Hypertrophy and hyperplasia are observed in severe obesity. Hyperplasia results from excessive proliferation of adipose precursor cells prior their differentiation. While the events controlling adipocyte differentiation have been largely explored, the mechanisms that direct multipotent mesenchymal progenitors towards the adipogenic lineage remain unknown. Key information regarding the origin of adipocytes is also lacking. The capacity of mouse embryonic stem (ES) cells to undergo adipocyte differentiation in vitro provides a powerful model to address these points (1). We have shown that activation of RA receptor beta y using the selective synthetic retinoid CD2314, was sufficient to induce differentiation of ES cells into adipocytes (2). Transcriptional profiling revealed that clusters of genes modulated by CD2314 in ES cells were enriched in neural crest-associated genes, suggesting that neuroectoderm rather than mesoderm was a source of adipocytes. Differentiation of ES cell-derived neuroepithelial progenitors into adipocytes confirmed neuroectoderm as a source of adipocytes in ES cell-derived cultures. Furthermore, mapping of neural crest derivatives in vivo using sox10-Cre transgenic mice demonstrated that in the cranial region adipocytes originated from the neural crest whereas in the trunk they originated from the mesoderm. Finally, we observed that in primary cultures, both cephalic and truncal neural crest cells from developing quail were able to differentiate into adipocytes (3). Altogether, these data indicate that adipocytes have an alternative origin in the neural crest. It is known that fat cells from different depots exhibit different properties and have different impact on the development of metabolic disorders. Whether adipocytes of different origin also differ in their biological functions remains to be investigated. In regard to humans, we have recently isolated multipotent stem cells from the crude stroma-vascular fraction of adipose tissue of young donors of both sex. These cells, we named hMADS for Multipotent Adipose-Derived Stem cells, display extensive self-renewal capacity in vitro, exhibit a normal diploid karyotype and maintain the capacity to undergo differentiation into functional mesenchymal cell types, such as adipocytes, osteoblasts, chondrocytes and skeletal myocytes (4). In vivo, entrapped within an appropriate scaffold, hMADS cells were able to form bone after subcutaneous injection into nude mice, suggesting that hMADS cells could be used for bone repair (5). Transplantation of hMADS cells into muscles of dystrophin-deficient mdx mice, an animal model of Duchene Muscular Dystrophy disease, resulted in substantial expression of human dystrophin, altogether showing the therapeutic potential of stem cells isolated from adipose tissue (4). As the maintenance of adipocyte precursor pools is regulated by self-renewal of stem cells, we have undertaken the identification of molecular events involved in proliferation and differentiation of hMADS cells. We have identified the occurrence of autocrine/paracrine FGF loop in the maintenance of self-renewal of adipose tissue-derived stem cells (6). It appears that 6-bromoindirubin-3′-oxime (BIO) or LiCl, two pharmacological inhibitors of GSK3, behave as both inhibitors of cell proliferation and of early steps of adipocyte differentiation (7). In contrast, purmorphamine, an activator of Hedgehog signaling, did not affect neither cell proliferation nor early steps of adipogenesis but impaired adipocyte maturation of hMADS cells (8). Finally, we see that activin A, a secreted protein composed of two βA-inhibin (Inhβa) subunits, is a critical early regulator of human adipose-derived stem cell adipogenesis. These pathways may represent potential targets for controlling adipocyte precursor pool under conditions were fat tissue formation is impaired.