Vascular endothelial growth factor (VEGF) exerts crucial functions during pathological angiogenesis and normal physiology. We have previously reported increased hematocrit (60–75%) and erythrocytosis after high-grade VEGF inhibition by diverse methods, including adenoviral expression of soluble VEGF receptor (VEGFR) ectodomains, recombinant VEGF Trap protein and the VEGFR2-selective antibody DC101. Erythrocytosis occurred in both mice and primate models in an erythropoietin (Epo)-dependent manner, with unexpected >40-fold induction of hepatic erythropoietin through a hypoxia- and HIF-1α-independent mechanism involving disruption of endothelial-hepatocyte cross-talk.

To identify candidate transcription factors required for VEGF inhibition-dependent hepatic Epo induction, we performed hepatic gene expression profiling, revealing up-regulation of HIF-2α target genes (Loxl2 and Cited2). Systemic VEGF inhibition resulted in hepatic stabilization of the HIF-2α protein without concomitant liver hypoxia. Conversely, hepatocyte-specific deletion of the Hif2a gene in Hif2aflox/flox mice strongly diminished the hematocrit elevation and hepatic Epo synthesis after systemic VEGF inhibition, indicating an essential role for HIF-2α. These results are consistent with other reports indicating potent regulation of hepatic Epo production by HIF-2α. In addition to the previously characterized Epo-dependent erythrocytosis, RBC mass monitoring in VEGF inhibitor-treated mice reveals that an erythrocytosis-independent mechanism also contributes to elevated hematocrit. VEGF inhibition initially moderately elevates hematocrit from a baseline of 47% to approximately 50–55% without a parallel increase in RBC mass or reticulocyte index. This initial hematocrit elevation (phase 1) occurs over the first 5 days of VEGF inhibition and is associated with decreased plasma volume (PV) as measured by Evans blue dye injection. Subsequently, a persistent erythrocytosis ensues, as reflected by increased RBC mass and reticulocytosis (phase 2) and is responsible for the vast majority of the erythrocytosis observed with stringent in vivo VEGF inhibition. Together, decreased PV (phase 1) with elevated erythrocytosis (phase 2) combine to maximally elevate hematocrit (near 75%) under stringent VEGF inhibition. Notably, phase 2 but not phase 1 hematocrit elevation and Epo induction were ablated by hepatocyte Hif2a gene deletion, again suggesting that phase 1 elevation arises through distinct mechanisms.

Clinical implications of our work include the use of VEGF inhibitors for the simultaneous treatment of malignancy and anemia as well as the use of hematocrit and/or Epo as surrogate markers for the efficacy of in vivo VEGF inhibition. Accordingly, we will summarize our initial attempts at prospective monitoring of these parameters in cancer patient populations treated with VEGF inhibitors.

Overall, these data highlight unexpected physiologic consequences of VEGF inhibition and indicate that VEGF is a previously unsuspected negative regulator of hepatic Epo synthesis and erythrocytosis via HIF-2α-dependent and -independent mechanisms.

Disclosures: No relevant conflicts of interest to declare.

Author notes

Corresponding author

Sign in via your Institution