Abstract
PDGFRb is a receptor tyrosine kinase found in cells of mesenchymal origin such as fibroblasts and pericytes. Activation of this receptor is dependent on paracrine ligand induction, and its preferred ligand, PDGFB, is released by neighboring epithelial and endothelial cells. While expression of both PDGFRb and PDGFB has been noted in patient breast tumors for decades, how PDGFB-to-PDGFRb tumor-stromal signaling mediates breast cancer initiation, progression, and metastasis remains unclear. To test this important research question, we developed a mouse model of mesenchymal-specific PDGFRb hyper-activation. PDGFRb mutant mammary glands exhibit increased tertiary side-branching and epithelial proliferation confirming a stromalspecific PDGFRb effect on neighboring epithelium during normal development. To test the effect of hyper-active mesenchymal PDGFRb on tumor progression, orthotopicmammarytumor growth wasevaluated. Interestingly, whileweobserved no significant difference in primary tumor growth, the incidence of brain metastases from the orthotopic site was significantly increased in the mutant animals. These findings were confirmed using experimental tail vein metastasis assays wherewealso observed prominent brain metastases in 50% of the PDGFRb mutant mice (n ¼ 5/10) with no brain lesions seen in controls (n ¼ 0/19). In both the orthotopic and tail vein assays, there was no difference in the incidence of lung or liver metastases in the mutant mice suggesting a pro-metastatic function for PDGFRb in the brain metastatic niche. To rule out dysfunction of the blood brain barrier contributing to the observed metastatic spread, we then intracranially injected mammary tumor cells, and as expected based on our metastasis assay, found that larger tumors formed in the brains of PDGFRb mutant mice versus controls. To our knowledge, these combined findings are the first example where genetic manipulation of the stroma increases breast cancer associated brain metastases (BCBM). Given that these pre-clinical data suggest that primary breast tumors expressing high PDGFB could preferentially metastasize to the brain, we analyzed PDGFB protein expression in a tissue microarray comprised of HER2-positive and triple negative breast cancer (TNBC) primary tumors (total n ¼ 425). While high PDGFB did not correlate with site-independent metastatic recurrence, it was prognostic of brain metastasis, mirroring our mouse data. Evaluation of PDGFB in a small cohort of matched primary breast tumors with associated brain (n¼5) and lung metastases (n¼2) revealed intensePDGFBstaining in 100% of the brain metastases, but only 50% of the lung metastases. These findings further suggest that high primary tumor PDGFB expression defines a subset of breast cancer patients predisposed to brain metastases and that these patients may benefit from therapeutic inhibition of PDGFRb signaling. To test this pre-clinically, we treated mice harboring intracranial tumors with the PDGFR specific inhibitor crenolanib. Excitingly, crenolanib treatment significantly inhibited the brain tumor burden in these mice. Combined, our findings to date (1) advocate that primary tumor expression of PDGFB is a novel prognostic biomarker for the development of BCBM and (2) support clinical trial evaluation of PDGFR inhibitors for the prevention and treatment of BCBM. Ongoing studies are evaluating howthePDGFRb-expressing mesenchymal cells within the brain promote a pro-metastatic niche.
Keywords: PDGFRb, PDGF-B, Tumor microenvironment, Breast cancer.
DOI: 10.14343/JCSCR.2018.6e1001