The murine lineage of preadipocytes NIH3T3-L1 was obtained from the National Bank of Cells, Federal University of Rio de Janeiro, Brazil. To keep preadipocytes in their undifferentiated state, we cultured the cells with medium containing Dulbecco's Modified Eagle Medium (DMEM) (Gibco) with 4.5 g/L of glucose (Invitrogen) supplemented with penicillin (100 U/mL) and streptomycin (100 μg/mL) (Gibco) and with 10% of bovine serum (Invitrogen). Preadipocytes (1.05 × 10⁴ cells/cm² of adherence area) were seeded in polystyrene culture plates (Biofil). Four days after the seeding (day 0) medium was replaced by the differentiation-inducting medium containing DMEM with 4.5 g/L of glucose, penicillin (100 U/mL), streptomycin (100 μg/mL), 10% of fetal bovine serum, 1 μM of dexamethasone (Sigma Aldrich), 0.5 mM of isobutylmethylxantine (IBMX) (Sigma Aldrich) and 0.3 units of insulin/mL (Regular Humulin—Lily). The differentiation induction medium was maintained for 3 days and then replaced by the maturation medium, containing DMEM with 4.5 g/L of glucose, penicillin (100 U/mL), streptomycin (100 μg/mL), 10 % of fetal bovine serum and 0.3 units/mL of insulin. Seventy-five percent of the maturation medium was renewed every 2–3 days.

For the experiments in which cells were differentiated in the presence of leptin, the first supplementation with leptin [4 or 40 nM, doses approximately matched to the circulating levels of leptin in obese and morbidly obese individuals (23–25)] occurred at day−3 (1 day after plating) and leptin was added at every medium renewal.

To investigate the ability of leptin to induce preadipocyte differentiation in the absence of insulin, insulin was replaced by leptin at day 0 and differentiation was performed as aforementioned.

Subcutaneous and retroperitoneal fat pads were isolated from male C57Bl/6J mice of 6 weeks of age. In order to achieve a considerable amount of stem cells to expand and differentiate we pooled right and left tissue depots from three animals for each experimental replicate. For the experiments using LepR-deficient mouse, we pooled the right and left tissue depots (subcutaneous, retroperitoneal and epididymal) from one BKS.Cg-m⁺/⁺Lepr^db/JUnib male mouse (UNICAMP, Campinas, SP, Brazil). Tissues were washed with phosphate-buffered saline (PBS) solution with penicillin (100 U/mL) and streptomycin (100 μg/mL) (Sigma) and 5 μg/mL of ciprofloxacin (Sigma) and then cut into small pieces. These pieces were rewashed with the same solution and incubated for 2 h with 2 mg/mL of collagenase (Sigma) in a volume of 2.5 mL per mg of tissue with constant shaking at 37°C. Collagenase was inactivated by doubling the volume with media containing antibiotics and 10% fetal bovine serum (Life Sciences). Digested tissues were then filtered in cell strainers of 100 μm pore size and centrifuged at 500 × g for 7 min. The superior fraction which contained mainly lipids and mature adipocytes was discarded and the pellet was resuspended in PBS plus penicillin (100 U/mL) and streptomycin (100 μg/mL), 5 μg/mL of ciprofloxacin, filtered again in cell strainers of 40 μm pore size and centrifuged at 500 × g for 7 min. The pellet of stromal vascular cells was then resuspended in culture media containing DMEM with 4.5 g/L glucose, penicillin (100 U/mL) and streptomycin (100 μg/mL), 5 μg/mL of ciprofloxacin, and 20% of fetal bovine serum (Life Sciences) and cultured. Cells were expanded 3–4 times before plating. All animal procedures were approved by the Committee of Ethics in Animal Research L011.2015.

Stromal vascular cells expanded up to two times were labeled with ASCs' positive (CD44, CD29, CD106, and CD105) and negative (MHC-class II, CD11b, CD31, CD45, and CD144) markers. Cells were incubated (30 min) with FITC-conjugated anti-CD45 (eBioscience, cat 12-1051-81, dilution 1:20); -CD31 (eBioscience, cat 11-0311-81, dilution 1:20) and -MHC class II (eBioscience, cat: 11-5320-82, dilution 1:20); APC-conjugated anti-CD11b (BD Pharmingen, cat 553312, dilution 1:20), and PE-conjugated anti-CD29 (eBioscience, cat 12-0291-81, dilution 1:20) or -CD105 (eBioscience, cat 12-1051-81, dilution 1:10). For evaluation of CD106 expression, cells were incubated (30 min) with rat anti-mouse CD106 (eBioscience, cat 14-1061-81, dilution 1:10) followed by 30 min incubation with Alexa Fluor® 546-conjugated anti-rat IgG (Molecular Probes, cat: A-11081, dilution 1:250); unbound antibodies were washed out and cells were incubated (30 min) with FITC-conjugated anti-CD45, -CD31, and -MHC class II, and APC-conjugated anti-CD11b. For evaluation of CD44 (eBioscience, cat 11-0441-81, dilution 1:20) expression, cells were incubated (30 min) with unconjugated rat antibodies against CD45 (BD Biosciences, cat: 550539, dilution 1:10) and CD144 (eBioscience, cat: 16-1441-85, dilution 1:20) followed by 30 min incubation with AlexaFluor 546-conjugated anti-rat IgG; unbound antibodies were washed out and cells were incubated (30 min) with FITC-conjugated anti-CD44 and APC-conjugated anti-CD11b antibodies. Cells incubated with isotype-matched IgG conjugated with the same fluorochromes or unconjugated IgG followed by incubation with the secondary antibody were used as a negative control. Cells were acquired in a Beckman Coulter CytoFLEX S using CytExpert software and analyzed using FlowJo v10 software. For analysis, cells were gated by the exclusion of leucocytes and endothelial cells markers (CD45, MHC class II, CD11b, CD31, and CD144) and the expression of ASCs markers evaluated as shown in Supplementary Figure 3.

Cells were fixed for 15 min with formaldehyde 3.7 %, washed with buffered saline, and stained with BODIPY^TM 493/503 (ThemoFisher Scientific) for 30 min and DAPI (ThemoFisher Scientific) for 5 min. Images were acquired with the microscope Olympus BX60 and analyzed with the software Fiji (26) version 1.49 m (National Institutes of Health, USA) with Java version 1.6.0_24 (64-bit). We developed a macro to analyze the total Bodipy stained area (green) in each field adjusting the same parameters of color balance, contrast, background, and noise. Images were processed so that the threshold setting for quantifying the total and relative area of Bodipy staining excluded most of the interferences from the image acquisition. A different macro was developed for the counting of nuclei numbers in each field (DAPI—blue). Then, total Bodipy stained area was normalized by the number of cells in each field and the mean of these measurements was plotted for each group.

Cells were washed with phosphate-buffered saline (PBS) solution and then subjected to lysis directly by 95°C-heated Laemmli buffer with concomitant cell scraping from the wells. Protein content was separated by 10–15% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to nitrocellulose membranes. Membranes were blocked with 5% non-fat dried milk diluted in Tris-Buffered Saline supplemented with 0.1% Tween 20 (Sigma) (TBS-T) for 1 h before incubation overnight with primary antibodies against PLIN1 (Cell Signaling, cat 3470, dilution 1:500), JAK2 (Cell Signaling, cat 3230, dilution 1:1,000), pTyr1007/1008 JAK2 (Cell Signaling, cat 3771, dilution 1:1,000), pTyr389 S6K (Cell Signaling, cat 9205, dilution 1:1,000), S6K (Cell Signaling, cat 2708, dilution 1:1,000), pTyr705 STAT3 (Cell Signaling, cat 9131, dilution 1:1,000), STAT3 (Cell Signaling, cat 9139, dilution 1:1,000), PPARγ (Santa Cruz, cat sc-7196, dilution 1:1,000), SREBP1 (Santa Cruz, cat sc-13551, dilution 1:1,000), and for 1–2 h with anti-β-actin (Sigma, cat A2228, dilution 1:10,000) or CAV-1 (Santa Cruz, cat sc-894, dilution 1:2,000). Membranes were then washed with TBS-T and proteins were detected using fluorescent dye-conjugated secondary antibodies (anti-mouse, IRDye 800CW, cat 926-32210 or anti-rabbit, IRDye 680RD cat 926-68071, LI-COR) or horseradish peroxidase-conjugated secondary antibodies (anti-rabbit, cat PI-1000 or anti-mouse, cat PI-2000; Vector).

Supernatants were harvested from cultured preadipocytes and adipocytes or differentiating ASCs at different timepoints and the levels of leptin, adiponectin, TNF-α, IL-10, IL-6, KC/CXCL1, and MCP-1were measured using standard ELISA protocols according to manufacturer's instructions (R&D Systems).

Statistics were performed using GraphPad Prism (San Diego, CA) version 6.05. All numeric variables were tested for normal distribution using the Kolmogorov-Smirnov normality test. For comparison among three or more groups, we used One-way ANOVA with Dunnet's post-test to compare differentiated cells with preadipocytes, or Tukey's post-test to compare leptin-treated with untreated cells among groups following a normal distribution. In cases of non-parametric distribution among three or more groups, we used One-way ANOVA, followed by Kruskal-Wallis' test with Dunn's correction. To analyze the effect of rapamycin over the effect of leptin we used Two-way ANOVA with Sidak's post-test to locate the differences among groups. For comparison between two groups we used the Mann-Whitney U-test for non-parametric distribution. For comparisons when the groups were normalized by the control group (thus containing repeated value) we used the Kolmogorov-Smirnov non-parametric test.

3T3-L1 cells progressively differentiated into adipocytes up to 17 days after the induction of adipogenesis (day 0) with increased formation of lipid droplets accompanied by enhanced secretion leptin and adiponectin (Supplementary Figures 1A–D). Briefly, the adipocyte differentiation protocol (Figure 1A) consisted of plating cells at day−4, followed by administration of the adipogenic medium (insulin, IBMX, dexamethasone) at day 0—so called for being the start-point of adipogenesis in vitro—which was maintained until day 3. Afterwards, cells were incubated in insulin-containing medium and kept in this medium until the end of the assay (Figure 1A) as described in more details in section Materials and Methods. In order to evaluate the effects of leptin in the adipogenic process, 3T3-L1 cells were incubated with leptin supplementation from the first day post-plating (day−3) until the end of the differentiation protocol (day 17). The cells were analyzed at the preadipocyte stage (day−1) and at the three stages of adipocyte maturation (days 3, 10, or 17) (Figure 1A). Leptin significantly enhanced lipid accumulation in preadipocytes, showing an anticipation of this adipogenic feature when compared to the control cells, without leptin (Figures 1B,C). Leptin lead to increased expression of the lipid droplet protein PLIN1 in adipocytes (Figure 1D). In addition, preadipocytes and adipocytes cultured with leptin exhibited increased levels of the adipogenesis- and lipogenesis-related factors PPARγ, SREBP1C and Caveolin-1 (CAV-1), mainly in preadipocytes and adipocytes at the first stages of differentiation (Figure 2D; Supplementary Figure 2 shows the replicates of all Western blots). These results show a synergistic proadipogenic effect of leptin and insulin.

Leptin binds to its LepRb receptor and activates JAK2, which can activate STAT3 and PI3K/AKT/mTOR pathways (9, 10, 15). To dissect leptin signaling pathways in adipocytes, differentiated adipocytes were stimulated with leptin for 20 min and the phosphorylated proteins of each pathway were analyzed through Western blot. We observed increased phosphorylation of JAK2, AKT, and STAT3 in adipocytes after stimulation with leptin (Figures 2A,B; Supplementary Figure 2), confirming the activation of the two branches of the leptin signaling pathway in adipocytes. Leptin also induced the phosphorylation of p70 S6 kinase (S6K), a substrate of mTOR (Figure 2C). Considering the importance of mTOR signaling in the induction of adipogenesis (11, 12, 27, 28), we asked whether leptin can modulate this pathway in undifferentiated cells (Figure 2D). We found that leptin also activated the mTOR pathway in preadipocytes, as evidenced by the increased phosphorylation of p70 S6K (Figure 2D). To gain insights into the role of mTOR signaling in leptin-induced adipogenesis, we stimulated preadipocytes with leptin in the presence of rapamycin, a selective mTOR inhibitor, and evaluated the biogenesis of lipid droplets after 48 h. Rapamycin treatment completely abolished the leptin-induced biogenesis of lipid droplets in preadipocytes (Figures 2E,F), indicating that the pro-adipogenic effects of leptin depend on mTOR activation.

Adipose tissue is an important source of cytokines and chemokines, which are collectively called adipokines (29). We observed increased production of TNF-α in 3T3-L1 control cells as adipogenesis progressed (Figures 3A,B; Supplementary Figure 2). Treatment with leptin further increased TNF-α expression by preadipocytes (day−1) and immature adipocytes (day 3) (Figure 3A; Supplementary Figure 2). In addition, treatment with leptin (40 nM) also induced the secretion of TNF-α and adiponectin at earlier stages of differentiation (Figures 3B,C), which is consistent with the anticipation of adipogenesis described above. Also, leptin treatment decreased the secretion of the anti-inflammatory cytokine IL-10 (Figure 3D). The levels of the chemokines KC/CXCL1 and MCP-1/CCL2, on the other hand, were not modulated by leptin treatment (Figures 3E,F). These results suggest that, in addition to adipogenesis, leptin induces a proinflammatory cytokine balance in 3T3-L1 adipocytes.

Insulin is considered essential for the proper differentiation of preadipocytes due to the induction of cell growth and fatty acid synthesis and is an important constituent of adipogenic cocktails (22, 30). Considering the proadipogenic effects of leptin observed above, we investigated whether leptin, in the absence of insulin, would be sufficient for the induction of adipogenesis in 3T3-L1 cells. As shown in Figure 4, 3T3-L1 cells cultured without insulin or leptin supplementation showed impaired lipid droplets' formation (Figures 4A,B) and reduced expression of the adipogenesis-related proteins PLIN1, SREBP1C, PPARγ, and CAV-1 (Figure 4C; Supplementary Figure 2) compared to insulin-differentiated cells. Importantly, replacement of insulin by leptin during 3T3-L1 differentiation recovered the biogenesis of lipid droplets and the expression of adipogenesis-related proteins (Figures 4A–C; Supplementary Figure 2). These data indicate that leptin exerts proadipogenic effects independently of insulin.

Next, we investigated whether leptin has proadipogenic effects in ASCs obtained from the stromal vascular fraction of subcutaneous and retroperitoneal fat pads from C57Bl/6J mice. Characterization of ASCs is described in Methods and shown in Supplementary Figure 3. Similar to 3T3-L1 cells, leptin also synergized with insulin and anticipated the differentiation of ASCs into adipocytes (Figure 5). As shown in Figures 5A,B, concomitant treatment of retroperitoneal—but not subcutaneous—ASCs with leptin and insulin increased lipid accumulation at day 5 of differentiation compared to control cells cultured with insulin only. Morphologically, ASCs cultured with insulin had already achieved maximal lipid accumulation at day 12, so there was no observable effect of leptin in potentiating insulin-induced lipid droplet biogenesis at this time point (Figures 5A,B). On the other hand, induction of adipogenesis and anticipation of differentiation were evidenced by the increased expression of CAV-1, PPARγ, SREBP1C and/or PLIN1 at days 5 and 12 in ASCs obtained from both depots (Figures 5C,D; Supplementary Figure 2). As shown with 3T3-L1, leptin treatment was able to modulate the expression of the adipogenic markers in ASCs. These data show that leptin is able to potentiate insulin-induced ASCs differentiation into adipocytes. In addition, leptin treatment induced a proinflammatory cytokine profile in differentiated ASCs by increasing the levels of TNF-α and IL-6 without affecting IL-10 levels (Figure 6).

As previously mentioned, adipocyte differentiation can be achieved by the combination of adipogenesis induction factors, with insulin being considered indispensable for this purpose (30). We then investigated the ability of leptin to support ASCs commitment to adipocytes in the absence of insulin. As shown in Figure 7, leptin induces lipid accumulation at day 12 of differentiation in subcutaneous and retroperitoneal ASCs when compared to the cells cultured without leptin or insulin supplementation (Figures 7A–C). Also, stimulation with leptin alone supported the expression of the adipogenesis-related proteins CAV-1, PLIN-1, PPAR-γ, and/or SREBP1C in ASCs from both adipose tissue depots (Figures 7D,E; Supplementary Figure 2). In addition, leptin did not induce adipogenesis in ASCs from LepRb-deficient (db/db) mouse, albeit these cells were still able to differentiate in response to insulin (Supplementary Figures 4A,B). Also, leptin signaling through mTOR pathway was absent in ASCs from db/db mice (Supplementary Figure 4C). Our data show that leptin supports the induction of adipocyte differentiation through LepRb signaling, even in the absence of insulin.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.