Will the real EPC please stand up?

In this issue of Blood, Yoder and colleagues take a step forward in resolving the existing controversy on the origin and functional definition of endothelial progenitor cells (EPCs). The authors demonstrate that the mononuclear cell fraction of peripheral blood harbors 2 clonally distinct progenitor populations, one with true endothelial differentiation potential in vitro and in vivo and another myeloid precursor, lacking this potential.

Since their initial description almost 10 years ago,¹  endothelial progenitor cells (EPCs) have met with great enthusiasm given the therapeutic promise they hold, as well as much confusion and controversy evidenced by the growing body of literature discussing their origin, molecular signature, and endothelial differentiation capacity in vitro and in vivo.²  The prospect of being able to supply/recruit functional endothelial precursors in ischemic patients as an alternative for/in addition to existing surgical or angiogenic growth factor treatments for revascularization has prompted many researchers to engage in testing the therapeutic potential of these cells in many animal species and patients. However, contradictory findings have been reported concerning the nature of EPCs and their actual contribution to blood vascular endothelium, not only in an ischemic context but also in other processes involving neovascularization, such as tumor growth. Several researchers have highlighted that EPCs phenotypically resemble monocytes and may even be derived from them.^3,4  Many early studies have claimed a significant—up to 50%—direct contribution of EPCs to newly formed vessels, while more recent studies have demonstrated that vascular incorporation of EPCs (and bone marrow–derived cells in general) is minimal and have suggested that their vascular potential may be mainly—if not exclusively—trophic, in that they supply angiogenic growth factors to host vascular cells.⁵ 

The controversy probably finds its origin in a combination of factors, some of which are related to variations in the research protocol and others that may be out of control of investigators. As an example of the latter, given the formidable heterogeneity of the endothelium throughout the body, it is likely that direct vascular contribution may not be as efficient in every vascular bed or that not every organ is as accessible/permissive for EPCs recruited from the peripheral blood—hence the variable direct contribution of EPCs. However, it is becoming clear that a large part of the variability comes from the differences in cell populations used in different studies because of perhaps small but biologically significant differences in the methodology to derive and culture them. Here, Yoder and colleagues clearly demonstrate that, starting from the same source (namely, peripheral blood mononuclear cells), 2 fundamentally different cell populations can be generated by varying the culture conditions. The simple variable that determined the nature of the resulting cell population was whether adherent or nonadherent cells were propagated in culture. When the adherent fraction was kept, colonies grew out late, which the authors call endothelial colony-forming cells, or ECFCs. Such ECFCs have not only the molecular but also functional characteristics of endothelial cells (ECs) in vitro and in vivo. If the nonadherent cells were replated and kept in culture, cell colonies emerged earlier with a mixed molecular signature that includes EC markers, but with functional characteristics of monocyte/macrophages, rather than ECs. These cells are referred to as endothelial cell colony-forming units, or CFU-ECs.

This is not the first study reporting the similarities of EPCs with monocytes.^3,4  However, in the current study the authors systematically “dissect out” the origin and in vitro and in vivo functional characteristics of these “monocyte-like EPCs” (CFU-ECs) as well as their true endothelial counterparts (ECFCs). Using clonal studies, the authors definitively prove that CFU-ECs are clonally related to hematopoietic stem cells, but not to ECFCs. CFU-ECs not only have monocyte/macrophage surface marker expression but also have myeloid colony-forming capacity and give rise to functional macrophages. Of most importance, unlike ECFCs, CFU-ECs were not able to incorporate into vessels in vivo, which may well explain why some previous reports have failed to find vascular integration of presumed EPCs following transplantation.

In summary, this study provides functional criteria for the definition of EPCs, or as termed by the investigators, ECFCs, and resolves some of the ongoing confusion in the field of postnatal vasculogenesis. The most important feature, after all, for a cell to be called an EPC is its ability to differentiate into an EC and directly participate in vessel growth. While ECFCs comply by these rules in a matrix implantation model, additional studies will be needed to determine if ECFCs behave similarly in an ischemic environment and contribute to new vessel formation. By all means, the authors' note that it is time to clearly define and characterize cell populations does not apply to EPCs only, but is an important take-home message for everyone involved in (stem) cell research.