Obesity is associated with chronic low-level inflammation, especially in fat tissues, which contributes to insulin resistance and type 2 diabetes mellitus (T2DM). Protein inhibitor of activated STAT 1 (PIAS1) modulates a variety of cellular processes such as cell proliferation and DNA damage responses. Particularly, PIAS1 functions in the innate immune system and is a key regulator of the inflammation cascade. However, whether PIAS1 is involved in the regulation of insulin sensitivity remains unknown. Here, we demonstrated that PIAS1 expression in white adipose tissue (WAT) was downregulated by c-Jun N-terminal kinase in prediabetic mice models. Overexpression of PIAS1 in inguinal WAT of prediabetic mice significantly improved systemic insulin sensitivity, whereas knockdown of PIAS1 in wild-type mice led to insulin resistance. Mechanistically, PIAS1 inhibited the activation of stress-induced kinases and the expression of nuclear factor-κB target genes in adipocytes, mainly including proinflammatory and chemotactic factors. In doing so, PIAS1 inhibited macrophage infiltration in adipose tissue, thus suppressing amplification of the inflammation cascade, which in turn improved insulin sensitivity. These results were further verified in a fat transplantation model. Our findings shed light on the critical role of PIAS1 in controlling insulin sensitivity and suggest a therapeutic potential of PIAS1 in T2DM.

Type 2 diabetes mellitus (T2DM), accounting for 90% of all cases of diabetes, is one of the most prevalent chronic diseases worldwide (1). During the last 3 decades, the number of adults with diabetes has doubled, rising from 153 million in 1980 to ∼366 million in 2011 (2,3). T2DM is associated with hyperglycemia, which results from insulin resistance mainly in peripheral tissues (4). Insulin resistance is caused by many risk factors, ranging from genes to the environment, with obesity as the most common contributor. Most individuals with early-onset T2DM are obese, and an inverse linear relationship exists between BMI and the age at which T2DM is diagnosed (5,6).

A key mechanism underlying obesity-driven insulin resistance is chronic inflammation in adipose tissue, which is characterized by the accumulation of tissue immune cells (7). Macrophages are one of the primary immune cell types and crucial in triggering inflammation in obese adipose tissue (810). In lean, the adipose tissue macrophages mainly exhibit an anti-inflammatory effect. In obesity, however, adipocytes, along with adipose tissue macrophages, produce a wide range of proinflammatory mediators, including chemokine (C-C motif) ligand 2 (CCL2), interleukin (IL)-1β, IL-6, and tumor necrosis factor-α (TNF-α) (811), which in turn activate key regulators of inflammation such as c-Jun N-terminal kinase (JNK) and the inhibitor of nuclear factor (NF)-κB kinase/NF-κB–signaling pathway (12). These stress kinases then phosphorylate insulin receptor (IR) and insulin receptor substrate 1 (IRS1) on inhibitory serine residues to impair insulin signaling in adipocytes (13,14). In addition, transcription factors, such as NF-κB, further induce the expression of inflammatory cytokines, forming a positive feedback of inflammation response (15). Therefore, inflammation is a crucial event in obesity-induced insulin resistance, and identifying the key regulator for the coordination between inflammation and insulin-signaling pathways may not only clarify the mechanisms underlying insulin resistance but also provide new strategies for T2DM treatment.

Mammalian protein inhibitor of activated STAT (PIAS) proteins are a family of transcriptional regulators that possess small ubiquitin-like modifier (SUMO) E3 ligase activity, which consist of four members: PIAS1, PIAS2, PIAS3, and PIAS4 (16,17). PIAS1 was initially named for its ability to interact with and inhibit signal transducer and activator of transcription (STAT) 1 (18), but subsequent investigations have clarified that PIAS1 regulates a variety of transcription factors through distinct mechanisms, including promoting SUMOylation of target proteins and blocking the DNA-binding activity of transcription factors (19). PIAS1 plays a critical role in such cellular processes as cell proliferation (20), cell differentiation (21,22), and immune responses (2325). In the innate immune system, PIAS1 occupies the binding sites of STAT 1 and NF-κB on the promoters of their target genes to block expression of proinflammatory factors such as TNF-α and macrophage inflammatory protein 2 (MIP2). PIAS1−/− mice show increased protection against pathogenic infection and increased serum levels of proinflammatory cytokines (26). Thus, PIAS1 is a key regulator in inflammation responses of innate immunity; however, whether PIAS1 modulates the pathogenesis of insulin resistance remains unknown.

Our previous report clarified that PIAS1 restricted adipocytes differentiation by inhibiting CCAAT–enhancer-binding protein-β (22); however, the function of PIAS1 in mature adipocytes is not clear yet. In the current study, we found that PIAS1 was downregulated in the inguinal white adipose tissue (iWAT) of prediabetic models, including leptin-deficient (ob/ob), leptin receptor-deficient (db/db), and high-fat diet (HFD) mice. Ectopic expression of PIAS1 in prediabetic iWAT significantly activated the insulin-signaling pathway and improved systemic insulin sensitivity. Further studies in adipocytes found that PIAS1 suppressed expression of proinflammatory cytokines, such as CCL2, MIP2, and TNF-α, which in turn inhibited macrophage infiltration in adipose tissue, thereby attenuating the inflammatory cascade and increasing insulin sensitivity. Our results demonstrated the critical role of PIAS1 in modulating insulin sensitivity via inhibition of the inflammation response in adipose tissue.

Animals

Male C57BL/6J mice, ob/ob mice, db/db mice, and their control littermates were purchased from the Model Animal Research Center of Nanjing University. These mice were maintained on a normal chow diet (NCD). To produce HFD mice, 6-week-old C57BL/6J mice were fed an HFD (51% kcal from fat) for 5 weeks or 12 weeks, with NCD mice as the control. All studies involving animal experimentation were approved by the Fudan University Shanghai Medical College Animal Care and Use Committee and followed the National Institutes of Health guidelines on the care and use of animals.

Reagent and Adipocyte Differentiation

Recombinant murine TNF-α was purchased from PeproTech (Rocky Hill, NJ). Palmitate was purchased from Sigma-Aldrich (St. Louis, MO). JNK inhibitor and NF-κB inhibitor were from Selleck Chemicals.

At 2 days postconfluence (designated day 0), 3T3-L1 were subjected to adipogenic differentiation as previously described (22). Adipocytes phenotype appeared on day 3 and reached maximum by day 8 postadipogenic induction. PIAS1 overexpression or knockdown assays were performed on day 5 postinduction, and cells were harvested on days 8 and 9.

Generation and Administration of Recombinant Adenovirus

Recombinant adenovirus (Ad) for PIAS1 overexpression was generated using the ViraPower Adenoviral Expression System (Invitrogen, Carlsbad, CA), with LacZ recombinant adenovirus as the negative control. Recombinant Ad for PIAS1 knockdown was produced through BLOCK-iT Adenoviral RNAi Expression System (Invitrogen). The adenoviral expression vector pAd/BLOCK-iT encoding short hairpin RNA (shRNA) of PIAS1 was constructed, with shRNA for LacZ as the control. The sequences (5′ to 3′) for shRNAs were shPIAS1: CACCTTATTATTGACGGGTTGTTTA; and shLacZ: AATTTAACCGCCAGTCAGGCT. Recombinant Ad was produced and amplified in 293A cells and purified using Ad purification kits (Sartorius, Göttingen, Germany). Purified Ad was injected twice a week subcutaneously adjacent to iWAT in 8-week-old mice for 2 weeks. For HFD mice, Ad administration was conducted in 18-week-old mice for 2 weeks. On day 4 after Ad administration, these mice were subjected to further studies.

Metabolic Parameter Measurement

For the glucose tolerance test (GTT), mice were injected intraperitoneally with d-glucose (2 mg/g body weight) after an overnight fast, and tail blood glucose levels were monitored every 0.5 h using a glucometer monitor (Roche). For the insulin tolerance test (ITT), mice were injected intraperitoneally with human insulin (Eli Lilly) (0.75 mU/g body weight) after a 4 h fast, and tail blood glucose was monitored every 0.5 h. Levels of plasma insulin were detected by an ELISA kit (Mercodia). Concentrations of CCL2, MIP2, and TNF-α in iWAT or in the medium from adipocytes were measured by an ELISA kit (USCN Life Science Inc.).

2-Deoxyglucose Uptake Assay

After iWAT-specific Ad-PIAS1 or Ad-shPIAS1 administration, as described above, iWAT and epididymal WAT (eWAT) were extracted from mice and allowed to recover in Krebs-Ringer buffer containing 11.1 mmol/L glucose for 1 h. After the initial recovery period, the fat depots were pretreated with or without insulin for 0.5 h, followed by incubation with 200 μmol/L 2-(N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino)-2-deoxyglucose (2-NBDG; Invitrogen) for 2 h in the absence or presence of insulin. After 2-DG uptake, the depots were rinsed and lysed. Fluorescence was measured in an Envision fluorescence microplate reader and normalized to the total protein concentration.

Western Blotting and Antibodies

Western blotting analysis was performed as previously described (22). For detection of the insulin-signaling pathway in 3T3-L1 adipocytes, cells were harvested after stimulation with insulin (100 nmol/L) for 5 min. For measurement of insulin signaling in iWAT, the iWAT was dissected from mice 10 min after injection with insulin. During dissection, lymph nodes were removed from iWAT. The primary antibodies used for Western blotting were IR-β, phosphorylated (p)–IR-β (Tyr1150/1151), p-IRS1 (Tyr895), AKT, p-AKT (Ser473), p-JNK (Thr183/Tyr185), p–NF-κB p65 (Ser536), p–extracellular signal–related kinase 1/2, and p–p38 mitogen-activated protein kinase from Cell Signaling Technology. NF-κB p65 and JNK were from Santa Cruz Biotechnology, Inc.; PIAS1 was from Abcam; and β-actin was from Sigma-Aldrich.

RNA Extraction and Quantitative PCR

RNA extraction and quantitative (q)PCR analysis were conducted as previously described (22). Results are presented as means ± SD from three independent experiments. Primers used are listed in Supplementary Table 1.

RNA Interference

Synthetic small interfering (si)RNA for PIAS1, JNK1, JNK2, and siRNA-negative control (siNC) were synthesized by Invitrogen and are listed in Supplementary Table 2. 3T3-L1 cells were transfected with siRNA on day 5 after adipogenic induction.

Chromatin Immunoprecipitation

Chromatin immunoprecipitation (ChIP) analysis was conducted as previously described (27) using anti–NF-κB p65 (Santa Cruz Biotechnology, Inc.) or anti-PIAS1 antibody (Abcam), with rabbit IgG as a negative control. Immunoprecipitated DNA was purified and quantified by qPCR, with the DNA level in input sample as an endogenous control. The primers (5′ to 3′) for ChIP-qPCR were as follows: TNF-α: GCAGGTTCTGTCCCTTTCAC and AGTGCCTCTTCTGCCAGTTC; and MIP2: AGCGCAGACATCACTTCCTT and CTAGCTGCCTGCCTCATTCT.

Macrophage Migration Assay

Chemotaxis of Raw264.7 macrophages was measured using Transwell plates (Corning) with a pore size of 8.0 μm. Macrophages were loaded on the upper chamber with DMEM supplemented with 0.2% BSA. Adipocyte-conditioned media (ACM) were added to the lower chamber. The migration cells were stained with 0.1% crystal violet and counted manually using 5–6 randomly selected areas.

Fat Transplantation

After being treated with Ad-LacZ or Ad-PIAS1 for 2 weeks, iWATs from ob/ob mice were transplanted to 8-week-old male C57BL/6J mice, as previously described (28,29). Briefly, fat pads from donor mice were removed, cut into approximately 0.2-g slices, and kept in saline until transplantation. For each recipient mouse, 0.8 g of fat was transplanted into subcutaneous area (i.e., below the skin on the back of the host mice). The sham group had surgery, but no fat was transplanted. After recovery for 2 weeks, the indicated host mice underwent the following studies.

Statistics

All experiments were independently repeated at least three times. Results are presented as means ± SD. Differences between two groups were assessed using the unpaired two-tailed Student t test. Differences among more than two groups were assessed by ANOVA, with post hoc analysis for multiple comparisons. Differences were considered as significant when P < 0.05.

PIAS1 Was Downregulated in Adipose Tissues of Prediabetic State

To investigate the potential role of PIAS1 in insulin sensitivity, we first measured PIAS1 expression in several target tissues of insulin action, including liver, eWAT, iWAT, brown adipose tissue (BAT), and muscle (Fig. 1A and B). The expression of PIAS1 at the mRNA (Fig. 1A) and protein level (Fig. 1B) was relatively higher in iWAT than in other tissues, which promoted us to investigate the function of PIAS1 in iWAT. Further investigation showed that PIAS1 has a broad distribution in different cell types from adipose tissue, for instance, adipocytes and immune cells such as macrophages (Supplementary Fig. 1). We next detected PIAS1 expression in WAT of ob/ob, db/db, and 12-week HFD mice, all of which were hyperglycemic and exhibited insulin resistance. PIAS1 was significantly decreased in the iWAT and eWAT of these prediabetic mice (Fig. 1C and D). Interestingly, PIAS1 was also reduced in WAT of 5-week HFD mice (Fig. 1D), indicating that PIAS1 downregulation was an early event during the development of insulin resistance. We further evaluated PIAS1 expression in mature adipocytes and the stromal vascular fraction (SVF), and found that PIAS1 was decreased in the adipocytes and SVF of prediabetic mice (Fig. 1E).

Figure 1

PIAS1 expression was negatively correlated with the prediabetic state in the fat of mice . A: qPCR analysis of PIAS1 mRNA expression in mouse liver, eWAT, iWAT, BAT, and muscle tissues (n = 3). Data were normalized to the PIAS1 mRNA level in liver. B: Western blotting was used to detect PIAS1 protein level in the mouse liver, eWAT, iWAT, BAT and muscle tissues (n = 3), with quantification data by ImageJ software on the bottom. C: qPCR analysis showed PIAS1 mRNA in the iWAT and eWAT of prediabetic mice, including ob/ob mice (n = 4) and their wild-type (WT) littermates (n = 4), db/db mice (n = 4) and their WT littermates (n = 4), NCD mice (n = 3), and HFD mice (n = 3). Data were normalized to the PIAS1 mRNA level in the control mice of each group. D: Western blotting analysis of PIAS1 protein level in the iWAT and eWAT of ob/ob mice (n = 3), db/db mice (n = 3), HFD mice (n = 3), and their control mice (top) and quantification of relative PIAS1 level normalized to β-actin (bottom). E: Western blotting analysis of PIAS1 protein level in the SVF or adipocytes fraction from the iWAT and eWAT of obese and normal mice. *P < 0.05; **P < 0.01; ***P < 0.001. w, weeks.

Figure 1

PIAS1 expression was negatively correlated with the prediabetic state in the fat of mice . A: qPCR analysis of PIAS1 mRNA expression in mouse liver, eWAT, iWAT, BAT, and muscle tissues (n = 3). Data were normalized to the PIAS1 mRNA level in liver. B: Western blotting was used to detect PIAS1 protein level in the mouse liver, eWAT, iWAT, BAT and muscle tissues (n = 3), with quantification data by ImageJ software on the bottom. C: qPCR analysis showed PIAS1 mRNA in the iWAT and eWAT of prediabetic mice, including ob/ob mice (n = 4) and their wild-type (WT) littermates (n = 4), db/db mice (n = 4) and their WT littermates (n = 4), NCD mice (n = 3), and HFD mice (n = 3). Data were normalized to the PIAS1 mRNA level in the control mice of each group. D: Western blotting analysis of PIAS1 protein level in the iWAT and eWAT of ob/ob mice (n = 3), db/db mice (n = 3), HFD mice (n = 3), and their control mice (top) and quantification of relative PIAS1 level normalized to β-actin (bottom). E: Western blotting analysis of PIAS1 protein level in the SVF or adipocytes fraction from the iWAT and eWAT of obese and normal mice. *P < 0.05; **P < 0.01; ***P < 0.001. w, weeks.

Unlike PIAS1, however, other genes of PIAS family did not exhibit good correlation with the prediabetic state (Supplementary Fig. 2). Briefly, no significant differences were found in the expression of other PIAS genes between wild-type and ob/ob WAT, except for PIAS2 in eWAT (Supplementary Fig. 2A and B), whereas increased PIAS2, PIAS3, and PIAS4 were observed in db/db mice (Supplementary Fig. 2C and D). These data together suggested a distinct function of PIAS1 compared with other PIAS proteins.

PIAS1 Promoted Insulin Sensitivity in Mature Adipocytes

Free fatty acid and inflammatory cytokines, such as TNF-α, are two main mediators of obesity-induced diabetes (30). We therefore detected PIAS1 expression in 3T3-L1 mature adipocytes upon TNF-α or palmitate treatment, with the result that expression of PIAS1 gradually declined depending on the dose and length of treatment (Fig. 2A–C), which was consistent with the results observed in prediabetic mice.

Figure 2

PIAS1 regulated the insulin-signaling pathway in adipocytes. On day 8 of adipogenic induction, 3T3-L1 adipocytes were treated with 0, 10, 20, and 40 ng/mL TNF-α (A) or 0, 100, 200, and 400 μmol/L palmitate (B) for 12 h. Western blotting analysis of PIAS1 protein level upon induction at the indicated dose is shown. C: 3T3-L1 adipocytes were incubated with 20 ng/mL TNF-α. After treatment with TNF-α for the indicated time, cells were harvested, and PIAS1 expression was measured by Western blotting. On day 5 after adipogenic induction, PIAS1 expression was artificially altered in 3T3-L1 adipocytes. For knockdown of PIAS1 expression, adipocytes were transfected with siPIAS1 (D); for overexpression of PIAS1, adipocytes were infected with Ad-PIAS1 (E). On day 8 after adipogenic induction, the indicated cells were treated with 20 ng/mL TNF-α for 12 h, followed by 100 nmol/L insulin stimulation for 5 min. These cells were then subjected to Western blotting assay to determine the protein level of p-IR-β, total (t)–IR-β, p-IRS1, p-AKT, t-AKT, PIAS1, and β-actin (D and E), with quantification of these proteins on the right. t, total. *P < 0.05; **P < 0.01; ***P < 0.001.

Figure 2

PIAS1 regulated the insulin-signaling pathway in adipocytes. On day 8 of adipogenic induction, 3T3-L1 adipocytes were treated with 0, 10, 20, and 40 ng/mL TNF-α (A) or 0, 100, 200, and 400 μmol/L palmitate (B) for 12 h. Western blotting analysis of PIAS1 protein level upon induction at the indicated dose is shown. C: 3T3-L1 adipocytes were incubated with 20 ng/mL TNF-α. After treatment with TNF-α for the indicated time, cells were harvested, and PIAS1 expression was measured by Western blotting. On day 5 after adipogenic induction, PIAS1 expression was artificially altered in 3T3-L1 adipocytes. For knockdown of PIAS1 expression, adipocytes were transfected with siPIAS1 (D); for overexpression of PIAS1, adipocytes were infected with Ad-PIAS1 (E). On day 8 after adipogenic induction, the indicated cells were treated with 20 ng/mL TNF-α for 12 h, followed by 100 nmol/L insulin stimulation for 5 min. These cells were then subjected to Western blotting assay to determine the protein level of p-IR-β, total (t)–IR-β, p-IRS1, p-AKT, t-AKT, PIAS1, and β-actin (D and E), with quantification of these proteins on the right. t, total. *P < 0.05; **P < 0.01; ***P < 0.001.

On the basis of the negative correlation between PIAS1 expression and the prediabetic state, we hypothesized that PIAS1 might be a potential regulator of insulin sensitivity. To test the presumption, we examined the effect of PIAS1 knockdown or overexpression on the insulin-signaling pathway by performing RNA interference for PIAS1 (siPIAS1) or infecting adipocytes with Ad-PIAS1. As expected, PIAS1 protein was significantly reduced by siPIAS1 (Fig. 2D) and effectively increased by Ad-PIAS1 (Fig. 2E). Oil red O staining indicated that neither knockdown nor overexpression of PIAS1 affected the lipid accumulation (Supplementary Fig. 3A and B), suggesting that altered expression of PIAS1 in adipocytes had no effect on adipogenesis. Importantly, phosphorylation of IR-β, IRS1, and AKT induced by insulin was significantly reduced by siPIAS1. In PIAS1-deficient cells, TNF-α could hardly further inhibit the insulin signaling (Fig. 2D). To rule out an off-target effect, another two sets of siPIAS1 were also used, with similar results (Supplementary Fig. 3C). Consistently, PIAS1 overexpression significantly rescued the impaired insulin signaling caused by TNF-α (Fig. 2E). Taken together, these results indicated that PIAS1 significantly augmented insulin sensitivity in adipocytes.

PIAS1 in iWAT Improved Systemic Glucose Tolerance and Insulin Sensitivity

To further explore the effect of PIAS1 on insulin sensitivity in vivo, we ectopically expressed PIAS1 in prediabetic mice by infecting Ad-PIAS1 into iWAT. Western blotting showed that PIAS1 expression was specifically increased in iWAT but not in other tissues such as eWAT and liver (Supplementary Fig. 4A). PIAS1 overexpression did not affect the body weight (Supplementary Fig. 4B–D) or iWAT weight (Supplementary Fig. 4E and F), suggesting that short-term overexpression of PIAS1 had little effect on fat development. Of note, blood glucose levels were decreased by PIAS1 overexpression under fasting and fed conditions (Fig. 3A–C); however, the plasma insulin level was not affected (Fig. 3A). In addition, GTT and ITT showed that forced expression of PIAS1 in iWAT significantly improved systemic glucose tolerance and insulin sensitivity in prediabetic mice (Fig. 3D–F) but had no effect on plasma insulin level during the GTT (Fig. 3D and F), implying that the improved insulin sensitivity by PIAS1 is what contributed to the decreased blood glucose. The insulin-signaling pathway in iWAT was consistently augmented by PIAS1 overexpression, as indicated by increased insulin-stimulated phosphorylation of IR-β, IRS1, and AKT (Fig. 3G and H). However, insulin signaling in eWAT was not significantly affected by PIAS1 overexpression in iWAT (Supplementary Fig. 4G). To better clarify the function of PIAS1 in regulating insulin sensitivity, we conducted a 2-DG uptake assay and found that PIAS1 overexpression in iWAT significantly promoted glucose uptake of iWAT upon insulin stimulation but had little effect on eWAT (Fig. 3I).

Figure 3

Forced expression of PIAS1 in iWAT improved systemic insulin sensitivity. Male ob/ob, db/db, and HFD mice were infected with Ad-LacZ or Ad-PIAS1 twice a week for 2 weeks via subcutaneous injection adjacent to iWAT, followed by measurement of blood glucose and serum insulin level (AC), performance of GTT and ITT assay (DF), examination of the insulin-signaling pathway (G), and evaluation of glucose uptake (I) on day 4 after Ad administration. A: Blood glucose (left) and serum insulin level (right) were monitored in ob/ob mice (n = 4/group). Blood glucose levels were measured in db/db mice (B) and HFD mice (C) (n = 4–5/group). GTT and ITT assays were conducted in ob/ob mice (D), db/db mice (E), and HFD mice (F) (n = 4–5/group), with insulin level measured as well during GTT (middle of D and F). Two-way repeated-measures ANOVA was used to assess the difference between groups. G: These mice were killed after insulin stimulation (0.75 mU/g body weight) for 10 min via intraperitoneal injection. Western blotting analysis of the protein level of p–IR-β, total (t)–IR-β, p-IRS1, p-AKT, t-AKT, PIAS1, and β-actin in iWAT of ob/ob mice (left) and HFD mice (right). H: Quantification of relative p-protein levels normalized to respective total kinase protein content or β-actin. I: The iWAT and eWAT from HFD mice subcutaneously treated with Ad-PIAS1 or Ad-LacZ (n = 3/group) were extracted and subjected to 2-DG uptake assay in the absence and presence of insulin stimulation. Data were normalized to the Ad-LacZ–treated group in basal condition. All of these data were representative of at least three independent experiments, with the number of mice included in each group of each experiment indicated. *P < 0.05; **P < 0.01; ***P < 0.001.

Figure 3

Forced expression of PIAS1 in iWAT improved systemic insulin sensitivity. Male ob/ob, db/db, and HFD mice were infected with Ad-LacZ or Ad-PIAS1 twice a week for 2 weeks via subcutaneous injection adjacent to iWAT, followed by measurement of blood glucose and serum insulin level (AC), performance of GTT and ITT assay (DF), examination of the insulin-signaling pathway (G), and evaluation of glucose uptake (I) on day 4 after Ad administration. A: Blood glucose (left) and serum insulin level (right) were monitored in ob/ob mice (n = 4/group). Blood glucose levels were measured in db/db mice (B) and HFD mice (C) (n = 4–5/group). GTT and ITT assays were conducted in ob/ob mice (D), db/db mice (E), and HFD mice (F) (n = 4–5/group), with insulin level measured as well during GTT (middle of D and F). Two-way repeated-measures ANOVA was used to assess the difference between groups. G: These mice were killed after insulin stimulation (0.75 mU/g body weight) for 10 min via intraperitoneal injection. Western blotting analysis of the protein level of p–IR-β, total (t)–IR-β, p-IRS1, p-AKT, t-AKT, PIAS1, and β-actin in iWAT of ob/ob mice (left) and HFD mice (right). H: Quantification of relative p-protein levels normalized to respective total kinase protein content or β-actin. I: The iWAT and eWAT from HFD mice subcutaneously treated with Ad-PIAS1 or Ad-LacZ (n = 3/group) were extracted and subjected to 2-DG uptake assay in the absence and presence of insulin stimulation. Data were normalized to the Ad-LacZ–treated group in basal condition. All of these data were representative of at least three independent experiments, with the number of mice included in each group of each experiment indicated. *P < 0.05; **P < 0.01; ***P < 0.001.

We then depleted PIAS1 expression in wild-type mice by infecting Ad-shPIAS1 into iWAT and found that PIAS1 knockdown specifically occurred in iWAT (Supplementary Fig. 5A). No obvious change was observed in body weight (Supplementary Fig. 5B) and iWAT weight (Supplementary Fig. 5C) upon PIAS1 knockdown. Importantly, PIAS1-depleted mice developed insulin resistance. Briefly, deficiency of PIAS1 in iWAT led to much higher blood glucose (Fig. 4A), impaired glucose tolerance and insulin sensitivity (Fig. 4B), attenuated the insulin-signaling pathway (Fig. 4C), and reduced glucose uptake in iWAT (Fig. 4D). Collectively, PIAS1 in iWAT regulated insulin sensitivity in vivo.

Figure 4

Deficiency of PIAS1 in iWAT gave rise to impaired insulin sensitivity. Male C57BL/6J mice were infected with Ad-shLacZ or Ad-shPIAS1 twice weekly for 2 weeks via subcutaneous injection adjacent to iWAT, followed by a test of insulin sensitivity on day 4 after Ad administration. A: Blood glucose was measured under fasting and fed conditions (n = 5–6/group). B: GTT and ITT analysis (n = 5–6/group). Two-way repeated-measures ANOVA was used to assess the difference between groups. C: Western blotting analysis of the protein level of p–IR-β, total (t)–IR-β, p-IRS1, p-AKT, t-AKT, PIAS1, and β-actin in iWAT after insulin stimulation. D: The iWAT and eWAT from mice treated with Ad-shPIAS1 or Ad-shLacZ (n = 3–4/group) were extracted and subjected to 2-DG uptake assay in the absence and presence of insulin stimulation. Data were normalized to the Ad-shLacZ–treated group in the basal condition. The data shown are representative of at least three independent experiments, with the number of mice included in each group of each experiments indicated. *P < 0.05; **P < 0.01; ***P < 0.001.

Figure 4

Deficiency of PIAS1 in iWAT gave rise to impaired insulin sensitivity. Male C57BL/6J mice were infected with Ad-shLacZ or Ad-shPIAS1 twice weekly for 2 weeks via subcutaneous injection adjacent to iWAT, followed by a test of insulin sensitivity on day 4 after Ad administration. A: Blood glucose was measured under fasting and fed conditions (n = 5–6/group). B: GTT and ITT analysis (n = 5–6/group). Two-way repeated-measures ANOVA was used to assess the difference between groups. C: Western blotting analysis of the protein level of p–IR-β, total (t)–IR-β, p-IRS1, p-AKT, t-AKT, PIAS1, and β-actin in iWAT after insulin stimulation. D: The iWAT and eWAT from mice treated with Ad-shPIAS1 or Ad-shLacZ (n = 3–4/group) were extracted and subjected to 2-DG uptake assay in the absence and presence of insulin stimulation. Data were normalized to the Ad-shLacZ–treated group in the basal condition. The data shown are representative of at least three independent experiments, with the number of mice included in each group of each experiments indicated. *P < 0.05; **P < 0.01; ***P < 0.001.

PIAS1 Inhibited Inflammatory Infiltration in iWAT

In obesity, inflammation in adipose tissue is a major contributor to insulin resistance (7). We therefore determined a potential role of PIAS1 in adipose tissue inflammation. PIAS1 overexpression in iWAT of prediabetic mice significantly decreased the mRNA level of serial of inflammatory cytokines, such as CCL2, IL-1β, TNFα, and MIP2, as well as F4/80, a critical macrophage marker (Fig. 5A and B). To determine which fraction of iWAT was significantly regulated by PIAS1, we isolated SVF and mature adipocytes from iWAT and found that proinflammatory genes were decreased in SVF and mature adipocytes (Fig. 5C). Consistently, PIAS1 overexpression decreased the protein level of CCL2, TNF-α, and MIP2 in iWAT (Fig. 5D), whereas PIAS1 knockdown increased them (Fig. 5E). Importantly, these phenomena were restricted to iWAT. Expressions of inflammatory genes in eWAT were not affected by PIAS1 alteration in iWAT (Supplementary Fig. 4H and I and Fig. 5D).

Figure 5

PIAS1 regulated inflammatory responses in iWAT. qPCR analysis of the mRNA level of the inflammation-related genes in the iWAT of ob/ob (A) and HFD mice (B) (n = 3–4/group). The data were normalized to the mRNA level in Ad-LacZ– treated group. C: The iWAT of ob/ob mice infected with Ad-LacZ or Ad-PIAS1 was separated into SVF (left) and mature adipocytes (right). qPCR analysis of the mRNA level of the inflammation-related genes in SVF and mature adipocytes (n = 3/group). The data were normalized to the mRNA level in Ad-LacZ–treated group. ELISA assay was used to measure the protein level of CCL2, MIP2, and TNF-α in the iWAT of HFD mice (n = 3/group) upon PIAS1 overexpression (D) or in the iWAT of wild-type mice upon PIAS1 knockdown (E). Data are presented as levels of cytokines per microgram of total extractable protein. ChIP-qPCR assay was done to test binding activity of NF-κB and PIAS1 on the promoters of TNF-α and MIP2 in the iWAT of ob/ob (F) and HFD mice (G). The data are expressed as the percentage of input and normalized to the Ad-LacZ–treated group. H: Western blotting analysis of the p-JNK, total (t)-JNK, p-p65, and t-p65 in the iWAT of ob/ob mice upon PIAS1 overexpression. I: Western blotting analysis of the p-p65 and p-JNK in the iWAT of wild-type C57BL/6J mice upon PIAS1 knockdown. J: IHC staining for F4/80 was conducted in the indicated iWAT of ob/ob mice (top) and HFD mice (bottom) (scale bar: 100 μm). The data shown are representative of at least three independent experiments, with the number of mice included in each group of each experiments indicated. IHC, immunohistochemistry. *P < 0.05; **P < 0.01.

Figure 5

PIAS1 regulated inflammatory responses in iWAT. qPCR analysis of the mRNA level of the inflammation-related genes in the iWAT of ob/ob (A) and HFD mice (B) (n = 3–4/group). The data were normalized to the mRNA level in Ad-LacZ– treated group. C: The iWAT of ob/ob mice infected with Ad-LacZ or Ad-PIAS1 was separated into SVF (left) and mature adipocytes (right). qPCR analysis of the mRNA level of the inflammation-related genes in SVF and mature adipocytes (n = 3/group). The data were normalized to the mRNA level in Ad-LacZ–treated group. ELISA assay was used to measure the protein level of CCL2, MIP2, and TNF-α in the iWAT of HFD mice (n = 3/group) upon PIAS1 overexpression (D) or in the iWAT of wild-type mice upon PIAS1 knockdown (E). Data are presented as levels of cytokines per microgram of total extractable protein. ChIP-qPCR assay was done to test binding activity of NF-κB and PIAS1 on the promoters of TNF-α and MIP2 in the iWAT of ob/ob (F) and HFD mice (G). The data are expressed as the percentage of input and normalized to the Ad-LacZ–treated group. H: Western blotting analysis of the p-JNK, total (t)-JNK, p-p65, and t-p65 in the iWAT of ob/ob mice upon PIAS1 overexpression. I: Western blotting analysis of the p-p65 and p-JNK in the iWAT of wild-type C57BL/6J mice upon PIAS1 knockdown. J: IHC staining for F4/80 was conducted in the indicated iWAT of ob/ob mice (top) and HFD mice (bottom) (scale bar: 100 μm). The data shown are representative of at least three independent experiments, with the number of mice included in each group of each experiments indicated. IHC, immunohistochemistry. *P < 0.05; **P < 0.01.

NF-κB is known as a proinflammatory transcriptional factor that drives the expression of inflammatory genes (31). We then performed ChIP assay, observing that overexpression of PIAS1 inhibited the binding of p65 NF-κB to the endogenous promoters of TNF-α and MIP2, which was accompanied by increased occupation of PIAS1 on these promoters (Fig. 5F and G). This might be a major cause for the impaired expression of proinflammatory-related genes. Moreover, PIAS1 overexpression inhibited activation of JNK and p65 in prediabetic iWAT, as indicated by their phosphorylation modification (Fig. 5H), both of which were proinflammatory, whereas PIAS1 knockdown was able to promote their activation (Fig. 5I). More importantly, ectopic expression of PIAS1 significantly decreased macrophage infiltration into iWAT, as indicated by F4/80 staining (Fig. 5J). This was most likely due to the impaired CCL2 level shown above, which is a key chemokine for macrophages. These data collectively suggested that PIAS1 attenuated the inflammation response in iWAT of obesity.

PIAS1 in Adipocytes Inhibited Macrophage Migration In Vitro

We then determined the effect of adipocytes PIAS1 on macrophage infiltration in vitro. We detected the inflammation cascade in 3T3-L1 mature adipocytes, with the results that PIAS1 overexpression inhibited the mRNA level of proinflammatory cytokines and chemokines (Fig. 6A), whereas PIAS1 knockdown promoted them (Fig. 6B) as well as the activation of p65 and JNK (Fig. 6C). Consistently, these cytokines secreted by adipocytes were significantly inhibited by PIAS1 overexpression and promoted by PIAS1 knockdown (Fig. 6D). In addition, we found that PIAS1 inhibited the binding of p65 NF-κB to the promoters of TNF-α and MIP2, even under the condition of TNF-α induction (Fig. 6E). TNF-α was considered as a mediator that links the level of CCL2 to PIAS1 (Supplementary Fig. 6). To further clarify the role of PIAS1 in macrophage migration, we detected ACM-induced chemotaxis in Transwell plates. Migration of Raw264.7 macrophages to Ad-PIAS1 ACM was significantly decreased compared with Ad-LacZ ACM, whereas siPIAS1 ACM induced much more macrophage chemotaxis than did siNC ACM (Fig. 6F). These results together emphasized the crucial role of adipocyte PIAS1 in inhibiting macrophage chemotaxis.

Figure 6

PIAS1 in the adipocytes suppressed the chemotaxis of macrophages. On day 5 after adipogenic induction, PIAS1 was silenced by siPIAS1 or ectopically expressed by Ad-PIAS1. The indicated cells were then harvested on day 8 postinduction and subjected to further studies. A: qPCR analysis of the mRNA level of inflammation-related genes. Data were normalized to the mRNA level in the cells treated with Ad-LacZ. B: qPCR analysis of the mRNA level of inflammation-related genes. Data were normalized to the mRNA level in the cells treated with siNC. C: The indicated cells were harvested, and Western blotting was conducted to measure the level of the indicated proteins. D: Cellular supernatant of the indicated cells was collected and then subjected to ELISA assay to detect the concentration of inflammatory cytokines. E: The indicated cells were harvested after stimulation with TNF-α for 12 h. ChIP qPCR assay was performed to measure binding activity of NF-κB on the promoters of TNF-α and MIP2. The data were normalized to the IgG control. F: The ACM, prepared by incubating the indicated adipocytes in DMEM supplemented with 0.2% BSA for 12 h, were added to the lower chamber. Macrophages were loaded on the upper chamber and allowed to migrate for 12 h at 37°C. The migratory cells on the lower surface of the membrane were counted manually using 5–6 randomly selected areas. ERK, extracellular signal–related kinase; iNOS, inducible nitric oxide synthase; MAPK, mitogen-activated protein kinase. *P < 0.05; **P < 0.01; ***P < 0.001.

Figure 6

PIAS1 in the adipocytes suppressed the chemotaxis of macrophages. On day 5 after adipogenic induction, PIAS1 was silenced by siPIAS1 or ectopically expressed by Ad-PIAS1. The indicated cells were then harvested on day 8 postinduction and subjected to further studies. A: qPCR analysis of the mRNA level of inflammation-related genes. Data were normalized to the mRNA level in the cells treated with Ad-LacZ. B: qPCR analysis of the mRNA level of inflammation-related genes. Data were normalized to the mRNA level in the cells treated with siNC. C: The indicated cells were harvested, and Western blotting was conducted to measure the level of the indicated proteins. D: Cellular supernatant of the indicated cells was collected and then subjected to ELISA assay to detect the concentration of inflammatory cytokines. E: The indicated cells were harvested after stimulation with TNF-α for 12 h. ChIP qPCR assay was performed to measure binding activity of NF-κB on the promoters of TNF-α and MIP2. The data were normalized to the IgG control. F: The ACM, prepared by incubating the indicated adipocytes in DMEM supplemented with 0.2% BSA for 12 h, were added to the lower chamber. Macrophages were loaded on the upper chamber and allowed to migrate for 12 h at 37°C. The migratory cells on the lower surface of the membrane were counted manually using 5–6 randomly selected areas. ERK, extracellular signal–related kinase; iNOS, inducible nitric oxide synthase; MAPK, mitogen-activated protein kinase. *P < 0.05; **P < 0.01; ***P < 0.001.

PIAS1 Regulated Insulin Sensitivity in the Fat Transplantation Model

To further verify the role of PIAS1 in regulating the inflammation response and insulin sensitivity, we established a fat transplantation model in which Ad-PIAS1– or Ad-LacZ–treated iWATs from ob/ob mice were transplanted into host mice (Fig. 7A). The three groups did not differ significantly in body weight (Fig. 7B), fasting blood glucose level (Fig. 7C), or insulin level (Fig. 7D). However, blood glucose under the fed condition was higher in the transplantation group, but there were few differences between the iWAT–Ad-LacZ and iWAT–Ad-PIAS1 transplantation groups (Fig. 7D). Remarkably, systemic glucose tolerance and insulin sensitivity were impaired in the iWAT–Ad-LacZ transplantation group compared with sham, primarily due to the bad fat from ob/ob, whereas iWAT–Ad-PIAS1 transplantation mice had improved glucose tolerance and enhanced insulin sensitivity compared with the iWAT–Ad-LacZ group (Fig. 7E). Insulin signaling was consistently better activated in the iWAT–Ad-PIAS1 fat graft than in the iWAT–Ad-LacZ graft (Fig. 7F). In addition, inflammation infiltration was decreased in the iWAT–Ad-PIAS1 fat graft compared with the iWAT–Ad-LacZ graft, as indicated by the expression of inflammation-related genes, such as CCL2, TNFα, and F4/80, and F4/80 staining (Fig. 7G and H).

Figure 7

PIAS1 modulated insulin sensitivity in fat transplantation models. A: Schematic design of fat transplantation. After 2-week Ad-LacZ or Ad-PIAS1 treatment, iWATs from ob/ob mice were transplanted into the subcutaneous area (i.e., below the skin on the back of host mice) of wild-type C57BL/6 mice, with sham (surgery, but no fat was transplanted) as a control. After recovery for 2 weeks, the indicated host mice were subjected to the following studies: body weight (B), fasting blood glucose (C), and blood glucose (left) and insulin level (right) under fed condition (D) were evaluated (n = 4/group). E: GTT and ITT assays were conducted among three groups (n = 4/group), with two-way repeated-measures ANOVA post hoc analysis to assess the difference. F: Indicated fat grafts were dissected from transplantation mice upon insulin stimulation and subjected to Western blotting to measure the level of p-IRS1 and p-AKT. The fat grafts were used to evaluate the inflammation response as indicated by the mRNA level of inflammation-related genes (G) (n = 4/group) and the F4/80 IHC staining (H) (scale bar: 100 μm). HE, hematoxylin and eosin; IHC, immunohistochemistry. Data are presented as means ± SD. For transplantation groups vs. sham group: #P < 0.05; for iWAT–Ad-PIAS1 transplantation group vs. iWAT–Ad-LacZ transplantation group: *P < 0.05; **P < 0.01.

Figure 7

PIAS1 modulated insulin sensitivity in fat transplantation models. A: Schematic design of fat transplantation. After 2-week Ad-LacZ or Ad-PIAS1 treatment, iWATs from ob/ob mice were transplanted into the subcutaneous area (i.e., below the skin on the back of host mice) of wild-type C57BL/6 mice, with sham (surgery, but no fat was transplanted) as a control. After recovery for 2 weeks, the indicated host mice were subjected to the following studies: body weight (B), fasting blood glucose (C), and blood glucose (left) and insulin level (right) under fed condition (D) were evaluated (n = 4/group). E: GTT and ITT assays were conducted among three groups (n = 4/group), with two-way repeated-measures ANOVA post hoc analysis to assess the difference. F: Indicated fat grafts were dissected from transplantation mice upon insulin stimulation and subjected to Western blotting to measure the level of p-IRS1 and p-AKT. The fat grafts were used to evaluate the inflammation response as indicated by the mRNA level of inflammation-related genes (G) (n = 4/group) and the F4/80 IHC staining (H) (scale bar: 100 μm). HE, hematoxylin and eosin; IHC, immunohistochemistry. Data are presented as means ± SD. For transplantation groups vs. sham group: #P < 0.05; for iWAT–Ad-PIAS1 transplantation group vs. iWAT–Ad-LacZ transplantation group: *P < 0.05; **P < 0.01.

Role of JNK Signaling in Regulation of PIAS1 Expression

The results above indicated that PIAS1 was downregulated in the inflammatory and prediabetic state, and overexpression of PIAS1 alleviated obesity-induced insulin resistance. Next, we sought to identify the signaling pathway that regulated PIAS1 expression. The NF-κB– and JNK-signaling pathways were two key mediators downstream of the inflammatory cytokines (7). We therefore determined whether PIAS1 could be regulated by NF-κB or JNK and found that the JNK inhibitor SP600125 promoted the expression of PIAS1, whereas the NF-κB inhibitor had little effect (Fig. 8A). In addition, PIAS1 expression was gradually enhanced, depending on the dose and length of SP600125 treatment (Fig. 8B and C), and inhibition of JNK by SP600125 reversed the downregulation of PIAS1 caused by TNF-α treatment (Fig. 8D). There are three different JNK genes: JNK1, JNK2, and JNK3. JNK1 and JNK2 have a broad distribution, whereas JNK3 is mainly localized in neurons rather than adipose tissues (32). We therefore transfected siRNAs specific for JNK1 or JNK2 into 3T3-L1 adipocytes, separately or together. Both siJNK1 and siJNK2 could effectively reduce their expression and had no mutual influence (Fig. 8E). We found that both siJNK1 and siJNK2 increased the PIAS1 protein level (Fig. 8F) and that when they were used together, PIAS1 expression was further augmented and even rescued impaired PIAS1 expression caused by TNF-α treatment (Fig. 8G). We then evaluated the regulation of PIAS1 by JNK in vivo. Consistently, PIAS1 expression in iWAT of ob/ob mice was significantly increased by SP600125 (Fig. 8H). Moreover, we found that JNK was much more activated in iWAT of prediabetic mice as indicated by the increased p-JNK (Fig. 8I and J), which was negatively correlated with the expression of PIAS1. All these results suggested that JNK was a vital mediator for PIAS1 downregulation.

Figure 8

PIAS1 was downregulated by JNK signaling in the prediabetic state. A: The 3T3-L1 mature adipocytes were treated with JNK inhibitor SP600125 and NF-κB inhibitor 4,6-quinazolinediamine, N4-[2-(4-phenoxyphenyl)ethyl] (QNZ) for 12 h, with DMSO as a control. Western blotting analysis of the protein level of PIAS1. B: The 3T3-L1 mature adipocytes were treated with JNK inhibitor SP600125 at the indicated dose. Western blotting analysis of the protein level. C: The 3T3-L1 mature adipocytes were treated with JNK inhibitor SP600125 (20 μmol/L) for the indicated time. Western blotting analysis of the protein level at the indicated time points. D: The 3T3-L1 mature adipocytes were pretreated with SP600125 (20 μmol/L) for 1 h, then incubated with TNF-α (20 ng/mL) for 24 h. Cells were harvested and subjected to Western blotting to test the protein level of PIAS1. siJNK1 or siJNK2 was transfected into 3T3-L1 mature adipocytes, separately or together. E: qPCR analysis of JNK1 and JNK2 expression. F and G: Western blotting analysis of PIAS1 level. H: SP600125 was injected once every other day subcutaneously adjacent to iWAT in ob/ob mice for 1 week, with DMSO as control. Western blotting was used to detect the PIAS1 expression in the indicated iWAT. I: Western blotting analysis of the p-JNK level in the iWAT of ob/ob, db/db, and HFD mice and their respective control. J: Quantification of p-JNK level in the iWAT of ob/ob, db/db, and HFD mice and their respective controls. K: In the prediabetic state, PIAS1 was significantly decreased by proinflammtory factors in adipocytes via activated JNK. Downregulated PIAS1 triggered augmented NF-κB and JNK signaling, which produce much more inflammatory cytokines and induce macrophage infiltration, thereby impairing insulin sensitivity. ATM, adipose tissue macrophages. *P < 0.05; **P < 0.01.

Figure 8

PIAS1 was downregulated by JNK signaling in the prediabetic state. A: The 3T3-L1 mature adipocytes were treated with JNK inhibitor SP600125 and NF-κB inhibitor 4,6-quinazolinediamine, N4-[2-(4-phenoxyphenyl)ethyl] (QNZ) for 12 h, with DMSO as a control. Western blotting analysis of the protein level of PIAS1. B: The 3T3-L1 mature adipocytes were treated with JNK inhibitor SP600125 at the indicated dose. Western blotting analysis of the protein level. C: The 3T3-L1 mature adipocytes were treated with JNK inhibitor SP600125 (20 μmol/L) for the indicated time. Western blotting analysis of the protein level at the indicated time points. D: The 3T3-L1 mature adipocytes were pretreated with SP600125 (20 μmol/L) for 1 h, then incubated with TNF-α (20 ng/mL) for 24 h. Cells were harvested and subjected to Western blotting to test the protein level of PIAS1. siJNK1 or siJNK2 was transfected into 3T3-L1 mature adipocytes, separately or together. E: qPCR analysis of JNK1 and JNK2 expression. F and G: Western blotting analysis of PIAS1 level. H: SP600125 was injected once every other day subcutaneously adjacent to iWAT in ob/ob mice for 1 week, with DMSO as control. Western blotting was used to detect the PIAS1 expression in the indicated iWAT. I: Western blotting analysis of the p-JNK level in the iWAT of ob/ob, db/db, and HFD mice and their respective control. J: Quantification of p-JNK level in the iWAT of ob/ob, db/db, and HFD mice and their respective controls. K: In the prediabetic state, PIAS1 was significantly decreased by proinflammtory factors in adipocytes via activated JNK. Downregulated PIAS1 triggered augmented NF-κB and JNK signaling, which produce much more inflammatory cytokines and induce macrophage infiltration, thereby impairing insulin sensitivity. ATM, adipose tissue macrophages. *P < 0.05; **P < 0.01.

In the current study, we propose a model in which PIAS1 plays a role in controlling insulin sensitivity by inhibiting the inflammatory cascade (Fig. 8K). In the prediabetic state, PIAS1 was significantly downregulated by proinflammatory factors in adipocytes as a consequence of activated JNK. Decreased PIAS1, however, gave rise to augmented NF-κB and JNK signaling, which produces many more inflammatory cytokines, thereby inducing macrophage infiltration. All of these inflammatory cascades caused by PIAS1 downregulation further led to insulin resistance. Importantly, PIAS1 overexpression in iWAT was able to improve obesity-induced insulin resistance. Accordingly, the reciprocal regulation between PIAS1 and the inflammatory response constitutes a feedback loop likely to determine systemic insulin sensitivity. Our study is the first to demonstrate a novel role of PIAS1 in the regulation of insulin sensitivity.

Adipose tissue is a highly active metabolic site of insulin action. In particular, chronic inflammation in adipose tissue is an essential cause of obesity-induced insulin resistance (30,33). Adipocytes in the insulin-sensitive adipose tissue of lean subjects secret anti-inflammatory cytokines (34). As an individual becomes obese, hyperplasia and hypertrophy occur in adipocytes, causing the release of inflammatory cytokines and chemokines that attract proinflammatory macrophages to adipose tissue (35). These cytokines secreted by adipocytes and macrophages further activate JNK and NF-κB, leading to amplification of the inflammation cascade. Insulin resistance occurs as a result (7). A notable observation in the current study was that modulation of PIAS1 expression only in iWAT is sufficient to alter whole-body insulin sensitivity. The glucose uptake assay in iWAT supported our hypothesis. Hence, we thought that metabolic improvement in iWAT could facilitate the relief of systemic metabolic dysfunction. Interestingly, PIAS1 overexpression also promoted the insulin-signaling pathway in cultured adipocytes in which macrophages were absent. This is primarily due to the autocrine effect of adipocytes. Briefly, insulin signaling of adipocytes was attenuated by the inflammatory cytokines secreted by the adipocytes.

In addition to its expression level, phosphorylation of PIAS1 plays a key role in the inflammation cascade. Ser90 and Ser522 are two known residues that could be phosphorylated. p-PIAS1, but not the Ser90A mutant, rapidly interacts with the promoters of NF-κB target genes in macrophages, thereby inhibiting expression of these inflammatory genes (24). Moreover, phosphorylation of PIAS1 on Ser522 promotes its transrepression activity on NF-κB and inhibits inflammation (36). Thus, phosphorylation of PIAS1 is critical to suppress the inflammation cascade. It is interesting to investigate whether phosphorylation of PIAS1 facilitates insulin sensitivity through its anti-inflammatory function.

The regulation of PIAS1 expression has seldom been studied so far. The current study determined that PIAS1 was aberrantly expressed in prediabetic mice and that proinflammatory factors could downregulate PIAS1 expression through JNK. Interestingly, because PIAS1 overexpression was able to inhibit the activation of JNK, a feedback loop might exist between PIAS1 and JNK regulation. The JNK-signaling pathway is a key mediator of metabolic stress responses caused by obesity, which is activated in the iWAT of obesity-induced diabetes (37,38). However, whether JNK directly modulates PIAS1 expression is still unclear and needs further investigation.

In addition to being a transcriptional suppressor, PIAS1 functions as a SUMO E3 ligase and promotes the SUMOylation of proteins, which in turn affects the location, stability, or activity of the target proteins (39,40). Some proteins in the insulin-signaling pathway could be modified by SUMO. For example, the oncogene AKT, a mediator of growth-factor signaling, is essential in cell proliferation and tumorigenesis (41). The SUMOylation of AKT, which has recently been reported, is required for its kinase activity and is essential for cell growth and tumorigenesis as well (42). It is possible that PIAS1 could directly induce SUMOylation of the proteins in the insulin-signaling pathway, such as IR or IRS, thereby modulating insulin sensitivity. However, we conducted an endogenous coimmunoprecipitation assay, and no obvious interaction between PIAS1 and IR or IRS was observed (data not shown). More experiments, such as coimmunoprecipitation assays to detect the interaction between exogenously expressed proteins and SUMOylation assays, should be performed to examine the possibility.

The increasing prevalence of T2DM is really a large burden globally, heightening an urgent need to investigate diabetes pathogenesis and develop efficient strategies to prevent and control diabetes. Our findings identified PIAS1 as a key contributor of insulin sensitivity. Upregulation of PIAS1 improved insulin resistance caused by obesity. Thus, PIAS1 may be a potential diagnostic and therapeutic target for T2DM.

See accompanying article, p. 3984.

Funding. This research was partially supported by National Key Basic Research Project grants 2011CB910201 and 2013CB530601, The State Key Program of National Natural Science Foundation grants 31030048C120114 and 81390350 (to Q.-q.T.), and National Natural Science Foundation grant 31000603 (to L.G.). The department is supported by Shanghai Leading Academic Discipline Project B110 and 985 Project 985 III-YFX0302.

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

Author Contributions. Ya.L. and X.G. designed and conducted the experiments, analyzed data, and wrote the manuscript. X.D., Yu.L., S.-r.Z., and X.-b.W. participated in conducting the experiments. L.G. and Y.-j.D. reviewed the manuscript and offered critical advice. S.-w.Q., H.-y.H., C.-j.X., W.-P.J., and X.L. participated in research discussion. Q.-q.T. directed the project and reviewed the manuscript. Q.-q.T. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

1.
Danaei
G
,
Finucane
MM
,
Lu
Y
, et al.;
Global Burden of Metabolic Risk Factors of Chronic Diseases Collaborating Group (Blood Glucose)
.
National, regional, and global trends in fasting plasma glucose and diabetes prevalence since 1980: systematic analysis of health examination surveys and epidemiological studies with 370 country-years and 2.7 million participants
.
Lancet
2011
;
378
:
31
40
[PubMed]
2.
Lyssenko
V
,
Laakso
M
.
Genetic screening for the risk of type 2 diabetes: worthless or valuable
?
Diabetes Care
2013
;
36
(
Suppl. 2
):
S120
S126
[PubMed]
3.
Palermo
A
,
Maggi
D
,
Maurizi
AR
,
Pozzilli
P
,
Buzzetti
R
.
Prevention of type 2 diabetes mellitus: is it feasible
?
Diabetes Metab Res Rev
2014
;
30
(
Suppl. 1
):
4
12
[PubMed]
4.
Wilmot
E
,
Idris
I
.
Early onset type 2 diabetes: risk factors, clinical impact and management
.
Ther Adv Chronic Dis
2014
;
5
:
234
244
[PubMed]
5.
Bergman
RN
,
Stefanovski
D
,
Kim
SP
.
Systems analysis and the prediction and prevention of Type 2 diabetes mellitus
.
Curr Opin Biotechnol
2014
;
28
:
165
170
[PubMed]
6.
Bi
Y
,
Wang
T
,
Xu
M
, et al
.
Advanced research on risk factors of type 2 diabetes
.
Diabetes Metab Res Rev
2012
;
28
(
Suppl. 2
):
32
39
[PubMed]
7.
Harford
KA
,
Reynolds
CM
,
McGillicuddy
FC
,
Roche
HM
.
Fats, inflammation and insulin resistance: insights to the role of macrophage and T-cell accumulation in adipose tissue
.
Proc Nutr Soc
2011
;
70
:
408
417
[PubMed]
8.
Lumeng
CN
,
Deyoung
SM
,
Bodzin
JL
,
Saltiel
AR
.
Increased inflammatory properties of adipose tissue macrophages recruited during diet-induced obesity
.
Diabetes
2007
;
56
:
16
23
[PubMed]
9.
Weisberg
SP
,
McCann
D
,
Desai
M
,
Rosenbaum
M
,
Leibel
RL
,
Ferrante
AW
 Jr
.
Obesity is associated with macrophage accumulation in adipose tissue
.
J Clin Invest
2003
;
112
:
1796
1808
[PubMed]
10.
Xu
H
,
Barnes
GT
,
Yang
Q
, et al
.
Chronic inflammation in fat plays a crucial role in the development of obesity-related insulin resistance
.
J Clin Invest
2003
;
112
:
1821
1830
[PubMed]
11.
Lagathu
C
,
Yvan-Charvet
L
,
Bastard
JP
, et al
.
Long-term treatment with interleukin-1beta induces insulin resistance in murine and human adipocytes
.
Diabetologia
2006
;
49
:
2162
2173
[PubMed]
12.
Schenk
S
,
Saberi
M
,
Olefsky
JM
.
Insulin sensitivity: modulation by nutrients and inflammation
.
J Clin Invest
2008
;
118
:
2992
3002
[PubMed]
13.
Arkan
MC
,
Hevener
AL
,
Greten
FR
, et al
.
IKK-beta links inflammation to obesity-induced insulin resistance
.
Nat Med
2005
;
11
:
191
198
[PubMed]
14.
Hirosumi
J
,
Tuncman
G
,
Chang
L
, et al
.
A central role for JNK in obesity and insulin resistance
.
Nature
2002
;
420
:
333
336
[PubMed]
15.
Suganami
T
,
Tanimoto-Koyama
K
,
Nishida
J
, et al
.
Role of the Toll-like receptor 4/NF-kappaB pathway in saturated fatty acid-induced inflammatory changes in the interaction between adipocytes and macrophages
.
Arterioscler Thromb Vasc Biol
2007
;
27
:
84
91
[PubMed]
16.
Palvimo
JJ
.
PIAS proteins as regulators of small ubiquitin-related modifier (SUMO) modifications and transcription
.
Biochem Soc Trans
2007
;
35
:
1405
1408
[PubMed]
17.
Rytinki
MM
,
Kaikkonen
S
,
Pehkonen
P
,
Jääskeläinen
T
,
Palvimo
JJ
.
PIAS proteins: pleiotropic interactors associated with SUMO
.
Cell Mol Life Sci
2009
;
66
:
3029
3041
[PubMed]
18.
Liu
B
,
Liao
J
,
Rao
X
, et al
.
Inhibition of Stat1-mediated gene activation by PIAS1
.
Proc Natl Acad Sci U S A
1998
;
95
:
10626
10631
[PubMed]
19.
Shuai
K
.
Regulation of cytokine signaling pathways by PIAS proteins
.
Cell Res
2006
;
16
:
196
202
[PubMed]
20.
Puhr
M
,
Hoefer
J
,
Neuwirt
H
, et al
.
PIAS1 is a crucial factor for prostate cancer cell survival and a valid target in docetaxel resistant cells
.
Oncotarget
2014
;
5
:
12043
12056
[PubMed]
21.
Diao
Y
,
Wang
X
,
Wu
Z
.
SOCS1, SOCS3, and PIAS1 promote myogenic differentiation by inhibiting the leukemia inhibitory factor-induced JAK1/STAT1/STAT3 pathway
.
Mol Cell Biol
2009
;
29
:
5084
5093
[PubMed]
22.
Liu
Y
,
Zhang
YD
,
Guo
L
, et al
.
Protein inhibitor of activated STAT 1 (PIAS1) is identified as the SUMO E3 ligase of CCAAT/enhancer-binding protein β (C/EBPβ) during adipogenesis
.
Mol Cell Biol
2013
;
33
:
4606
4617
[PubMed]
23.
Jennewein
C
,
Kuhn
AM
,
Schmidt
MV
, et al
.
Sumoylation of peroxisome proliferator-activated receptor gamma by apoptotic cells prevents lipopolysaccharide-induced NCoR removal from kappaB binding sites mediating transrepression of proinflammatory cytokines
.
J Immunol
2008
;
181
:
5646
5652
[PubMed]
24.
Liu
B
,
Yang
Y
,
Chernishof
V
, et al
.
Proinflammatory stimuli induce IKKalpha-mediated phosphorylation of PIAS1 to restrict inflammation and immunity
.
Cell
2007
;
129
:
903
914
[PubMed]
25.
Pascual
G
,
Fong
AL
,
Ogawa
S
, et al
.
A SUMOylation-dependent pathway mediates transrepression of inflammatory response genes by PPAR-gamma
.
Nature
2005
;
437
:
759
763
[PubMed]
26.
Liu
B
,
Mink
S
,
Wong
KA
, et al
.
PIAS1 selectively inhibits interferon-inducible genes and is important in innate immunity
.
Nat Immunol
2004
;
5
:
891
898
[PubMed]
27.
Guo
L
,
Huang
JX
,
Liu
Y
, et al
.
Transactivation of Atg4b by C/EBPβ promotes autophagy to facilitate adipogenesis
.
Mol Cell Biol
2013
;
33
:
3180
3190
[PubMed]
28.
Gavrilova
O
,
Marcus-Samuels
B
,
Graham
D
, et al
.
Surgical implantation of adipose tissue reverses diabetes in lipoatrophic mice
.
J Clin Invest
2000
;
105
:
271
278
[PubMed]
29.
Tran
TT
,
Yamamoto
Y
,
Gesta
S
,
Kahn
CR
.
Beneficial effects of subcutaneous fat transplantation on metabolism
.
Cell Metab
2008
;
7
:
410
420
[PubMed]
30.
Glass
CK
,
Olefsky
JM
.
Inflammation and lipid signaling in the etiology of insulin resistance
.
Cell Metab
2012
;
15
:
635
645
[PubMed]
31.
Sanz
Y
,
Moya-Pérez
A
.
Microbiota, inflammation and obesity
.
Adv Exp Med Biol
2014
;
817
:
291
317
[PubMed]
32.
Bogoyevitch
MA
.
The isoform-specific functions of the c-Jun N-terminal Kinases (JNKs): differences revealed by gene targeting
.
BioEssays
2006
;
28
:
923
934
[PubMed]
33.
Seijkens
T
,
Kusters
P
,
Chatzigeorgiou
A
,
Chavakis
T
,
Lutgens
E
.
Immune cell crosstalk in obesity: a key role for costimulation?
Diabetes
2014
;
63
:
3982
3991
[PubMed]
34.
Bełtowski
J
.
Adiponectin and resistin--new hormones of white adipose tissue
.
Med Sci Monit
2003
;
9
:
RA55
RA61
[PubMed]
35.
Bruun
JM
,
Lihn
AS
,
Pedersen
SB
,
Richelsen
B
.
Monocyte chemoattractant protein-1 release is higher in visceral than subcutaneous human adipose tissue (AT): implication of macrophages resident in the AT
.
J Clin Endocrinol Metab
2005
;
90
:
2282
2289
[PubMed]
36.
Heo
KS
,
Chang
E
,
Takei
Y
, et al
.
Phosphorylation of protein inhibitor of activated STAT1 (PIAS1) by MAPK-activated protein kinase-2 inhibits endothelial inflammation via increasing both PIAS1 transrepression and SUMO E3 ligase activity
.
Arterioscler Thromb Vasc Biol
2013
;
33
:
321
329
[PubMed]
37.
Li
H
,
Yu
X
.
Emerging role of JNK in insulin resistance
.
Curr Diabetes Rev
2013
;
9
:
422
428
[PubMed]
38.
Liu
G
,
Rondinone
CM
.
JNK: bridging the insulin signaling and inflammatory pathway
.
Curr Opin Investig Drugs
2005
;
6
:
979
987
[PubMed]
39.
Ghisletti
S
,
Huang
W
,
Ogawa
S
, et al
.
Parallel SUMOylation-dependent pathways mediate gene- and signal-specific transrepression by LXRs and PPARgamma
.
Mol Cell
2007
;
25
:
57
70
[PubMed]
40.
Liu
B
,
Shuai
K
.
Targeting the PIAS1 SUMO ligase pathway to control inflammation
.
Trends Pharmacol Sci
2008
;
29
:
505
509
[PubMed]
41.
Porta
C
,
Paglino
C
,
Mosca
A
.
Targeting PI3K/Akt/mTOR Signaling in Cancer
.
Front Oncol
2014
;
4
:
64
[PubMed]
42.
Li
R
,
Wei
J
,
Jiang
C
, et al
.
Akt SUMOylation regulates cell proliferation and tumorigenesis
.
Cancer Res
2013
;
73
:
5742
5753
[PubMed]

Supplementary data