Regulators of human adipose-derived stem cell self-renewal

Adipose tissue is an alternative source of mesenchymal stem cells and human adipose-derived stem cells (ASCs) display an attractive and substantial therapeutic potential when transplanted in animal models. To this end, an understanding of ASC biology is necessary and the knowledge of mechanisms that maintain ASCs in an undifferentiated state with no loss of differentiation potential during ex vivo expansion represents a crucial step. However, these mechanisms remain to be identified because appropriate human cellular models are scant. In this review we will describe a cellular model isolated from human adipose tissue displaying all the features of stem cells. Then, we will focus on the identification of intrinsic and extrinsic factors regulating the balance between human ASC proliferation and differentiation. We will point out the role of factors secreted by undifferentiated ASCs, such a FGF2, activin A, BMP4, Hedgehog molecules and secreted by adipose tissue macrophages. Finally, we will outline the role of miRNAs in these processes.

Am J Stem Cell 2012;1(1):42-47 cells [6].They exhibit the characteristics of mesenchymal stem cells, i.e., the capacity to selfrenew and to differentiate into several cell types at the clonal level.Cells can be expanded ex vivo for more than 160 population doublings (i.e., around 30 passages) while maintaining a normal diploid karyotype.Expanded hMADS cells are then able to differentiate under serumfree adipogenic condition into cells able to exhibit characteristics of human fat cells.More recently, hMADs cells have been described as a faithful model to study human fat cell metabolism [7,8].These data indicate that expanded hMADS cells enter the adipose lineage at a high rate and differentiate into cells that display a unique combination of properties similar, if not identical, to those of native human adipocytes.Thus, they provide a unique model to analyse human adipose tissue physiopathology.In this regard, they were further used to assess the molecular mechanisms underlying drug-induced adipose tissue disorders such as lipodystrophy as observed in HIV-infected patients receiving HAART therapy [9,10].After ex vivo expansion, hMADS cells endow the ability to undergo differentiation into adipocytes, osteoblasts, and chondrocytes at the single cell level [6,11] without any loss of their therapeutic potential.Actually, transplantation of hMADS cells into mdx mouse, an animal model for Duchenne muscular dystrophy, results in substantial expression of human dystrophin on a long-term basis and engraftment takes place in nonimmunocompromised animals [6].When transplanted with a scaffold, hMADS cells are able to form ectopic bone in mouse [12].
Altogether hMADS cells appear to be a powerful cellular model to investigate human ASC selfrenewal and differentiation.

FGF2, activin A and BMP4
The morphology of hMADS cells changes from a spindle-shaped morphology to a flat one during ex vivo expansion.This morphological change is accompanied by a change in cell proliferation ability (at passage 15 the doubling time is around 2 days, and becomes 4 days at passage 20) and by a loss of differentiation potential.Interestingly, Zaragosi et al. reported that the increase in doubling time was concomitant with a decrease of FGF2 secreted by undifferentiated hMADS cells [11].As cells express also FGF type 1 receptor, the high-affinity receptor for FGF2, it has been proposed that FGF pathway plays an autocrine/paracrine role in ASC self-renewal.In support of this hypothesis the addition of exogenous FGF2 was able to sustain proliferation and adipogenic potential over passages.It is interesting to note that EGF, PDGF or FGF10 were not able to mimic FGF2 effects.Activation of ERK1/2 pathway is required to mediate FGF2 effects on hMADS cell proliferation.However, inhibition of MEK1 reduced the clonogenic potential of hMADS cells but did not affect their differentiation potential, indicating that the ERK1/2 signalling pathway is partly involved in FGF2-mediated self-renewal.FGF1 has also been reported to stimulate proliferation of human adipose progenitors and subsequently to increase their capacity to undergo differentiation [13].The role of FGF pathway in the maintenance of ASC pool in adipose tissue remains to be investigated, but recently it has been proposed that FGFs could play a role in expansion of adipose tissues in obese patients [14].
Activin A, a member of the TGFβ family, is secreted by human undifferentiated ASCs isolated from various fat depots of donors of different ages.Its expression dramatically decreases as ASCs undergo differentiation into adipocytes.Activin A is not only a marker of undifferentiated cells but plays also a functional role in differentiation and proliferation.Sustained activation of activin A pathway promotes hMADS cell proliferation and impaires adipocyte differentiation, whereas its inhibition decreases proliferation and promotes differentiation.These effects are mediated via C/EBPβ and Smad2 pathways in an autocrine/paracrine manner [15].Therefore, it has been proposed a model in which FGF2 and activin A regulate ASC proliferation and differentiation.However, it was surprising to observe that in contrast to FGF2 expression, activin A expression increases during inhibition of hMADS cell self-renewal (Figure 1).Therefore, a critical question remains opened: does expansion of ASCs with FGF2 or with activin A maintain their differentiation potential?Indeed, treatment of FGF2-expressing hMADS cells only during proliferation with PD173074, a specific FGF receptor inhibitor, decreases dra-Am J Stem Cell 2012;1(1):42-47 matically their differentiation potential, indicating that FGF pathway is required for both the maintenance of proliferation and of differentiation potential.In contrast, expansion of hMADS cells with activin A induces a myofibroblast-like phenotype.In addition, exposure of hMADS cells to activin A promotes osteogenic differentiation at the expense of adipogenic differentiation (unpublished data).This observation needs to be further investigated and suggests that activin A pathway could change the fate of undifferentiated ASCs.Altogether, these data indicate that both FGF2 and activin A promote ASC proliferation but might impact the differentiation potential of expanded hMADS cells in different ways (Figure 2).These data point out the importance to analyse in details the cellular and molecular events induced by factors regulating proliferation of ASCs and the consequences on their behaviour.BMP4, which is as activin A another member of TGFβ super-family, regulates hASC proliferation in an autocrine and dose-dependent manner while maintaining their multipotent property [16].Therefore, there is an emerging role of TGFβ pathway in self-renewal and differentiation of ASCs, as recently reviewed by Zamani and Brown [17].

Adipose tissue macrophage-secreted factors
Investigating FGF2 and activin A gene expression in the obesity context revealed new regulators of ASC proliferation and differentiation.Obesity is associated with the presence of a higher number of hypertrophic adipocytes, with new macrophages recruited into adipose tissue [18,19], and with an increased proportion of ASCs exhibiting proliferative potential [20].This latter observation strongly suggests an impor-  tant contribution of obese microenvironment is inducing ASC self-renewal.Interestingly, macrophages isolated from obese adipose tissues secrete factors that both inhibit differentiation of ASCs and stimulate expression of activin A and FGF2 genes in hMADS cells [15].These observations fit well with a model proposing that macrophages play a key role in self renewal of ASCs in part through activin A and FGF2.Macrophage factors involved in the regulation of hMADS cell self renewal are unknown so far.However, TNF-α could be one of them because it mimics the effects of macrophage-conditioned medium on ASC proliferation and differentiation.In addition, TNFa stimulates expression of FGF2 and of activin A (Wdziekonski, Villageois and Dani, unpublished data).Besides TNFa other macrophages factors, such as inflammatory cytokines or Wnt molecules [21,22] are likely to be involved in proliferation and differentiation of human ASCs.Although the phenotype of ASCs expanded in the presence of macrophages factors remains to be investigated in details, these observations suggest that macrophages constitute determinant niche components for ASC self-renewal.

Hedgehog pathway
Hedgehog (Hh) pathway affects also selfrenewal and differentiation of ASCs.Indeed, Hh signalling decreases during adipocyte and os-teoblast differentiation of hMADS cells.Moreover, its activation inhibits both adipocyte maturation, producing ill-differentiated adipocytes that are insulin resistant [23], and osteogenic differentiation [24].Interestingly, there is a basal level of Hh signaling in undifferentiated cells that appears necessary for the maintenance of hMADS cell proliferation and clonogenic capacity, probably through regulation of pRB phosphorylation and cyclin A expression [25].However, in contrast with FGF pathway, inhibition of Hh signalling during proliferation did not alter the differentiation potential of hMADS cells.miRNAs MicroRNAs also emerged as important players in adipogenesis (see for recent review [26]) and in ASC self-renewal and differentiation.Dicer is an RNase III-family nuclease critical for miRNA generation.Knockdown of this enzyme using RNA interference, compromises severely hMADS cell proliferation and differentiation (Figure 3, unpublished data).This observation is in agreement with the impairment of ASC survival consequent to disruption of the miRNA processing machinery reported recently by Sun Kim et al. [27].Deep sequencing of small RNAs expressed in undifferentiated and differentiating hMADS cells indicated an up-regulation on the miR-30 family during adipogenic differentia- Functional analysis revealed that inhibition of the miR-30 family blocks adipogenesis while it promotes osteogenesis.Then, it has been additionally showed that Runx2 targeting is, at least in part, responsible for miR30 positive effects on adipocyte differentiation.In this context it is interesting to remind that the transcription factor Runx2 is expressed in undifferentiated hMADS cells and that represents the major regulator of osteogenesis [28].Altogether, these data support a model in which miR30 regulates the fate of hMADS cells via the modulation Runx2.
In conclusion, all together these observations reveal the complexity of ASC maintenance and behaviour.FGF2 and activin A appear as common regulators governing self-renewal of human embryonic and adult stem cells [29,30].Other factors such as Wnt, Hh and BMPs, more readily implicated in self-renewal of neural stem cells [31,32] and bone-marrow mesenchymal stem cells [33] play also a key role in ASC biology.Therefore, ASCs seem to combine regulatory signalling pathways of embryonic and tissue -specific adult stem cell self-renewal.The influence of the niche is illustrated by the susceptibility of these cells to macrophage-secreted factors that are present in the adipose tissue of obese patients.A better knowledge of these factors and of their impact in a long-term exposure to ASCs is crucial to properly maintain these cells ex vivo for regenerative medicine and also to investigate their role in the pathological expansion of the adipose mass.

Figure 1 .
Figure 1.Cell morphology and potential of ASC to proliferate and to differentiate change during ex vivo expansion.In the same time, secretion of FGF2 and of activin A decreases and increases respectively.

Figure 2 :
Figure 2: Autocrine/paracrine loops regulating proliferation and differentiation of human ASCs.FGF2 and activin A are secreted by undifferentiated ASCs and promote proliferation of ASCs via ERK and Smad2 pathways respectively.FGF2 maintains adipogenic potential of undifferentiated ASCs whereas activin A change their phenotype.

Figure 3 .
Figure 3. Role of miRNAs in proliferation and differentiation of ASCs.Knockdown of Dicer reduces the proliferation and adipocyte differentiation of hMADS cells.Similar effects were observed with the knockdown of miR30.Adipocytes are stained with Oild Red O for lipid droplets.