A population of fibro/adipogenic but non-myogenic progenitors located between skeletal muscle fibers was recently discovered. The aim of this study was to determine the extent to which these progenitors differentiate into fully functional adipocytes. The characterization of muscle progenitor-derived adipocytes is a central issue in understanding muscle homeostasis. They are considered as being the cellular origin of intermuscular adipose tissue that develops in several pathophysiological situations. Here fibro/adipogenic progenitors were isolated from a panel of 15 human muscle biopsies on the basis of the specific cell-surface immunophenotype CD15+/PDGFRα+CD56−. This allowed investigations of their differentiation into adipocytes and the cellular functions of terminally differentiated adipocytes. Adipogenic differentiation was found to be regulated by the same effectors as those regulating differentiation of progenitors derived from white subcutaneous adipose tissue. Similarly, basic adipocyte functions, such as triglyceride synthesis and lipolysis occurred at levels similar to those observed with subcutaneous adipose tissue progenitor-derived adipocytes. However, muscle progenitor-derived adipocytes were found to be insensitive to insulin-induced glucose uptake, in association with the impairment of phosphorylation of key insulin-signaling effectors. Our findings indicate that muscle adipogenic progenitors give rise to bona fide white adipocytes that have the unexpected feature of being insulin-resistant.

The differentiation of multipotent cells into undesirable lineages is a significant risk factor when performing cell therapy. In muscular diseases, myofiber loss can be associated with progressive fat accumulation that is one of the primary factors leading to decline of muscular strength. Therefore, to avoid any contribution of injected multipotent cells to fat deposition, we have searched for a highly myogenic but nonadipogenic muscle-derived cell population. We show that the myogenic marker CD56, which is the gold standard for myoblast-based therapy, was unable to separate muscle cells into myogenic and adipogenic fractions. Conversely, using the stem cell marker CD34, we were able to sort two distinct populations, CD34(+) and CD34(-), which have been thoroughly characterized in vitro and in vivo using an immunodeficient Rag2(-/-)gamma(c) (-/-) mouse model of muscle regeneration with or without adipose deposition. Our results demonstrate that both populations have equivalent capacities for in vitro amplification. The CD34(+) cells and CD34(-) cells exhibit equivalent myogenic potential, but only the CD34(-) population fails to differentiate into adipocytes in vitro and in vivo after transplantation into regenerative fat muscle. These data indicate that the muscle-derived cells constitute a heterogeneous population of cells with various differentiation potentials. The simple CD34 sorting allows isolation of myogenic cells with no adipogenic potential and therefore could be of high interest for cell therapy when fat is accumulated in diseased muscle.