Loss of Ron receptor signaling leads to reduced obesity, diabetic phenotypes and hepatic steatosis in response to high-fat diet in mice

Abstract

The Ron receptor tyrosine kinase is a heterodimeric, membrane-spanning glycoprotein that participates in divergent processes, including proliferation, motility, and modulation of inflammatory responses. We observed male C57BL/6 mice with a global deletion of the Ron tyrosine kinase signaling domain (TK^−/−) to be leaner compared with control (TK^+/+) mice under a standard diet. When fed a high-fat diet (HFD), TK^−/− mice gained 50% less weight and were more insulin sensitive and glucose tolerant than controls. Livers from HFD TK^−/− mice were considerably less steatotic and weighed significantly less than TK^+/+ livers. Serum cytokine levels of HFD TK^−/− mice were also significantly altered compared with TK^+/+ mice. Fewer and smaller adipocytes were present in the TK^−/− mice on both control and HFD and were accompanied by diminished adiponectin and peroxisome proliferator-activated receptor-γ expression. In vitro adipogenesis experiments suggested reduced differentiation in TK^−/− embryonic fibroblasts (MEFs) that was rescued by Ron reconstitution. Likewise, signal transducer and activator of transcription (STAT)-3 phosphorylation was diminished in TK^−/− MEFs but was increased after Ron reconstitution. The adipogenic inhibitors, preadipocyte factor 1 and Sox9, were elevated in TK^−/− MEFs and increased in both groups after STAT3 silencing. In total, these studies document a previously unknown function for the Ron receptor in mediating HFD-induced obesity and metabolic dysregulation.

originally a challenge primarily for the western world, obesity has become a worldwide problem in developed and developing countries (15). The World Health Organization estimates that 1.4 billion adults were overweight [body mass index (BMI) >25] and 500 million obese (BMI >30) in 2008, a twofold increase since 1980. Over 2.8 million deaths per year are attributed to being overweight or obese (47). Concurrently, interest in the underlying mechanisms of fat storage and related metabolic disorders such as type 2 diabetes has increased.

The Ron tyrosine kinase receptor (MST1R) is a heterodimeric, membrane-spanning glycoprotein with a mass of ∼185 kDa. Ron is expressed predominantly on macrophages and epithelial cells and directs divergent processes, including proliferation, motility, and modulation of inflammatory responses. Known pathways activated by Ron include phosphatidylinositol 3-kinase (PI3K)/protein kinase B (Akt), β-catenin, Ras, MAPK, and JAK/signal transducer and activator of transcription (STAT). Ron and c-Met share 34% homology and are the only members of their receptor family in mammals. Overexpression of Ron in epithelial cells is associated with cancer of the breast, stomach, colon, pancreas, kidney, and prostate (reviewed in Refs. 1 and 39). The only known ligand for Ron is hepatocyte growth factor-like protein (HGFL), alternatively called macrophage-stimulating protein, which is produced primarily in the liver and secreted into the blood stream in an inactive form that is activated locally by a proteolytic event (9, 43). Ron has also been shown to bind and cross talk with other cell surface receptors, e.g., c-Met, platelet-derived growth factor receptor, epidermal growth factor receptor (EGFR), and insulin-like growth factor-I receptor (IGF-IR) (7, 17, 23, 29, 32).

Analysis of available National Center for Biotechnology Information databases for tissue-specific Ron mRNA expression revealed abundant Ron mRNA in white adipose tissue (WAT) and differential expression in WAT between high- and low-weight-gaining C57BL/6 mice fed a high-fat diet (HFD) (24). We compared weights of male C57BL/6 mice with a global deletion of the intracellular tyrosine kinase region of Ron (TK^−/−) (41) with wild-type (TK^+/+) mice and found the TK^−/− mice to be slightly, but significantly, leaner than their wild-type counterparts on a standard rodent diet. To determine whether Ron may play a role in diet-induced obesity (DIO) and associated metabolic pathologies, we fed male TK^+/+ and TK^−/− mice a HFD. We show here that HFD TK^−/− mice gained 50% less weight, were more insulin sensitive, and had only modest liver steatosis compared with control TK^+/+ mice. After 19 wk of HFD, serum IL-6, macrophage inflammatory protein (MIP)-1β, interferon (IFN)-γ, and granulocyte macrophage colony-stimulating factor (GM-CSF) cytokine levels were surprisingly elevated in the leaner TK^−/− mice compared with TK^+/+ mice, with higher cytokines generally associated with increased obesity and insulin resistance. Additionally, adipocyte size analysis demonstrated markedly smaller adipocytes in the TK^−/− white fat, suggesting a defect in adipogenesis, since no evidence of increased adipocyte cell death was observed. In vitro experiments revealed adipogenic inhibition in TK^−/− embryonic fibroblasts accompanied by decreased STAT3 phosphorylation and increased preadipocyte factor 1 (Pref-1) and Sox9 expression, known inhibitors of adipogenesis (13). This report demonstrates for the first time that the Ron receptor acts as a potent driver of obesity and related phenotypes and thus presents a novel potential target in the treatment of obesity and metabolic syndromes.

MATERIALS AND METHODS

Mice.

C57BL/6 mice with a targeted deletion of the Ron tyrosine kinase signaling domain (TK^−/−) were generated as previously described (41). Wild-type (TK^+/+) and TK^−/− mice were maintained on regular mouse chow until 8 wk of age then fed a high-fat, lard-based diet with ∼45% calories from fat (D12451; Test Diet, Richmond, IN) or regular mouse chow (control) ad libitum. For tissue harvest, mice were killed by CO₂ asphyxiation followed by cervical dislocation. All mice were maintained under specific pathogen-free conditions and handled in accordance with protocols approved by the Institutional Animal Care and Use Committee of the University of Cincinnati.

RNA and real-time quantitative PCR.

RNA was isolated using TRI Reagent (Molecular Research Center, Cincinnati, OH). One microgram of RNA was converted to cDNA with the high-capacity RNA-to-cDNA kit using random primers according to manufacturer's instructions (Applied Biosystems, Foster City, CA). Real-time PCR was performed using FastStart SYBR Green (F. Hoffmann-La Roche, Nutley, NJ). The following genes and corresponding sequences were chosen: INFγ (5′-GATATCTCGAGGAACTGGCAAAA-3′; 5′-CTTCAAAGAGTCTGAGGTAGAAAGAGATAAT-3′), adiponectin (5′-ACGTCATCTTCGGCATGACT-3′; 5′-CTCTAAAGATTGTCAGTGGATCTG-3′), peroxisome proliferator-activated receptor (PPAR)-γ (5′-TCTTCCATCACGGAGAGGTC-3′; 5′-GATGCACTGCCTATGAGCAC-3′), Pref-1 (5′-GTAACTGCCCCTGGCTGTGT-3′; 5′-CTCCAGGTCCACGCAAGTTC-3′), Sox9 (5′-AAGCTCTGGAGGCTGCTGAA-3′; 5′-TTCGGCCTCCGCTTGTC-3′), CCAAT/enhancer-binding protein-β (CEBPB) (5′-CAACCTGGAGACGCAGCACAAG-3′; 5′-GCTTGAACAAGTTCCGCAGGGT-3′), and F4/80 (5′-GAGATTGTGGAAGCATCCGAGAC-3′; 5′-GATGACTGTACCCACATGGCTGA-3′). All others were described by Kulkarni et al. (25). 18S ribosomal RNA was used for normalization.

Luminex assay.

Serum samples were analyzed for cytokines, chemokines, and metabolic hormones with MILLIPLEX Mouse Cytokine/Chemokine 32 Plex and Metabolism Multiplex Assays (EMD Millipore, Billerica, MA) using a Luminex 100/200 multiplex instrument (Luminex, Austin, TX).

Adipocyte isolation.

Epididymal adipose tissue was minced with a razor blade in a sterile petri dish, passed through a 1-mm tissue sieve, and digested with collagenase by rotating at 120 revolutions/min for 30 min at 37°C. Adipocytes were floated by centrifugation at 50 g and retrieved. Preadipocytes were separated from the stromal vascular pellet by differential centrifugation.

Metabolic measurements and telemetry.

Metabolic measurements were performed using an Oxymax indirect calorimetry system (Columbus Instruments, Columbus, OH). Mice were individually housed with a 12:12-h light-dark cycle at 22–24°C. Metabolic rate (V̇o₂), respiratory quotient (RQ, ratio of V̇co₂/V̇o₂), and activity (counts) were evaluated over 24 h. To monitor body temperature, animals received telemetry implants (TA-F20; Data Sciences International, New Brighton, MN). Under anesthesia, a small incision was made in the flank, and the capsule was inserted in the abdominal cavity and anchored with suture to the muscle wall. All incisions were closed with suture and tissue adhesive. Animals were allowed to recover for at least 7 days before the onset of testing.

Intestinal fat absorption test.

Intestinal fat absorption was measured by the method described by Jandacek et al. (16). Briefly, mice were fed a semisynthetic diet containing 5% sucrose polybehenate, a nonabsorbable fat additive, in a semisynthetic diet containing absorbable fat ad libitum for 3 days. Two or more fecal pellets were collected on days 3 and 4. Fat absorption was calculated from the ratios of behenic acid to other fatty acids in the diet and in the feces as analyzed by gas chromatography of fatty acid methyl esters after extraction with hexane.

Glucose and insulin tolerance tests.

Mice were fasted overnight (16 h) before glucose tolerance measurements. d-Glucose (1.5 g/kg body wt; Amresco, Solon, OH) was injected intraperitoneally using a 20% wt/vol glucose solution. Insulin tolerance was tested by intraperitoneal injection of 1 U insulin/kg body wt (Humulin-R; Eli Lilly, Indianapolis, IN) on mice fasted 3 h. Blood glucose was measured from a tail nick with a Bayer Contour glucometer (Bayer HealthCare, Tarrytown, NY).

Lipid profile.

Serum cholesterol, phospholipid, triglyceride, and free fatty acid levels were determined by colorimetric assays at the Mouse Metabolic Phenotyping Center at the University of Cincinnati.

Liver evaluation.

Steatosis was scored semiquantitatively by a blinded observer assigning a value of one to five based on the degree of macrovesicular accumulation: a value of one representing <5% macrovesicles; two, 5–20%, three, 21–40%; four, 41–60%, and five being >60%. Triglycerides were measured in whole liver tissue by colorimetric assay according to the manufacturer's instructions (Cayman Chemical, Ann Arbor, Michigan).

Adipocyte cell number and size.

Epididymal and inguinal adipose tissue were removed from control diet or HFD mice and weighed. Approximately 0.1 g of fat from each fat pad was washed with PBS, weighed, and then fixed in osmium tetroxide (EMS, Hatfield, PA) as described (5). Briefly, after 3 days of fixation in 3% osmium tetroxide, adipose tissue was transferred to 9% NaCl for 24 h. Next, NaCl was removed, and 10 ml of 8 M urea containing 0.154 NaCl and 0.01% Triton X-100 was added. Tissue was stored for 1 wk with occasional shaking to disassociate adipocytes and passed through a 250-μm filter, and cells in 10 μl (∼100–300) were counted and sized from photographs using NIH ImageJ software (35). Total adipocyte number was determined by multiplying the cell count times the volume and total fat pad mass. For graphing purposes, cell area was grouped by 1,000 μm², i.e., 1,000–1,999, 2,000–2,999, etc.

Adipogenesis.

Mouse embryonic fibroblasts (MEFs) were cultured from 13.5- to 14.5-day embryos. Cells were seeded at 100,000 cells/well in 24-well plates and grown to confluency in DMEM with 10% FBS, sodium pyruvate, glutamine, and gentamicin (basal medium). After 2 days, adipogenesis was initiated with serum-free adipogenesis medium (14) plus 500 μM isobutyl methylxanthine (Sigma, St. Louis, MO) for 2 days. Next, the medium was changed to basal medium plus 1 μM insulin for 2 days followed by basal medium for 6 days. Cells were fixed in 10% formalin and stained with 0.3% Oil Red O (ORO) in 60% isopropanol. After washing with 60% isopropanol and water, ORO was extracted with 100% isopropanol, and absorbance was measured at 500 nm.

Terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick-end labeling analysis.

Paraffin-embedded sections (4 μm) of WAT were analyzed for cell death by terminal deoxynucleotidyl transferase-mediated nick end labeling (TUNEL) staining using fluorescein deoxyuridine triphosphate and antifluorescein antibody coupled to horseradish peroxidase according to the manufacturer's instructions (Roche, Mannheim, Germany). For each section, the number of TUNEL-positive cells per total number of cells in three random ×20-objective high-powered fields was counted by an investigator blinded to treatment group.

Immunoblotting.

Cells or tissues were lysed in RIPA buffer, and protein concentrations were determined using the Pierce BCA assay (Thermo Fisher Scientific, Waltham, MA). Antibodies to Ron β (C-20; Santa Cruz Biotechnology, Santa Cruz, CA), phospho-STAT3 (Y705) and STAT3 (Cell Signaling, Danvers, MA), and actin (C-4; Seven Hills Bioreagents, Cincinnati, OH) were used according to the manufacturer's instructions. Peroxidase-conjugated secondary antibodies were diluted 1:40,000 (Jackson ImmunoResearch, West Grove, PA). Antibody detection was performed according to the manufacturer's instructions with ECL Plus (GE Healthcare Life Sciences, Piscataway, NJ) and exposed to film.

Lentiviral vectors.

Sigma-Aldrich MISSION Lenti-short-hairpin RNA (shRNA) for STAT3 (TRCN0000071456, target sequence 5′-CCTGAGTTGAATTATCAGCTT-3′) was purchased from Cincinnati Children's Hospital Medical Center Library Core. For Ron reconstitution, full-length mRON cDNA was inserted in a pCDH lentiviral expression vector (System Biosciences, Mountain View, CA) at the Xba-1, Not-1 sites. Vectors were packaged into transducible particles at the Cincinnati Children's Hospital Medical Center Viral Vector Core, and virus was administered to MEFs in the presence of 4 μg/ml polybrene (Sigma) for 4 h.

Statistics.

Data are expressed as means ± SE. Statistical significance was determined by t-tests for unpaired comparisons or two-way ANOVA for curve comparison with GraphPad Prism software (GraphPad, San Diego, CA). Significance was set at P < 0.05.

RESULTS

TK^−/− mice are leaner under control and HFD.

To investigate the biological significance of Ron in DIO and associated metabolic syndrome pathologies, TK^−/− mice and wild-type (TK^+/+) mice were examined. Slight differences in weight were observed between male mice fed standard rodent chow (control diet). Eight-week-old mice were of similar size, but, by 16 wk, the groups were significantly different. At 24 wk, while body length was unchanged between the groups (data not shown), the average weight of male TK^−/− mice was 3 grams less than TK^+/+ mice (Fig. 1A), and epididymal and retroperitoneal fat pads were substantially smaller (Fig. 1B). TK^+/+ and TK^−/− mice were fed a high-fat, lard-based diet (45% fat calories) and, as shown in Fig. 1, C and D, male TK^−/− mice were dramatically leaner than TK^+/+ mice and weighed 25% less after 8 wk. Female TK^−/− mice appeared smaller, but not significantly so because of higher variability in weight gain (data not shown). Because male C57BL/6 mice have been reported to be more prone to DIO and metabolic syndrome than female mice (31), we focused on the male cohort for the remainder of this study. Brown fat and epididymal and inguinal fat pads were significantly smaller in the TK^−/− mice (Fig. 1E). Total food intake was similar under the control diet and dropped under the HFD, but more so with the TK^−/− mice (Fig. 1F). Food intake normalized to body weight was not different between genotypes (data not shown). While we cannot exclude the notion that the less obese phenotype observed in the TK^−/− mice is dependent on less food consumption on a HFD, the weight gain and fat depots in the TK^−/− mice were significantly less than those of the TK^+/+ mice under control diet conditions (Fig. 1, A and B), when food consumption is similar between genotypes. Intestinal fat absorption efficiency was determined by gas chromatography of fecal pellets and was not statistically different (Fig. 1G), although the TK^−/− group was slightly lower and more variable. A tissue survey of Ron mRNA expression on both diets is shown in Fig. 1H. Ron in epididymal WAT increased 2.5-fold after 19 wk of HFD but decreased significantly in the colon. Expression was unchanged in muscle and liver.

Fig. 1.

C57BL/6 mice with a global deletion of the Ron tyrosine kinase signaling domain (TK^−/−) are leaner than control Ron (TK^+/+) mice. A: time course comparing body weights of TK^+/+ mice with TK^−/− mice fed control diet (n = 11–18 mice). B: epididymal (epi) and retroperitoneal (RP) fat pad weights in TK^−/− mice compared with TK^+/+ mice fed standard chow (n = 11 mice). C: growth curve of TK^+/+ and TK^−/− mice fed a high-fat diet (HFD) for 16 wk starting at 8 wk of age (n = 8 mice). D: TK^+/+ and TK^−/− male mice fed a HFD for 10 wk. E: fat pad weights of TK^+/+ and TK^−/− mice after 10 wk of HFD. BAT, brown adipose tissue; ing, inguinal subcutaneous. F: average HFD food intake per mouse per day up to 7 wk (n = 7 mice). G: fat absorption test comparing control diet-fed TK^+/+ and TK^−/− mice (n = 6 mice). H: Ron mRNA expression in selected tissues comparing control diet with HFD-fed mice. Expression is relative to 18S, and TK^+/+ expression is normalized to 1 (n = 3 mice). I: respiratory quotient (V̇co₂/V̇o₂, RQ) under feeding and fasting conditions. Left, control diet (Ctl); right, 4 wk of HFD (TK^+/+ 0.83 ± 0.004 vs. TK^−/− 0.78 ± 0.009, n = 8 mice). J: volume of oxygen (ml/min) consumed under feeding and fasting conditions normalized to body weight. Control diet on left, HFD on right (n = 8 mice). All n numbers reflect per genotype. *P < 0.05.

Energy expenditure measurements, body temperature, and cage activity.

Next, we compared cage activity and energy expenditure under both diets for 24 h. Under the control diet, activity appeared higher in the TK^−/− mice, but was not significantly different; however, temperature was ∼0.8°C higher in the TK^−/− mice at 4–5 and 22 h. Unexpectedly, activity and body temperature converged after 4 wk of HFD (data not shown). Indirect calorimetry was performed on eight mice of both genotypes to measure energy expenditure. Control diet mice displayed no difference in the RQ or volume of oxygen consumed between groups. Under HFD, the RQ (Fig. 1I) and total oxygen consumed (Fig. 1J), when normalized to body weight, were significantly higher in the HFD TK^−/− mice. Overall, the energy expenditure measurements point to altered metabolism in the HFD TK^−/− mice.

TK^−/− mice are more glucose tolerant and insulin sensitive with less liver steatosis than TK^+/+ mice.

C57BL/6 mice fed a HFD are prone to obesity-related metabolic syndrome pathologies such as glucose intolerance, insulin resistance, and fatty liver. To examine the role of Ron signaling in such metabolic syndromes, control and HFD TK^+/+ and TK^−/− mice were challenged with either glucose or insulin, and blood glucose levels were measured temporally. As shown in Fig. 2A, HFD TK^−/− mice tolerated glucose much better than HFD TK^+/+ mice, and, remarkably, insulin sensitivity in TK^−/− mice did not change after 19 wk of HFD (Fig. 2B). Interestingly, there was no difference between basal insulin levels in control diet groups; however, under HFD, the serum insulin level in TK^−/− mice remained low in contrast to a 4.5-fold increase in TK^+/+ mice (Fig. 2C). Concentrations of C-peptide, a marker of insulin synthesis and release (21), were comparative to insulin levels (Fig. 2D).

Fig. 2.

Ron TK^−/− mice are more glucose tolerant and have less insulin resistance and liver steatosis than wild-type mice. A: ip glucose tolerance test on fasted mice (n = 8 mice). B: insulin resistance test by ip injection. y-Axis is % change of glucose from basal level (n = 8 mice). C: fasted insulin levels as determined by ELISA (n = 4 control, 6 HFD mice). D: serum C-peptide levels of control and HFD (19 wk) TK^+/+ and TK^−/− mice determined by Luminex assay (n = 6 mice). E: hematoxylin and eosin (H&E) staining of representative sections of liver from TK^+/+ and TK^−/− mice fed a control diet or HFD, respectively. Bar = 50 μm. F: semiquantitative steatosis scores of livers from TK^+/+ and TK^−/− mice fed a HFD (n = 8 mice). G: triglyceride (TG) content of TK^+/+ and TK^−/− livers under control and HFD determined by colorimetric assay (n = 4 control, 8 HFD mice). H: comparative liver weights of TK^+/+ and TK^−/− mice fed a control and HFD normalized to body weight (n = 8 mice). I: serum lipid profile of TK^+/+ and TK^−/− mice fed a control and HFD by colorimetric assay. CH, cholesterol; PL, phospholipid (n = 6 mice). All n numbers reflect per genotype. *P < 0.05.

Hepatic steatosis is associated with obesity (6). Consequently, livers from TK^+/+ and TK^−/− mice were evaluated by overall weight and histologically by heamatoxylin and eosin staining. Figure 2E shows representative liver sections from mice fed a HFD for 19 wk. Steatosis is evident in livers from both groups of HFD mice; however, TK^−/− livers were notably less damaged and were 40% smaller than the enlarged TK^+/+ livers (Fig. 2, F and H, respectively). Steatosis scores were confirmed by tissue triglyceride measurement (Fig. 2G). Comparison of lipid profiles (Fig. 2I) showed that total cholesterol and phospholipids increased for both HFD groups but significantly less so in the TK^−/− mice. Triglyceride levels were not different between genotypes. Similarly, no differences were observed in free fatty acid levels (data not shown).

Serum inflammatory cytokines are more highly elevated in TK^−/− mice under HFD.

Under standard diet, TK^−/− serum cytokine levels are low and similar to control mice (26). Obesity is characterized by a chronic condition of low-grade inflammation. Serum cytokine concentrations were compared by Luminex assay from HFD TK^+/+ and TK^−/− mice. Whereas serum leptin and resistin levels were not significantly different (Fig. 3A), the leaner, more insulin-sensitive TK^−/− mice displayed significantly higher serum IL-6, IFNγ, MIP-1β, and GM-CSF compared with TK^+/+ mice (Fig. 3B). A tissue analysis of IL-6 expression levels by real-time quantitative (RT-q) PCR showed increased expression in TK^−/− brain tissue (Fig. 3C). WAT was a major source of cytokine mRNA expression (Fig. 3C), and, as shown in Fig. 3D, levels of IL-10 and tumor necrosis factor-α were lower in TK^−/− WAT. No difference in F4/80 expression was observed between groups, suggesting similar macrophage infiltration into WAT.

Fig. 3.

HFD Ron TK^−/− mice exhibit altered cytokine profiles. A: serum leptin and resistin levels of HFD mice were similar between genotypes as determined by Luminex assay (n = 6 mice). B: serum cytokine levels of HFD mice as determined by Luminex assay (n = 3 mice). C: real-time quantitative (RT-q) PCR tissue survey of HFD mice for IL-6 expression relative to 18S (n = 3–4 mice). White adipose tissue (WAT) contained the highest message level of IL-6 compared with other tissues with relative levels across samples shown. D: interferon (IFN)-γ, IL-6, IL-10, tumor necrosis factor (TNF)-α, and F4/80 mRNA levels relative to 18S in HFD WAT (n = 5 mice). TK^+/+ values are normalized to 1. Mice were fed a HFD for 19 wk starting at 8 wk of age. All n numbers reflect per genotype. *P < 0.05.

TK^−/− mice have modified adipose cellularity with fewer and smaller adipocytes.

To examine the decreased fat pad size in TK^−/− mice, we quantified adipocyte number and cross-sectional area from control diet and HFD (fed 10 wk) epididymal WAT. Under the control diet, the proportional frequency of the smallest adipocytes was higher in TK^−/− compared with TK^+/+ WAT, with cells of area 1,000–2,000 μm² comprising 36.5% (±6.1) vs. 21.3% (±4.3), respectively, of the population. Under the HFD, a pronounced peak at 4,000–5,000 μm² was observed in TK^−/− adipocytes compared with a much broader peak for TK^+/+ cells, with a median at 11,000–12,000 μm² (Fig. 4A). The number of epididymal adipocytes increased appreciably in the TK^+/+ mice when fed a HFD but did not significantly change in TK^−/− mice (Fig. 4B) (6).

Fig. 4.

Ron TK^−/− mice have fewer and smaller adipocytes. A: size distribution of adipocytes from TK^+/+ and TK^−/− mice fed a control diet (n = 8 mice) or HFD (n = 3 mice) for 10 wk. B: total number of adipocytes from epididymal and inguinal fat pads from TK^+/+ and TK^−/− mice fed a control diet or HFD (n = 4 mice). C: relative adiponectin and peroxisome proliferator-activated receptor (PPAR)-γ mRNA expression in epididymal WAT from TK^+/+ and TK^−/− mice on HFD by RT-qPCR (n = 8 mice). Expression is relative to 18S, and TK^+/+ values are normalized to 1. D: relative Ron and hepatocyte growth factor-like protein (HGFL) mRNA expression in purified epididymal WAT adipocytes from TK^+/+ mice fed a control diet or HFD (n = 3 mice). Expression is relative to 18S, and TK^+/+ values are normalized to 1. E: terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick-end labeling analysis (TUNEL) staining of representative WAT from mice fed a HFD for 10 wk. Bar = 100 μm. Arrows show representative staining. F: percentage of positive TUNEL-stained nuclei to total nuclei (n = 3 ×20 fields from 3 mice). All n numbers reflect per genotype. *P < 0.05.

Adiponectin is an adipokine secreted by adipocytes that promotes insulin sensitivity and lipid oxidation (49) and is a marker for differentiation. An autocrine role for adiponectin in adipogenesis and proliferation has been reported in 3T3-L1 preadipocytes (8), and lower gene expression was found in WAT from individuals demonstrating impaired adipogenesis (27). As seen in Fig. 4C, adiponectin mRNA expression in epididymal WAT was significantly lower in TK^−/− compared with TK^+/+ HFD mice. Similarly, expression of PPARγ, a major adipogenic transcription factor that is highly induced in WAT during adipogenesis (38), was 2.5-fold less in TK^−/− compared with TK^+/+ mice (Fig. 4C). To compare adipocyte Ron expression without interference from macrophages or preadipocytes, we purified adipocytes from control and HFD epididymal WAT. Ron expression was modestly increased in the purified HFD adipocytes; however, expression of its ligand, HGFL, doubled under the HFD (Fig. 4D). TUNEL staining of HFD WAT sections was equivalent between groups, suggesting that increased adipocyte death is not a cause of the reduced TK^−/− fat pad size (Fig. 4, E and F). Furthermore, the diminution of adipocyte size in HFD WAT sections is apparent by histological analysis (Fig. 4E). Taken together, the shift toward smaller and fewer adipocytes and decreased PPARγ and adiponectin expression as well as no difference in adipocyte death suggests a deficiency of adipogenesis in the TK^−/− mice.

TK^−/− MEFs have impaired adipogenesis associated with decreased STAT3 signaling and increased Pref-1 and Sox9 expression.

To analyze Ron protein expression in preadipocytes undergoing adipocytic conversion, the well-characterized 3T3-L1 cell line was used (10). Western blot analysis showed increased Ron protein expression at 16 h and at 5 days after treatment with adipogenic media (Fig. 5A), suggesting a role in early differentiation. To investigate whether a defect in adipogenesis might contribute to the reduced adiposity of the TK^−/− mice, TK^+/+ and TK^−/− MEFs were grown to confluence and treated with an adipogenic cocktail, and after 8 days accumulated lipid was stained with ORO. As shown in Fig. 5B, TK^−/− MEFs exhibited impaired differentiation as judged by the comparison of ORO staining and extracted stain quantitation, as well as reduced PPARγ expression during conversion (Fig. 5C). Ron loss did not affect survival of the TK^−/− MEFs during differentiation, since the plates were equally confluent after 8 days of differentiation (Fig. 5B, as judged by crystal violet staining). Ron mRNA expression in TK^+/+ MEFs was quantitated by RT-qPCR at 24, 48, and 96 h after adipogenic initiation and increased 2.5-fold, peaking at 48 h (Fig. 5C). To provide further evidence of the role of Ron in adipogenesis, TK^−/− MEFs were transduced with a lentiviral Ron expression construct or empty vector control, and treated with adipogenic media (Fig. 5D). Ron reconstitution allowed for enhanced ORO staining (Fig. 5D). Ron reconstitution was confirmed by examining mRNA and protein expression levels (Fig. 5E) and significantly increased lipid accumulation over the empty vector control as determined by extracted ORO stain quantification (Fig. 5F).

Fig. 5.

Adipogenesis is impaired in the absence of Ron. A: immunoblot showing Ron protein expression by Western analysis in 3T3-L1 cells during initiation of adipogenesis at 0, 4, and 16 h of differentiation and 5 and 14 days (d). Actin is shown as a loading control. Representative Oil Red O (ORO) staining of 3T3-L1 cells at 5 and 14 days after initiation of adipogenesis (right). B: ORO staining of mouse embryonic fibroblasts (MEFs) 8 days after adipogenesis initiation. Crystal violet staining is below for comparison of cell number. Right, quantification of extracted ORO stain by absorbance measurement at 500 nm (n = 4 wells and is representative of n = 3 independent experiments). C: relative PPARγ mRNA expression by RT-qPCR in MEFs over 64 h of differentiation (n = 6 wells from at least 2 independent experiments). Right, relative Ron mRNA expression by RT-qPCR in TK^+/+ MEFs undergoing adipogenesis. The 0 time point is normalized to 1 (average of 3 experiments). D: ORO staining of control (Ctl), vector-infected [empty vector (e.v.)], or Ron-reconstituted TK^−/− MEFs treated as in B. Crystal violet staining on the right is for comparison of cell number (data are representative of at least 3 independent experiments) E: top, RT-qPCR (n = 3 wells) for reconstituted Ron and e.v. in TK^−/− MEFs normalized to TK^+/+ expression. Bottom, Western blot showing reconstituted Ron expression. Two independent samples from different experiments are shown. F: quantification of extracted ORO from D (n = 4 wells and is representative of 3 independent experiments). 18S was the internal control for all RT-qPCR. All n numbers reflect per genotype. *P < 0.05.

Ron has been shown to activate STAT3 in myeloid cells (11, 48) and the JAK/STAT pathway. Most notably, STAT5 has been implicated in adipogenesis (33). More recently, STAT3 has been reported to mediate adipocyte differentiation upstream of PPARγ (42, 50), and inhibition of STAT3 suppressed adipogenesis in 3T3-L1 cells (18). No differences were detected in pSTAT5 in TK^+/+ vs. TK^−/− MEFs (data not shown). However, STAT3 (Tyr⁷⁰⁵) phosphorylation was strikingly reduced in TK^−/− MEFs (Fig. 6A), which was restored (Fig. 6B) after lentiviral re-expression of Ron. HGFL stimulation of TK^+/+ MEFs induced rapid phosphorylation of STAT3 as shown in Fig. 6C, suggesting a direct Ron effect. Treatment of TK^+/+ and TK^−/− MEFs with 10 ng/ml of recombinant IL-6 showed equal pSTAT3 signaling by Western analysis in both groups, suggesting that IL-6 signaling is intact in TK^−/− MEFs (data not shown). Interestingly, when STAT3 expression was silenced by transduction with a short-hairpin RNA (shSTAT3), expression of Pref-1, a potent inhibitor of adipogenesis (37), and the Pref-1 target Sox9 (45) increased in both MEF genotypes and were 2.5- and 1.5-fold higher in TK^−/− MEFs (Fig. 6, E and F, respectively). The shSTAT3 cells were not viable under adipogenic conditions. However, Pref-1 and Sox9 expression was significantly higher in untreated confluent and differentiating TK^−/− MEFs compared with TK^+/+ MEFs (Fig. 6, G and H, respectively). Purified TK^−/− preadipocytes also exhibited increased Pref-1 and Sox9 expression compared with control cells and were associated with decreased CEBPβ, an early adipogenic event protein (Fig. 6, I and J, respectively) (4).

Fig. 6.

Signal transducer and activator of transcription (STAT)-3 signaling is reduced in TK^−/− MEFs and is associated with increased preadipocyte factor 1 (Pref-1) and Sox9 expression. A: immunoblot for phospho (p)-STAT3 from nonconfluent TK^+/+ and TK^−/− MEFs (n = 3 wells). B: immunoblot for pSTAT3 in TK^−/− MEFs transduced with a lentiviral expression vector for murine Ron or empty vector control (n = 3 wells). C: representative Western blot showing increasing pSTAT3 in 5-h-starved TK^+/+ MEFs after treatment with HGFL. Total STAT3 and actin are loading controls. D: representative Western blot showing STAT3 knockdown by lentiviral shSTAT3 treatment in nonconfluent MEFs. Total STAT3 and actin are loading controls. E and F: relative Pref-1 and Sox9 mRNA expression, respectively, by RT-qPCR in nonconfluent TK^+/+ and TK^−/− MEFs treated with shStat3 virus or control virus (n = 3 wells). G and H: relative Pref-1 and Sox9 mRNA expression, respectively, by RT-qPCR in TK^+/+ and TK^−/− MEFs during adipogenesis (n = 3 wells). I: Pref-1 and Sox9 mRNA expression in cultured, purified preadipocytes from TK^+/+ and TK^−/− mice. J: relative CCAAT/enhancer-binding protein-β (CEBPB) expression is lower in cultured, purified preadipocytes from TK^−/− mice vs. control mice. Graphs are representative of 3 separate experiments, and the internal control was 18S. All n numbers reflect per genotype. *P < 0.05.

DISCUSSION

In this study, we have shown that Ron TK^−/− mice are leaner than Ron TK^+/+ mice under standard rodent diet and dramatically so under HFD, and are the first to demonstrate a function for the Ron receptor in relation to obesity. Body length was unaffected by the lack of Ron such that general growth stunting was not a factor. The difference in weight gain was primarily attributable to smaller epididymal and inguinal fat pads and less hepatosteatosis in the TK^−/− mice. The difference in adiposity was associated with limited adipocyte size expansion in the TK^−/− mice. Liver steatosis and enlargement were severe in the TK^+/+ mice, whereas the TK^−/− mice were mildly steatotic. Whether this is a direct effect of Ron signaling in liver has not yet been determined. We previously demonstrated a hepatic-protective phenotype in TK^−/− mice wherein Ron functioned to limit NF-κB signaling in hepatocytes (36). Currently, the role of NF-κB activation in the progression of fatty liver is not clear. Obesity is associated with low-grade inflammation, whether as a failed defense system or as a byproduct of adipocyte cell lineage is unknown (2, 3). Elevated inflammatory cytokines, such as IL-6 and IFNγ, are associated with obesity and insulin resistance, although the role of IL-6 in metabolic syndrome is dubious (28, 30). For example, IL-6-overexpressing mice were protected from DIO (34), and IL-6 knockout mice unexpectedly grew obese with age (40), but normal mice infused with IL-6 had increased hepatic insulin resistance (19, 22). When challenged with inflammatory agents, Ron-deficient mice are known to be immunologically dysregulated (41, 48). Serum cytokines were elevated in both groups of mice after 19 wk of HFD, with IL-6, IFNγ, MIP-1β, and GM-CSF being more so in the TK^−/− serum. IFNγ and IL-6 mRNA expression were similar in whole WAT from both mice groups although higher IL-6 mRNA expression was observed in the brain and trending higher levels in the spleens of TK^−/− mice. Increased macrophage infiltration was not detected, and 24-h body temperature and activity were not different between the two groups under HFD. It is unclear if the higher cytokines in the TK^−/− mice serum are a secondary effect due to HFD feeding for 19 wk or whether important cytokine changes possibly manifested at different temporal time points may be critical for the phenotypes observed. Further experiments with conditional or inducible deletions of Ron could help discern the contribution of macrophages and (pre) adipocytes to the leaner phenotype.

The differences in the adipocyte size profiles led us to explore a cell autonomous defect in adipogenesis as a possible contributor to the leaner phenotype. Even under control diet conditions, the relative abundance of small adipocytes was increased in TK^−/− epididymal WAT. In vitro experiments show for the first time that Ron mRNA and protein expression are upregulated during adipogenesis. Furthermore, experiments with MEFs suggest that functional Ron signaling may contribute to differentiation and lipid accumulation given the adipogenesis defect we observed in TK^−/− MEFs. Alternatively, it is possible that the smaller and fewer adipocytes we observed in the TK^−/− mice may reflect enhanced adipocyte apoptosis and fat wasting, although our TUNEL staining does not support this conclusion. While the mechanisms behind this in vivo effect require further study, cell surface receptors and receptor tyrosine kinases such as the insulin receptor and IGF-IR have been shown to be key drivers of adipogenesis. The involvement of other receptor tyrosine kinases is less clear. In vitro, EGFR has been shown to negatively and positively influence adipocyte differentiation (12, 44). Heterodimerization with whole Ron cannot be definitely ruled out; however, we were unable to show cross-signaling or coimmunoprecipitation between Ron and IGF-IR in 3T3-L1 cells and MEFs using HGFL and IGF-I as stimuli (data not shown). Ron has been shown to activate several pathways that modify adipogenesis, including PI3K/AKT, MAPK, and STAT3. Activated STAT3 (pSTAT3) was diminished in the TK^−/− MEFs and was partially restored with Ron reconstitution along with increased lipid accumulation. In a search for inhibitor candidates, we found higher mRNA expression of Pref-1 and its target Sox9 in differentiating TK^−/− MEFs. Pref-1 is a cell surface molecule exclusively produced by preadipocytes that becomes an adipogenic inhibitor when cleaved into a soluble form (46). In MEFs undergoing adipogenesis, Pref-1 expression is initially increased before decreasing after several days (20). Treatment of MEFs with STAT3 shRNA increased Pref-1 and Sox9 expression, more so in the TK^−/− MEFs. The data suggest that STAT3 may be a negative regulator of Pref-1 expression, and this may be a mechanism by which Ron promotes adipogenesis, although further study is needed to fully support this claim. While Ron may be an important factor regulating adipocyte number and/or differentiation, the effects observed in the TK^−/− mice on HFD, such as the decreased steatosis, suggest the phenotypes may be a result of adipocyte cell autonomous and noncell autonomous effects. Therefore, additional experimentation is needed to discern the cell type-specific functions of the Ron receptor in mediating obesity-related phenotypes.

Taken together, our data suggest that Ron is important in HFD-induced obesity and may also play a role as an adipogenic promotion factor. WAT Ron expression during the intake of fat calories is associated with increased PPARγ and adiponectin expression and weight gain as well as the development of hyperinsulinemia. In vitro, we have shown that Ron is expressed in the mouse preadipocyte cell line 3T3-L1; in MEFs, is upregulated during terminal differentiation into adipocytes; and may contribute to adipogenesis through STAT3. Further investigation of the role of Ron in adipogenesis and adipocyte homeostasis may lead to novel therapeutic approaches for obesity and metabolic syndromes.

GRANTS

This work was supported in part by National Institutes of Health Grants R01-DK-73552 (S. E. Waltz), R01-CA-125379 (S. E. Waltz), F31-CA-165767 (A. M. Paluch), T32-CA-117846 (S. E. Waltz and N. E. Brown), P30-DK-078392, and U24-DK-059630 (Mouse Metabolic Phenotyping Core) as well as BX-000803 (S. E. Waltz) from the Veterans Affairs Merit Program.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

Author contributions: W.D.S. and S.E.W. conception and design of research; W.D.S., N.E.B., and A.M.P. performed experiments; W.D.S., N.E.B., A.M.P., and S.E.W. analyzed data; W.D.S., N.E.B., A.M.P., and S.E.W. interpreted results of experiments; W.D.S. and N.E.B. prepared figures; W.D.S. drafted manuscript; W.D.S., N.E.B., A.M.P., and S.E.W. approved final version of manuscript; N.E.B., A.M.P., and S.E.W. edited and revised manuscript.