Suppressor of cytokine signaling 1 (SOCS-1) and SOCS-3 cause insulin resistance through inhibition of tyrosine phosphorylation of insulin receptor substrate proteins by discrete mechanisms

Abstract

Insulin resistance is a pathophysiological component of type 2 diabetes and obesity and also occurs in states of stress, infection, and inflammation associated with an upregulation of cytokines. Here we show that in both obesity and lipopolysaccharide (LPS)-induced endotoxemia there is an increase in suppressor of cytokine signaling (SOCS) proteins, SOCS-1 and SOCS-3, in liver, muscle, and, to a lesser extent, fat. In concordance with these increases by LPS, tyrosine phosphorylation of the insulin receptor (IR) is partially impaired and phosphorylation of the insulin receptor substrate (IRS) proteins is almost completely suppressed. Direct overexpression of SOCS-3 in liver by adenoviral-mediated gene transfer markedly decreases tyrosine phosphorylation of both IRS-1 and IRS-2, while SOCS-1 overexpression preferentially inhibits IRS-2 phosphorylation. Neither affects IR phosphorylation, although both SOCS-1 and SOCS-3 bind to the insulin receptor in vivo in an insulin-dependent fashion. Experiments with cultured cells expressing mutant insulin receptors reveal that SOCS-3 binds to Tyr960 of IR, a key residue for the recognition of IRS-1 and IRS-2, whereas SOCS-1 binds to the domain in the catalytic loop essential for IRS-2 recognition in vitro. Moreover, overexpression of either SOCS-1 or SOCS-3 attenuates insulin-induced glycogen synthesis in L6 myotubes and activation of glucose uptake in 3T3L1 adipocytes. By contrast, a reduction of SOCS-1 or SOCS-3 by antisense treatment partially restores tumor necrosis factor alpha-induced downregulation of tyrosine phosphorylation of IRS proteins in 3T3L1 adipocytes. These data indicate that SOCS-1 and SOCS-3 act as negative regulators in insulin signaling and serve as one of the missing links between insulin resistance and cytokine signaling.

FIG. 1.

Increased expression levels of SOCS-1 and SOCS-3 in insulin-sensitive tissues in insulin-resistant states. (a) Induction of SOCS-1 and SOCS-3 in liver and muscle by endotoxemia. The top panels show representative results of RT-PCR. In the bottom panels the data are quantified and each bar represents the means ± standard errors (n = 6) of the SOCS expression standardized to GAPDH expression. *, P < 0.05 for 0 h versus 1 h of treatment; **, P < 0.05 for 0 h versus 4 h of treatment. (b) Upregulation of SOCS-1 and SOCS-3 expression in db/db mice. The top panels show representative results of RT-PCR, and in the bottom panels each bar represents the means ± standard errors (n = 4) of the SOCS expression standardized to GAPDH expression. *, P < 0.05 for control versus db/db mice.

FIG. 2.

Endotoxemia impairs insulin signaling at early steps. (a) Tyrosine phosphorylation of IR and its substrates. The immunoprecipitates (IP) with αIR, αIRS-1, or αIRS-2 from liver lysates with or without insulin stimulation were immunoblotted (IB) with αPY (top panels) or the same antibody (bottom panels). (b) PI 3-kinase activity associated with phosphotyrosine complexes. The top panel shows a representative result. In the bottom panel the data are quantified and each bar represents the means ± standard errors (n = 4). *, P < 0.05 for LPS negative versus LPS positive with insulin treatment. (c) Insulin-induced Akt activity. Each bar represents the means ± standard errors (n = 4). *, P < 0.05 for LPS negative versus LPS positive without insulin treatment; **, P < 0.05 LPS negative versus LPS positive with insulin treatment.

FIG. 3.

Increased SOCS-1 or SOCS-3 inhibits tyrosine phosphorylation of IRS proteins and downstream signaling in liver. (a) Comparable expression of SOCS-1 and SOCS-3 introduced by adenovirus injection to those induced by LPS treatment. The panels show representative results of RT-PCR using RNA from liver with the indicated treatment. (b) Liver contents of SOCS-1 and SOCS-3. The immunoprecipitates (IP) with αSOCS-1, αSOCS-3, or αFLAG from liver lysates were immunoblotted (IB) with the same antibody. (c) Tyrosine phosphorylation of IRS-1, IRS-2, and IR. The immunoprecipitates with αIRS-1, αIRS-2, or αIR antibodies from liver lysates with or without insulin stimulation were immunoblotted with αPY (top) or the same antibody (middle). (d) Insulin-induced PI 3-kinase activity. p85 subunit of PI 3-kinase bound to tyrosine-phosphorylated proteins (top) and PI 3-kinase activity (middle) were estimated by using the immunoprecipitates with αIRS-1, αIRS-2, or αPY from liver lysates with or without insulin stimulation. The top and middle panels show representative results; in the bottom panels each bar represents the means ± standard errors (n = 3). *, P < 0.05 for LacZ versus SOCS-1; **, P < 0.05 for LacZ versus SOCS-3 adenovirus treatment.

FIG. 4.

SOCS-1 and SOCS-3 bind to distinct sites of the IR and inhibit phosphorylation of IRS-1 and IRS-2 in different fashions. (a) Insulin-dependent binding of SOCS-1 and SOCS-3 to the IR in vivo. The immunoprecipitates (IP) with αPY (left), αSOCS-1 (middle), or αSOCS-3 (right) from liver or muscle lysates with or without insulin stimulation were immunoblotted (IB) with αPY (top) or αIR (bottom). (b) SOCS-3, but not SOCS-1, binds to Tyr960 of the IR. The immunoprecipitates with αSOCS-1 (top) or αSOCS-3 (middle) from CHO-IR cells or CHO-960F cells were immunoblotted with αIR, and the immunoprecipitates with αPY (bottom) were immunoblotted with αPY. (c) SOCS-1, but not SOCS-3, binds to the kinase domain of the IR. Total cell lysates (left lane), the pulled-down proteins with GST-TY (middle lane), or the phosphorylated GST-TY (pGST-TY, right lane) were blotted with αFLAG. Ins., insulin.

FIG. 5.

Differential inhibitory effects of SOCS-1 and SOCS-3 on IRS-1 and IRS-2 phosphorylation. (a) Effects of overexpression of SOCS-1 or SOCS-3 on the phosphorylation of IR, IRS-1, and IRS-2. The immunoprecipitates (IP) with αIRS-1, αIRS-2, or αIR from Fao cells infected with the indicated adenovirus were immunoblotted (IB) with αPY or the same antibody, and the lysates were immunoprecipitated with αFLAG. (b) Time course of inhibitory effects of SOCS-1 and SOCS-3 on IRS-1 and IRS-2 phosphorylation. The immunoprecipitates with αIRS-1, αIRS-2, or αIR from Fao cells infected with the indicated adenovirus stimulated with insulin (Ins.) for the indicated period were immunoblotted with αPY. Results were expressed as means ± standard errors (n = 3) of the percent maximal phosphorylation in control cells treated with LacZ. *, P < 0.05 for LacZ versus SOCS-1; **, P < 0.05 for LacZ versus SOCS-3 adenovirus treatment.

FIG. 6.

Inhibitory effects of SOCS-1 and SOCS-3 on biological responses to insulin in muscle cells and adipocytes. (a) Increased SOCS-1 or SOCS-3 inhibits glycogen synthase activity in L6 myotubes. (b) Increased SOCS-1 or SOCS-3 inhibits glucose transport activity in 3T3L1 adipocytes. The left panels show representative results of RT-PCR using RNA with the indicated treatment, and in the right graphs each bar represents the means ± standard errors (n = 4). *, P < 0.05 for LacZ versus SOCS-1; **, P < 0.05 for LacZ versus SOCS-3 adenovirus treatment.

FIG. 7.

Effects of antisense treatment of SOCS-1 or SOCS-3 on TNF-α-induced inhibition of IRS phosphorylation. The immunoprecipitates with αSOCS-1 (a, top row), αSOCS-3 (b, top row), αIRS-1 (a and b, middle rows), and αIRS-2 (a and b, bottom rows) from 3T3L1 adipocytes transfected with the indicated oligonucleotide treated with the indicated reagent were immunoblotted with the same antibody (for αSOCS-1 and αSOCS-3) or αPY (for αIRS-1 and αIRS-2). In the graphs, each bar represents the means ± standard errors (n = 4). *, P < 0.05 for C3 (Ins. + TNF) versus AS3 (Ins. + TNF); **, P < 0.05 for C1 (Ins. + TNF) versus AS1 (Ins. + TNF) or C3 (Ins. + TNF) versus AS3 (Ins. + TNF). Ins., insulin.

FIG. 8.

Model of discrete mechanisms of inhibition of insulin signaling by SOCS-1 and SOCS-3 induced by proinflammatory cytokines.