Tumor Progression Locus 2 (TPL2) Regulates Obesity-Associated Inflammation and Insulin Resistance

Human insulin (Novolin R) and recombinant macrophage colony stimulating factor (M-CSF) were purchased from Novo Nordisk (Denmark) and enScript (Piscataway, NJ), respectively. Antibodies recognizing ERK (9102), phospho-ERK (9101), JNK (9252), phospho-JNK (9251), AKT (9272), and phospho-Akt (9271) were obtained from Cell Signaling (Danvers, MA), and a MAC-2 (CL8942AP) antibody was obtained from Cedarlane (Burlington, ON, Canada). Horseradish peroxidase-conjugated secondary antibodies and bicinchoninic acid protein assay reagents were obtained from Thermo Scientific (Rockford, IL). Enhanced chemiluminescence reagent was purchased from PerkinElmer Life Sciences (Boston, MA). Immunohistochemistry was performed by VectaStain Elite ABC kit (Vecta Laboratory, Burlingame, CA). Cell culture products were obtained from Gibco (Carlsbad, CA), and chemical reagents were purchased from Sigma-Aldrich (St. Louis, MO).

Bone-marrow cells were isolated from femoral and tibial bone marrow of WT and TPL2KO mice and differentiated to macrophages in the presence of M-CSF (17,32). Media for the culture of BMDM was composed of 80% DMEM, 20% heat-inactivated FBS, 2.0 g/L glucose, 100 μmol/L nonessential amino acid, 2 mmol/L l-glutamine, 100 ng/mL M-CSF, and penicillin/streptomycin. The progenitor cells were proliferated and differentiated to mature macrophages for 5 days, at which point adherent viable cells were harvested and plated in six-well plates.

Experiments were conducted in a viral pathogen–free facility at the Jean Mayer-U.S. Department of Agriculture Human Nutrition Research Center on Aging at Tufts University in accordance with Institutional Animal Care and Use Committee guidelines. WT and TPL2KO mice generated from heterozygous matings were used as homozygous breeding pairs to generate WT and KO experimental mice. Genotyping of animals was verified at weaning and again when mice were killed. Male 6-week-old mice were individually caged and fed a normal diet (ND) with 17% calories from fat (Harlan #7012) or HFD with 60% calories from fat (Research Diets D12492) for 17 weeks. Mice were killed by CO₂ narcosis, followed by cardiac puncture. Tissues were dissected, weighed, and snap frozen for analysis of protein or gene expression, or were fixed, embedded in paraffin, and sectioned for histologic analysis.

Fasted (overnight) plasma insulin was measured by ELISA (Crystal Chem, Downers Grove, IL). A hyperinsulinemic-euglycemic clamp was performed for 120 minutes, after an overnight fast, using a primed/continuous infusion of human insulin (126 pmol/kg prime, 18 pmol/kg/min infusion) as described (33). During the clamp, plasma glucose was maintained at basal concentrations. Rates of basal and insulin-stimulated whole-body glucose fluxes and tissue glucose uptake were determined as described (33). To assess peripheral tissue insulin sensitivity, mice were fasted 6 h and then received an intraperitoneal injection of saline (basal) or insulin (stimulated, 0.75 μU/kg). Mice were killed 10 minutes after injection, and blood and adipose, liver, and muscle tissues were collected and snap frozen until Western blotting analyses.

Protein from the frozen tissue samples was isolated as described previously (34). Briefly, gonadal adipose and liver were processed in 5% SDS lysis buffer and muscle was pulverized in liquid nitrogen and processed in a 1% Triton X-100 buffer. Proteins were separated by SDS-PAGE and transferred to Hybond enhanced chemiluminescence nitrocellulose membrane and incubated with the indicated antibody and horseradish peroxidase-coupled anti-species antibodies. Proteins were visualized by enhanced chemiluminescence and quantified by densitometry.

Adipose tissue RNA was extracted using Qiagen Lipid Mini kits and liver RNA was isolated using Qiagen Mini kits according to manufacturer’s instructions. RNA was quantified and checked for purity using the Nanodrop spectrophotometer (Nanodrop 1000, Wilmington, DE). cDNA was generated from 1 μg of RNA, and real-time quantitative PCR was performed using SYBR Green (Applied Biosystems 7300, Carlsbad, CA). Fold-changes were calculated as 2^–ΔΔCT, with cyclophilin A or cyclophilin B used as the endogenous control.

Liver triglyceride content was determined using a modified protocol described by Schwartz and Wolins (35). Briefly, frozen liver was extracted in PBS-10 mmol/L EDTA buffer, and protein concentration was determined by bicinchoninic acid protein assay. Samples were assayed in duplicate, and 25 μg of protein extract was subjected to organic extraction. Triglyceride was quantified using a colorimetric enzyme-linked kit (Sigma).

Gonadal adipose and liver tissue were fixed in 4% formaldehyde overnight, embedded in paraffin, sectioned, and stained with hematoxylin and eosin (H&E). Digital images were acquired with an Olympus DX51 light microscope (Center Valley, PA). Adipocyte death was quantified by identification of crown-like structures (CLS) within histologic sections of epididymal adipose tissue. The percentage of CLS ([number dead adipocytes/number total adipocytes] × 100) was calculated and used for comparison among experimental groups. Adipocyte volume was calculated from cross-sectional area obtained from perimeter tracings using Image J software (Sun Microsystems, Santa Clara, CA) (13).

Data are presented as mean ± SE. Data were determined to have a normal distribution with equal variance, and statistical differences were determined by t test or ANOVA using the Tukey least significant differences test for post hoc determination using Minitab software (University Park, PA). Significance was set at P ≤ 0.05.

The 6-week-old male TPL2KO and WT mice were fed ND or HFD for 17 weeks. Although the HFD increased growth rate and final body weight compared with the ND, we observed no difference in these variables between WT and TPL2KO mice receiving the same diet (Supplementary Fig. 1 and Table 1). Consistent with a lack of difference in body weight, food intake and fat pad mass were similar between obese WT and TPL2KO mice (Table 1). We did observe a small but significant reduction in the liver weight of TPL2KO mice fed an HFD. Obesity is typically associated with decreased insulin sensitivity and elevated circulating concentrations of glucose and insulin. Despite having a similar body weight and fat pad mass, 6-h fasting blood glucose and overnight fasting plasma insulin concentrations were decreased in obese TPL2KO mice, suggesting an improved insulin sensitivity, which is further investigated subsequently. Although we did not observe a difference in blood glucose concentrations after an overnight fast (data not shown), a recent report indicated that in mice, blood glucose concentrations after a 6-h fast are a better representation of blood glucose levels throughout the day (36).

Consistent with an improved metabolic state, obese TPL2KO mice had reduced plasma concentrations of the proinflammatory cytokines TNF-α and IL-6 (Supplementary Table 1). Plasma concentrations of nonesterified fatty acid, leptin, and resistin were not significantly different between obese WT and TPL2KO mice (Supplementary Table 1). We observed no differences in plasma glucose, insulin, or proinflammatory cytokine concentrations in ND-fed WT and TPL2KO mice (data not shown). In summary, obese TPL2KO mice had reduced levels of circulating cytokines and improvements in glucose-insulin homeostasis compared with obese WT mice.

TPL2 is located downstream of IKK-β and regulates MAPK activation, both of which have been implicated in obesity-associated inflammation. Therefore, we investigated the effects of TPL2 ablation on adipose tissue inflammation. A hallmark of obesity-associated adipose tissue inflammation is the localization of macrophages around dead adipocytes to form CLS (6,13). Importantly, the presence of CLS in the adipose tissue of obese TPL2KO mice was reduced by more than 70% (Fig. 1A and C). The observed reduction in CLS occurred independent of a difference in adipocyte size (Fig. 1B) or fat pad mass (Table 1). Consistent with the observed reduction in CLS, gene expression of the general macrophage marker F4/80 was attenuated in the adipose tissue of obese TPL2KO mice (Fig. 2). We observed a similar reduction in the expression of CD11c, a marker for a subset of proinflammatory immune cells that have been demonstrated to be major contributors to the insulin resistance associated with obesity (17,32). In accordance with a reduced inflammatory phenotype, relative mRNA expression of the cytokine TNF-α and the chemokine monocyte chemoattractant protein-1 (MCP-1) were also markedly reduced in obese TPL2KO adipose tissue.

Gene expression of TLR2 and TLR4 was upregulated with obesity in WT mice, and this increase was diminished in obese TPL2KO mice (Fig. 2). Furthermore, lack of TPL2 signaling in adipose tissue attenuated obesity-induced activation of ERK and JNK, which is consistent with our observation of reduced inflammatory gene expression (Fig. 3). In summary, the adipose tissue of obese TPL2KO mice had significantly reduced immune cell infiltration, MAPK signaling, and inflammatory mediators that have been implicated in local and systemic insulin resistance.

Increased hepatic inflammation and steatosis are prominent characteristics of obesity and are risk factors for the development of insulin resistance. Histology suggested that livers of obese WT mice contained more lipid than livers of obese TPL2KO mice (Fig. 4A), and analysis by organic extraction confirmed liver triglyceride content was 30% greater in obese WT mice (Fig. 4B). This reduction in liver triglyceride may explain the small reduction in liver weight observed between obese WT and TPL2KO mice (Table 1). Consistent with our observed differences in liver triglyceride values, several genes associated with lipogenesis and lipid storage, such as PPARγ, GPAT, and ADRP, were significantly decreased in the livers of obese TPL2KO mice compared with their obese WT counterparts (Fig. 4C). Additional analysis of the livers revealed reduced gene expression of macrophage markers (CD11b, CD11c), TLR2, and the cytokines TNF-α and IL-6 in the obese TPL2KO mice compared with obese WT controls (Fig. 4D).

We also observed a reduction in suppressor of cytokine signaling (SOCS)-1 expression and a trend for reduced SOCS-3 expression in obese TPL2KO mice. These data support the notion of an improved liver phenotype, because research has shown expression of SOCS-1 and SOCS-3 is increased in the livers of obese insulin-resistant animals (37,38).

We did not observe an effect of obesity or lack of TPL2 signaling on activation of ERK and JNK in the liver (Supplementary Fig. 2). However, these studies examined ERK and JNK activation in the entire liver, not in specific cell types, such as macrophages or Kupffer cells, in which there may have been an effect. Overall, obese TPL2KO mice have reduced triglyceride accumulation in their livers, reduced expression of SOCS, and decreased mRNA expression of lipogenic and inflammatory genes.

It is generally well accepted that BMD immune cells contribute significantly to the inflammation that occurs with obesity (4,39). Systemic concentrations of fatty acids and endotoxin (lipopolysaccharide) are increased with obesity and are able to induce an inflammatory response through TLR4 activation (16). To better understand our observation of reduced inflammation in obese TPL2KO mice, we cultured BMDM from WT and TPL2KO mice to determine the consequences of TLR activation. Macrophages from WT and TPL2KO mice were differentiated and treated with or without the TLR agonist lipopolysaccharide for 3 h. We observed that in response to lipopolysaccharide treatment, TPL2KO BMDM expressed significantly less TNF-α, IL-6, and MCP-1 mRNA compared with BMDM isolated from WT mice (Fig. 5A). Consistent with the gene expression data, TPL2KO BMDM secreted less TNF-α and IL-6 protein, and there was a trend for reduced MCP-1 secretion (Fig. 5B). The difference in secreted TNF-α between WT and TPL2KO macrophages was significantly more robust than the difference in TNF-α mRNA, a finding consistent with a role for TPL2 in the post-transcriptional regulation of this cytokine (28). In the absence of lipopolysaccharide stimulation, basal secretion of these factors was below the minimum detectable limits of the assays (data not shown). These in vitro studies are consistent with our in vivo observations and confirm a role for TPL2 in modulating the expression of cytokines and chemokines in response to an inflammatory stimulus.

To determine the physiologic consequences of TPL2 ablation on glucose-insulin homeostasis in obese mice, we performed hyperinsulinemic euglycemic clamps on obese WT and TPL2KO mice (Fig. 6; Supplementary Table 2). The infusion rate of exogenous glucose during the clamp was approximately 100% greater in obese TPL2KO mice (22.7 ± 2 mg/kg/min) compared with obese WT mice (11.6 ± 1.6 mg/kg/min; Fig. 6B; Supplementary Table 2). This enhanced insulin sensitivity in obese TPL2KO mice reflects enhanced insulin suppression of hepatic glucose output (89% suppression in obese TPL2KO compared with 30% in obese WT; Fig. 6C) as well as enhanced insulin-stimulated whole-body glucose uptake (Fig. 6D; Supplementary Table 2). The latter is reflected in the enhanced insulin-stimulated glucose uptake in skeletal muscle, which was significantly (80%) increased in obese TPL2KO mice compared with obese WT mice (Fig. 6E). In contrast with its salutary effects on liver and muscle insulin sensitivity, TPL2 ablation had no effect on insulin-stimulated glucose uptake in adipose tissue (Fig. 6E).

We next asked if enhanced hepatic and skeletal muscle insulin sensitivity in obese TPL2KO mice was associated with enhanced insulin signaling, measured as the magnitude of Akt phosphorylation in response to an insulin bolus. Insulin-stimulated Akt phosphorylation was greater in skeletal muscle and adipose tissue of obese TPL2KO mice than in obese WT mice, whereas liver Akt was activated to a similar extent regardless of TPL2 status (Supplementary Fig. 3). Inflammatory gene expression and ectopic triglyceride/metabolite accumulation are both implicated in the compromised peripheral insulin signaling of obesity (40,41). In the current study, enhanced Akt phosphorylation in skeletal muscle of TPL2KO mice was not associated with significantly reduced skeletal muscle inflammatory gene expression (data not shown) but did correlate with a significant reduction in triglyceride content compared with obese WT mice (Supplementary Fig. 4). Interestingly, reduced muscle triglyceride in obese TL2KO mice was observed in the absence of reduced basal or insulin-suppressed plasma free fatty acids (data not shown; see 21discussion).

In summary, these studies provide strong evidence of a protected metabolic phenotype in obese TPL2KO mice and are consistent with our data demonstrating decreases in local and systemic inflammation and reduced accumulation of triglyceride in liver and skeletal muscle.

The serine/threonine kinase TPL2 integrates inflammatory inputs from the MAP kinase and IKK/NF-κB signaling pathways (24–26) and is upregulated in the inflamed adipose tissue of obese mice and humans (31). These observations suggest that TPL2 actions promote the inflammatory and metabolic complications of obesity. The current study used a diet-induced obesity paradigm with TPL2KO and WT mice to provide the first demonstration that TPL2 deletion reduces peripheral inflammation, hepatic steatosis, and improves whole-body insulin resistance in the obese state. A key finding of our study was that whole-body insulin sensitivity was improved in obese TPL2KO mice compared with obese WT controls. The observed improvement in glucose-insulin homeostasis was due to increased glucose uptake in skeletal muscle and also increased suppression of hepatic glucose output.

Importantly, the improvement in whole body insulin sensitivity in obese TPL2KO mice occurred in the absence of any reductions in adipose depot mass, suggesting that an alternative mechanism promotes enhanced insulin sensitivity in TPL2KO mice. Specifically, our results suggest that the improved metabolic profile of TPL2KO mice reflects the salutary effects of TPL2 gene ablation on inflammation.

Two key components of the TPL2-mediated signaling axis in nonhematopoetic and hematopoietic cells and tissues are the MAP kinases ERK and JNK (26,27), both of which are implicated in obesity-associated insulin resistance. JNK and ERK activity are both increased in adipocytes from obese patients with diabetes, and this MAPK activation has been linked to the development of insulin resistance, either directly or indirectly, by increased production of downstream inflammatory mediators such as TNF-α and IL-6 (8,42). Phosphorylation of both JNK and ERK was increased in the adipose tissue of WT mice in response to diet-induced obesity, and this increase was attenuated by ≥50% in TPL2KO mice.

Adipose tissue inflammation is largely due to the proinflammatory actions of recruited BMD immune cells (4–6). In both rodents and humans (43–45), the metabolic complications of obesity are associated with the accumulation of proinflammatory macrophages (F4/80⁺ cells) that express the dendritic cell marker CD11c in adipose tissue (13,46). These macrophages selectively localize to CLS surrounding dead and dying adipocytes (6). Genetic ablation of CD11c⁺ cells results in significant reductions in CLS and improves metabolic profiles in obese mice (17,45). Our data suggest that TPL2 is a key regulator of CD11c⁺ macrophage infiltration and inflammation in obese adipose tissue. Gene expression of F4/80 and CD11c were dramatically decreased in the adipose tissue of obese TPL2KO mice, suggesting reduced infiltration or retention of CD11c⁺ macrophages. Consistent with this observation, the frequency of CLS and mRNA levels of MCP-1 and TNF-α were both significantly reduced. Our data do not rule out the possibility that reduced adipose tissue inflammation in TPL2KO mice reflects altered inflammatory responses of T-cells (47) or other adipose tissue immune cells.

TPL2 is a master regulator of ERK-dependent gene expression downstream of TLRs in hematopoietic cells, including macrophages (25,48). Thus, in addition to its effects on adipose tissue macrophage recruitment, TPL2 ablation is likely to have direct effects on macrophage inflammatory gene expression in obesity. We therefore determined the effect of TPL2 ablation on lipopolysaccharide-induced IL-6, MCP-1, and TNF-α expression in BMDM from WT and TPL2KO mice. We observed reduced secretion of TNFα, MCP-1, and IL-6 from TPL2KO BMDM. These data are in line with the gene expression data from the adipose tissue or liver, or both, of obese WT and TPL2KO mice and suggest TPL2 signaling in immune cells is an important factor in the initiation and maintenance of obesity-induced peripheral inflammation.

In addition to reducing adipose tissue immune cell infiltration and inflammation, ablation of TPL2 signaling in obese mice reduced hepatic lipid accumulation and inflammation. Decreased hepatic inflammation and lipid accumulation are both associated with improved insulin action in the liver. Hepatic inflammation is linked to enhanced triglyceride deposition in the liver; therefore, TPL2 ablation may reduce hepatic steatosis by directly reducing hepatic inflammation, for example, in Kupffer cells.

Alternatively our observations are consistent with studies reporting positive associations between visceral adipose tissue inflammation and hepatic inflammation in obese humans (49) and mice (13). Future studies using tissue-specific TPL2 ablation will be required to determine if the observed reduction in hepatic inflammation is a result of reduced adipose tissue inflammation or is caused by a lack of TPL2 signaling in the liver.

Accumulation of triglyceride and active lipid species (i.e., diacylglycerols) has been implicated in the pathogenesis of insulin resistance (40,41) and could be a common denominator in the improved metabolic profile observed in liver and muscle. Triglyceride accumulation was reduced in the liver and skeletal muscle in obese TPL2KO mice compared with obese WT mice. Consistent with reduced ectopic triglyceride accumulation, the clamp studies demonstrated increased sensitivity to insulin’s actions in liver and skeletal muscle. Insulin-stimulated Akt phosphorylation was enhanced in skeletal muscle but not in the liver of obese TPL2KO mice. This discrepancy between clamp and Akt data presumably reflects differential responses to an acute insulin bolus versus the integrated physiologic response to a steady-state insulin infusion during the clamp. It is intriguing that reduced ectopic triglyceride was observed in obese TPL2KO mice without a relative reduction in basal or clamped plasma free fatty acid concentrations. Future studies will address the mechanism(s) by which TPL2 reduces ectopic triglyceride accumulation in the liver and skeletal muscle and its role in improving glucose-insulin homeostasis.

In summary, the current study demonstrates a significant role for TPL2 in obesity-induced inflammation and metabolic dysfunction. TPL2 deletion in the obese state prevents MAPK activation, proinflammatory cytokine expression, and the accumulation of immune cells in peripheral tissues, resulting in reduced hepatic steatosis and improved insulin sensitivity. The recent report of increased TPL2 activity in human obesity (31) and the results of the current study suggest that TPL2 is a promising therapeutic target for the inflammatory and metabolic complications of obesity.

This work was supported by grants from the American Diabetes Association (R01-DK-082574) to A.S.G.; the Robert C. and Veronica Atkins Foundation Grant to A.S.G.; the U.S. Department of Agriculture, Agricultural Research Service, under agreement No. 58-1950-7-70 to A.S.G.; R01-DK-O74979 to M.S.O.; and in part from T32-DK-007651 to J.W.P., U24-DK-076169 to G.I.S., and R01-DK-40936 to G.I.S. Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the U.S. Department of Agriculture.

No potential conflicts of interest relevant to this article were reported.

J.W.P. and Y.L. researched data and wrote the manuscript. G.I.S. and V.T.S. contributed to discussion and reviewed and edited the manuscript. M.J.J. researched data, contributed to discussion, and reviewed and edited the manuscript. E.C. and C.X. researched data. P.N.T. and M.S.O. contributed to discussion and reviewed and edited the manuscript. A.S.G. wrote the manuscript and contributed to discussion.

The authors thank the following individuals at the JM-USDA Human Nutrition Research Center on Aging at Tufts University: Dr. Don Smith and the CBU staff for their assistance with care of the animals, and Dr. Katherine Strissel, Dr. Sajid Hussain, and Dr. Victoria Vieira for help with scientific discussion.