CD146: a new partner for VEGFR2

In this issue of Blood, Jiang et al identify the cell adhesion molecule CD146 as novel co-receptor for vascular endothelial growth factor receptor 2 (VEGFR2).¹ 

The cell adhesion molecule CD146 was first described in 1987 because of its expression on malignant melanocytes and was correlated with a negative prognosis of melanoma patients. Based on sequence homology analysis, CD146 could be identified as a member of the cell adhesion molecules (CAMs) of the immunoglobulin superfamily. Therefore, it is also known as M-CAM or Mel-CAM (melanoma cell adhesion molecule). Not long afterward, CD146 was identified as a cell-surface antigen of endothelial cells. CD146 became popular as marker of circulating endothelial cells (CECs), which are increased in pathologic conditions such as cardiovascular diseases, inflammation, or cancer. Through the use of activating and inhibiting antibodies, involvement of several intracellular pathways in CD146 signaling including focal adhesion kinase or p38 kinase was identified. In this regard, CD146 mediates cell-cell interactions and migration of endothelial cells.² 

Jiang et al demonstrate the direct binding between CD146 and VEGFR2 in coimmunoprecipitation experiments. Furthermore, they found that the interaction takes place in the extracellular protein domain as the antibody AA98, which recognizes an extracellular CD146 epitope could block the interaction between VEGFR2 and CD146. Introduction of a mutation in this protein domain confirmed this observation. In addition, evidence was provided that the interplay of CD146 with VEGFR2 is mandatory for functional VEGFR2 signaling. Using an anti-CD146 antibody or CD146 siRNA, VEGF-induced phosphorylation of VEGFR2 was suppressed in human umbilical vein endothelial cells. Furthermore, inhibition of CD146 resulted in abrogation of the downstream cascade of p38 and Akt signaling, whereas ERK signaling was not affected by anti-CD146 antibody or CD146 siRNA.

VEGFR2 mediates the full range of VEGF responses in endothelial cells including proliferation, regulation of survival, migration, and vascular tube formation. VEGFR2 has multiple tyrosine phosphorylation sites, which may explain its manifold biologic functions.³  For instance, phosphorylation of Tyr¹¹⁷⁵ results in activation of protein kinase C and downstream induction of the ERK pathway leading to cell proliferation.⁴  On the other hand, upon phosphorylation of tyrosine residue Tyr¹²¹⁴, tyrosine kinase Fyn is activated, which results in subsequent activation of Cdc42 and MAP kinase p38 inducing reorganization of the actin cytoskeleton and thus enhanced cell migration.⁵  Although the exact molecular mechanisms of how stimulation of VEGFR2 induces diverging downstream signals have not yet been elucidated in detail, ligand diversity and availability as well as interaction with co-receptors might explain most of these effects. VEGF-A has at least 9 different splicing forms that induce distinct cellular functions due to different binding affinities to their receptors or extracellular matrix components.³  Chen et al previously reported that VEGFR2 signaling induced from soluble versus matrix-bound VEGF resulted in distinct molecular activation patterns. Upon exposure to matrix-bound VEGF, clustering and internalization of VEGFR2 were potentiated compared with soluble VEGF, which resulted in prolonged VEGFR2 phosphorylation of tyrosine residue Tyr¹²¹⁴ and thus extended p38 signaling.⁶  In addition, interaction with co-receptors is essential for functional signaling of many tyrosine kinases. Co-receptors identified for VEGFR2 include neuropilin-1, the hyaluronic acid receptor CD44, vascular endothelial cadherin, and β integrins.⁷ 

In addition to the extracellular binding of CD146 to VEGFR2, Jiang et al show that intracellular CD146 signaling is also indispensable for VEGF-induced signal transduction. The intracellular tail of CD146 binds to ezrin-radixin-moesin (ERM) proteins that cross-link actin filaments with plasma membranes, therefore mediating cytoskeleton remodeling. Through mutation of the CD146/ERM binding site, VEGF-induced downstream signaling was abrogated independently of the VEGFR2 phosphorylation. Similar observations were made for interaction of CD44 with VEGFR2 or c-Met. For a functional downstream cascade, VEGF- or hepatocyte growth factor–induced signaling requires interaction of the cytoplasmic tail of CD44 with ERM proteins. The latter connect the protein complex to the cytoskeleton, which functions as scaffold for downstream molecules.⁸  Obviously, VEGFR2 and CD146 are part of such a signalosome in which co-receptors not only facilitate ligand binding to the kinase receptors but also enable intracellular signaling via connection with ERM proteins (see schematic overview in figure).

Jiang et al expand their in vitro studies to preclinical mouse models. In this regard, they could show that microvessel density in matrigel plugs mixed with VEGF was reduced in mice with conditional knockout of CD146 in endothelial cells compared with wild-type mice. In addition, in a preclinical pancreas carcinoma xenograft model, additive therapeutic efficacy of combined treatment with the anti-VEGF antibody bevacizumab and the anti-CD146 antibody AA98 was noted accompanied by a reduction in microvessel density on tumor sections in the combined treatment group compared with the single agent groups.

Antiangiogenic therapy targeting the VEGF pathway is an established therapeutic modality in several human malignancies including renal, colorectal, or lung cancer. But this therapeutic approach still has limitations. The responses are often not durable and several cancer types, such as pancreatic carcinoma, are not efficiently targetable.⁹  The results presented by Jiang et al raise the hope that the clinical efficacy of anti-VEGF therapy can be augmented by anti-CD146–directed treatment in the future.