Role of Erythropoietin as an Angiogenic Factor and Target in Cancer.

Erythropoietin is a pleiotropic cytokine with diverse non-hematopoietic functions. Epo and its receptor EpoR have been implicated in the physiologic angiogenesis that occurs in the mouse embryo, female genital tract and during wound healing. Epo is involved in the pathologic angiogenesis of diabetic retinopathy but the role of Epo in tumor angiogenesis, a pathologic process that is essential for tumor progression, is not established. We investigated the hypothesis that Epo may play an important role in tumor cell-induced angiogenesis and progression. Fluorescently-labeled rodent mammary carcinoma cells were implanted in dorsal skin-fold window chambers in nu/nu mice, a model that allows direct, non-invasive, serial visualization and real-time assessment of tumor cells and neovascularization simultaneously during the initial stages of tumor formation. Tumor angiogenesis (vascular length density, VLD) and growth (tumor area) were measured serially over 8 days using intravital microscopy and computerized image analysis. Co-injection of Epo with implanted tumor cells significantly increased VLD by 78% compared to vehicle-injected controls (p<0.001 by repeated measures ANOVA and Bonferroni test, n=8 mice/group). This early proangiogenic effect was associated with significant stimulation of tumor growth by 66% (p<0.001). Implantation of R3230-GFP cells engineered to stably express a constitutively active EpoR mutant (R129C) resulted in significant stimulation of neovascularization by 76% (p<0.001, n=7) and increased tumor growth by 79% compared to empty vector-transfected control cells (p<0.001). To target endogenous Epo function in window chambers, recombinant soluble EpoR (sEpoR) or a neutralizing anti-Epo monoclonal antibody (mAb) were co-injected at the time of tumor cell implantation. By day 8, marked inhibition of neovascularization was observed in response to treatment with sEpoR (44% reduction) or mAb (47% reduction) compared to vehicle-injected controls (p<0.001, n=7). This anti-angiogenic effect was associated with significant decrease in tumor size by 37% (sEpoR) and 39% (mAb), respectively (p<0.001). Stable expression of a secreted Epo antagonist (Epo-R103A) protein in tumor cells was associated with a remarkable anti-angiogenic effect with 50% reduction in VLD (p<0.001, n=7) and near complete disappearance of tumor cells by day 8 compared to controls (p<0.001). Analysis of two independent single cell clones of each transfected cell line showed similar results. All transfected cell lines exhibited similar in vitro growth characteristics and cell cycle profile. To further assess in vivo tumor growth, cells were implanted orthotopically in the mammary fat pad of female nu/nu mice (10 animals/group). EpoR-R129C expression was associated with significantly increased tumor volume (531±41 mm³, n=10) compared to vector-transfected cells (232±46 mm³, n=7). EpoR-R129C tumors exhibited significantly increased numbers of positive cells for proliferation marker Ki67 and blood vessel marker CD31 per high power field (p=0.0006 and 0.0008, respectively). Remarkably, tumor growth was completely absent after mammary fat pad implantation of R3230-GFP cells secreting the antagonist R103A-Epo protein (n=10 mice) compared to controls (183±46 mm³, n=8). Taken together, these data indicate that 1)-Epo is an important angiogenic factor that modulates tumor cell-induced angiogenesis and 2)-Suppression of tumor angiogenesis and progression by Epo blockade suggests that Epo may constitute a potential target for the therapeutic modulation of angiogenesis in cancer.