Promise(s) of Mesenchymal Stem Cells as an In Vitro Model System to Depict Pre-Diabetic/Diabetic Milieu in WNIN/GR-Ob Mutant Rats

Background Development of model systems have helped to a large extent, in bridging gap to understand the mechanism(s) of disease including diabetes. Interestingly, WNIN/GR-Ob rats (Mutants), established at National Centre for Laboratory Animals (NCLAS) of National Institute of Nutrition (NIN), form a suitable model system to study obesity with Type 2 diabetes (T2D) demonstrating several secondary complications (cataract, cardiovascular complications, infertility, nephropathy etc). The present study has been carried out to explore the potent application(s) of multipotent stem cells such as bone marrow mesenchymal stem cells (BM-MSCs), to portray features of pre-diabetic/T2D vis-à-vis featuring obesity, with impaired glucose tolerance (IGT), hyperinsulinemia (HI) and insulin resistance (IR) seen with Mutant rats akin to human situation. Methodology/Principal Findings Primary cultures of BM-MSCs (third passage) from Mutants, its lean littermate (Lean) and parental control (Control) were characterized for: proliferation markers, disease memory to mark obesity/T2D/HI/IR which included phased gene expression studies for adipogenic/pancreatic lineages, inflammatory markers and differentiation ability to form mature adipocytes/Insulin-like cellular aggregates (ILCAs). The data showed that BM-MSCs from Mutant demonstrated a state of disease memory, depicted by an upregulated expression of inflammatory markers (IL-6, TNFα); increased stem cell recruitment (Oct-4, Sox-2) and proliferation rates (CD90+/CD29+, PDA, ‘S’ phase of cell cycle by FACS and BrdU incorporation); accelerated preadipocyte induction (Dact-1, PPARγ2) with a quantitative increase in mature adipocyte formation (Leptin); ILCAs, which were non-responsive to high glucose did confer the Obese/T2D memory in Mutants. Further, these observations were in compliance with the anthropometric data. Conclusions Given the ease of accessibility and availability of MSCs, the present study form the basis to report for the first time, application of BM-MSCs as a feasible in vitro model system to portray the disease memory of pre-clinical/T2D with IR - a major metabolic disorder of global concern.


Introduction
Diabetes mellitus is a debilitating metabolic syndrome arising either due to (i) autoimmune destruction of the pancreatic beta cells (b-cells) resulting in absolute insulin deficiency (Type 1 diabetes (T1D)), or (ii) impaired glucose uptake attributed to peripheral resistance in target organs such as muscle and adipose tissue, as well as (iii) b-cell exhaustation in chronic conditions (Type 2 diabetes (T2D)). More recent studies document diabetes as a state of profound and consistent oxidative stress/chronic inflammatory condition [1] evidenced by increased levels of free radicals [1,2], with a concomitant decrease in antioxidant status in situ [3]. Interestingly, compared to all other organs/cells in the body, b-cells of pancreas are highly vulnerable to oxidative damage on account of its inherently weak antioxidant systems [4] causing impaired b-cell functions, resulting in an imbalance of glucose homeostasis vis-à-vis altered cellular microenvironment.
Cellular microenvironment/niche is a critical unit, patterned to orchestrate the integration of intrinsic factors with extrinsic cues in a spatiotemporal manner for the maintenance of cellular integrity and metabolism [5]. In addition, cells also rely on a variety of potent antioxidant defense mechanisms [6], and a close interaction of antioxidant molecules generated in actively growing and proliferating cells help to scavenge the reactive oxygen species [7]. This is significant in view of the recent findings to support diabetes as a state of oxidative stress [8], affecting various soluble and insoluble factors that form part of microenvironment/niche, affecting the stem cell pool.
Mesenchymal stem cells (MSCs) are a promising source of adult stem cells, well documented for their plasticity and differentiation potential to form multiple cell types in vitro and in vivo [9]. Recent studies have documented the immunomodulatory functions (i.e. secretion of various angiogenic and cytoprotective factors) [10,11] and their potential to improve healthcare by either augmenting body's own regenerative potential or development of new therapies [12]. A dysregulation in the microenvironment of MSCs from human subjects have shown for an altered surface marker expression, self-renewal capacities in several pathological conditions such as chronic renal failure [13], multiple myeloma [14], acute myeloid leukemia [15], rheumatoid arthritis [16] and T1D [17] and T2D [18], demonstrating upregulated expression of inflammatory cytokines for depicting the disease memory towards understanding the cellular and molecular basis of the alterations.
In the present study, we have investigated BM-MSCs (multipotent stem cells) to assess for the altered cellular milieu in WNIN/ GR-Ob (Mutant) rats which depict obesity, impaired glucose tolerance (IGT) with insulin resistance (IR) similar to the preclinical/clinical features of T2D seen in human subjects. To address these, we have calculated IR status, assessed for mesenchymal phenotype and proliferation rate, studied the gene expression profile in BM-MSCs to mark the disease memory depicting obesity, IGT with IR including the inflammatory status, and finally we have examined for differentiation/functional capabilities of BM-MSCs to form mature adipocytes and Insulin-like cellular aggregates (ILCAs). All these parameters have been compared between Mutants, their lean littermates (Lean) and parental controls (Control) under the same experimental conditions.

Anthropometric Measurements
Body weights were significantly higher in Mutant (450632.8 gms*/p = 0.03), compared to Lean (250.6619.9 gms) and Control (290.2620.8 gms). A concomitant increase in the adipose depot weights among Mutant (23.9264.85 gms*/ p = 0.04) was also significant compared to Lean (15.0563.1 gms) and Control (16.4162.9 gms). Values have been represented as Mean 6 Standard error (SE) from six animals per group.

Blood and Plasma Measurements
Published reports from these Mutants have shown a normoglycemic response with fasting blood glucose (FBG) in the range of 50-109 mg/dl [19].

Characterization of MSCs during in vitro Expansion
MSCs were seeded as single cells onto DMEM/F12+10% fetal bovine serum (FBS). By 2-5 days after isolation, spindle shaped adherent cells were observed. Non-adherent cells were removed during subsequent media changes. These adherent cells formed a monolayer by day 7 ( Figure 1A) and also showed the colony forming unit (CFU) featuring an MSC phenotype (insight in Figure 1A). Such confluent cultures largely comprised of MSCs which were positive for mesenchymal stem cell-specific cytosolic protein STRO-1 ( Figure 1B). Homogeneity of the population was further confirmed by characterization of the surface phenotype by flow Cytometry (FACS) and represented as median fluorescence intensity (MFI) as given in Figure 2P. Expression of mesenchymal specific surface proteins such as CD90 was higher in Mutant (47986787 p = 0.02) (Figure 2A), when compared to Lean (14316153) ( Figure 2B) and Control (28726763) ( Figure 2C) although, the expression of CD90 from Lean was of lesser magnitude than that from Control. A similar significant increase in expression of CD29 among Mutant (46196314) ( Figure 2D) compared to Lean (24346184, p = 0.02) ( Figure 2E) and Control (11786326) was also evident ( Figure 2F). This shows that the differential expression levels of both the surface markers (CD90 & CD29) showed a slight difference among the groups.
Interestingly, CD31 which was used as a negative marker for MSCs was almost absent in Control (61619) ( Figure 2I) as compared to Mutant (82616, p = 0.04) ( Figure 2G) and Lean (163614, p = 0.02) ( Figure 2H) which was significantly high. Figures 2J-L and 2M-O show the negative controls run parallely during the analysis for CD90/CD31 and CD29 respectively.

Proliferation Rate
A) Population doubling assay (PDA). MSCs at passage 3 were subjected to PDA as per our earlier published protocol [22]. MSCs from Mutant showed an increase in proliferation rates at all the time points (24,48, and 72 hrs) compared to Lean and Control. However at 96 hrs, PDA was statistically significant in Mutant (5.2760.13610 4 , p = 0.04) when compared to Lean (4.4760.16610 4 ) and Control (4.9060.11610 4 ) ( Figure 3A).

Gene Expression by RT-PCR
For Adipocyte lineage. mRNA expression for embryonic markers such as Oct-4 and Sox-2 showed an upregulation among MSCs from Mutant phenotype (p = 0.001) compared to its Lean and Control ( Figure 4A). A similar increase was also noted in Mutants with other markers such as Dact-1 (mesenchymal marker), preadipocyte factor 1 (Pref-1, a preadipocyte marker) ( Figure 4B) and peroxisome proliferator activator receptor gamma 2 (PPARc2) (a master adipocyte transcription factor for adipogenesis), compared to Control (Mutant, Lean. Control) as given in Figure 4C. The transcript levels of mature adipocyte markers such as Leptin, Adiponectin ( Figure 4D) and CCAAT/enhancerbinding protein alpha (C/EBPa) (adipocyte transcription factor) ( Figure 4C) were decreased in Mutant and Control compared to Lean. However, the difference in these expression levels was not statistically significant. In addition, transcript levels of transforming growth factor beta 1 (TGFb1) involved in osteogenesis was significantly decreased (p = 0.02) among Mutant compared to Lean and Control ( Figure 4C). Interestingly, expression of glucose transporter-4 (GLUT-4) and Insulin receptor substrate 1 (IRS-1), which primarily participate in peripheral glucose uptake, were significantly decreased (p = 0.001) in Mutant implicating for an impaired glucose uptake in these Mutants compared to Lean and Control ( Figure 4E).
For pancreatic lineage. Since these Mutants demonstrate obesity with IGT and IR, we assessed mRNA expression levels of key genes of the b-cell lineage in BM-MSCs amongst the three groups. Interestingly, BM-MSCs from Mutant and Lean showed significantly upregulated expression of Insulin (Mutant.Lean) compared to Control (p = 0.02). On the other hand, expression of both pancreatic duodenum homeobox-1 (PDX-1) -a pancreatic transcriptional factor and glucose transporter-2 (GLUT-2) -a pancreatic glucose transporter were significantly decreased in BM-MSCs of Mutant compared to Lean and Control suggesting for an IGT in Mutants ( Figure 4F).

Inflammatory and Stress Markers
Thiobarbituric acid reacting species (TBARS) as a measure of global oxidative stress was measured in plasma (levels in circulation) from Mutant, Lean and Control. Levels of Plasma TBARS were found to be significantly higher in Mutant (2.7060.59* nM/ml) compared to Lean (2.5960.28 nM/ml) and Controls (2.1560.34 nM/ml). mRNA levels of inflammatory cytokines, IL-6 and TNFa were significantly upregulated among BM-MSCs of Mutant, followed by Lean and Control ( Figure 5A). Supporting findings such as increased localization of TNFa ( Figure 5C) and endoplasmic reticulum (ER) stress protein RL-77 ( Figure 5E), further reinstated the inflammatory state in BM-MSCs from Mutant which has also been expressed as relative fluorescence units (RFU) ( Figure 5D&F). The increased TNFa immunolocalization among Mutants compared to Lean and Control was however not statistically significant. Confirmatory results by FACS analysis for TNFa ( Figure 5B) did support for the increase in in situ stress of BM-MSCs from Mutant (2.860.5%), compared to Lean (1.060.2%) and Control (0.460.03%) phenotypes. Data have been compiled from three independent experiments performed in duplicates.

Differentiation of BM-MSCs to Mature Adipocytes
MSCs from Mutant, Lean and Control phenotypes were subjected to adipogenic induction as described earlier [23]. At the end of a 15 day protocol, the oval or near spherical cells filled with lipid droplets that stained positive with oil red ( Figure 6A), were quantitated for the mature adipocytes formed using ACT2U  software (Nikon, Japan). As indicated in Figure 6B, the number of mature adipocytes formed were much higher from Mutant (49.2565.6%) compared to Lean (35.4662.5%) and Control (38.563.2%).

Differentiation of BM-MSCs to ILCAs
MSCs from Mutant, Lean and Control were also studied for the formation of ILCAs using the three-step protocol published earlier [24]. Briefly, addition of serum free medium (SFM)-A initiated the process of differentiation which was evidenced by a change in morphology of MSCs. On addition of SFM-B in the second step, cluster-like formation was evident and maturation of these clusters was appreciable with the addition of SFM-C containing maturation factors such as glucagon-like peptide 1 (GLP-1), Nicotinamide and high concentrations of Taurine (i.e. 3 mM). After 10 days of initiation, the newly formed ILCAs showed deep crimson red color when stained with dithizone (DTZ), which has been used as an islet specific marker and reflects maturation of the clusters. Amongst the three groups, BM-MSCs from Mutant showed an increased formation of ILCAs compared to Lean and Control ( Figure 7). ILCAs obtained from Mutant, Lean and Control were next measured for their insulin secretory response at both basal (5.5 mM) and high glucose challenge (17.5 mM) similar to our published data [20]. The fact that these Mutants were hyperinsulinemic was evident by their higher basal insulin levels in Mutant (50.7160.13 mU/L/mg protein) and Lean (42.2360 [20].

Discussion
In the present study we showed that BM-MSCs from Mutant rats portray features of obesity with IGT/IR/hyperinsulinemia (HI) vis-à-vis lean body mass, and higher fat mass as compared to its Lean and Control and mimic the disease state in vivo of preclinical/clinical T2D with IR seen in human subjects [25]. The present strain of Mutants showed an incomplete dominant mutation exhibiting three phenotypes -homozygous lean (+/+), heterozygous carrier (+/2) and homozygous obese (2/2) in a 1:2:1 Mendelian ratio. All the three phenotypes could be identified phenotypically as early as 35 days by the presence of a 'kinky tail', a trait unique to the WNIN obese mutant rats that co-segregates with obese and carrier but not lean [19], suggesting for an early onset of degenerative-like changes/symptoms in these Mutants. Interestingly, these Mutants demonstrated most of the secondary complications associated with T2D such as onset of cataract, cancers, infertility, etc., making them a good animal model to study the pathophysiological state underlying obesity and/obesityinduced diabetes [19,[26][27][28].
Microenvironment/niche plays a pivotal role towards maintenance of cellular homeostasis, cellular integrity and functions. Yet, the niche may also induce pathologies by imposing aberrant function to the cellular components observed in diseased states such as diabetes [29,30], myelomas [31], neoplastic cells [32] and bone loss associated with T1D [17] etc. Inflammation and inflammatory milieu are known to be the underlying mechanism(s) operating in these disease states. Our present data on BM-MSCs isolated from Mutant phenotype showed an upregulated expression of (i) inflammatory genes -IL-6 and TNFa ( Figure 5A) and (ii) stress-related proteins TNFa and RL-77 ( Figure 5B-F) with a proportional increase in circulating TBARS in plasma, suggestive of cellular stress and inflammatory condition in the progenitor pool, similar to studies reported with BM-MSCs from diabetic [33] and obese -diabetic subjects [34] advocating the role of BM-MSCs in disease pathology. IR is the chief abnormality underlying adult onset degenerative diseases like diabetes, cardiovascular diseases (CVD), hypertension etc., and there has been increasing evidence to suggest causative links between inflammation and development of IR [1,3] akin to our data. It has also been shown that TNFa can induce IR in human skeletal muscle by altering the insulin signaling pathway providing a unique system for molecular dissection in vitro [35]. The physiological source of increased circulating levels of TNFa in T2D remains unclear; however it is believed to be a major cytokine involved in the exchange between adipose tissue and muscle with increased levels in obesity and T2D possibly contributing to IR in skeletal muscle [36].
Fat is an integral part of the BM microenvironment [37,38] and marrow adipogenesis has shown to be an integral part of adipocyte homeostasis and appears to be the major contributor as a nonadipose origin for the increased risk of visceral adiposity [39,40]. Further, reports of Gimble et al., [41] suggest that non-adipose derived adipocytes, such as BM-progenitors form an important contributor towards the increased visceral fat accumulation seen mostly in post-menopausal women who are at an increased cardiometabolic risk as compared to pre-menopausal women [42].
In similar lines we have demonstrated for an increased adipogenesis from BM-MSCs of Mutant rats (1.6-1.8 fold) as compared to Leans and Controls, examined as a phased gene expression (ESCR MSCR Pre-adipocyteR Mature adipocyte). Substantiating these have been a sequalae of events, which were more noted from BM-MSCs of Mutants showing an increase in: a) recruitment of stem cells (Oct-4 and Sox-2) ( Figure 4A), b) 'S' phase of cell  [44,45]. A higher tendency for adipocytes derived from BM to home among the visceral adipose depots have been reported by Gimble et al., [41] and Gambacciani et al., [46] and visceral adiposity is a known contributor to a range of secondary complications of T2D [47] attributed to its lipolytic/lipotoxic nature [48,49]. This along with increased pro-inflammatory adipokine (IL-6 and TNFa) and decreased adiponectin secretion (causing decreased insulin sensitivity) [50,51] have been a known predictor for T2D with IR. In similar lines, the observed increase in adipocyte formation among BM-MSCs (Mutant) vis-à-vis with an increased expression of IL-6 and TNFa (pro-inflammatory adipokine) underline for visceral fat accumulation [39] in these Mutant rats as compared to its Lean and Control [19] and correlate positively with obesity, IR and IGT.
Obese phenotype memory has also been appreciated in these Mutants with an upregulated expression in PPARc2, inflammatory cytokine production (IL-6 and TNFa) and with a relative low leptin expression among BM-MSCs, compared to its Lean, which marks for the disease memory of tissue [18,52] suggesting for visceral adiposity/visceral fat mass accumulation in these Mutants [19]. Supporting data such as, an increased expression of Dact-1 (a Wnt/b-catenin antagonist) and PPARc2 ( Figure 4B&C) and decreased TGFb1 expression ( Figure 4C) in BM-MSCs, does indicate for accelerated adipocytic induction with a proportional decrease in anti-osteoblastic recruitment, which underlines for early onset osteoporotic changes in Mutants [19] similar to humans [42]. In fact, PPARc2 is not only a master transcription factor of adipogenesis, but also decides the progenitor commitment either for an adipocyte or an osteocyte formation [37,43], since both adipocytes and osteocytes share a common progenitor [41] and maintenance of optimum adipocyte number in the BM is controlled by a balanced pro-adipocytic (increased activity of PPARc2) and an anti-osteoblastic MSC allocation (decreased TGF-b/Bone Morphogenic Protein (BMP), Wnt/b-catenin, and Insulin growth factor-1 (IGF-I) signaling pathways), in addition to C/EPBa which is responsible for the maintenance of fully differentiated phenotype of adipocytes [53].
T2D accounts for approximately 90-95% of all existing diabetic cases and is a major health concern. Interestingly, we have shown an increase in the number of ILCAs obtained from BM-MSCs derived from Mutants, suggestive of obese/obese-diabetic environment similar to the findings of Rosen et al., [39] and Sun et al., [54]. In support of these, we have also demonstrated a significantly upregulated expression of insulin (hyperinsulinemic) and a decreased PDX-1, (a transcription factor necessary for pancreatic development and b-cell maturation) and GLUT-2 (pancreatic glucose transporter) ( Figure 4F) implicating for a hyperglycemic environment [18,34,55]. Such an ectopic expression of these genes in BM-MSCs possibly can be attributed to body's adaptive mechanism in the management of hyperglycemia, since HI (increased insulin levels in circulation) precedes onset of IR and T2D [25,27]. Despite increase in the formation of ILCAs, the ability to respond to high glucose challenge was impaired from Mutant and Lean as compared to Control (Table 2). This phenomenon could be associated to IR in peripheral tissues [25,42] with relatively insufficient secretion of insulin, resulting in pancreatic b-cell dysfunction with severity and progress of the disease/due to the blunting effects of hyperglycemia [56], or by feedback suppression for a prevailing HI [57]. Nevertheless, our present findings for an altered insulin secretion in ILCAs obtained from BM-derived of Mutant and to some extent in Lean, does suggest for an altered cellular milieu [19], similar to the prediabetic/T2D subjects with IR [25]. Keeping in view of the promise(s) of this Mutant model system in T2D, our studies are in progress to delineate the mechanism(s) underlining Obesity/IGT/ HI/IR, which form important predisposing factors precipitating in several micro-and macro-vascular complications associated with chronic T2D.

Conclusions
Given the ease of accessibility and availability of MSCs, our present study forms the basis to report for the first time, that MSCs could be explored as a feasible in vitro stem cell model system to portray disease memory in situ (altered milieu) of pre-diabetes/ T2D, associated with Obesity, IGT, HI with IR, a major metabolic disorder of global concern.   oxygen, humidity and ventilation till sacrificed. All studies were performed under identical conditions for all the three phenotypes. Animals were subjected to a 16 h fasting prior to euthenization to normalize for the differences in feeding patterns.

Anthropometric and Plasma Measurements
Anthropometric measurements such as body weights and adipose depot weights were assessed for all animals used for the study and represented as Mean 6 SE (n = 6). Subcutaneous and visceral (intraperitoneal) adipose depots were dissected from the Mutant, Lean and Control phenotypes during sacrifice, weighed and represented as Mean 6 SE gms total adipose depot weight (subcutaneous+visceral) as per published protocols [58].
Blood (2-3 ml) was collected by retro-orbital vein puncture from Mutant, Lean and Control rats in pro-coagulant Vacutainers (Genire Bio, Germany) and plasma was stored appropriately till analysis. Blood glucose was measured during blood collection with a glucometer by the glucose oxidase method using OneTouch Horizon glucose measurement strips (Johnson and Johnson Ltd, Milpitas, CA, USA) and values were represented as mg/dl. Plasma insulin was measured using High Range Rat Insulin ELISA Kit (Mercodia, Sweden) as per the manufacturer's protocol and values were represented as mU/ml [20] from six animals performed in duplicates. IR indices such as fasting glucose to insulin ratio (FGIR), Homeostasis model of assessment for insulin resistance (HOMA-IR) and Quantitative Insulin Sensitivity Check Index (QUICKI) have been calculated using mathematical formulas FGIR = FPG/FPI; HOMA-IR = FPG6FPI/2430 and QUICKI = 1/[Log(FPG)+ Log(FPI)], where FPG is Fasting Plasma Glucose in mg/dl and FPI is Fasting Plasma Insulin in mU/ml as reported in literature [21].
Thiobarbituric acid reacting species (TBARS) as a measure of global oxidative stress/lipid peroxidation was estimated in plasma by spectrophotometric measurement of MDA released at 532 nM as per previously published protocol [59] and values were represented as Mean 6 SE nM/ml from six animals performed in duplicates.

Isolation and Expansion of BM-MSCs
Animals were euthanized by CO 2 asphyxiation and femur bones were collected under sterile conditions in DMEM/F12 (GIBCO, USA) supplemented with antibiotics (Himedia Laboratories Pvt Ltd, India). BM-MSCs were isolated as per the previously published protocol [60]. Briefly, under sterile conditions metaphyseal ends of femur were cut and marrow was flushed using a 2 ml syringe into DMEM/F12 supplemented with antibiotics. The blood clots were then dispensed uniformly, washed and centrifuged (3-4 times) at 1800 rpm for 10 minutes. The pellet obtained was an enriched fraction of mononuclear cells which was treated with RBC lysis buffer (Tris-Cl+NH 4 Cl), washed and seeded at a density of 5-7610 3 cells/cm 2 in DMEM/F12+10% FBS onto T 25 culture flask (Corning, USA) in a humified chamber at 37uC and 5% CO 2 and cultured till confluence. Cells were passaged every 5-6 days using 0.25% trypsin-EDTA (Sigma, USA) and a split ratio of 1:2 was adopted. Viability of the cultures was routinely monitored by trypan blue exclusion (TBE) as well as by MTT assay [22].

Proliferation Rate
Population doubling assay (PDA). To compare the in vitro expansion rate, proliferation rates of MSCs from Mutant, Lean and Control phenotypes were determined by seeding at a density of 1610 4 cells at passage 3 in 35 mm dish (BD, San Jones, USA) and were maintained in DMEM/F12+10% FBS in a humified chamber at 37uC and 5% CO 2 for a period of 24, 48, 72, and 96 hrs. Cells were trypsinised at the end of culture duration (24,48,72 and 96 hrs) and TB excluded cells were counted using Neubar's Hemocytometer [22]. PDA has been represented as Mean 6 SE (n = 6) (as multiples of 1610 4 ) from three independent experiments performed in duplicates as a function of time.
Bromodeoxy Uridine (BrdU) incorporation studies. To label the cells in S phase, the primary cultures of BM-MSCs (passage 3) were incubated overnight in growth medium containing 10 mM BrdU salt (a structural analogue of nucleotide Thymidine, that gets incorporated into dsDNA during 'S' phase in place of thymidine in actively proliferating cells) as per the published method [61]. Incorporated BrdU is immunostained with anti-BrdU antibody and proliferation index was calculated as Proliferation index = (No. of BrdU positive nuclei/Total number of nuclei) x100 per field and values have been represented as Mean 6 SE (n = 6) from three independent experiments performed in duplicates.

Immunostaining
MSCs from Mutant, Lean and Control phenotypes at passage 3 were seeded at a density of 1610 4 cells per coverslip, and after attaining the confluency (70%), cells were fixed with freshly prepared 4% paraformaldehyde, permeabilized using chilled 50% methanol (v/v in water), blocked with 4% horse serum and processed for immunolocalization similar to Kiran et al., [29].

Flow Cytometric Analysis
Cell cycle analysis. MSCs isolated from Mutant, Lean and Control phenotypes at passage 3 were analyzed for cell cycle. Briefly 1.0610 6 cells/ml were fixed in 75% ethanol, washed in PBS and incubated with propidium iodide (3 g/L) containing RNase. The percentage of cells in G0-G1, S and G2-M phases was calculated using 488 nm red laser fluorescence and analyzed in BD FACS AriaII (BD, San Jose, CA). Peak fluorescence was gated to discriminate aggregates. The cells were analyzed for cell cycle analysis using FACS Diva Software in FACS Aria II (BD, San Jose, CA) [22].

Semi-quantitative Reverse Transcription Polymerase Chain Reaction
Gene expression for adipogenic/pancreatic/ inflammatory markers. Isolation of RNA, synthesis of cDNA and gene amplification was carried out as per our published protocol [62]. Briefly, RNA was isolated from cultured MSCs (1.0610 6 cells at passage 3) using TriReagent (Sigma, USA). 4 mg of RNA was used for cDNA synthesis using Enhanced Avian Reverse Transcriptase Enzyme (eAMV-RT) (Sigma, USA) and 2 ml of cDNA template was amplified using gene sequences given in Table 3. 5 ml of the amplified products were electrophoresed on 1.2% ethidium bromide stained agarose gel, visualized in GelDoc (BioRad, Italy) and densities of amplicons reflecting quantity of the gene were measured as intensity per square mm using Quanti-tyOne software (BioRad, Italy). Expression levels of genes were normalized against the house keeping gene b-actin and values from three independent experiments performed in duplicates have been indicated (Mean 6 SE).

Differentiation Potential of MSCs
Adipocyte formation. Confluent monolayer of MSCs (passage 3) were trypsinized, seeded at a density of 1.0610 6 cells per 35 mm dish and cultured in cell culture medium till confluence. MSCs were subjected to adipogenic induction, as per the manufacturer's instructions (Chemicon, USA) [23]. At the end of 15 days, cells were fixed in 3% paraformaldehyde and stained with Oil-Red stain to visualize the formation of lipid-droplet filled adipocytes. Adipogenic induction potential was quantitated by counting the number of Oil-red stained cells (indicating mature adipocyte formation) from six random fields using ACT2U software under light microscope (Nikon, Japan) at magnifications indicated using scale bars and values have been represented as Mean 6 SE percentage of oil red positive cells from three independent experiments performed in duplicates.

Insulin Secretion Assay
Insulin secretion assay (ISA) was performed to evaluate the functional response of ILCAs formed from Mutant, Lean and Control Phenotypes. Briefly, 100 ILCAs formed from MSCs were subjected to insulin secretion at basal (5.5 mmol/L) followed by high glucose challenge (17.5 mmol/L) in Krebs Ringer Bicarbonate (KRB) buffer. Insulin secreted into the buffer was estimated using High Range Rat Insulin ELISA Kit (Mercodia, Sweden) and values were represented as mU/L insulin/mg protein [20].

Statistical Analyses
Descriptive statistics (Mean, SD) were calculated for all the samples and values have been represented as Mean 6 SE. To compare between the phenotypes analysis of variance (ANOVA) followed by post hoc tests (LSD/Dunnett's C test) were carried out based on Levene's test for equality of error variances. p,0.05 was considered significant for all the tests. All the statistical analyses were performed using SPSS vs 15.0 software.