Pancreastatin-Dependent Inflammatory Signaling Mediates Obesity-Induced Insulin Resistance
Abstract
Chromogranin A knockout (Chga-KO) mice exhibit enhanced insulin sensitivity despite obesity. Here, we probed the role of the chromogranin A–derived peptide pancreastatin (PST: CHGA273–301) by investigating the effect of diet-induced obesity (DIO) on insulin sensitivity of these mice. We found that on a high-fat diet (HFD), Chga-KO mice (KO-DIO) remain more insulin sensitive than wild-type DIO (WT-DIO) mice. Concomitant with this phenotype is enhanced Akt and AMPK signaling in muscle and white adipose tissue (WAT) as well as increased FoxO1 phosphorylation and expression of mature Srebp-1c in liver and downregulation of the hepatic gluconeogenic genes, Pepck and G6pase. KO-DIO mice also exhibited downregulation of cytokines and proinflammatory genes and upregulation of anti-inflammatory genes in WAT, and peritoneal macrophages from KO mice displayed similarly reduced proinflammatory gene expression. The insulin-sensitive, anti-inflammatory phenotype of KO-DIO mice is masked by supplementing PST. Conversely, a PST variant peptide PSTv1 (PST-NΔ3: CHGA276–301), lacking PST activity, simulated the KO phenotype by sensitizing WT-DIO mice to insulin. In summary, the reduced inflammation due to PST deficiency prevented the development of insulin resistance in KO-DIO mice. Thus, obesity manifests insulin resistance only in the presence of PST, and in its absence obesity is dissociated from insulin resistance.
Introduction
The chromogranin A (human CHGA/mouse Chga) proprotein (1–4) undergoes proteolysis and gives rise to bioactive peptides including the antihypertensive catestatin (CHGA352–372) (5–8) and the diabetogenic pancreastatin (PST: CHGA250–301) (9–12). We have shown that Chga-deficient mice (Chga-KO) are obese, hyperadrenergic, and hypertensive. They display elevated levels of circulating leptin and catecholamines but lower levels of interleukin (IL)-6 and Mcp-1 (11,13–16). Despite these abnormalities, Chga-KO mice exhibit enhanced insulin sensitivity (11), a phenotype masked by supplementing PST. PST regulates hepatic insulin signaling through conventional (c) PKC and Srebp-1c (11). Increased plasma PST levels in diabetic populations correlate with insulin resistance (10). Similarly, increased circulating levels of PST in diet-induced obesity (DIO) and db/db mice are associated with insulin resistance. Despite high levels of plasma leptin and catecholamines, Chga-KO mice are obese owing to peripheral leptin and catecholamine resistance (17).
Since normal chow diet (NCD)-fed Chga-KO mice displayed increased insulin sensitivity (11), we hypothesized that Chga-KO mice may be able to maintain insulin sensitivity when exposed to the dysglycemic stress of a high-fat diet (HFD). The hallmarks of insulin resistance in DIO mice are obesity, hyperinsulinemia, and increased inflammation (18–22). Suppression of inflammation in DIO mice can improve insulin sensitivity (23–25). For example, rosiglitazone can improve inflammation and insulin sensitivity in DIO mice without reducing obesity significantly (23–25). Chga-KO mice are obese and presumably would become more obese after HFD feeding. Here, we address the following questions: 1) Does HFD suppress the insulin-sensitive phenotypes of Chga-KO mice? 2) Does PST deficiency in Chga-KO mice confer protection against HFD-induced insulin resistance? 3) Is endogenous PST a proinflammatory factor that is responsible for or contributes to the development of insulin resistance after HFD feeding?
Research Design and Methods
Animals
Male wild-type (WT; control) and Chga-KO mice on a stable mixed genetic background (50% 129svJ; 50% C57BL/6J), for >50 generations were used. In experiments shown in Figs. 1F and 8H, mice were on a C57BL/6J background. Animals were kept in a 12-h dark/light cycle. The institutional animal care and utilization committee approved all the procedures.
Chga-KO mice display elevated body weight gain on NCD and HFD. A: Body weights of WT and Chga-KO mice (with mixed genetic background) on NCD were taken every week starting from week 7 until week 22 (two-way ANOVA: strain, P < 0.0001; age, P < 0.0001; interaction, P < 0.004; n = 9). B: Initial weight at week 7 and final weight at week 22 were compared. C: Similarly, body weights of WT and Chga-KO mice on HFD were taken every week starting from week 7 until week 22 (two-way ANOVA: strain, P < 0.0001; age, P < 0.0001; interaction, P = 0.16; n = 13). D: Initial weight at week 7 and final weight at week 22 were compared. E: Effects of PST administration to WT and Chga-KO mice on NCD for 2 weeks (from week 19 to week 21) on body weight gain are shown. Sal, saline. F: Plasma PST levels were measured in 6-month-old WT-NCD, WT-DIO, or obese-diabetic db/db mice (all mice with C57BL/6 background) (n = 6). G: Four-month-old WT and Chga-KO mice with mixed genetic background were fed NCD or HFD for 12 weeks. Plasma leptin levels were measured by ELISA (n = 7). *P < 0.05, **P < 0.01, ***P < 0.001 for comparisons between the indicated groups.
Diets
Mice were fed ad libitum for 12–16 weeks with an HFD (60% of calories from fat, cat. no. D12492; Research Diets, Inc.). Some mice were also fed an NCD (14% of calories from fat; LabDiet 5P00).
Synthetic Peptides
WT human PST (CHGA273–301: PEGKGEQEHSQQKEEEEEMAVVPQGLFRG-amide) and PST variant PSTv1 (also known as PSTNΔ3; CHGA276–301: KGEQEHSQQKEEEEEMAVVPQGLFRG-amide) were synthesized with the solid-phase method and purified by reverse-phase high-performance liquid chromatography to at least 95% homogeneity (GenScript Corporation, Piscataway Township, NJ).
Glucose Tolerance and Insulin Tolerance Tests and Clamp Studies
For glucose tolerance tests (GTTs), glucose (1 mg/g i.p.) was injected (IP-GTT) (at time zero) or gavaged orally (O-GTT) after a 12-h fast. Tail vein glucose levels were measured at −30, 0, 15, 30, 60, 90, and 120 min. For insulin tolerance tests (ITTs), insulin (0.4 mU/g i.p.) was injected (IP-ITT) and blood glucose levels were measured at the indicated time points. Chronic PST (5 µg/g/day) and PSTv1 (10 µg/day) treatments were carried out for 15 days. For acute treatment, either peptides or saline was injected 30 min before glucose or insulin injection. Synergy Software Kaleidagraph, version 4.0, calculated the area under the curve (AUC) for each line curve.
The hyperinsulinemic-euglycemic clamp protocol was adapted from the description by Kim (26) and previously described by us (11). We calculated insulin-stimulated glucose disposal rate (IS-GDR) as follows: IS-GDR = GDR during the clamp – basal GDR. This value represents a measurement of the increase in GDR from the basal value due to insulin infusion. Since basal hepatic glucose production (HGP) = basal GDR, we calculated the total GDR during the clamp as glucose infusion rate (GIR) + HGP (during the clamp). During the clamps, insulin was infused at a constant rate of 12 mU/kg/min. The suppression of HGP by insulin was calculated as described by Kim (26) from these data.
Immunoblotting
Tissues were homogenized in a buffer containing phosphatase and protease inhibitors as previously described (11,17). Homogenates were subjected to SDS-PAGE and immunoblotted. Primary antibodies for phosphorylated and total Akt, AMP-activated protein kinase (AMPK), FoxO1, and Jun NH2-terminal kinase (JNK) were from Cell Signaling Technology (Beverly, MA). Actin and Srebp-1 antibodies were from Santa Cruz Biotechnology.
Preparation and Culture of Intraperitoneal Macrophages
Peritoneal macrophages were isolated after thioglycollate (3% solution in water) stimulation as described in detail by Zhang, Goncalves, and Mosser (27). Isolated macrophages were cultured in DMEM with 10% FBS for 48 h with daily medium changes. Cells were then serum starved overnight in DMEM with low FBS (0.5%) and then exposed to saline, PST (100 nmol/L), or lipopolysaccharide (LPS) (100 ng/mL) for 4 h. At the end of incubation, the cultures were subjected to RNA extraction.
In Vitro Chemotaxis Assay
In vitro chemotaxis assays were performed as previously described (28). Briefly, thioglycollate-activated peritoneal macrophages were isolated from 8-week-old male C57BL/6J mice and pretreated with the indicated concentrations of PST for 2 h at 37°C in serum-free RPMI. For migration analysis, 200,000 peritoneal macrophages were placed in each of the upper chamber of a 4-µm polycarbonate filter (24-transwell format; Corning, Tewksbury, MA), whereas compounds resuspended in serum-free RPMI at the indicated concentrations were placed in the lower chamber. After 3 h of migration at 37°C, cells were fixed in 4% paraformaldehyde, stained with DAPI, and counted under the microscope.
Real-Time PCR
RNA from tissues was extracted using a kit (RNeasy Plus; Qiagen, Valencia, CA). After DNase digestion, 200 ng RNA was transcribed into cDNA in a 20-μL reaction using a High Capacity cDNA Archive kit, and this cDNA was amplified. PCR (25-μL reactions) contained 5 μL cDNA, 2× SYBR Green PCR Master Mix, and 400 nmol/L of each primer. Differences in the threshold cycle (ΔCt) number between the target gene, the housekeeping gene Gapdh, and the ribosomal protein gene 36B4 were used to calculate differences in expression. Primers sequences are provided as a supplement.
Determination of Plasma Cytokines
Plasma cytokine concentrations were determined by a multiplex system from Quansys Biosciences (Logan, UT).
Statistics
Data are expressed as mean ± SEM. Statistical analyses were performed using Student t tests, as well as one- and two-way ANOVA followed by Dunnett post hoc test when appropriate. Statistical significance was concluded at P < 0.05. Statistics were computed with the program InStat (GraphPad Software, San Diego, CA).
Results
Effects of PST Deficiency and Supplementation on Obesity
The deficiency of Chga proprotein causes obesity in Chga-KO mice. At 7 weeks of age on an NCD, Chga-KO mice were heavier than WT mice (31.28 g vs. 23.08 g). From weeks 7 to 22, there was an additional increase in body weight gain compared with WT mice (from 23.08 g to 32.53 g for WT mice vs. 31.28 g to 45.96 g for Chga-KO mice) during the same period of growth (Fig. 1A and B). A similar trend holds true for mice on an HFD. After 15 weeks of HFD feeding, Chga-KO DIO (KO-DIO) mice gained more weight than WT-DIO mice (from 24.4 g to 44.84 g for WT-DIO vs. from 30.72 g to 58.59 g for Chga-KO mice) (Fig. 1C and D). PST-deficient KO-DIO mice also ate more on an HFD (2.94 vs. 2.11 g/day, P < 0.0001) than WT-DIO mice. PST administration to Chga-KO mice on NCD for 2 weeks did not change body weight (Fig. 1E). With the increase in obesity, plasma levels of PST were elevated in WT-DIO mice, similar to the levels seen in obese and diabetic db/db mice (Fig. 1F). Since obesity in Chga-KO mice occurred in the absence of PST expression and PST supplementation did not change body weight, we concluded that PST is not involved in the development of obesity. Plasma leptin level increased in both WT-DIO and KO-DIO mice compared with NCD-fed mice (Fig. 1G), but the level in KO-DIO was the highest owing to 1) the increase in adipose mass (compared with WT-DIO mice) and 2) the absence of PST-mediated negative regulation of leptin production (29).
Obese Chga-KO Mice Display Improved Glucose Tolerance and Insulin Sensitivity, Dependent Upon the Absence of PST
Despite increased obesity, KO-DIO mice maintained insulin sensitivity as shown by GTT and ITT (Fig. 2A–F). Both IP-GTT (Fig. 2A and B) and O-GTT (Fig. 2C and D) demonstrated improved glucose tolerance, whereas IP-ITT (Fig. 2E and F) showed increased insulin sensitivity in KO-DIO mice. Decreased AUC for corresponding GTT (Fig. 2B and D) and ITT (Fig. 2F) established that KO-DIO mice were more glucose tolerant and insulin sensitive than WT-DIO mice. Of note, fasting glucose levels (at time zero) in KO-DIO mice were consistently lower than in WT-DIO mice (Figs. 2A, C, and E and 3A and C). It appears that the lack of PST in KO-DIO mice helps to maintain their insulin sensitivity on HFD, as this phenotype was abolished upon chronic supplementation of PST (Fig. 2E and F). Similarly, when PST was supplemented to WT or Chga-KO mice on an NCD, PST caused insulin resistance (Fig. 3E and F). These data emphasize the point that PST treatment causes insulin resistance without causing obesity and that obesity occurring in the absence of PST does not induce insulin resistance. These results in mice with a mixed genetic background were also corroborated in mice with a C57BL/6J background (Supplementary Fig. 1).
Chga-KO mice on HFD display improved glucose tolerance and insulin sensitivity, dependent upon the absence of PST. A–F: WT-DIO and KO-DIO male mice with mixed genetic background were fasted for 12 h and subjected to IP-GTT (two-way ANOVA: strain, P < 0.0001; time, P < 0.0001; interaction, P < 0.045; n = 8) (A), O-GTT (two-way ANOVA: strain, P < 0.0001; time, P < 0.0001; interaction, P = 0.09; n = 8) (C), or IP-ITT (two-way ANOVA for treatment [WT], P < 0.0001; time, P < 0.0001; interaction, P = 0.56; two-way ANOVA for treatment [KO], P < 0.0001; time, P < 0.0001; interaction, 0.43; n = 9) (E), and AUC for glucose excursions were determined (B, D, and F). E: WT-DIO and KO-DIO mice were treated with (5 µg/g body wt i.p.) for 15 days, fasted for 12 h after the last day of injection, and subjected to GTT. GTT and the corresponding AUC are shown in E and F, respectively. Body weight–matched WT-DIO and KO-DIO mice were fasted for 12 h and subjected to clamp studies to determine GIR (G), GDR (H), IS-GDR (I), and % suppression of HGP (J) (n = 8–9). Fasting (12 h) basal levels of plasma insulin before (K) and after (L) chronic PST treatment (5 µg/g body wt/day i.p. for 15 days) (n = 6) are shown. “&” indicates comparison between WT-DIO+saline (Sal) and KO-DIO+saline. “$” indicates comparison between WT-DIO+saline and WT-DIO+PST. “#” indicates comparison between KO-DIO+saline and KO-DIO+PST. “*” indicates comparison between the groups as shown. *$#P < 0.05. **&&##$$P < 0.01. ***###$$$P < 0.001.
Treatment with the PST variant PSTv1 improves, whereas PST worsens, glucose tolerance and insulin sensitivity in WT-DIO and WT-NCD mice. WT-DIO mice were treated with saline or PSTv1 (10 µg/g body wt/day) for 15 days. Weight-matched DIO mice were fasted for 12 h and subjected to IP-GTT (two-way ANOVA: treatment, P < 0.0001; time, P < 0.0001; interaction, P = 0.17, n = 6) (A) and IP-ITT (two-way ANOVA: treatment, P < 0.0001; time, P < 0.0001; interaction, P = 0.62, n = 9) (C), and AUCs for glucose excursions were determined (B and D). WT-NCD mice were treated with saline (Sal) or PST (5 µg/g body wt) for 15 days, fasted for 12 h, and subjected to IP-ITT (two-way ANOVA: treatment, P < 0.0001; time, P < 0.0001; interaction, P < 0.0004; n = 8) (E). The corresponding AUC is shown in F. WT-NCD mice were treated with saline or PSTv1 (10 µg/g body wt) for 15 days, fasted for 12 h, and subjected to IP-ITT (two-way ANOVA: treatment, P < 0.0001; time, P < 0.0001; interaction, P < 0.02, n = 8) (G). The corresponding AUC is shown in H. Chr, chronic. *P < 0.05, **P < 0.01, ***P < 0.001 for comparisons between the indicated groups.
The GTT and ITT findings of insulin sensitivity in KO-DIO mice were reinforced by hyperinsulinemic-euglycemic clamp studies, where KO-DIO mice displayed an increased GIR and suppressed HGP by insulin. The basal GDR was similar between WT-DIO and KO-DIO mice, but IS-GDR was higher in KO-DIO compared with WT-DIO mice (Fig. 2G–J). Fasting plasma insulin concentrations in KO-DIO mice were lower than in WT-DIO mice (Fig. 2K). Since supplementation of PST raised fasting glucose (Fig. 2E and F) and insulin levels in KO-DIO mice over WT-DIO levels (Fig. 2L), our results support the notion that the deficiency of PST in Chga-KO mice prevented the development of hyperglycemia and hyperinsulinemia in KO-DIO mice.
Effects of PST, Much Like Anti-Insulin, Were Reversed by a PST-Variant Peptide
To combat anti-insulin–like effects of PST, we created a variant, PSTv1, by deleting three N-terminal residues of native PST. PSTv1 treatment of WT-DIO mice (10 µg/g/day) for 15 days lowered fasting plasma glucose levels (Fig. 3A and C) as well as improved glucose tolerance (Fig. 3A and B) and insulin sensitivity (Fig. 3C and D), possibly by competing against endogenous PST. While chronic PST supplementation to both WT and Chga-KO mice on an NCD induced insulin resistance (Figs. 3E and F and 4A–D), chronic PSTv1 improved insulin sensitivity (Fig. 3G and H). In NCD-fed Chga-KO mice (KO-NCD), the hyperglycemic effect of PST was attenuated by concomitant injection of PSTv1 (Fig. 4A and B). Moreover, chronic supplementation of PSTv1 reversed the effects of chronic PST treatment (Fig. 4C and D). Of note, PSTv1 alone had no effect on Chga-KO mice, presumably because of the lack of endogenous PST with which PSTv1 could compete (Fig. 4A–D). These results demonstrate PST deficiency to be the primary reason for insulin sensitivity and protection against diet-induced insulin resistance in Chga-KO mice.
The PST variant PSTv1 blocks the ability of PST to suppress the glucose tolerance of KO-NCD mice, and PST treatment reversed reduced gluconeogenic and lipid metabolic gene expression in KO-DIO mice. Acute (Acu) effects: KO-NCD mice were fasted for 12 h and treated with saline (Sal), PST (5 µg/g body wt), PSTv1 (10 µg/g body wt), or PST+PSTv1 for 30 min (−30 min) before injection of glucose (0 min) for IP-GTT (two-way ANOVA for treatment [saline vs. PST], P < 0.0001; time, P < 0.0001; interaction, P < 0.002; two-way ANOVA for treatment [PST vs. PSTv1], P < 0.0001; time, P < 0.0001; interaction, P < 0.005; n = 7) (A). The corresponding AUC is shown in B. Chronic (Chr) effects: KO-NCD mice were treated with saline, PST (5 µg/g body wt), PSTv1 (10 µg/g body wt), or PST+PSTv1 for 15 days, fasted for 12 h after the last of injection, and subjected to GTT (two-way ANOVA for treatment [saline vs. PST], time, P < 0.0001; interaction, P < 0.05; two way ANOVA for treatment [PST vs. PSTv1], P < 0.0001; time, P < 0.0001; interaction, P < 0.01; n = 7) (C). The corresponding AUC is shown in D. E: Liver glycogen at basal and during clamp in DIO mice (n = 6). F: Chronic effects of PST (5 µg/g body wt for 15 days) on expression of hepatic gluconeogenic genes (Pepck and G6pase) in DIO mice (n = 6). G: Chronic effects of PST (5 µg/g body wt for 15 days) on expression of hepatic lipid metabolic (Acc, Pparα, Cpt-1, Acox-1, and Srebp-1) genes in DIO mice (n = 6). “$” indicates comparison between saline and PST. “#” indicates comparison between PSTv1 and PST+PSTv1. “*” indicates comparison between the groups as shown. *$#P < 0.05. **$$##P < 0.01. ***$$$P < 0.001.
Effects of PST Deficiency and HFD on Metabolic Regulation and Insulin Signaling
Consistent with our metabolic characterization described thus far, we found that KO-DIO mice displayed increased hepatic glycogen storage and reduced gluconeogenic gene expression. Under the hyperinsulinemic condition during clamps, KO-DIO mice assimilated more glycogen in the liver than WT-DIO mice (Fig. 4E), which helped remove circulating glucose. The mRNA levels of gluconeogenic genes (Pepck and G6pase) in the liver were significantly lower in KO-DIO mice (Fig. 4F) and were reversed by treatment with PST (Fig. 4F). In addition, analysis of lipid metabolism gene expression revealed no difference in the mRNA levels of genes for lipogenesis (Acc and Srebp-1) and lipid oxidation (Pparα, Cpt-1, and Acox) between WT-DIO and KO-DIO mice (Fig. 4G), whereas PST supplementation increased the expression of Acox-1 and Cpt-1.
Among the metabolic signals that were significantly improved in Chga-KO mice were Akt and Srebp-1c, which are likely to be involved in producing the insulin-sensitive phenotype of these animals. We analyzed phosphorylation and activation of Akt, an important component of the insulin signaling pathway, by Western blot. Akt phosphorylation in muscle (Fig. 5A) and white adipose tissue (WAT) (Fig. 5B) was not significantly increased by insulin in WT-DIO mice (due to insulin resistance) but was robustly stimulated by insulin in KO-DIO mice (Fig. 5A and B). In the liver, insulin significantly enhanced Akt phosphorylation in KO-DIO mice when compared with the saline-treated controls (Fig. 5C). The fold response to insulin was reduced in WT-DIO livers (because of insulin resistance) but was higher in KO-DIO livers compared with WT-DIO livers (Fig. 5C). In contrast, PST supplementation to KO-DIO mice completely inhibited stimulation of phosphorylation by insulin (Fig. 5A–C).
KO-DIO mice display improved insulin (Ins)-induced Akt and AMPK phosphorylation, dependent upon the absence of PST. Weight-matched saline (Sal)-treated WT-DIO or KO-DIO and PST-treated KO-DIO mice were fasted for 12 h, injected with insulin (0.4 mU/g body wt i.p.) for 10 min, and killed for tissue collection. Tissues were homogenized, and lysates were subjected to SDS-PAGE and immunoblotted for phosphorylated (P)-Akt in muscle (n = 4) (A), WAT (n = 4) (B), and liver (n = 4) (C) and for phosphorylated AMPK in muscle (D), WAT (E), and liver (F). A.U., arbitrary units; T-, total. *P < 0.05, **P < 0.01, ***P < 0.001 for comparison between the indicated groups.
In addition to Akt, we analyzed the activity of other signaling pathway components, including AMPK, FoxO1, and Jnk. AMPK phosphorylation was improved in the muscle and WAT of KO-DIO mice (but not in the liver) (Fig. 5D–F), which might have contributed to an increase in glucose disposal and decrease in inflammation in KO-DIO mice. However, PST supplementation to KO-DIO mice did not change the overall pattern of AMPK signaling. As expected from increased Akt activation, the insulin-stimulated phosphorylation of FoxO1, downstream of Akt (82 kDa pFoxO1 and 95 kDa pFoxO1), in KO-DIO mice was higher in WAT and the liver than corresponding insulin-treated WT-DIO tissues (Fig. 6A and B). However, PST supplementation modulated FoxO1 signaling in WAT but not in the liver (Fig. 6A and B). In addition, the basal phosphorylation of p46 and p54 Jnk was lower in KO-DIO WAT compared with WT-DIO, suggesting persistent activation of inflammatory signals in WT-DIO mice (Fig. 6C and D). Interestingly, insulin seems to stimulate these negative Jnk signals in WAT for self-regulation. However, PST treatment did not alter hepatic AMPK, FoxO1, or Jnk signaling, suggesting additional PST-independent regulatory mechanisms.
FoxO1, Srebp-1, and JNK signaling is altered in KO-DIO mice. Weight-matched saline (Sal)-treated WT-DIO or KO-DIO and PST-treated KO-DIO mice were fasted for 12 h, injected with insulin (Ins) (0.4 mU/g body wt i.p.) for 10 min, and killed for tissue collection. Tissues were homogenized, and lysates were subjected to SDS-PAGE and subsequently immunoblotted to detect phospho–p95-FoxO1 and phospho–p82-FoxO1 signals in WAT (n = 4) (A) and phospho–p82-FoxO1 signals in liver (n = 4) (B), phospho–p54-Jnk and phospho–p46-Jnk signals in WAT (n = 4) (C) and liver (n = 4) (D), and p125–Srebp-1 and p-68–Srebp-1 (mature) signals in liver (n = 4) (E). A.U., arbitrary units; P-, phosphorylated; T-, total. *P < 0.05, **P < 0.01, and ***P < 0.001 for comparison between the indicated groups.
Consistent with reduced gluconeogenesis, the expression level of mature Srebp-1c p68 protein in the liver of insulin-treated KO-DIO mice was higher than in the WT-DIO controls (Fig. 6E), suggesting that the processing of the precursor p125 to mature p68 was facilitated by insulin in KO-DIO livers. Compared with WT-DIO mice, basal levels of p125 and p68 were also elevated in the liver (Fig. 6E) of KO-DIO mice. PST treatment reversed the expression pattern of p68 in KO-DIO livers by raising basal levels and decreasing insulin sensitivity (Fig. 6E). As a result, PST treatment decreased the p68-to-p125 ratio compared with untreated KO-DIO livers. In other words, PST treatment reduced insulin-stimulated processing of the Srebp-1c precursor p125 to mature p68 in KO-DIO livers.
PST Modulates Adipose Tissue Inflammation
The increased activation of Akt and AMPK in PST-deficient KO-DIO mice led us to wonder whether these mice may display reduced inflammation, which is regulated by these factors (30–40). Indeed, mRNA levels of the proinflammatory cytokines IL-1β, Tnfα, IL-6, and Mcp-1 were reduced in adipose tissue of KO-DIO mice compared with WT-DIO mice (Fig. 7A). Consistent with this trend, the expression of the proinflammatory genes Cd11c and IL-12p40 (Fig. 7B) and iNos (Fig. 7C) were also lower in KO-DIO adipose tissue than in WT-DIO mice. Accordingly, PST treatment caused increased expression of the proinflammatory genes IL-1β, Tnfα, IL-6, Mcp-1, and iNos (Fig. 7A–C). Conversely, expression of anti-inflammatory genes such as Arg-1, IL-10, Mgl-1, Mgl-2, and Ym1 in adipose tissue was higher in KO-DIO mice than WT control animals (Fig. 7D and E). PST treatment significantly reduced the expression of Arg-1 and IL-10, and displayed a trend toward downregulating Ym-1, but had no effect on Mgl-1 or Mgl-2 (Fig. 7D and E). The adipose tissue expression of Tnfα, IL-6, and Mcp-1 genes was reduced in KO mice in either NCD or DIO conditions compared with WT, although DIO increased levels in both backgrounds (Fig. 7F). We speculate that the subdued state of inflammation found in KO mice contributes to their improved insulin sensitivity.
PST modulates adipose tissue inflammation. A group of KO-DIO mice was injected with saline (Sal) or PST (5 µg/g body wt) for 15 days. Weight-matched saline-treated WT-DIO and KO-DIO and PST-treated KO-DIO mice were fasted for 12 h and killed. Blood was collected to measure plasma cytokine levels. Tissues were subjected to RNA extraction, cDNA preparation, and RT–quantitative PCR analysis for cytokines (n = 6) (A), proinflammatory genes (Cd11c and IL-12p40, n = 6 [B], and iNos, n = 6 [C]), and anti-inflammatory genes (Arg-1 and IL-10, n = 6 [D], and Mgl-1, Mgl-2, and Ym1, n = 6 [E]). Plasma cytokine levels are shown in G and H (n = 5). Chr, chronic. *P < 0.05, **P < 0.01, and ***P < 0.001 for comparison between the indicated groups.
Corresponding to reduced inflammatory gene expression, several cytokines were detected at lower levels in the plasma of KO-DIO compared with WT-DIO mice. Levels of IL-12p70, interferon-γ, Mip-1, IL-6, and Kc showed significant decreases, whereas Mcp-1 levels showed a trend to be lower, and tumor necrosis factor (Tnf)α levels were unchanged (Fig. 7G and H). PST treatment of KO-DIO mice raised plasma levels of IL-12p70 and Mcp-1 but had no effect on interferon-γ, IL-6, Mip-1α, Kc, and Tnfα (Fig. 7G and H).
Overall, expression of most proinflammatory markers was reduced, whereas anti-inflammatory markers were elevated in KO-DIO mice compared with WT-DIO mice, likely due to deficiency of PST and other Chga-derived peptides.
PST Promotes Macrophage Inflammation and Chemotaxis
As PST influenced inflammatory gene expression in adipose tissue, we investigated its role in macrophages. Elevated numbers of adipose tissue macrophages are associated with obesity, increased inflammation, and reduced insulin sensitivity (41–44). Chga mRNA expression was identified in peritoneal macrophages of WT mice (Fig. 8A). LPS caused robust expression of the proinflammatory gene iNos (∼100-fold) in both WT and Chga-KO macrophages (Fig. 8B). This was our reference point for comparing PST effects on expression of other inflammatory markers. The basal expression of iNos, Tnfα, Mcp-1, IL-6, and IL-12p40 was lower in Chga-KO macrophages, and PST treatment caused a smaller increase in expression than in WT macrophages (Fig. 8B–F). Specifically, PST increased the expression of Tnfα and Mcp-1 but had no effect on the expression of IL-6, IL-12p40, and iNos genes in Chga-KO macrophages (Fig. 8C–G). Developmental adjustments in Chga-KO mice may have lowered the potential for induction of these genes by PST. It is also possible that supplementation of other Chga-derived peptides may be necessary to produce the full complement of expression in Chga-KO mice.
PST promotes macrophage inflammation and chemotaxis. Peritoneal macrophages were isolated from WT-NCD and KO-NCD after thioglycollate injection. After 4-h exposure to saline, LPS (100 ng/mL), or PST (100 nmol/L), RNAs were extracted and cDNAs were prepared for quantitative PCR analyses of Chga (n = 8) (A), iNos (n = 8) (B), Tnfα (n = 8) (C), Mcp-1 (n = 8) (D), IL-6 (n = 8) (E), IL-12p40 (n = 8) (F), and iNos (n = 8) (G) genes. Effects of LPS stimulation on expression of iNos are shown in B, and the effects of PST are shown in C–G. H: Effects of PST on chemotaxis of activated macrophages from NCD-fed C57BL/6J mice (n = 5). Sal, saline. *P < 0.05, **P < 0.01, ***P < 0.001 for comparison between the indicated groups.
Consistent with induction of inflammatory gene expression, PST induced chemotaxis of macrophages in vitro. Using peritoneal macrophage cultures, we compared the chemotactic ability of the chemokine Mcp-1 versus PST. PST treatment led to significantly higher levels of chemotaxis compared with saline-treated controls, although with lower potency than Mcp-1 (chemotaxis induced by 10 nmol/L PST was 44% that induced by 10 nmol/L Mcp-1) (Fig. 8H).
In summary, these results reinforce the notions that 1) PST alone negatively regulates insulin sensitivity in several tissues, and as a result, PST supplementation can produce insulin resistance, and 2) along with another Chga-produced peptide, PST upregulates proinflammatory signals that are lacking in Chga-KO mice.
Discussion
PST deficiency renders NCD-fed Chga-KO mice more insulin sensitive than WT mice (11). A guanosine triphosphate (GTP)-binding protein–coupled receptor–mediated signaling pathway leading to activation of conventional DAG/Ca2+-dependent PKC and downregulation of mature Srebp-1c (p68) is thought to play a role in the anti-insulin effects of PST (11). Recently, we observed that PST could also modulate endoplasmic reticulum (ER) stress by interacting with binding immunoglobin protein (BiP)/78 kDa glucose-regulated protein (GRP78) (45). Feeding an HFD creates obesity, leading to hyperinsulinemia and inflammation (18–22). PST-deficient KO-DIO mice are more obese than WT-DIO mice but remain more insulin sensitive as assessed by GTT, ITT, and clamp studies (Fig. 2). Insulin sensitivity was compromised by PST replacement, suggesting that the absence of PST in Chga-KO mice not only enhanced insulin action but also prevented further damage by dietary fat.
The most important aspect of this study is that obesity can be dissociated from insulin resistance as long as inflammation is suppressed. The presence of supraphysiological levels of PST can reconnect obesity with insulin resistance by inducing inflammation. In the absence of PST, animals could be insulin sensitive despite obesity. This is reminiscent of rosiglitazone-treated WT-DIO mice, which are insulin sensitive but obese (23–25).
PST and Insulin Resistance
We can now provide answers to the questions we raised in the introduction: 1) HFD feeding does not suppress the insulin-sensitive phenotype of Chga-KO mice. 2) PST deficiency in Chga-KO mice provides protection against HFD-induced insulin resistance. And 3) PST is an endogenous proinflammatory factor that affects insulin action independent of obesity. Obesity caused by dietary fat cannot induce inflammation and insulin resistance in the absence of PST and another Chga-produced peptide.
It appears that PST deficiency provides benefits to obese mice by 1) enhancing hepatic glycogen storage and decreasing glucose production through stimulation of Akt signaling and suppression of gluconeogenic genes via increased expression of Srebp-1c proteins (46,47), 2) increasing glucose disposal by muscle via increased AMPK and insulin-stimulated Akt signaling (Figs. 2 and 5), and 3) suppressing macrophage-mediated inflammation in adipose tissue (Figs. 7 and 8). Moreover, PST deficiency may decrease ER stress (31).
PST deficiency also increased insulin-stimulated pFoxO1-p95 signals in WAT of KO-DIO mice, which was reversed after PST supplementation (Fig. 6A). Since increased phosphorylation of FoxO1 (inactive form) in WAT favors adipogenesis through PPARγ (48,49), our results are consistent with the greater adiposity of Chga-KO mice.
PST and Inflammation
The increased Akt and AMPK activities in WAT in KO-DIO mice (Fig. 5) could potentially provide the benefit of dampening the inflammation induced by an HFD. This is because both Akt and AMPK can suppress the inflammatory responses of resident macrophages in adipose tissue (30–35,40) and modulate cytokine production by adipocytes, macrophages, and neutrophils (32,36–39). Our results corroborate this notion and suggest that supraphysiological levels of PST can impose an inflammatory burden and thereby disrupt glucose homeostasis. Our profiling of pro- and anti-inflammatory markers and cytokines suggests that the adipose tissue of KO-DIO mice was less inflamed than that of WT-DIO mice. Since PST treatment of peritoneal macrophage cultures (obtained from WT and Chga-KO mice on NCD) increased the expression of Tnfα, Mcp-1, IL-6, IL-12p40, and iNos (Fig. 8C–G), this suggests a direct effect of PST on the behavior of peritoneal macrophages. As expected, the expression of these genes was much lower in Chga-KO macrophages. Treatment of Chga-KO macrophages with PST elevated their expression but not to the extent of PST-treated WT macrophages (Fig. 8C–G). Moreover, PST provoked macrophage chemotaxis in vitro (Fig. 8H). It is possible that other Chga-derived peptides also contributed to the regulation of inflammatory gene expression, leading to a partial PST-stimulated response in Chga-KO macrophages compared with WT macrophages. However, in terms of insulin resistance, PST alone can fully account for the insulin resistance conferred to KO-DIO mice. It is possible that the diabetogenic effects of a supraphysiological level of PST may result from the combined stimulation of two different pathways, inflammation and ER stress, as we have shown that PST reduces the adaptive unfolded protein response through binding to the ER chaperone, GRP78 (31).
Insulin Sensitization by Antagonism of PST Action
Finally, to implicate a direct in vivo role of PST in the regulation of insulin sensitivity, we injected WT-DIO mice with the PST variant, PSTv1, to block the effect of native PST (which is upregulated in DIO mice). PSTv1 lacks the first three N-terminal residues of native PST and was designed to block PST-mediated inhibition of glucose uptake and leptin secretion in 3T3-L1 adipocytes (G.K.B., T. Pasqua, S. Talukdar, J.R.G., S.K.M., unpublished data). Chronic PSTv1 treatment lowered fasting plasma glucose levels in WT-DIO mice and improved glucose tolerance (Fig. 3A and B) and insulin sensitivity (Fig. 3C and D). These results suggested that in WT-DIO mice where the level of PST is high, PSTv1 administration could compete with the native PST and thereby phenocopy Chga-KO mice. From this perspective, PSTv1 could serve as an antidiabetes agent.
Conclusion
The deficiency of PST in Chga-KO mice provides an anti-inflammatory environment leading to prevention of HFD-induced insulin resistance. As a result, insulin sensitivity is maintained even in obese KO-DIO mice. Thus, obesity manifests insulin resistance only in the presence of PST, and in its absence obesity is dissociated from insulin resistance. We visualize the chain of events after suppression of PST action that contributes to improved glucose homeostasis as 1) promotion of insulin signaling and suppression of inflammatory cytokine production through increased PI-3-K/Akt/FoxO1 signaling and 2) suppression of hepatic gluconeogenesis by increased expression of mature Srebp-1c. Future studies will be directed to determine whether administration of PSTv1 to diabetic animals will improve insulin sensitivity by reducing PST-induced inflammation and ER stress.
Article Information
Funding. This work was supported by grants from the Department of Veterans Affairs (5I01BX000323-04 to S.K.M. and 5I01BX000702 to N.-W.C.) and S.K.M.'s personal fund.
Duality of Interest. No potential conflicts of interest relevant to this article were reported.
Author Contributions. G.K.B. researched data and wrote the manuscript. M.L., E.A., J.A.S., J.R.G., J.W., and C.U.V. researched data. N.-W.C. participated in discussion and edited the manuscript. D.T.O. participated in discussion and reviewed and edited the manuscript. S.K.M. conceived the idea and researched and analyzed the data; made the graphics; and reviewed and edited the manuscript. S.K.M. and G.K.B. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Footnotes
This article contains Supplementary Data online at http://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db13-1747/-/DC1.
- Received November 15, 2013.
- Accepted July 14, 2014.
- © 2015 by the American Diabetes Association. Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered.
References
In this Issue
