Biomarkers for Anti-angiogenic Therapy

Abstract

Predictive biomarkers for anti-angiogenesis agents have remained elusive. The development of bevacizumab and other vascular endothelial growth factor (VEGF) targeted drugs has been slowed by this limitation and the true potential of these agents will not be realized until better biomarkers have been discovered. The development of hypertension is correlated with response to VEGF agents. This was initially noted in trials of renal cell carcinoma in which a strong connection was seen between onset of hypertension and response to VEGF targeted drugs.1An association with development of hypertension and clinical benefit was also seen in the landmark non-small cell lung cancer (NSCLC) trial E4599, which randomized patients with advanced NSCLC to receive firstline carboplatin/paclitaxel with or without the VEGF targeted antibody bevacizumab.2 This trial demonstrated a substantial increase in response rate (more than double), increased progression-free survival (PFS) and a statistically significant improvement in overall survival (OS) (10.3 compared to 12.3 months) in the bevacizumab arm, with onset of hypertension associated with more pronounced benefit. However, surrogates for the development of hypertension have yet to be discovered and this response can therefore only be assessed after initiation of therapy. Serum and plasma markers, including circulating VEGF levels, have been studied extensively. In E4599, high baseline plasma VEGF levels were found to correlate with higher response to bevacizumab but did not predict for a survival benefit.3 E4599 investigators also looked at basic fibroblast growth factor (bFGF), soluble intercellular adhesion molecule (ICAM) and E-selectin, both at baseline and during therapy. Baseline ICAM levels were shown to be prognostic but not predictive as patients on both arms with low baseline ICAM had a higher response rate and better OS (p = 0.00005) compared to those with high ICAM. In a NSCLC study of the VEGF receptor tyrosine kinase inhibitor (VEGFR-TKI) vandetanib, it was noted that VEGF levels predicted for PFS with lower levels associated with a better PFS.4 In other malignancies the role of baseline circulating VEGF-A is more promising as a predictive marker for bevacizumab benefit. In a large phase III gastric cancer trial of bevacizumab (AVAGAST), the investigators noted that high plasma VEGF-A levels were associated with a poor prognosis in all patients, but benefit from bevacizumab was more pronounced in those patients as well.5 A trend toward higher response and survival was seen with the addition of bevacizumab in patients with high baseline VEGF-A levels, but not in those with low levels. Similar differential survival benefit from bevacizumab was seen in both breast and pancreatic cancer patients with high, but not low, baseline plasma VEGF-A levels indicating the importance of further study of this biomarker. The technique used in those analyses, which quantifies soluble low molecular weight isoforms of VEGF-A, is the first truly predictive biomarker of bevacizumab activity, but these results have yet to be validated in other tumor types. The gastric trial also showed some potential predictive value with neuropilin-1 (low levels predicting for OS benefit). Genetic variants in VEGF pathway members have also been explored as potentially predictive biomarkers. A subset of patients with pancreatic or renal cell carcinoma enrolled on two separate phase III clinical trials of bevacizumab therapy were studied for single nucleotide polymorphisms (SNPs) in many genes related to angiogenesis. After investigating over 100 SNPs, a single SNP in VEGFR-1 was significantly associated with PFS for patients receiving bevacizumab in both trials, and additionally OS in the pancreatic cancer study.6It is felt that this SNP leads to increased VEGFR-1 expression and subsequently to increased downstream VEGFR-1 signaling. In E4599, SNPS in VEGF, VEGFR-1, ICAM-1, and epidermal growth factor were analyzed without a conclusive result, though one VEGF polymorphism did seem to correlate with survival.7 Further work on VEGF(R) SNPs continues. Investigators have also studied circulating endothelial cells, but as they are difficult to quantify and not clearly of predictive (rather than prognostic) value, it is unclear how far this field of research will develop. Profiling of multiple plasma markers has led to identification of marker signatures that have been associated with response to VEGFR-TKIs. In an analysis of 33 NSCLC patients treated with the VEGFR-TKI pazopanib, the cytokine and angiogenic factor (CAF) profiles were associated with tumor shrinkage. This was primarily true for changes in soluble VEGFR-2 and IL-4 and with baseline levels of IL-12 and hepatocyte growth factor (HGF).8 However, the cytokines of interest have been variable with other VEGFRTKIs in analyses by the same group. 9,10 Ongoing clinical trials of anti-angiogenic agents must incorporate exploratory and confirmatory biomarker analyses to help reach better understanding. The adjuvant NSCLC bevacizumab trial E1505 is collecting tumor samples, as well as plasma and serum at baseline, during, and after completion of chemotherapy, and these samples should allow for validation of some of the biomarkers under investigation in an adjuvant NSCLC population.