Abstract
5034
Neuropilin-1 (NRP-1) is an isoform-specific receptor for vascular endothelial growth factor (VEGF) and placenta growth factor (PlGF) in endothelial cells and for class 3 semaphorins (SEMA3) in the nervous system. In breast cancer cell lines, NRP-1 expression through its binding to VEGF165 has been associated with cell survival, invasion and migration. In contrast to endothelial cells, where NRP-1 is thought to work mainly as a co-receptor for VEGF with VEGF-receptor 2 (VEGFR-2), NRP-1 is most often reported as the only functional VEGF receptor in breast cancer cell lines. This also contrasts with studies in breast cancer tissues in which both VEGFR-1 and VEGFR-2 expression are reported. We investigated NRP-1 epithelial expression in breast cancer and analyzed its co-expression with VEGFR-1, VEGFR-2, VEGF, PlGF, SEMA3A, through immunohistochemistry staining of a breast tissue microarray containing 16 benign breast lesions, 12 in situ breast carcinomas, 107 invasive breast cancers. Almost no invasive cancer was negative for NRP-1 expression, while 38% were scored as ∼1+, 33% as 2+ and 28% as 3+. Conversely, no benign lesion was 3+, while 50% were negative, 42% 1+ and only 8% 2+. Among in situ cancers, 9% were negative,45% 1+, 27% 2+ and 18% 3+. The differences between benign breast, in situ breast carcinomas and invasive breast carcinomas were highly significant (p=0.00003). The staining for PlGF, with an antibody recognizing both isoforms, didn’t show significant differences for the three kind of lesions. Although SEMA3A expression was weakest of all, a significant (p=0.012) trend of increasing expression from benign breast lesions, to in situ carcinomas to invasive carcinomas was observed. VEGF, VEGFR-1 and VEGFR-2 all showed significantly increased expression in invasive breast cancers compared to benign breast lesions, although the prevalence of staining varied in the in situ breast lesions. All the stainings significantly correlated with each other (p≤0.0001), but with different scores. NRP-1 expression was most strictly correlated with VEGFR-2 (r=0.66) and VEGF (r=0.53) expression. SEMA3A expression was correlated with VEGF and VEGFR-2 as well, while PlGF expression correlated best with VEGFR-1, of which it is the ligand. These results show that NRP-1 is at least moderately expressed in over 60% of malignant breast epithelium, in significantly higher amounts than in benign and pre-invasive breast lesions. As NRP-1 expression correlates with VEGFR-2 and VEGF, NRP-1 may function in breast cancers as in endothelial cells, as a co-receptor with VEGFR-2 for VEGF. These data suggest that both malignant epithelial cells and endothelial cells may be targets of anti-angiogenesis therapies in cancer.
[Proc Amer Assoc Cancer Res, Volume 47, 2006]