Cholesterol-Dependent Modulation of Stem Cell Biomechanics: Application to Adipogenesis

Abstract

Cell mechanics has been shown to regulate stem cell differentiation. We have previously reported that altered cell stiffness of mesenchymal stem cells can delay or facilitate biochemically directed differentiation. One of the factors that can affect the cell stiffness is cholesterol. However, the effect of cholesterol on differentiation of human mesenchymal stem cells remains elusive. In this paper, we demonstrate that cholesterol is involved in the modulation of the cell stiffness and subsequent adipogenic differentiation. Rapid cytoskeletal actin reorganization was evident and correlated with the cell's Young's modulus measured using atomic force microscopy. In addition, the level of membrane-bound cholesterol was found to increase during adipogenic differentiation and inversely varied with the cell stiffness. Furthermore, cholesterol played a key role in the regulation of the cell morphology and biomechanics, suggesting its crucial involvement in mechanotransduction. To better understand the underlying mechanisms, we investigated the effect of cholesterol on the membrane–cytoskeleton linker proteins (ezrin and moesin). Cholesterol depletion was found to upregulate the ezrin expression which promoted cell spreading, increased Young's modulus, and hindered adipogenesis. In contrast, cholesterol enrichment increased the moesin expression, decreased Young's modulus, and induced cell rounding and facilitated adipogenesis. Taken together, cholesterol appears to regulate the stem cell mechanics and adipogenesis through the membrane-associated linker proteins.

1. Introduction

Mechanical properties of human mesenchymal stem cells (hMSCs) are critical for differentiation, potentially impacting the efficacy of regenerative therapeutics [1]. This is because not only the mechanical properties of hMSCs change with differentiation in response to the soluble factors [2] but also they are known to function as biomarkers in signaling transduction [3,4]. The mechanical properties of stem cells can affect the cells' physical interactions with the surrounding extracellular matrix. The cell's elastic modulus has been shown to increase with cells spreading on a high-density matrix of fibronectin, but decrease in cells that are morphologically round on a low matrix protein density, thus potentially influencing the mechanical signal transduction in cell differentiation [5,6]. Moreover, alterations in the cell's mechanical properties are associated with disease or malignant transformation [7–11]. It appears that the cellular biomechanics can potentially be used as diagnostic biomarkers for different states of disease, transformation, or differentiation.

As one of the major lipid components in the cell membrane, cholesterol has been known to play an important role that is not only limited to change the membrane structure and altering the function of membrane proteins but also remodeling the cytoskeleton. Multiple studies have shown that depletion of the membrane cholesterol can increase the cell stiffness [12–15] by generation of contractile forces by the cells in extracellular matrix and strengthening the membrane–cytoskeleton adhesion, although the mechanisms are not fully known. Our previous work showed that stem cell elasticity was dependent on a family of linker proteins—ezrin, radixin, and moesin (referred to as ERM proteins). For example, inhibiting the cell spreading by knocking down the ERM proteins either prevented or delayed the intended osteogenesis [16]. A recently published paper indicates a correlation between caveolin-1 and osteogenesis [17]. Such a correlation is thought to provide a negative feedback mechanism through cholesterol-rich lipid rafts. In this study, however, the role of predominantly membrane-bound cholesterol on the biomechanics of hMSCs and adipogenic stem cell differentiation was quantitatively determined. Specifically, we used atomic force microscopy (AFM) to characterize changes in the cell elasticity, and then siRNA was applied to transiently suppress the ERM protein expressions. Our results indicate moesin is critical in the signal transduction pathway to affect the underlying cytoskeleton that regulates the stem cell mechanics and adipogenic differentiation.

2. Materials and Methods

2.1. Cell Culture and Adipogenic Differentiation.

Bone marrow-derived hMSCs were purchased (Lonza, Walkersville, MD) and cultured in complete medium of DMEM (Invitrogen, Carlsbad, CA) supplemented with 2% glutamine, 15% fetal bovine serum (Atlantic Biologicals, Lawrenceville, GA), and 1% antibiotics/antimycotics (Invitrogen; final concentrations: penicillin 100 U/ml, streptomycin 100 μg/ml, and amphotericin B 0.25 μg/ml). Cells were incubated in a humidified atmosphere containing 95% air and 5% CO₂ at 37 °C.

For adipogenic differentiation, hMSCs at passages between 5 and 7 were plated at the density of 2 × 10⁴ cells/cm² on coverslips placed in culture dishes with culture medium. After 1 day, cells were switched to the adipogenic induction medium, composed of complete DMEM with 1 μM dexamethasone, 10 μg/ml insulin, 200 μM indomethacin, and 0.5 mM isobutylmethylxanthine (Sigma-Aldrich, St. Louis, MO).

2.2. Transfection of Ezrin, Radixin, and Moesin siRNA.

RNA interference was used to reduce the ezrin/moesin/radixin expression. Cells were seeded on a coverglass, incubated overnight, and transfected with 300 pmol of either a pool of three short interfering RNAs (siRNAs) (Santa Cruz Biotechnology, Santa Cruz, CA) or specific to human ezrin, moesin, or radixin on two consecutive days in regular complete culture medium to achieve stable transfect. siRNA were washed, and cells were placed in the adipogenic medium. In each set of experiments, cells transfected with nonspecific scrambled siRNA were used as control.

2.3. Cytoskeleton Disruption and Cytochemical Staining of Cholesterol.

To explore contribution of the cytoskeleton structure to the cellular elasticity during adipogenic differentiation, cells were treated with 20 μM cytochalasin D at 37 °C for 30 min. Samples were used for AFM measurements (see below) and fluorescent imaging. To visualize actins, cells were fixed in 3.7% formaldehyde and permeabilized in cold (−20 °C) acetone for 3 min. Intracellular actin filaments of normal and treated cells were stained with rhodamine phalloidin (1:100 dilution, Invitrogen) for 30 min at room temperature. Nuclei of cells were stained by DAPI (1:3000 dilution, Invitrogen) for 3 min and rinsed thoroughly with phosphate buffered saline (PBS) before imaging.

2.4. Atomic Force Microscopy Micro-Indentation.

Young's modulus of the cell was quantitatively determined using an atomic force microscope (Novascan Technologies, Ames, IA) mounted on an inverted Nikon microscope (Model TE2000-S) [18]. A micro-indenter of 10 μm in diameter was placed over the cytoplasmic region of the cell between the nucleus and the cell periphery and thereby avoiding the cell's perinuclear region. It was slowly lowered onto the cell surface and caused an indentation depth up to 500 nm (∼10% of the cell height) [18]. The spring constant of silicone nitride cantilever was 0.12 N/m. About 70–100 cells were probed under different experimental conditions. The force–distance curves were collected and analyzed using the simple Hertz model [19]. Such analysis relates the loading force (F) and the indentation depth (δ) by

$$F=\frac{4}{3}\frac{E}{\left(1-{\upsilon }^2\right)}{\delta }^{\frac{3}{2}}\sqrt{R}$$ (1)

where ν is the Poisson's ratio, R is the radius of the spherical indenter, and E is the local Young's elastic modulus. Treating the cell as incompressible material [20], the cellular Poisson ratio was assumed to be 0.5. Least square minimization algorithm was applied to estimate the local Young's modulus. The average Young's modulus for each experimental condition was calculated and statistically analyzed at the level of p < 0.05.

2.5. Modulation of Cholesterol Level.

To examine the effect of cholesterol on the elasticity of hMSCs, cells were first treated with MβCD (methylated β-cyclodextrin, Sigma) saturated with cholesterol or MβCD alone, then subjected to AFM measurement. Membrane cholesterols were enriched or depleted following the published protocols [21]. Briefly, cholesterol was dissolved in a given amount of chloroform:methanol (1:1 vol.:vol.) in a glass tube, and the solvent was evaporated. Then, a 5 mM MβCD solution in DMEM medium without serum was added to the dried cholesterol to yield the molar ratio of MβCD:cholesterol equal to 8:1, which allowed MβCD to be saturated with cholesterol [22]. The tube was vortexed, sonicated, and incubated overnight in a bath at 37 °C. To enrich or deplete membrane of hMSCs, 2 ml MβCD with cholesterol or MβCD without cholesterol were added to each sample of hMSCs seeded on coverslip, incubated for 60 min at 37 °C, and washed three times with Hanks' buffer before AFM measurements or immunostaining. To examine the impact of the membrane cholesterol on stem cell differentiation, depletion of the membrane cholesterol was carried out immediately before the induction of adipogenic differentiation.

2.6. Fluorescent Staining and Microscopy.

Fluorescent filipin III was used to stain the membrane cholesterol. This probe was specifically chosen due to its ability to fluoresce and bind to cholesterol in the membranes (Santa Cruz Biotechnology). Briefly, cells seeded on glass coverslips were fixed with 3.7% formaldehyde, rinsed with PBS, and incubated with 50 μg/ml filipin III in PBS solution for 30 min at room temperature. Cells were then washed in PBS and filipin III-stained cholesterols were visualized by fluorescent microscopy. To quantify adipogenic differentiation, intracellular lipid droplets were stained in PBS containing 1% LipidTox (Invitrogen) and 0.1% Saponin (Sigma-Aldrich). The ERM linker proteins were visualized by immunostaining. After the cholesterol levels were enriched or depleted as described above, the samples were then fixed in 3.7% formaldehyde and permeabilized in cold acetone for 3 min. Nonspecific binding sites were blocked using a 5% bovine serum albumin solution for 1 h at room temperature, and then cells were treated with mouse monoclonal antibodies against moesin at 1:100 dilution. Following incubation overnight at 4 °C, cells were further incubated with fluorescein isothiocyanate-conjugated goat antimouse IgG at room temperature for 1 h. Actin and nucleus were costained with rhodamine phalloidin and DAPI at the end of these treatments. Samples were imaged by a laser scanning confocal system (Radiance 2001MP, Bio-Rad, Hercules, CA).

2.7. Flow Cytometry.

To quantitatively assess the membrane cholesterol, cells were fixed and stained with 50 μg/ml fresh filipin III in PBS for 1 h in dark at room temperature. The samples were washed with PBS three times, then the cholesterol level was determined and analyzed by a FACStar Plus (Becton Dickinson) equipped with a 360 nm Coherent Enterprise Argon laser (Santa Clara, CA). Emissions were collected via a 640-nm dichroic long-pass filter with 424/44 nm band-pass filter.

2.8. Lipid Staining and Western Blotting.

For lipid staining, cells were washed and fixed in 10% formalin for 1 h at room temperature. After washing, the cells were stained with oil red O solution (0.5% in isopropyl alcohol) for 30 min, washed again, and imaged using brightfield microscopy.

Protein expression and phosphorylation were determined by standard western blot techniques. Briefly, after modulation of the hMSC membrane cholesterol by exposure to MβCD or cholesterol-saturated MβCD (or vehicle controls), cell lysates were collected, and then protein concentrations were adjusted to be similar across samples. Identical volumes for each sample were subsequently separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. After transfer to nitrocellulose membranes (Bio-Rad), nonspecific binding was reduced by blocking with 5% bovine serum albumin in PBS for 1 h. Membranes were then incubated with primary antibodies for ezrin, radixin, or moesin phosphorylated at Thr567, Thr564, or Thr558, respectively (corresponding to the active state of these proteins) (Cell Signaling, Beverly, MA). After washing, incubation was then performed with secondary antibodies linked to horseradish peroxidase, and protein expression was then detected using enhanced chemiluminescence. Detected bands were scanned using the Gel Doc XR system (BioRad), and densiometric quantification was performed using BioRad Quantity One software. Relative expression levels were normalized to the corresponding control samples.

2.9. Statistical Analysis.

Data are presented as mean ± SEM. Experimental conditions were compared to the untreated control cells, and pairwise comparisons were considered. Differences between two selected groups were determined using an unpaired student's t test at a significance level of α = 0.05. The differences were considered statistically significant at p < 0.05.

3. Results and Discussion

3.1. Adipogenic Differentiation.

Human mesenchymal stem cells undergoing biochemically directed adipogenic differentiation demonstrated various morphologies. Three-color composite fluorescent images were constructed to verify the extent of adipogenesis at day 7 of differentiation by visualizing lipid droplets (green), F-actins (red), and nuclei (blue) (Fig. 1). While MSC differentiation may be statistical in nature, the elongated morphology of MSCs appears to indicate a reduction in the expression of lipid droplets. The altered morphological changes should remodel the microfilaments from thick stress fibers to less prominent shorter fibers without preferred orientation [23] and began to define the cortical actin cytoskeleton pattern. Actin cytoskeletal reorganization would be expected to alter the cellular biomechanics and could be correlated with the expression of lipid droplets.

Fig. 1.

Composite images of hMSCs undergoing adipogenesis at day 7. hMSCs were captured and fluorescently visualized. Actins (red), lipid droplets (greed), and nuclei (blue). These images were recorded using a 60×/1.4 NA microscope objective. Bar = 20 μm.

3.2. Time-Dependent Changes in Cholesterol During Human Mesenchymal Stem Cells Differentiation.

The cholesterol level is an important factor that is involved in the regulation of the cellular elasticity. In control and undifferentiated hMSCs, the membrane cholesterol was imaged using the filipin III staining. By overlaying with brightfield images that can identify the lipid droplets, the filipin III was found to be specific to the membrane-bound cholesterol and did not penetrate inside the cell (Figs. 2(a) and 2(b)). Flow cytometry was utilized to further quantify adipogenesis. Signals from the control, undifferentiated, and unlabeled hMSCs were gated out, and the gates for filipin III-positive were selected. The flow cytometry results demonstrate that, in comparison to undifferentiated control hMSCs (Fig. 3(a)), more cells were found filipin-positive at days 3, 7, and 10 of postinitiation of adipogenesis (Figs. 3(b)–3(d), respectively). Accordingly, the population histogram shifted progressively to the right (Fig. 3(e)), suggesting the level of membrane cholesterol was upregulated during adipogenesis. Quantitative analysis indicated that the fraction of filipin III-positive cells increased from 31% (day 0) to 65% (day 10) of adipogenesis (10,000–15,000 cells analyzed).

Fig. 2.

Membrane cholesterol accumulation during adipogenesis. The membrane cholesterol was stained with fresh filipin III (green) and imaged at day 0 (A) and day 7 (B), and overlaid with brightfield images. These images were recorded using a 60×/1.4 NA microscope objective. Bar = 20 μm.

Fig. 3.

Flow cytometry analysis of membrane cholesterol expression during adipogenesis. The membrane cholesterol was stained with fresh filipin and analyzed by a flow cytometer with 360-nm argon laser excitation. Emission was collected using a 424/44 nm bandpass filter. Panel (a), undifferentiated control hMSCs; panels (b)–(d), at days 3, 7, and 10 of postdifferentiation. Panel (e) shows comparison of histograms of filipin at different time points and demonstrates a progressive shift to the right.

3.3. Elasticity of Human Mesenchymal Stem Cells During Adipogenic Differentiation.

Atomic force microscopy (AFM, Fig. 4(a)) was used to measure the elastic modulus of hMSCs during adipogenesis. Consistent with our previous results [24], undifferentiated hMSCs were shown to be mechanically stiffer. During the first nine days of adipogenic differentiation, the cell's elastic modulus decreased by twofold from 3.0 to 1.5 kPa (Fig. 4(b)). Since AFM indentation measurements should reflect the actin cytoskeletal organization [16], changes in the cellular mechanics during adipogenesis should be consistent with actin reorganization. To confirm that the actin cytoskeleton is primarily responsible, cells were treated with cytochalasin D at days 7 and 12 to reorganize the actin stress fibers to actin globules (images not shown). The cytochalasin D treatment induced the cell's elastic modulus to substantially decrease to ∼0.7 kPa. We next assessed the impact of the membrane cholesterol on the cell elasticity. Using undifferentiated hMSCs, cholesterol was either depleted from or inserted to the cell membrane. After treating cells with 5 mM MβCD for 1 h to deplete cholesterol, the elastic modulus was statistically increased. In contrast, when cholesterol was added to the cell membrane, the cell's elastic modulus decreased significantly (Fig. 4(c)).

Fig. 4.

Change of Young's moduli during adipogenesis: (a) Actual image of an AFM cantilever used to determine the cellular mechanical property and (b) AFM results indicate continuous decrease in the cell stiffness during the first 12 days of adipogenic differentiation. When cytochalasin D was used at day 7 or day 12, a significant decrease in the cell stiffness was observed. The dotted line represents data fitting to an exponential function. (c) Depletion of the membrane cholesterol by MβCD caused the elastic modulus to significantly increase (e.g., stiffer cells), while enrichment of the membrane cholesterol reduced it significantly (e.g., softer cells).

Since the membrane cholesterol appears to regulate the stem cell biomechanics, it is plausible that it can also alter the intended adipogenic differentiation. To test this hypothesis, the membrane cholesterol was depleted as described, and the adipogenic soluble factors were added. The cholesterol depletion caused both a delay and a reduction in adipogenesis. In a control experiment at day 10 postinduction of adipodifferentiation, round adipocytelike cells were observed in response to the soluble factors and expressed lipid droplets (dark regions in Fig. 5(a)). When the cells were treated with MβCD to deplete cholesterol, the expression of lipid droplets was noticeably suppressed (Fig. 5(b)). In order to verify this finding from brightfield images, the lipid droplets were imaged using fluorescent labeling (Figs. 5(c) and 5(d)) and oil red O staining (Figs. 5(e) and 5(f)). Confirmatory results are consistent with the observation that the membrane cholesterol depletion interferes with the intended adipogenesis.

Fig. 5.

Effect of cholesterol depletion on adipogenic differentiation. hMSCs were induced to undergo adipogenesis using the soluble factors and without (control; panels (a), (c), (e)) and with (panels (b), (d), (f)) depleting membrane cholesterol. At day 10, brightfield images of adipocyte-like cells are shown in response to the soluble factors only (a) and with MβCD treatment (b). Corresponding fluorescence images of LipidTox labeling ((c) and (d)) and oil red O staining ((e) and (f)) are consistent with the observation that membrane cholesterol depletion interferes with the intended adipogenesis. The oil red O staining was adopted from a published protocol [25]. Images were taken using a 10× microscope objective. Bar = 100 μm.

3.4. Effect of Cholesterol on Ezrin, Radixin, and Moesin Linker Proteins.

We recently reported that transient knockout of the linker proteins (ezrin, radixin, and moesin) that physically associate the cytoskeleton to the cell membrane can delay and suppress adipogenic differentiation [23]. To understand how the membrane cholesterol imparts its effect, the involvement of the ERM proteins was examined. We first detected phosphorylated ERM proteins (p-ERM) by Western blot (Fig. 6(a)). The results suggest that cholesterol depletion induces a reduction in the p-moesin expression but increases the p-ezrin expression. AFM measurements were carried out to characterize the biomechanical effects of a cholesterol enrichment or deletion. As shown in Fig. 6(b), the ezrin and moesin regulate the cell elastic modulus but in an opposing manner. Specifically, when the moesin expression was downregulated, hMSCs became stiffer and, in an opposing manner, the cells became softer when the ezrin expression was knocked down, suggesting these two proteins might function competitively. It should be noted that radixin appears not to play a key role in the regulation of the cell biomechanics. For example, at day 12 of adipogenesis Young's modulus of MSCs treated with siRNA against radixin was measured 3.2 ± 0.3 kPa (compared to control cells, 3.7±0.2 kPa).

Fig. 6.

Modulation of p-ezrin and p-moesin linker proteins by cholesterol. (a) The p-ezrin and p-moesin expressions were analyzed by Western blot. A cholesterol depletion is observed to upregulate p-ezrin but downregulate p-moesin. A cholesterol enrichment seems to induce opposite result. (b) AFM measurements showed the effect of ezrin and moesin on the cell stiffness when cells were treated with siRNA directed against either ezrin or moesin. Knockdown of the ezrin and moesin caused the elastic modulus to decrease or increase, respectively, indicating that ezrin is responsible for stiffening the cell and, in the opposite manner, moesin is required to soften the cells.

One key question still remains; how does cholesterol affect the ERM proteins? we, therefore, modulated the level of membrane cholesterol and examined the expression of the ERM proteins. Based on immunostaining, we note that ezrin (Figs. 7(a)–7(c)) and moesin (Figs. 7(d)–7(f)) respond in different ways to the membrane cholesterol level. While cholesterol depletion upregulated p-ezrin and downregulated p-moesin, cholesterol enrichment caused the exact opposite. We further observed that cholesterol also determines the hMSC morphology. Unlike untreated control hMSCs (Figs. 7(a) and 7(d)), cholesterol-depleted hMSCs exhibited the jagged-edge cell shape that is similar to musculoskeletal cell phenotypes such as fibroblasts and osteoblasts (Figs. 7(b) and 7(e)). This would predict that the p-ezrin expression should have been upregulated in these cells, which is confirmed by immunostaining. With cholesterol enrichment, we would predict hMSCs should assume more rounded morphology with downregulated p-ezrin (Figs. 7(c) and 7(f)). It is interesting to observe that cholesterol enrichment not only induced rounded cell morphology but also substantially shrank the cells. Since we postulated that the ezrin and moesin work in the opposite manner, we treated hMSCs with cholesterol and examined the response of p-moesin. Cholesterol-depleted hMSCs again showed the jagged-edge morphology but the p-moesin expression was essentially nonexistent (Fig. 7(e)). In cholesterol-enriched hMSCs, the p-moesin expression would be significantly upregulated, as confirmed by immunostaining (Fig. 7(f)). Collectively, immunostaining data show clearly that both the ERM proteins and cytoskeleton-dependent cell morphology are modulated by the membrane cholesterol, thus providing additional evidence that cytoskeleton remodeling is potentially mediated by the ERM proteins-dependent signaling pathways. In addition, cholesterol depletion is found to increase the cell stiffness and upregulate p-ezrin, while cholesterol enrichment decreased the cell stiffness and upregulate p-moesin. Based on these observations, a postulate is formulated that ezrin is responsible for stiffening the cell and spread morphology and, in the opposite manner, moesin is required to soften the cell mechanically and induce rounded morphology. When hMSCs were transfected with specific ezrin siRNA or moesin siRNA, the cell's elastic modulus decreased with the ezrin suppression but increased with the moesin suppression, which is consistent with modulation of the cellular biomechanics induced by the membrane cholesterol.

Fig. 7.

Immunofluorescence images of ezrin and moesin in response to cholesterol treatment. Distribution of p-ezrin (green) in control hMSCs (a), in cells treated with MβCD (b) or cholesterol enriched (c). Distribution of p-moesin (green) in hMSCs under the same treatment as above ((d), control; (e), cholesterol depleted; and (f), cholesterol enriched). Actins and nuclei were costained with rhodamine-phalloidin (red) and DAPI (blue) in each image. Cholesterol appears to determine the cell morphology. Cholesterol-depleted hMSCs exhibited jagged-edge cell shape, while cholesterol enrichment induced rounded morphology and a reduction in the cell size. Bar = 50 μm.

4. Discussion

Cholesterol has been found to play an important role in determining the membrane fluidity [26–28] and cell's elasticity [10,29–32]. However, only a few studies have examined the effect of cholesterol on the stem cell mechanics and biochemically intended differentiation. Its impact on the stem cell differentiation and underlying cytoskeleton during differentiation remains to be fully elucidated. The goals of this study were, therefore, (i) to characterize the cholesterol expression and stem cell's elastic modulus during adipogenesis using hMSCs and (ii) to establish correlation between the cholesterol and ERM linker proteins. Elucidation of potential mechanisms would undoubtedly provide a better understanding for the stem cell mechanotransduction and may lead to an alternate pathway for regulating stem cell behaviors and responses.

To our knowledge, this study is the first systematic examination of the role of the membrane cholesterol during adipogenic differentiation. Accompanied by an increasing amount of lipid droplets (Fig. 1), adipogenesis is also correlated with increasing amounts of the membrane-bound cholesterol (Figs. 2 and 3). About 80–90% of cholesterol is typically found in the cell membrane [33–35]. Because filipin III preferentially binds to the membrane cholesterol, it is a useful probe to elucidate the role of membrane cholesterol involved in the lineage commitment of stem cells. Since cholesterol is known to play a key role in the regulation of the cellular mechanical properties, our results confirmed that cholesterol-dependent changes in the cellular mechanics are important in hMSCs by regulating their actin cytoskeleton remodeling. Independent of the membrane cholesterol level, disruption of the actin cytoskeleton induced a dramatic reduction in the cell's elastic modulus.

A novel discovery in our present study suggests a transduction mechanism that couples cholesterol and actin via the ERM linker proteins. The ERM linker proteins serve to physically couple the actin filaments to the cell membrane, and thus mediating actin–membrane linkage and regulating multiple signaling molecules [36,37]. The ERM proteins have also been implicated in manipulating the cellular mechanics and directional cell migration [23,38,39]. Although it is not clear whether the ERM proteins directly bind to cholesterol, the ERM proteins can anchor transmembrane proteins in the lipid rafts [40–42]. It is plausible then, as cholesterol is added or depleted, the lipid rafts may have been rearranged that could potentially affect the ERM proteins' coupling to actin, and subsequently regulate the cellular biomechanics and stem cell differentiation (Fig. 8). This is further supported by a reduction in the ERM proteins and actins observed in differentiated adipocytes as well as redistribution of the active ERM proteins to the membrane-associated pattern and actin reorganization to the cortical cytoskeleton. Alternatively, because the cytoskeleton tension has been shown to regulate the stem cell lineage specification [24,43,44], it is possible that ERM knockdown may affect the actin reorganization by releasing tension through the Rho signaling pathway and interferes with the intended stem cell differentiation. For example, cholesterol is now thought to stabilize lipid rafts [45]. Cholesterol depletion could, therefore, be hypothesized to interfere with internalization of lipid rafts and increase the Rho activity. This would not only increase the cellular tension but also activate the ERM proteins and the Rho-substrate Rho-kinase (ROCK), which enhances the cell stiffness, induce spread cellular morphology and inhibits adipogenesis. While this alternate mechanism remains to be validated, it is interesting to note that recent studies indeed suggest that modulation of the Rho/ROCK signaling may alter adipogenesis and adipocyte differentiation [46–49]. Moreover, connection between the ERM proteins and Rho signaling pathways has been well studied. The ERM proteins, for example, are closely associated with cytoskeletal reorganization, and there is direct interaction between ERM and ROCK [50]. More specifically, moesin is required in actin rearrangement and inhibiting the Rho pathway can prevent such rearrangement [51], suggesting that is a coupling mechanism between the ERM proteins and the Rho signaling. Several kinases, for example, including ROCK, have been implicated in phosphorylating ERM [52]. The Rho/ROCK pathway regulates actin organization and has been shown to influence hMSC differentiation. Suppression of RhoA/ROCK1 can enhance adipogenesis, while activation promotes osteogenesis [53], providing additional support to the working model proposed in Fig. 8. Details of the molecular interactions between the ERM proteins and cholesterol remain to be fully elucidated.

Fig. 8.

Proposed model for the role of cholesterol in adipogenesis. Based on our findings, we propose a working model in which cholesterol depletion or insertion can disrupt or reinforce the lipid rafts, respectively. Furthermore, cholesterol insertion leads to preferential upregulation of moesin, round cell morphology, a decrease in the cell stiffness and facilitated adipogenesis. In the opposite manner, cholesterol depletion leads to upregulation of ezrin, spread cell morphology, an increase in the cell stiffness and impeded adipogenesis.

In summary, we postulate that, during adipogenesis, an increase in the membrane-bound cholesterol is accompanied by downregulation of ezrin, which mechanically softens the cells and promotes adipogenic differentiation. In contrast, as the membrane cholesterol is depleted, moesin is preferentially downregulated which, in turn, causes the cells to mechanically stiffen and interferes with adipogenic differentiation. These novel findings shed a new light on the involvement of ERM proteins in stem cells during adipogenesis and the ERM protein-dependent mechanotransduction pathways that regulate adipogenic differentiation.

Funding Data

• NIH grants (CA113975 and HL083298; Funder ID: 10.13039/100000002).