in Lauren's classification. Together, our data demonstrate the potential usefulness of our vessel subtype gene signatures in identifying tumor subtypes that may benefit from this anti-angiogenic therapy. Citation Format: Jiangang Liu, Beverly L. Falcon, Janice A. Nagy, Shou-Ching S. Jaminet, Dan Li, Thompson N. Doman, Sudhakar Chintharlapalli, Bronek Pytowski, Mark T. Uhlik, Harold Dvorak, Laura Benjamin, Anthony S. Fischl. Bioinformatics analysis of an Ad-VEGF flank angiogenesis model identifies vessel subtype gene signatures: Implications for anti-VEGF therapy. [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 2359. doi:10.1158/1538-7445.AM2014-2359