Genome-scale systems biology analyses reveal potentially targetable signaling networks in breast cancer brain metastases

Abstract Background: Roughly 10-30% of metastatic breast cancer patients will develop brain metastases (BM) over the course of their disease. The prognosis for breast cancer patients with BM remains poor due to the deficiency of effective targeted therapies in this setting. A comprehensive characterization of deregulated signaling networks in breast cancer brain metastases would enable the delineation of pathways mediating metastases to the brain thus improving treatment options for this deadly disease. Using our novel systems biology paradigm (InFlo) that robustly infers deregulated signaling networks in individual tumor samples (Oncogene, 2017), we report genome-scale characterization of signaling network deregulations in breast cancer BM and matched primary tumor samples. Methods: Pathology records were searched for patients who underwent surgical resection of breast cancer BM and primary breast tumors (PB). Archival FFPE material and clinical data was obtained for a total of 76 breast cancer patients. Tumor DNA/RNA was extracted from 2 mm macrodissected FFPE cores followed by transcriptome profiling using Affymetrix HTA 2.0 arrays. Transcriptome profiles for a total of 133 FFPE samples included 72 BM and 31 PB, counting 31 BM/PB pairs from the same patient. PAM50 subtypes were derived for BM and PB samples, and InFlo was employed to interrogate genome scale signaling network deregulations in BM and PB as compared to normal breast controls. InFlo-derived signaling deregulations in the 72 BM were contrasted against the 31 PB using a Wilcoxon rank-sum test, followed by paired analyses in the matched 31 BM/PB pairs using the Wilcoxon signed-rank test. The immune microenvironment of BM/PB was assessed using model-based deconvolution of transcriptomic profiles, along with independent assessment of tumor infiltrating lymphocytes (TILs) in accordance with the International TILs Working Group standards. Results: We identified subtype switching between BM and matched PB in 5/31 pairs (16.1%) and loss of ER in 6/31 pairs (19.4%). Changes in receptor status were not associated with concurrent BM and only occurred in patients with two or more years between excision of the primary tumor and diagnosis/resection of BM. Consistent with prior studies, BM exhibited significantly lower TILs as compared to PB (P = 0.001) by both pathologist assessment and transcriptomic deconvolution. InFlo-based analyses revealed recurrent and subtype-independent hyperactivation of signaling networks in BM as compared to PB, including RhoA (62.5% of BM; P = 0.023) and anti-apoptotic BCL signaling (69.4% of BM; P = 0.003). In contrast, PPARα signaling was activated in 15.3% of BM (P = 0.003) and FGFR2 signaling was activated in 11.1% of BM (P = 0.005), with the Basal-like subtype exhibiting the highest rates of activation of these networks (PPARα, 21.7%; FGFR2, 17.4%). Activated HER3 signaling was observed in 13.9% of BM (P = 0.023) with the highest activation rates (28.6%) occurring within the Luminal A subtype. Conclusions: This first of its kind systems biology interrogation of genome-scale signaling network deregulations in BM in breast cancer reveals highly recurrent and subtype-specific pathway activations. The highly recurrent activation of pro-survival BCL-signaling specific to BM has high translational significance given recent positive developments of BH3 mimetics in the clinic. Furthermore, our observance of substantial differences in signaling network activities in brain metastases as compared to matched primary tumors points to the plasticity of breast cancer cells as they adapt to their metastatic niche. Overall, our study underscores the need to decode the molecular dependencies of metastatic lesions in order to develop effective therapeutic interventions targeting breast cancer metastases to the brain. Citation Format: Hannah Elizabeth Hill, Salendra Singh, Kristy Miskimen, Paula Silverman, Jill Barnholtz-Sloan, Andrew Sloan, Hannah Gilmore, Vinay Varadan. Genome-scale systems biology analyses reveal potentially targetable signaling networks in breast cancer brain metastases [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr PD8-09.