Circulating Levels of the Interferon-γ-Regulated Chemokines CXCL10/CXCL11, IL-6 and HGF Predict Outcome in Metastatic Renal Cell Carcinoma Patients Treated with Antiangiogenic Therapy.

Abstract

Simple SummaryWe have studied blood levels of cytokines/chemokines in patients with metastatic renal cell carcinoma treated with sunitinib or pazopanib, with the goal of identifying biomarkers that can predict efficacy and survival. We have found that high levels of CXCL10, CXCL11, HGF and IL-6 before treatment associate with poor prognosis in these patients. Moreover, these factors are correlated in patients with renal carcinoma, suggesting a coordinated expression and secretion. We have developed a prognostic signature including these factors that predicts very accurately prognosis. Our results may help defining better the group of renal cell carcinoma patients who may benefit from sunitinib/pazopanib.Sunitinib and pazopanib are standard first-line treatments for patients with metastatic renal cell carcinoma (mRCC). Nonetheless, as the number of treatment options increases, there is a need to identify biomarkers that can predict drug efficacy and toxicity. In this prospective study we evaluated a set of biomarkers that had been previously identified within a secretory signature in mRCC patients. This set includes tumor expression of c-Met and serum levels of HGF, IL-6, IL-8, CXCL9, CXCL10 and CXCL11. Our cohort included 60 patients with mRCC from 10 different Spanish hospitals who received sunitinib (n = 51), pazopanib (n = 4) or both (n = 5). Levels of biomarkers were studied in relation to response rate, progression-free survival (PFS) and overall survival (OS). High tumor expression of c-Met and high basal serum levels of HGF, IL-6, CXCL11 and CXCL10 were significantly associated with reduced PFS and/or OS. In multivariable Cox regression analysis, CXCL11 was identified as an independent biomarker predictive of shorter PFS and OS, and HGF was an independent predictor of reduced PFS. Correlation analyses using our cohort of patients and patients from TCGA showed that HGF levels were significantly correlated with those of IL-6, CXCL11 and CXCL10. Bioinformatic protein–protein network analysis revealed a significant interaction between these proteins, all this suggesting a coordinated expression and secretion. We also developed a prognostic index that considers this group of biomarkers, where high values in mRCC patients can predict higher risk of relapse (HR 5.28 [2.32–12.0], p < 0.0001). In conclusion, high plasma HGF, CXCL11, CXCL10 and IL-6 levels are associated with worse outcome in mRCC patients treated with sunitinib or pazopanib. Our findings also suggest that these factors may constitute a secretory cluster that acts coordinately to promote tumor growth and resistance to antiangiogenic therapy.