CommentaryPericytes are a type of perivascular cell associated with the microvasculature which as other mural cells (e.g. vascular smooth muscle cells (VSMC)) are needed to support the endothelium and control vascular tone [1]. In tumors, pericytes aberrantly invest the microvascular bed [2]. Indeed, pericytes are commonly deficient and loosely associated with blood vessel capillaries [2], observations that have been frequently linked to increase leakiness of blood vessels, metastasis and thus worse cancer prognosis [3][4][5]. Recent clinical and preclinical animal model studies support the view that improved vascular wall pericyte investment normalizes blood flow which improves drug delivery and inflammatory cell infiltration and decreases metastasis [3]. Therefore, a better understanding of pericyte function could lead to therapeutic strategies to normalize tumor vasculature.A major difficulty in studying pericytes is the absence of a specific method for their identification. Morphologic criteria, such as apposition with endothelial cells in microvessels are the most reliable current criteria to define pericytes [1]. However, this is not a reliable criterion in remodeling vasculature, especially in tumor where investment of pericytes is aberrant, e.g. loosely associated with the vascular wall [2]. In addition, this requires high-definition imaging, which does not allow for functional studies. Markers such as α-SMA, NG2, Desmin and PDGFRβ are often used to identify pericytes, although none of these individually is a bona fide marker that distinguishes pericytes from other mesenchymal cells such as the VSMC and fibroblasts, as both can have a perivascular location and express pericyte markers. In addition, the labile expression of these markers which varies during the differentiation of pericytes further hinders their identification. Therefore, it is crucial to develop new approaches to identify and isolate pericytes from tumors. CD146 (MCAM) has long been used as an endothelial marker, but a recent study demonstrated that a subset of CD146 + cells that are also CD45 − and CD34 − identify pericytes from various normal mouse and human tissues [6]. We have recently developed a similar FACS based strategy substituting CD34 by CD31 to identify and isolate mouse tumor pericytes defined as CD146 + /CD45 − /CD31 − /lo cells. To ascertain that we sorted bona fide pericytes of stromal origin and not cancer cells with a similar marker signature, we implanted tumor models (LLC and B16F10) into transgenic syngeneic mice ubiquitously expressing GFP or CFP. In addition, for some experiments we implanted LLC-YFP tumor cells in transgenic GFP or CFP mice to exclude engulfment events (YFP + GFP + CFP + cells) leading to misinterpretation of stromal identity. We validated by RT-qPCR for endothelial cell, pericyte and fibroblast markers the nature of the different GFP + or CFP + /CD45 -sorted populations by FACS relative to the expression of CD146 and CD31. In addition, we confirmed by immunostaining that a large fraction (80%) of...