PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptNIH Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Cell Metab. Author manuscript; available in PMC Jul 4, 2011.
Published in final edited form as:
PMCID: PMC2907185
NIHMSID: NIHMS204799
PKCζ-regulated inflammation in the non-hematopoietic compartment is critical for obesity-induced glucose intolerance
Sang Jun Lee,1 Ji Young Kim,1 Ruben Nogueiras,2 Juan F. Linares,1 Diego Perez-Tilve,2 Dae Young Jung,3 Hwi Jin Ko,3 Susanna M. Hofmann,2 Angela Drew,1 Michael Leitges,4 Jason K. Kim,3 Matthias H. Tschöp,2 Maria T. Diaz-Meco,1 and Jorge Moscat1,5
1Department of Cancer and Cell Biology, University of Cincinnati College of Medicine, 3125 Eden Avenue Cincinnati, OH 45267
2Metabolic Diseases Institute, Department of Medicine, University of Cincinnati College of Medicine, 3125 Eden Avenue Cincinnati, OH 45267
3Program in Molecular Medicine, and the Department of Medicine, Division of Endocrinology, Metabolism and Diabetes, University of Massachusetts Medical School, Worcester, Massachusetts 01605
4The Biotechnology Centre of Oslo, University of Oslo, N-0317, Oslo, Norway
5 Corresponding author: , PhD, Department of Cancer and Cell Biology, University of Cincinnati College of Medicine, The Vontz Center for Molecular Studies, 3125 Eden Avenue, Cincinnati, OH 45267, Tel: (513) 558 8419, jorge.moscat/at/uc.edu
Obesity-induced inflammation is critical for the development of insulin resistance. Here we show, that genetic inactivation of PKCζ in vivo leads to a hyperinflammatory state in obese mice that correlates with a higher glucose intolerance and insulin resistance. Previous studies implicated PKCζ in the regulation of type 2 inflammatory responses in T cells. By using ex vivo and in vivo experiments, here we demonstrate that although PKCζ is involved in the alternative (M2) activation of macrophages, surprisingly, PKCζ ablation in the non-hematopoietic compartment but not in the hematopoietic system is sufficient to drive inflammation and IL-6 synthesis in the adipose tissue, as well as insulin resistance. Experiments using PKCζ/IL-6 double-knockout mice demonstrated that IL-6 production accounts for obesity-associated glucose intolerance induced by PKCζ deficiency. These results establish PKCζ as a critical negative regulator of IL-6 in the control of obesity-induced inflammation in adipocytes.
Research on the molecular signaling pathways that control obesity and that regulate adipocyte development and differentiation, food intake, and energy combustion has attracted much attention due to the potential of this work to provide new strategies for the prevention and treatment of obesity, and associated diseases such as glucose intolerance and type 2 diabetes (Flier, 2004). Several of the signaling molecules that have been implicated in the pathophysiology of obesity were primarily known for their roles in the control of cell growth, differentiation, and apoptosis (Fajas et al., 2002; Flier, 2004; Hirosumi et al., 2002; Kim et al., 2004b; Matsumoto et al., 2003; Molero et al., 2004; Nakae et al., 2003; Suzawa et al., 2003; Um et al., 2004). However, recent studies highlight the importance of inflammation, characterized by infiltration of macrophages and the secretion of inflammatory cytokines in metabolically active tissues, in the induction of glucose intolerance and insulin resistance during obesity (Hotamisligil, 2006; Qatanani and Lazar, 2007; Sabio et al., 2008; Schenk et al., 2008; Shoelson et al., 2006). Adoptive transfer studies have suggested the primary role of the hematopoietic system in insulin resistance in high-fat diet (HFD)-induced obesity (Saberi et al., 2009; Solinas et al., 2007). Also, the ablation of macrophages normalizes insulin sensitivity in obesity (Patsouris et al., 2008), whereas genetic inactivation of PPARγ or PPARδ in macrophages leads to the inhibition of the alternative activation (also known as type 2 inflammation) of resident macrophages in liver and adipose tissue, which results in a tonic type 1 hyperinflammatory state, and subsequent glucose intolerance and impaired insulin responses in obese mice (Kang et al., 2008; Odegaard et al., 2008). Furthermore, other investigators have presented compelling in vivo evidence on the role of T lymphocytes and mast cells in this process (Feuerer et al., 2009; Liu et al., 2009; Nishimura et al., 2009; Winer et al., 2009). Therefore, a paradigm is emerging according to which signaling molecules controlling macrophage and/or T-cell function could be excellent therapeutic targets in type 2 diabetes.
Contrary to this model, recent data in JNK-deficient mice indicate that inflammatory signals from adipocytes, but not from the hematopoietic system, are the key players in inflammation-induced insulin resistance in obesity (Sabio et al., 2008). That is, recent elegant studies using bone-transfer chimeras and conditional JNK knockout mice show that secretion of IL-6 by adipocytes, but not by macrophages, is what is essential for the control of insulin resistance in obese mice (Sabio et al., 2008). It is likely that, depending on the signaling cascade, the hematopoietic system and/or the stroma play different roles in these processes. The identification of the signaling molecules utilized by the different cellular components involved in this phenomenon will be instrumental for the design of better therapies.
The atypical PKC isoform, PKCζ, has been implicated in the control of two critical inflammatory cascades (Moscat et al., 2006a; Moscat et al., 2009). Through its ability to phosphorylate ReIA, the transactivating subunit of the canonical NF-κB complex, PKCζ modulates κB-dependent transcription in fibroblasts activated by TNFα and IL-1 (Duran et al., 2003; Leitges et al., 2001). PKCζ also phosphorylates Jak1, a critical player in IL-4 signaling (Duran et al., 2004; Martin et al., 2005). This might account for its implication in Th2-dependent inflammation (Moscat et al., 2006b). Therefore, PKCζ could play a dual role in inflammation: (1) a pro-inflammatory role through the activation of NF-κB and type 1 responses; or (2) a critical regulatory role in the type 2 anti-inflammatory response through the IL-4/Stat6 cascade. As obesity leads to a tonic hyperinflammatory response characterized by an increase in type 1 vs. type 2 inflammation, we hypothesized that PKCζ might be critical in the regulation of the balance between the two types of inflammatory responses during obesity-induced inflammation and the induction of glucose intolerance and insulin resistance.
Impaired insulin sensitivity in HFD-fed PKCζ-/- mice
As glucose intolerance and insulin resistance are key features of the metabolic syndrome (Semenkovich, 2006; Shoelson et al., 2006), and inflammation plays an important role in this phenomenon (Hotamisligil, 2006; Shoelson et al., 2006), we have evaluated whether PKCζ deficiency would lead to increased inflammation and the ensuing systemic glucose intolerance and insulin resistance in the context of fat accumulation. To address this question we have used PKCζ-deficient mice. Figure 1A shows the genotype of PKCζ-/- mice, which lack the PKCζ allele as described previously (Leitges et al., 2001). Results shown in Figure 1B demonstrate that PKCζ is actually lost, as determined by immunoblotting in different tissues, and that PKCζ is abundantly expressed in WAT, at levels comparable to those of lung and liver, and to a lesser degree in BAT and muscle. Therefore, PKCζ is expressed in metabolically relevant tissues and its genetic ablation results in the expected lack of its protein expression. We then performed glucose tolerance tests (GTT) in WT and PKCζ-/- mice that had been fed a HFD or regular diet (RD) for 12 weeks. WT mice kept under HFD-fed conditions displayed reduced glucose clearance in response to a bolus of glucose, as compared with WT mice fed in parallel with RD (Figure 1C). PKCζ-/- mice fed with RD displayed similar glucose clearance to that of identically treated WT mice (Figure 1C), indicating that the loss of PKCζ does not affect glucose metabolism under normal conditions. However, glucose clearance was much more impaired in PKCζ-deficient mice when fed a HFD than in identically treated WT mice or RD-fed WT or PKCζ-/- mice (Figure 1C). These results indicate that, even though PKCζ does not play a major role in glucose clearance under RD-fed conditions, HFD feeding makes PKCζ-/- mice highly glucose intolerant. This could indicate a higher level of insulin resistance upon HFD feeding in PKCζ-/- mice. In keeping with this notion, although glucose clearance in response to insulin injection (ITT) was significantly impaired only at 15 min post injection in HFD-fed WT mice as compared to RD-fed WT mice, it was dramatically worsened at all time points in HFD-fed PKCζ-deficient mice as compared to identically treated WT mice (Figure 1D). This is indicative of insulin resistance in the HFD-fed PKCζ knockout mouse group. In contrast, there were no differences between WT and PKCζ-deficient mice in terms of ITT when they have been fed with RD (Figure 1D). Previous data indicated that the atypical PKCs could be involved in glucose metabolism and insulin action (Farese and Sajan, 2009). However, our GTT and ITT data demonstrate that, under normal RD feeding conditions, PKCζ does not play such a role in vivo and that, if any atypical PKC is involved in that process, it is more likely to be PKCλ/ι (Farese et al., 2007).
Figure 1
Figure 1
Development of insulin resistance in HFD-fed PKCζ-/- mice
These results suggest that, although PKCζ is not required for normal in vivo insulin action in RD-fed mice, its lack in HFD-fed mice leads to enhanced insulin resistance. In keeping with this, blood insulin levels were dramatically increased in HFD-fed PKCζ-deficient mice upon glucose injection, as compared to identically treated WT mice (Figure 1E). Consistently, Figure 1F demonstrates increased basal non-fasting blood insulin levels in HFD-fed PKCζ-/- mice. Fasting glucose levels were increased in HFD-fed PKCζ-/- mice as compared to identically fed WT mice or RD-fed WT and PKCζ-/- mice, although the latter also showed a noticeable, but not statistically significant, increase in levels of glucose as compared to WT controls (Figure 1G). It should be noted that the levels of gluconeogenic genes in WT and PKCζ-deficient livers did not differ when mice were fed a HFD (Figures 1H-1J). Although, the expression of these genes was reduced in PKCζ-/- livers as compared with WT in mice fed a RD (Figures 1H-1J), this alteration did not seem to affect the basal fasting glucose levels (Figure 1G), nor the response to GTT or ITT (Figures 1C and 1D).
All these results suggest that the loss of PKCζ exacerbates HFD-induced glucose intolerance and insulin resistance, but no differences were observed in body weight, fat content, or blood leptin levels in PKCζ-deficient mice fed a HFD, as compared with identically treated WT mice, although HFD feeding increased all these parameters as compared with RD feeding in WT mice (Figures 1K and 1L). The levels of circulating TAG, NEFA, and cholesterol were dramatically induced by HFD in WT mice as compared to RD-fed WT mice, and were significantly higher in HFD-fed PKCζ-deficient mice (Table S1).
To determine the effects of PKCζ deficiency on glucose metabolism in individual organs, we performed a 2-hr hyperinsulinemic-euglycemic clamp in conscious PKCζ-/- mice and WT littermates following HFD feeding (Kim et al., 2004a) (Figure 2A). The steady-state glucose infusion rates required to maintain euglycemia during the clamps were significantly reduced in HFD-fed PKCζ-/- mice as compared with HFD-fed WT mice. Insulin-stimulated whole-body glucose turnover rates were reduced by 40% in HFD-fed PKCζ-/- mice. Basal rates of hepatic glucose production (HGP) were not altered, but HGP rates during insulin clamping tended to be higher in HFD-fed PKCζ-/- mice, resulting in lower hepatic insulin action in these mice. Insulin-stimulated glucose uptake in skeletal muscle and heart were markedly reduced in HFD-fed PKCζ-/- mice. Glucose uptake in white adipose tissue was not significantly altered, but brown adipose tissue glucose uptake was reduced by more than 50% in HFD-fed PKCζ-/- mice. Taken together, these results indicate that PKCζ deficiency exacerbates insulin resistance in peripheral organs and liver following HFD.
Figure 2
Figure 2
Reduced glucose metabolism and insulin signaling in HFD-fed PKCζ-/- mice
To further characterize insulin resistance in these mice, they were injected with insulin for 10 min, after which they were sacrificed and extracts were prepared from liver, WAT, and muscle. Akt activation was determined by immunoblotting with an anti-phospho-Akt antibody. Interestingly, Akt activation was normal in liver and WAT of PKCζ-/- mice fed a RD (Figure 2B). There was a significant reduction in Akt activation in muscle from RD-fed PKCζ-/- mice, as compared with identically treated WT mice (Figure 2B), consistent with the evidence that PKCζ plays a role in Akt activation, at least in some tissues (Joshi et al., 2008). HFD feeding impaired Akt activation in liver and muscle but not in WAT in WT mice (Figure 2B). However, the most striking observation was the dramatic impairment of Akt activation in the liver of HFD-fed PKCζ-/- mice as compared with identically treated WT mice (Figure 2B). These data are consistent with a model whereby the loss of PKCζ increases insulin resistance in HFD-fed mice especially in the liver.
Role of PKCζ in obesity-induced inflammation
The fact that PKCζ-deficient mice fed a HFD showed enhanced insulin resistance and impaired glucose tolerance suggested that the lack of PKCζ might lead to enhanced inflammation, which has previously been shown to play a major role in obesity-induced insulin resistance and glucose intolerance. Therefore, in the next series of experiments we analyzed whether, in fact, PKCζ deficiency promotes inflammation during HFD-induced obesity. Adipose tissue and liver are the two tissues in which obesity-induced inflammation has been shown to develop and which play a major role in glucose metabolism. Interestingly, morphological analysis of white adipose (WAT, epididymal fat pads) (Figure 3A) and brown adipose tissue (BAT, interscapular depots) (Figure 3B) from HFD-fed WT and PKCζ-deficient mice demonstrate similar adipose tissue volume in both mouse genotypes, although they were significantly bigger than the same tissues from mice fed a RD (Figures 3A and 3B). Similar results were obtained when the weights of these tissues were compared (data not shown). Histological analysis of hematoxylin and eosin (H&E) stained tissue sections revealed increased adipocyte size in WAT from HFD-fed PKCζ-/- mice as compared to identically treated WT controls (Figure 3C lower panels, and Figure 3D), whereas no differences were apparent between the WAT from RD-fed mice of different genotypes (Figure 3C, upper panels, and Figure 3D). Although HFD feeding enhanced the adipocyte size of WT mice, as compared with RD feeding, the effect of HFD on PKCζ-deficient mice was greater than in WT mice (Figure 3C and 3D). More importantly, WAT from HFD-fed PKCζ-/- mice, but not from HFD-fed WT mice, or from RD-fed mice of either genotype displayed a dramatic increase in macrophage infiltration, as assessed by staining with anti-F4/80 antibody (Figure 3E). This observation was confirmed by RT-PCR analysis of CD68 and F4/80 mRNAs in these tissues (Figure 3F and 3G). These results indicate that feeding a HFD to PKCζ-deficient mice leads to an inflammatory response in WAT that is not apparent in identically treated WT mice.
Figure 3
Figure 3
Increased macrophage infiltration and hepatic steatosis in HFD-fed PKCζ-/- mice
Morphological and histological analysis revealed a significant increase in the size and weight of HFD-fed PKCζ-/- mouse livers, as compared to identically treated WT mice or RD-fed mice of both genotypes (Figure 3H and 3I). This correlated with enhanced lipid accumulation, as determined by histological analysis and triglyceride content, which was more apparent in the HFD-fed PKCζ-deficient mice, although it was also detected to a lesser extent in RD-fed PKCζ-/- mice (Figure 3J and 3K). We did not detect changes in liver triglyceride secretion (Figure S1A), or in the expression levels of, for example LDLR, (Figures S1B). Consistent with this fatty liver phenotype, we detected significantly increased hepatic expression levels of the adipogenic/lipogenic genes PPARγ, PGC1β, FAS, C/EBPβ and ACC without changes in the expression levels of SREBP1c or C/EBPβ in PKCζ-/- mice as compared with WT mice, with both genotypes fed a HFD (Figures S1C-S1I). These alterations were not apparent when livers of the two genotypes were compared under a RD feeding regime (Figures S1C-S1I). The increase in the levels of adipogenic/lipogenic enzymes in the liver of HFD-fed PKCζ-deficient mice might contribute to the fatty liver phenotype observed these mice. However, we did not detect macrophage infiltration in these livers (data not shown). These data demonstrate that the targeted deletion of PKCζ leads to selective WAT inflammation when mice are fed a HFD, which is not observed in WT mice. Therefore, the lack of PKCζ sensitizes the adipose tissue to inflammation driven by HFD intake, which correlates with fatty liver, glucose intolerance, and insulin resistance.
Recent data demonstrate that adipocytes secrete type 2 cytokines, such as IL-4, that skew adipose tissue macrophages (ATM) towards an anti-inflammatory type 2 (also known as M2) profile. In contrast to the pro-inflammatory type 1 profile (also known as M1), the M2 profile limits inflammation in vivo via reduced tissular secretion of pro-inflammatory cytokines (Gordon, 2003; Kang et al., 2008; Odegaard et al., 2008; Vats et al., 2006). This is important because an increased M1/M2 cytokine ratio has been proposed to be responsible for insulin resistance during obesity (Kang et al., 2008; Odegaard et al., 2008). Therefore, we tested the hypothesis that PKCζ deficiency could tilt the balance of adipose tissue polarization towards an M1 pro-inflammatory phenotype that could account for the higher insulin resistance and glucose intolerance observed in the HFD-fed PKCζ-deficient mice. To do this, we analyzed the expression of a series of inflammatory genes in WAT from HFD-fed WT and PKCζ-deficient mice. Interestingly, as shown in Table 1, WAT from HFD-fed PKCζ-/- mice displayed increased levels of the pro-inflammatory cytokines IL-6, TNFα, and IFNγ as compared with WAT from identically treated WT mice. In contrast, levels of the Th2 cytokines, IL-4 and IL-13, were unaltered (Table 1). HFD feeding did not affect the adipose tissue cytokine signature of WT mice (Table 1). Also, no changes, other than a decrease in IL-4 synthesis were detected when this parameter was analyzed in PKCζ-deficient adipose tissue as compared to WT under RD feeding conditions (Table 1). Therefore, HFD feeding leads to a hyperinflammatory state in PKCζ-/- WAT that might account for the glucose intolerance observed in these mice when fed a HFD. Because IL-4 is secreted by adipocytes and T-cells and targets resident macrophages to drive them into an M2 anti-inflammatory phenotype (Kang et al., 2008; Odegaard et al., 2008), our results suggest that PKCζ deficiency does not affect the secretion of IL-4 in HFD-fed mice, but that it might be important for its actions. Interestingly, this inflammatory phenotype is not apparent in liver (data not shown), indicating that the major contribution of PKCζ to obesity-induced inflammation is restricted to its function in adipose tissue. This cytokine profile could be consistent with a skewed M1/M2 balance of ATM toward an M1 type of hyperinflammatory response. To test that possibility, we measured NOS (M1 marker), and Arg1 and Mgl2 (M2 markers) transcript levels in WAT from HFD-fed WT and PKCζ-deficient mice. NOS mRNA levels were significantly increased in WAT from HFD-fed knockout mice, in keeping with the pro-inflammatory phenotype of PKCζ-deficient mice under these conditions (Table 1). Although little or no alterations were observed in the levels of Arg1 in WAT from HFD-fed knockout mice, as compared with identically treated WT mice, the levels of Mgl2 were significantly lower (Table 1), indicating a reduced M2 phenotype in PKCζ-/- WAT. As no alterations were observed in IL-4 (Table 1), these results indicate that IL-4 actions on the adipose tissue from HFD-fed PKCζ-/- mice are impaired, giving rise to a reduced M2 phenotype and the subsequent appearance of a pro-inflammatory M1 state.
Table 1
Table 1
White adipose tissue expression of inflammatory genes in age-matched wild-type (WT) and PKCζ KO mice (6 of each phenotype) fed either a regular diet (RD) for 5 months or an RD for two months followed by a high-fat diet (HFD) for 3months.
Cell-autonomous role of PKCζ in macrophage inflammatory responses
To investigate whether the role of PKCζ in M2 polarization can be detected in ex vivo cultures of macrophages and is thus cell autonomous, we prepared bone marrow-derived macrophages (BMDM) from WT and PKCζ-/- mice and analyzed their inflammatory gene signature by QRT-PCR. Importantly, we found that PKCζ-deficient macrophages had reduced levels of Arg1 and Mgl2, whereas the levels of the pro-inflammatory genes NOS, TNFα, IL-6, and IL-1β were increased, as compared with the WT controls (Table S2). These results could be interpreted to mean that PKCζ-deficient macrophages display an intrinsic and cell-autonomous tendency toward an M1-skewed phenotype. Therefore, we next determined whether the lack of PKCζ would impair the expression of bona fide M2 markers in response to IL-4 in ex vivo macrophage cultures. Figures S2A-S2E shows that exposure of WT BMDM to IL-4 triggered the expression of PPARγ and the M2 markers Arg1, Mgl2, Mgl1, and Mrc1. This activation was significantly and reproducibly inhibited in PKCζ-/- macrophages (Figure S2A-S2E), demonstrating that PKCζ plays a major role in IL-4 induced activation of the M2 phenotype. Moreover, exposure to LPS led to enhanced production of pro-inflammatory TNFα and IL-6 in PKCζ-/- BMDM as compared to identically treated WT cells (Figure S2F and S2G), consistent with the concept that PKCζ is necessary for a tuned M1/M2 balance. IL-4 is capable of inhibiting LPS-induced production of inflammatory cytokines (Odegaard et al., 2008). Interestingly, the ability of IL-4 to inhibit LPS induction of TNFα synthesis was likewise dramatically impaired in PKCζ-/- BMDM as compared with WT controls (Figure S2H). Figures S3A-S3E demonstrate that IL-4 signaling through activation of the Arg1 promoter (a marker of M2 macrophages), as well as the Jak1/Stat6 cascade activated by IL-4, is severely impaired in the macrophage cell line Raw267 in which PKCζ has been depleted by shRNAi, and in PKCζ-deficient BMDMs. Collectively, these results indicate that PKCζ is an important intermediary in the acquisition of an anti-inflammatory M2 phenotype by macrophages, acting at the level of IL-4 signaling. According to the prevailing model, impaired M2 differentiation could account for the pro-inflammatory state detected in PKCζ-/- adipose tissue (Table 1), which could, in turn, account for the insulin-resistant state detected in these mutant mice.
Obesity-induced inflammation and insulin resistance in mice lacking PKCζ in the non-hematopoietic compartment
Because the loss of PKCζ leads to increased production of pro-inflammatory cytokines in the adipose tissue of HFD-fed mice, as well as a cell-autonomous skewing of macrophage phenotype towards a more inflammatory state, we sought to determine whether the loss of PKCζ in the hematopoietic system is sufficient to phenocopy the insulin-resistant and glucose-intolerant phenotype of total PKCζ-deficient mice. Thus, we performed adoptive transfer experiments to generate chimeras with PKCζ specifically deleted in the hematopoietic or non-hematopoietic compartments. We reconstituted lethally irradiated PKCζ-deficient and WT mice with either WT or PKCζ-deficient bone marrow cells to generate PKCζ-/- mice with reconstituted WT bone marrow (PKCζ-/-+WT-BM), and WT mice reconstituted with WT bone marrow (WT+WT-BM), or with PKCζ-/- bone marrow (WT+PKCζ-/--BM) (Figure 4A). Flow cytometry showed that white blood cells were efficiently reconstituted and displayed the donor genotype after bone marrow transplantation (Figure 4B). To investigate reconstitution of resident macrophages, we used PCR to genotype different tissues, including BMDMs, 23 weeks after reconstitution. Genotyping of BMDM and parenchymal cells indicated that BMDM cells were almost fully derived from the donor bone marrow, whereas the parenchymal cells were those of the recipient (Figure 4C).
Figure 4
Figure 4
Impaired glucose tolerance in mice with PKCζ-/- deficiency in the non-hematopoietic compartment
The chimeras were fed a HFD for 20 weeks starting at 3 weeks after transplantation, after which the three chimeric lines displayed comparable body weights (Figure 4D). We next performed GTT in these mice and found that the lack of PKCζ in the non-hematopoietic compartment, but not in the hematopoietic compartment, was sufficient to induce glucose intolerance upon HFD feeding (Figure 4E). These were unexpected observations demonstrating that PKCζ deficiency does not need to be manifested in the hematopoietic system to promote glucose intolerance in HFD-fed mice, despite the cell-autonomous role of PKCζ in M2 inflammation in macrophages and the inflammatory response detected in PKCζ-deficient WAT from HFD-fed PKCζ knockout mice (Table S2 and Figures S2 and S3). When ITTs were performed with these chimeras, it was apparent that the lack of PKCζ in the hematopoietic compartment was not sufficient to trigger insulin resistance upon HFD feeding, whereas mice with deficiency of PKCζ in the non-hematopoietic compartment displayed a defective insulin response (Figure 4F). Morphological analysis of WAT demonstrated similar adipose tissue volume in the three chimeras (Figure 4G). Interestingly, mice with PKCζ deficiency in the non-hematopoietic compartment, but not in the hematopoietic system, show macrophage infiltration, as determined histologically (Figure 4H) and by RT-PCR (Figures 4I and 4J). These results indicate that the simple deletion of PKCζ in the non-hematopoietic system is sufficient to induce WAT inflammation. Consistent with this, these mice also had fatty liver (Figures 4K and 4L). Of note, deletion of PKCζ just in the hematopoietic compartment not only did not favor fatty liver induction upon HFD feeding, but even slightly reduced HFD-induced fatty liver (Figures 4K and 4L).
To identify the underlying cause of this intriguing observation, we determined the cytokine inflammatory profile of WAT from the HFD-fed PKCζ-/- chimeras. For this, we analyzed the mRNA levels of different inflammatory genes as described above. Interestingly, WAT from mice with WT stroma and PKCζ-deficient hematopoietic systems (WT+PKCζ-/--BM) showed no significant alterations in any of the inflammatory cytokines, although Arg1 levels were reduced and NOS levels were increased (Table S3). These results can be interpreted to mean that the loss of PKCζ in the hematopoietic compartment is not sufficient to trigger an inflammatory phenotype in WAT in obese mice, which correlates with normal glucose tolerance and insulin responsiveness in these chimeric mice (Figure 4E and 4F). The alterations in Arg1 and NOS levels observed in these mice suggest that the loss of PKCζ in the hematopoietic compartment of HFD-fed mice is sufficient to at least partially skew the macrophage linage towards M1, but that this is not enough to induce an inflammatory response that would result in glucose intolerance and insulin resistance.
The lack of PKCζ in the non-hematopoietic compartment (PKCζ-/-+WT-BM) led to increased IL-6 levels in HFD-fed chimeras with no changes in any of the other cytokines or macrophage differentiation markers (Table S3). This indicates that PKCζ in the non-hematopoietic compartment exerts an anti-inflammatory role in WAT during HFD-induced obesity. This increase in IL-6 mRNA levels in the adipose tissue correlated with significantly enhanced levels of circulating IL-6 in the chimeras with PKCζ deleted in the non-hematopoietic compartment (Figure 4M). As the chimeric mice lacking PKCζ in the non-hematopoietic compartment are glucose intolerant, these results suggest that IL-6 overproduction due to the lack of PKCζ in the non-hematopoietic compartment may be the conducive agent in the glucose intolerance manifested in these HFD-fed mice, and probably the only cytokine required to induce glucose intolerance in obese PKCζ-deficient mice.
IL-6 overproduction in HFD-fed PKCζ-deficient mice is necessary for glucose intolerance
Our data show that WAT from HFD-fed PKCζ-deficient mice have higher IL-6 mRNA levels than identically treated WT mice (Table 1). Consistent with this, PKCζ-/- mice fed a HFD, but not a RD, displayed increased levels of circulating IL-6 (Figure 4N). To address the potential role of PKCζ-controlled IL-6 production in glucose intolerance during obesity, we crossed PKCζ-/- mice with IL-6-/- mice to generate a double-knockout mouse line. These mice were fed a HFD, as above, and GTTs and ITTs were performed. Although IL-6-/- mice showed reduced body weight and a lower percentage change in fat weight (Figures 5A and 5B), WT, PKCζ-/-, and double-knockout mice displayed very similar values for both parameters (Figures 5A and 5B). Interestingly, IL-6 ablation inhibited the glucose intolerance of PKCζ-deficient mice in the double-knockout setting (Figure 5C). ITT experiments confirm that ablation of IL-6 in the PKCζ-deficient mouse line restores the glucose clearance in response to insulin to normal values (Figure 5D). Therefore, the enhanced IL-6 levels in HFD-fed PKCζ-/- mice are necessary and likely sufficient to induce glucose intolerance in obese mice, and that PKCζ is instrumental for the basal repression of IL-6 production by the non-hematopoietic system under conditions of obesity-induced inflammation. Consistent with the role of IL-6 in PKCζ-/--enhanced inflammation during obesity, we found that the loss of IL-6 in PKCζ-deficient WAT completely ablated the recruitment of macrophages (Figures 5E-5H). Interestingly, IL-6 ablation also reduced fatty liver induction by HFD feeding of PKCζ-deficient mice (Figure 5I). These results indicate that IL-6 regulation by PKCζ is a key event in HFD-induced glucose intolerance and insulin resistance.
Figure 5
Figure 5
Restored glucose tolerance in HFD-fed PKCζ/IL-6 double-knockout mice
Cell-autonomous effect of PKCζ on IL-4 repressed inflammation in adipocytes
To determine the mechanism whereby PKCζ regulates IL-6 production in the adipose tissue, we exposed WT and PKCζ-deficient primary adipocytes to lipopolysaccharide (LPS) in the presence or absence of IL-4, after which IL-6 levels were determined by RT-PCR. It is well established that IL-4 inhibits inflammatory functions, which might be the basis for the induction of the type 2 anti-inflammatory state during obesity (Kang et al., 2008; Odegaard et al., 2008). Results in Figure 6A demonstrate that the ability of IL-4 to inhibit LPS-induced IL-6 production was severely impaired in PKCζ-deficient cell cultures. These results are in keeping with a model in which PKCζ is required for IL-4 signaling in adipocytes to dampen proinflammatory activation during obesity. Interestingly, like in macrophages (Figure S3), the loss of PKCζ in adipocytes results in the inhibition of Stat6 activation by IL-4 (Figure 6B). This negative role of PKCζ in the synthesis of IL-6 was not only detected in primary adipocyte cultures activated by IL-4, but also in fat pads (kept in culture for 3 days) from HFD-fed PKCζ-deficient mice. Thus, as shown in Figures 6C and 6D, IL-6 secretion was significantly higher in cultures of epididymal and retroperitoneal fat from HFD-fed PKCζ-/- mice than in those of WT fat. IL-6-/- and double knockout fat pads produced no IL-6, as expected (Figures 6C and 6D). To further demonstrate that this is due to the lack of PKCζ, we cultured 3T3-L1 fully differentiated adipocytes, controls or cells with PKCζ levels reduced by siRNA for 3 days, after which the levels of IL-6 secretion were determined. As shown in Figure 6E, depletion of PKCζ levels resulted in increased production of IL-6. On the other hand, it has recently been shown that fatty acids, which are abundant lipids in adipocytes, can activate inflammation in these cells (Shi et al., 2006). The results in Figure 6F demonstrate that PKCζ-deficient primary adipocytes secrete more IL-6 than their WT controls when exposed in culture to myristic or palmitic acid. Collectively, these results demonstrate that PKCζ exerts a negative role on IL-6 secretion by adipose tissue and isolated adipocyte cultures. Results in Figure 6G show that PKCζ-/- precursors differentiate into adipocytes more efficiently than do WT controls. The reasons for these observations still need to be explored, as well as the potential role of increased lipid accumulation in PKCζ-/- adipocytes in the enhanced secretion of IL-6. Future studies will be designed to address these questions. Interestingly, PKCζ levels were severely reduced in WAT from HFD-fed mice (Figure 6H). This strongly suggests that obesity leads to reduction of PKCζ levels, which results in increased IL-6 production and the subsequent enhancement in fatty liver induction, glucose intolerance, and insulin resistance. In keeping with this notion, WAT from obese Ob/Ob mice also displayed reduced levels of PKCζ (Figure 6I).
Figure 6
Figure 6
Role of PKCζ in IL-6 production in adipocytes
The data reported here unveil a critical role for PKCζ in adipose tissue inflammation during HFD-induced obesity, by demonstrating that, in HFD-fed obese mice, the loss of PKCζ triggers the M1 profile and inhibits the M2 profile of inflammatory cytokines in adipose tissue but not in liver. As the fat content and body weight gain of WT and PKCζ-/- mice were equalized by feeding with a HFD, this strongly suggests that the anti-inflammatory role of PKCζ in this process is not secondary to obesity, but that is manifested when mice undergo the tonic basal inflammation characteristic of the obese state. This is a very important observation because an increased M1/M2 balance has been shown to trigger glucose intolerance and insulin resistance during obesity and could explain why obese PKCζ-/- mice are more glucose intolerant and insulin resistant than equally treated obese WT mice. Surprisingly, although PKCζ participates in the M1/M2 decision point in macrophages in a cell-autonomous manner in BMDM cultures, this does not appear to account for the systemic glucose-intolerant state of obese PKCζ-deficient mice, as inactivation of PKCζ specific to the non-hematopoietic system is sufficient to induce glucose intolerance in obese bone marrow-chimera mice. This correlates with increased IL-6 levels but, in contrast to PKCζ knockout mice, this increase occurs without changes in the levels of other inflammatory cytokines. Experiments with PKCζ/IL-6 double-knockout mice demonstrate that the simple ablation of IL-6 in PKCζ-/- mice is sufficient to reverse the glucose-intolerant phenotype of obese PKCζ-deficient mice. Interestingly, whereas some investigators found that the loss of IL-6 led to mature-onset obesity in animals kept on a RD (Wallenius et al., 2002), others did not (Di Gregorio et al., 2004). Here we show that, in a pure BL6 background, IL-6-/- mice display resistance to obesity induce by HFD, in agreement with previous data (Park et al.). However, the PKCζ/IL-6 double knockout mice display a body weight that is similar to that of the WT or PKCζ-deficient mice when all genotypes are fed a HFD. Therefore, the restoration of glucose tolerance and insulin responsiveness to normal values in the doubly mutant mice as compared to the PKCζ-/- mice cannot be ascribed to differences in weight gain.
Collectively, our findings demonstrate that, although the inactivation of PKCζ in both the hematopoietic and the non-hematopoietic compartments is necessary to drive a type 1 inflammatory response, the production of IL-6 solely in the PKCζ-deficient non-hematopoietic compartment is sufficient and necessary to induce glucose intolerance. These observations are consistent with recent data demonstrating that JNK regulation of IL-6 synthesis by adipocytes, but not by the hematopoietic system, is what accounts for obesity-induced glucose intolerance (Sabio et al., 2008). This is important because it shows that an inflammatory cytokine secreted by a non-hematopoietic cell type is essential during obesity-induced inflammation and its associated impact on glucose metabolism. Here we show that, despite the hyperactivation of inflammation in the obese PKCζ knockout mice, only the IL-6 produced by the non-hematopoietic systems is necessary and sufficient to promote glucose intolerance in these mice. In the case of PKCζ, the study of adipocyte primary cultures from knockout mice revealed a mechanism whereby PKCζ plays a critical role in the anti-inflammatory effects of type 2 cytokines, such as IL-4, in the function of adipocytes as inflammatory cells in obesity. This is reminiscent of our previously reported evidence showing that PKCζ’s role in IL-4 signal transduction is also important for another IL-4 function, Th2 differentiation and allergic airway inflammation (Duran et al., 2004; Martin et al., 2005). In this way, PKCζ emerges as a key player in several inflammatory processes in which an adequate balance in the response is required for normal homeostatic control of metabolism.
Mice
All mice used in the experiments were bred on a C57BL/6 background for more than twelve generations and all experiments involved comparisons among littermates. IL-6 knockout mice were purchased from The Jackson Laboratory. All procedures were approved by the Institutional Animal Care and Utilization Committees at the University of Cincinnati, in accordance with the NIH guide for the care and use of laboratory animals.
Glucose tolerance and insulin tolerance testing
To assess glucose tolerance and insulin sensitivity, mice were first subjected to overnight fasting and injected intraperitoneally with 2 g glucose/kg body wt (25% D-glucose (Sigma) in 0.9% saline) for the glucose tolerance test (GTT), and 0.75 U or 1.25 U insulin/kg body weight (100 U/ml; Humolog Pen, Lily, Indianapolis, IN) for the insulin tolerance test (ITT). Tail blood glucose levels (mg/dl) were measured by using a handheld glucometer (Accu-chek Aviva, Roche).
Generation of adoptive transfer chimeras
To generate chimeras, mice were administered 7 Gy and 4.75 Gy of ionizing radiation at 3-hr intervals and injected through the tail vein with 5×106 whole bone marrow cells prepared from WT or PKCζ-/- mice. Mice were maintained for 3 weeks on normal chow including doxycycline (200 mg/kg) to allow for bone marrow reconstitution. Mice were placed on a HFD for 20 weeks after termination of antibiotic treatment. To quantify reconstitution efficiency, we used congenic donors and recipients that differed at the Ly5.1/Ly5.2 locus. Whole blood cells were collected and analyzed by flow cytometric analysis of congenic Ly5.1 (CD45.2-FITC; BD Pharmingen, California) and Ly5.2 (CD45.1-PE; BD Pharmingen, California) markers. Genomic DNA was extracted from tissues or cultured bone marrow-derived macrophages and genotyped by PCR.
Additional Experimental Procedures can be found in Supplemental Information.
Supplementary Material
01
Acknowledgments
This work is funded by National Institutes of Health Grants NIDDK59630 (M.H.T.), NIDDK69987 (M.H.T.), NIDDK56863 (M.H.T.), R01AI072581 (J.M.), R01CA132847 (J.M.), R01DK088107, R01CA134530 (M.T.D.-M.), and R01DK80756 (J.K.K.), the American Diabetes Association (7-07-RA-80 to J.K.K.), the UMass Diabetes Endocrinology Research Center grant DK32520, and by the Marie Curie Foundation (R.N.). We are grateful for skilled research assistance by Nilika Chaudhary, Birgit Ehmer, and Lyndsey Bolanos. We thank Maryellen Daston for editing this manuscript, and Glenn Doerman for the art work.
Footnotes
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
  • Di Gregorio GB, Hensley L, Lu T, Ranganathan G, Kern PA. Lipid and carbohydrate metabolism in mice with a targeted mutation in the IL-6 gene: absence of development of age-related obesity. Am J Physiol Endocrinol Metab. 2004;287:E182–187. [PubMed]
  • Duran A, Diaz-Meco MT, Moscat J. Essential role of RelA Ser311 phosphorylation by zetaPKC in NF-kappaB transcriptional activation. The EMBO journal. 2003;22:3910–3918. [PubMed]
  • Duran A, Rodriguez A, Martin P, Serrano M, Flores JM, Leitges M, Diaz-Meco MT, Moscat J. Crosstalk between PKCzeta and the IL4/Stat6 pathway during T-cell-mediated hepatitis. The EMBO journal. 2004;23:4595–4605. [PubMed]
  • Fajas L, Landsberg RL, Huss-Garcia Y, Sardet C, Lees JA, Auwerx J. E2Fs regulate adipocyte differentiation. Dev Cell. 2002;3:39–49. [PubMed]
  • Farese RV, Sajan MP. Metabolic Functions of Atypical Protein Kinase C: “Good and Bad” Am J Physiol Endocrinol Metab. 2009 ajpendo.00608.02009.
  • Farese RV, Sajan MP, Yang H, Li P, Mastorides S, Gower WR, Jr, Nimal S, Choi CS, Kim S, Shulman GI, et al. Muscle-specific knockout of PKC-lambda impairs glucose transport and induces metabolic and diabetic syndromes. The Journal of clinical investigation. 2007;117:2289–2301. [PMC free article] [PubMed]
  • Feuerer M, Herrero L, Cipolletta D, Naaz A, Wong J, Nayer A, Lee J, Goldfine AB, Benoist C, Shoelson S, et al. Lean, but not obese, fat is enriched for a unique population of regulatory T cells that affect metabolic parameters. Nature medicine. 2009;15:930–939. [PMC free article] [PubMed]
  • Flier JS. Obesity wars: molecular progress confronts an expanding epidemic. Cell. 2004;116:337–350. [PubMed]
  • Gordon S. Alternative activation of macrophages. Nat Rev Immunol. 2003;3:23–35. [PubMed]
  • Hirosumi J, Tuncman G, Chang L, Gorgun CZ, Uysal KT, Maeda K, Karin M, Hotamisligil GS. A central role for JNK in obesity and insulin resistance. Nature. 2002;420:333–336. [PubMed]
  • Hotamisligil GS. Inflammation and metabolic disorders. Nature. 2006;444:860–867. [PubMed]
  • Joshi J, Fernandez-Marcos PJ, Galvez A, Amanchy R, Linares JF, Duran A, Pathrose P, Leitges M, Canamero M, Collado M, et al. Par-4 inhibits Akt and suppresses Ras-induced lung tumorigenesis. The EMBO journal. 2008;27:2181–2193. [PubMed]
  • Kang K, Reilly SM, Karabacak V, Gangl MR, Fitzgerald K, Hatano B, Lee CH. Adipocyte-derived Th2 cytokines and myeloid PPARdelta regulate macrophage polarization and insulin sensitivity. Cell Metab. 2008;7:485–495. [PMC free article] [PubMed]
  • Kim HJ, Higashimori T, Park SY, Choi H, Dong J, Kim YJ, Noh HL, Cho YR, Cline G, Kim YB, et al. Differential effects of interleukin-6 and -10 on skeletal muscle and liver insulin action in vivo. Diabetes. 2004a;53:1060–1067. [PubMed]
  • Kim JK, Fillmore JJ, Sunshine MJ, Albrecht B, Higashimori T, Kim DW, Liu ZX, Soos TJ, Cline GW, O’Brien WR, et al. PKC-theta knockout mice are protected from fat-induced insulin resistance. The Journal of clinical investigation. 2004b;114:823–827. [PMC free article] [PubMed]
  • Leitges M, Sanz L, Martin P, Duran A, Braun U, Garcia JF, Camacho F, Diaz-Meco MT, Rennert PD, Moscat J. Targeted disruption of the zetaPKC gene results in the impairment of the NF-kappaB pathway. Molecular cell. 2001;8:771–780. [PubMed]
  • Liu J, Divoux A, Sun J, Zhang J, Clement K, Glickman JN, Sukhova GK, Wolters PJ, Du J, Gorgun CZ, et al. Genetic deficiency and pharmacological stabilization of mast cells reduce diet-induced obesity and diabetes in mice. Nature medicine. 2009;15:940–945. [PMC free article] [PubMed]
  • Martin P, Villares R, Rodriguez-Mascarenhas S, Zaballos A, Leitges M, Kovac J, Sizing I, Rennert P, Marquez G, Martinez AC, et al. Control of T helper 2 cell function and allergic airway inflammation by PKC{zeta}; Proceedings of the National Academy of Sciences of the United States of America; 2005. pp. 9866–9871. [PubMed]
  • Matsumoto M, Ogawa W, Akimoto K, Inoue H, Miyake K, Furukawa K, Hayashi Y, Iguchi H, Matsuki Y, Hiramatsu R, et al. PKClambda in liver mediates insulin-induced SREBP-1c expression and determines both hepatic lipid content and overall insulin sensitivity. The Journal of clinical investigation. 2003;112:935–944. [PMC free article] [PubMed]
  • Molero JC, Jensen TE, Withers PC, Couzens M, Herzog H, Thien CB, Langdon WY, Walder K, Murphy MA, Bowtell DD, et al. c-Cbl-deficient mice have reduced adiposity, higher energy expenditure, and improved peripheral insulin action. The Journal of clinical investigation. 2004;114:1326–1333. [PMC free article] [PubMed]
  • Moscat J, Diaz-Meco MT, Albert A, Campuzano S. Cell Signaling and Function Organized by PB1 Domain Interactions. Molecular cell. 2006a;23:631–640. [PubMed]
  • Moscat J, Diaz-Meco MT, Wooten MW. Of the atypical PKCs, Par-4 and p62: recent understandings of the biology and pathology of a PB1-dominated complex. Cell death and differentiation. 2009;16:1426–1437. [PubMed]
  • Moscat J, Rennert P, Diaz-Meco MT. PKCzeta at the crossroad of NF-kappaB and Jak1/Stat6 signaling pathways. Cell death and differentiation. 2006b;13:702–711. [PubMed]
  • Nakae J, Kitamura T, Kitamura Y, Biggs WH, 3, Arden KC, Accili D. The forkhead transcription factor Foxo1 regulates adipocyte differentiation. Dev Cell. 2003;4:119–129. [PubMed]
  • Nishimura S, Manabe I, Nagasaki M, Eto K, Yamashita H, Ohsugi M, Otsu M, Hara K, Ueki K, Sugiura S, et al. CD8+ effector T cells contribute to macrophage recruitment and adipose tissue inflammation in obesity. Nature medicine. 2009;15:914–920. [PubMed]
  • Odegaard JI, Ricardo-Gonzalez RR, Red Eagle A, Vats D, Morel CR, Goforth MH, Subramanian V, Mukundan L, Ferrante AW, Chawla A. Alternative M2 activation of Kupffer cells by PPARdelta ameliorates obesity-induced insulin resistance. Cell Metab. 2008;7:496–507. [PMC free article] [PubMed]
  • Park EJ, Lee JH, Yu GY, He G, Ali SR, Holzer RG, Osterreicher CH, Takahashi H, Karin M. Dietary and genetic obesity promote liver inflammation and tumorigenesis by enhancing IL-6 and TNF expression. Cell. 140:197–208. [PMC free article] [PubMed]
  • Patsouris D, Li PP, Thapar D, Chapman J, Olefsky JM, Neels JG. Ablation of CD11c-positive cells normalizes insulin sensitivity in obese insulin resistant animals. Cell Metab. 2008;8:301–309. [PMC free article] [PubMed]
  • Qatanani M, Lazar MA. Mechanisms of obesity-associated insulin resistance: many choices on the menu. Genes & development. 2007;21:1443–1455. [PubMed]
  • Saberi M, Woods NB, de Luca C, Schenk S, Lu JC, Bandyopadhyay G, Verma IM, Olefsky JM. Hematopoietic cell-specific deletion of toll-like receptor 4 ameliorates hepatic and adipose tissue insulin resistance in high-fat-fed mice. Cell Metab. 2009;10:419–429. [PMC free article] [PubMed]
  • Sabio G, Das M, Mora A, Zhang Z, Jun JY, Ko HJ, Barrett T, Kim JK, Davis RJ. A stress signaling pathway in adipose tissue regulates hepatic insulin resistance. Science (New York, NY) 2008;322:1539–1543. [PMC free article] [PubMed]
  • Schenk S, Saberi M, Olefsky JM. Insulin sensitivity: modulation by nutrients and inflammation. The Journal of clinical investigation. 2008;118:2992–3002. [PMC free article] [PubMed]
  • Semenkovich CF. Insulin resistance and atherosclerosis. The Journal of clinical investigation. 2006;116:1813–1822. [PMC free article] [PubMed]
  • Shi H, Kokoeva MV, Inouye K, Tzameli I, Yin H, Flier JS. TLR4 links innate immunity and fatty acid,Äìinduced insulin resistance. The Journal of clinical investigation. 2006;116:3015–3025. [PMC free article] [PubMed]
  • Shoelson SE, Lee J, Goldfine AB. Inflammation and insulin resistance. The Journal of clinical investigation. 2006;116:1793–1801. [PMC free article] [PubMed]
  • Solinas G, Vilcu C, Neels JG, Bandyopadhyay GK, Luo JL, Naugler W, Grivennikov S, Wynshaw-Boris A, Scadeng M, Olefsky JM, et al. JNK1 in hematopoietically derived cells contributes to diet-induced inflammation and insulin resistance without affecting obesity. Cell Metab. 2007;6:386–397. [PubMed]
  • Suzawa M, Takada I, Yanagisawa J, Ohtake F, Ogawa S, Yamauchi T, Kadowaki T, Takeuchi Y, Shibuya H, Gotoh Y, et al. Cytokines suppress adipogenesis and PPAR-gamma function through the TAK1/TAB1/NIK cascade. Nature cell biology. 2003;5:224–230. [PubMed]
  • Um SH, Frigerio F, Watanabe M, Picard F, Joaquin M, Sticker M, Fumagalli S, Allegrini PR, Kozma SC, Auwerx J, et al. Absence of S6K1 protects against age- and diet-induced obesity while enhancing insulin sensitivity. Nature. 2004;431:200–205. [PubMed]
  • Vats D, Mukundan L, Odegaard JI, Zhang L, Smith KL, Morel CR, Wagner RA, Greaves DR, Murray PJ, Chawla A. Oxidative metabolism and PGC-1beta attenuate macrophage-mediated inflammation. Cell Metab. 2006;4:13–24. [PMC free article] [PubMed]
  • Wallenius V, Wallenius K, Ahren B, Rudling M, Carlsten H, Dickson SL, Ohlsson C, Jansson JO. Interleukin-6-deficient mice develop mature-onset obesity. Nature medicine. 2002;8:75–79. [PubMed]
  • Winer S, Chan Y, Paltser G, Truong D, Tsui H, Bahrami J, Dorfman R, Wang Y, Zielenski J, Mastronardi F, et al. Normalization of obesity-associated insulin resistance through immunotherapy. Nature medicine. 2009;15:921–929. [PMC free article] [PubMed]