Vascular Endothelial Growth Factor and Vascular Endothelial Growth Factor–Related Proteins

Under normal and disease conditions, VEGF is a critical regulator of angiogenesis and lymphangiogenesis. On a cellular level, VEGF is persistently upregulated in ccRCC as a direct result of VHL gene inactivation. Given the relationship between VEGF and RCC tumorigenesis, the prognostic value of circulating VEGF levels and response to VEGF-targeted therapy has been extensively evaluated with conflicting results ( Table 2 ). Biomarker analysis from two sunitinib trials demonstrated that low baseline levels of soluble VEGFR-3 and VEGF-C were found to be favorably correlated with longer progression-free survival (PFS). Conversely, in the AVOREN phase III trial comparing interferon (IFN)-α alone with IFN-α plus bevacizumab, the PFS benefit observed in the bevacizumab arm was not significantly different between patients with baseline VEGF levels above or below the median. In a separate phase III randomized controlled trial, higher pretreatment levels of VEGF were associated with a trend toward improved PFS in sorafenib-treated patients compared with placebo ( P = .096). The same study investigated the predictive significance of changes in VEGF and soluble VEGFR-2 after 3 or 12 weeks of treatment with sorafenib; however, there was no association identified. This lack of association was similarly found between decreases in circulating VEGF-2 and response in patients treated with sunitinib. On the contrary, biomarker analysis from a phase II sunitinib study revealed that patients with objective tumor responses experienced substantially larger changes in VEGF, soluble VEGFR-2, and soluble VEGFR-3 levels compared with those patients exhibiting tumor progression ( P <.05). The conflicting nature of the results to date may be a direct consequence of inconsistent biomarker detection methods in addition to the effect that therapy may have on the ability to accurately measure the concentrations of drug targets.

Cytokine and Angiogenic Factors

Numerous cytokines and other proteins involved in the angiogenic cascade have also been evaluated as biomarkers for response to VEGF-targeted therapies (see Table 2 ). Using data from phase II and phase III pazopanib trials, the authors identified interleukin (IL)-6, IL-8, VEGF, osteopontin (OPN), E-selectin, and hepatocyte growth factor (HGF) to be associated with tumor shrinkage and PFS. In the validation set of samples from the phase III trial, patients in the pazopanib arm with elevated concentration of IL-8 ( P = .006), OPN ( P = .0004), HGF ( P = .010), and tissue inhibitor of metalloproteinase-1 ( P = .006) had shorter PFS than did those with low concentrations. They also found that elevated levels of IL-6 were correlated with improved PFS in the pazopanib-treated arm as compared with placebo arm ( P = .009). Additional analysis from this study generated a seven-factor angiogenic signature (IL-6, IL-8, HGF, OPN, tissue inhibitor of metalloproteinase-1, VEGF, and E-selectin) and patients were categorized into high and low groups, which were then correlated with PFS. They found that patients with higher signature scores were associated with significantly shorter PFS in the placebo ( P = .001) and pazopanib ( P = .001) arms. In a separate investigation of cytokine and angiogenic factors as biomarkers predicting therapeutic response to sorafenib, two distinct clusters of patients emerged that were characterized either by elevated levels of proangiogenic or proinflammatory factors. A panel of six baseline mediators (OPN, VEGF, carbonic anhydrase IX [CAIX], collagen IV, VEGFR-2, and tumor necrosis factor–related apoptosis-inducing ligand) was then correlated with PFS after sorafenib. Patients negative for the six-marker signature benefitted from an improved PFS (hazard ratio [HR], 0.20 vs 2.25 in the signature negative vs positive, respectively; P = .0002).

Inflammatory Markers

Several circulating markers of inflammation have been investigated as prognostic markers for advanced RCC treated with targeted agents (see Table 2 ). In a small series of 41 patients with mRCC treated with sunitinib, C-reactive protein (CRP) was considered as a possible biomarker for therapeutic response. They found that patients with normal CRP levels had a significantly higher partial response plus stable disease rate (84.6% vs 35.7%; P = .002) and significantly longer PFS (median, 19.0 vs 6.0 months; P = .036) than patients with an elevated level of CRP. CRP was identified as an independent predictor of objective response ( P = .016) on multivariate analysis. Nonelevated levels of CRP (<8 mg/L) was also found to be an independent predictor of improved overall survival (OS) ( P = .003) in another small study (N = 52) that included treatment with both sunitinib and sorafenib. A larger retrospective report evaluating 200 patients who received sunitinib as first-line therapy substantiated the previous findings in smaller subsets of treated patients. They found that elevated baseline CRP levels conferred a clear disadvantage with respect to PFS (8 months vs 25 months; HR, 2.48) and OS (12 months vs 50 months; HR, 3.17). Increasing baseline CRP levels was independently associated with inferior PFS in a multivariate model accounting for the variables included in the International Metastatic RCC Database Consortium model (HR, 1.14 for each doubling in CRP).

Neutrophil-to-lymphocyte ratio (NLR) is an inflammatory response marker that has demonstrated prognostic value in several cancer types including RCC. This ratio has recently been examined as a tool for predicting response to VEGF-targeted agents in patients with mRCC. The first study analyzed 109 sunitinib-treated patients and compared pretreatment NLR with posttreatment outcomes. They found NLR less than three to be associated with improved PFS (HR, 0.285; P <.001) and OS (HR, 0.38; P = .043). Similarly, another analysis correlated pretreatment NLR with PFS and OS in 100 patients. They demonstrated an improved PFS ( P = .009) and OS (0.004) after VEGF-directed therapy in patients with a lower baseline NLR (≤3.04). The median OS was 16 months versus 29 months in patients with NLR greater than 3.04 versus less than or equal to 3.04, respectively ( P = .004). This association was further corroborated in a recent analysis that demonstrated high baseline NLR to be correlated with lower response rates to TKIs.

Additional circulating proteins have been studied as potential biomarkers, although in a limited capacity. In a small prospective predictive marker trial of 13 patients with mRCC on TKI therapy, the authors found that the plasma granulocyte-macrophage colony–stimulating factor concentrations were significantly higher in the therapy responsive group when compared with the patients who either progressed or remained stable posttreatment ( P = .012). In a biomarker subanalysis from a phase II randomized controlled trial comparing sunitinib dose schedules, the authors correlated drug efficacy with selected serum markers. Of the 45 proteins evaluated, only two of the proteins showed statistically significant correlations with tumor response (complete and partial response vs stable and progressive disease): lower angiopoietin-2 concentrations ( P = .0215) and higher matrix metalloproteinase-2 concentrations ( P = .0180). Lastly, in a phase III sunitinib trial, elevated lactate dehydrogenase was found to be independently associated with shorter PFS (HR, 1.575) and OS (HR, 2.01); however, the fact that lactate dehydrogenase is a robust prognostic marker in advanced RCC independent of treatment calls these findings into question.