<em>PBRM1</em> loss in clear cell renal cell carcinoma leads to a proangiogenic phenotype via the CXCL5/CXCR2 axis that can be targeted by CXCR2 inhibition

Abstract

Immune checkpoint blockade (ICB) and anti-angiogenic tyrosine kinase inhibitors (TKI) have substantially improved the outcomes of metastatic clear cell renal cell carcinoma (ccRCC). Nevertheless, the anti-angiogenic TKIs targeting the VEGF/ VEGFR axis result only in temporary antitumor response, and most ccRCCs become ultimately TKI-resistant. Mutations of polybromo-1 (<em><strong>PBRM1</strong></em>) occur in about one-third of ccRCC. It is well known that <em><strong>PBRM1</strong></em>-mutated tumors are highly vascularized. This project aims to dissect the molecular mechanism driving the proangiogenic phenotype of <em><strong>PBRM1</strong></em>-mutated ccRCC. In an unbiased approach, pharmacological inhibition of PBRM1 in ccRCC cell line leads to strong upregulation of the CXCR2 ligands <em><strong>CXCL5</strong></em> and <em><strong>CXCL6</strong></em>. Accordingly, our TCGA in silico analysis revealed that <em><strong>PBRM1</strong></em>-mutant ccRCC exhibited enhanced <em><strong>CXCL1</strong></em>, <em><strong>CXCL2</strong></em>, <em><strong>CXCL3</strong></em>, and <em><strong>CXCL5</strong></em> expression; furthermore, patients who had high CXCL5 mRNA expression are associated with worse prognosis in TCGA-KIRC. We established CRISPR-Cas9 induced polyclonal <em><strong>PBRM1</strong></em>-knockouts in the ccRCC cell lines 786O and Caki1. We also identified elevated secretion of CXCL1, CXCL2, CXCL5, CXCL6, and CXCL8, all CXCR2-activating chemokines in <em><strong>PBRM1</strong></em>-loss cell lines. Exposing the cells to pro-inflammatory cytokines (TNF&alpha;, and IL-17A) leads to hyper-induction of these chemokines. <em><strong>PBRM1</strong></em>-KO conditioned-media (CM) significantly promoted HUVECs viability and culminated in the enhanced phosphorylation of GSK-3β compared to <em><strong>PBRM1</strong></em>-wild-type CM. The addition of CXCR2 blockade to ccRCC not only suppressed the inhibition of GSK3β signaling, but also the proliferation of HUVECs. The association between <em><strong>PBRM1</strong></em>-deficiency and an enhanced tumor angiogenesis was recapitulated using the spheroid sprouting assay, a three-dimensional (3D) in vitro angiogenic model. Augmented angiogenic sprouting was detected in HUVEC-spheroids which were incubated with supernatant from <em><strong>PBRM1</strong></em>-loss cells. CXCL5 was identified as one of the drivers of the pro-angiogenic phenotype of <em><strong>PBRM1</strong></em>-mutated tumors. Of note, enhanced sprouting capacity of the supernatant of <em><strong>PBRM1</strong></em>-KO ccRCC cell lines can be selectively suppressed by CXCR2 inhibitor and CXLC5 blocking antibody. The therapeutic potential of CXCL5/CXCR2 signaling pathway disruption in ccRCC were shown in the CAM assay, underscoring that targeting CXCL5/CXCR2 axis is a promising approach for ccRCC treatment, specifically in <em><strong>PBRM1</strong></em>-defective ccRCC.