Pid1 induces insulin resistance in both human and mouse skeletal muscle during obesity

Abstract

Obesity is associated with insulin resistance and abnormal peripheral tissue glucose uptake. However, the mechanisms that interfere with insulin signaling and glucose uptake in human skeletal muscle during obesity are not fully characterized. Using microarray, we have identified that the expression of Pid1 gene, which encodes for a protein that contains a phosphotyrosine-interacting domain, is increased in myoblasts established from overweight insulin-resistant individuals. Molecular analysis further validated that both Pid1 mRNA and protein levels are increased in cell culture models of insulin resistance. Consistent with these results, overexpression of phosphotyrosine interaction domain-containing protein 1 (PID1) in human myoblasts resulted in reduced insulin signaling and glucose uptake, whereas knockdown of PID1 enhanced glucose uptake and insulin signaling in human myoblasts and improved the insulin sensitivity following palmitate-, TNF-α-, or myostatin-induced insulin resistance in human myoblasts. Furthermore, the number of mitochondria in myoblasts that ectopically express PID1 was significantly reduced. In addition to overweight humans, we find that Pid1 levels are also increased in all 3 peripheral tissues (liver, skeletal muscle, and adipose tissue) in mouse models of diet-induced obesity and insulin resistance. An in silico search for regulators of Pid1 expression revealed the presence of nuclear factor-κB (NF-κB) binding sites in the Pid1 promoter. Luciferase reporter assays and chromatin immunoprecipitation studies confirmed that NF-κB is sufficient to transcriptionally up-regulate the Pid1 promoter. Furthermore, we find that myostatin up-regulates Pid1 expression via an NF-κB signaling mechanism. Collectively these results indicate that Pid1 is a potent intracellular inhibitor of insulin signaling pathway during obesity in humans and mice.

Figure 1.

Microarray Analysis and qPCR Validation of Differentially Expressed Genes between IS and IR Human Myoblasts. A, Gene expression changes between IS and IR human myoblasts. A (left panel), heat map representing genes that are differentially expressed between IS and IR human myoblasts as determined through microarray (fold change ≥ 1.5 and P < .05); the color scheme, which represents fold change, is indicated at the bottom of the heat map. A (right panel) qPCR analysis of the mRNA expression of 60 selected genes from microarray data; the color scheme, which represents fold change, is indicated at the bottom of the heat map. The data show fold change ± SEM normalized to GAPDH from 2 independent experiments. B, qPCR analysis of PID1 mRNA expression in IS and IR human myoblasts (n = 7 per group). Bars represent mean fold change in PID1 gene expression ± SEM normalized to GAPDH from 3 independent experiments. P < .01 (**). C, Immunoblot (IB) analysis of PID1 protein levels in IS and IR human myoblasts (n = 7 per group). The levels of tubulin were assessed to ensure equal loading of protein. The graphs displays densitometry analysis of PID1 protein levels normalized to tubulin and expressed as arbitrary units (A.U). The data represent the mean ± SEM of triplicate experiments. P < .001 (***).

Figure 2.

Differential Expression of Human and Mouse PID1/Pid1 Isoforms. A (upper panel) Schematic showing protein alignment of the 4 different human PID1 isoforms. A (lower panel), Schematic showing protein alignment of the 2 different mouse Pid1 isoforms. Alignments were generated using the UniProt protein alignment tool (http://www.uniprot.org/uniprot/Q7Z2X4). The amino acid number along with UniProt protein IDs and gene symbol are illustrated. B, qPCR analysis of human PID1 isoform 1 (hIso1), isoform 2 (hIso2), isoform 3 (hIso3), and isoform 4 (hIso4) mRNA expression levels in hMb15 myoblasts and myotubes. Bars represent mean relative change in human PID1 gene expression ± SEM normalized to GAPDH from triplicate experiments. P < .05 (*) and P < .01 (**). C, qPCR analysis of human PID1 hIso2 and hIso4 mRNA expression in IS and IR human myoblasts. Bars represent mean fold change in human PID1 gene expression ± SEM normalized to GAPDH from triplicate experiments. P < .01 (**). D, qPCR analysis of mouse Pid1 isoform 1 (mIso1) and isoform 2 (mIso2) mRNA expression in C2C12 myoblasts (MB), Myotubes (MT), primary myoblasts (1^° MB), and primary myotubes (1^° MT) derived from WT mice (C57BL/6J). Bars represent mean relative change in mouse Pid1 gene expression ± SEM normalized to GAPDH from triplicate experiments. P < .05 (*) and P < .01 (**). E, qPCR analysis of mouse Pid1 isoform 1 (mIso1) and isoform 2 (mIso2) mRNA expression in different tissues isolated from WT mice (C57BL/6J) (n = 3). Bars represent mean relative change in muse Pid1 gene expression ± SEM normalized to GAPDH from triplicate experiments. F, qPCR analysis of mouse Pid1 isoform 1 (mIso1) and isoform 2 (mIso2) mRNA expression in biceps femoris (BF), extensor digitorum longus (EDL), quadriceps (Quad), soleus (Sol), gastrocnemius (Gas), and tibialis anterior (TA) muscles collected from WT mice (C57BL/6J) (n = 3). Bars represent mean relative change in mouse Pid1 gene expression ± SEM normalized to GAPDH from triplicate experiments.

Figure 3.

Pid1 is Up-Regulated during HFD and High-Glucose-Induced Obesity and IR in Mice. qPCR analysis of mouse Pid1 isoform 1 (mIso1) and isoform 2 (mIso2) mRNA expression in adipose (A), muscle (Gas) (B) and liver (C) tissues isolated from WT mice (C57BL/6J) following 4, 9, 12, 15, and 19 weeks of either HFD or CD feeding (n = 8 per group). Bars represent mean fold change in mouse Pid1 gene expression ± SEM normalized to GAPDH from triplicate experiments. P < .05 (*) and P < .01 (**). D, IB analysis of Pid1 protein levels in adipose, muscle (Gas), and liver tissues isolated from WT mice (C57BL/6J) following 4, 9, 12, 15, and 19 weeks of either HFD or CD feeding. The levels of tubulin were assessed to ensure equal loading of protein. qPCR analysis of mouse Pid1 isoform 1 (mIso1) and isoform 2 (mIso2) mRNA expression in adipose (E), muscle (Gas) (F), and liver (G) tissues isolated from WT mice (C57BL/6J) injected with either saline or 2 mg/g BW glucose (n = 8 per group) for 12 weeks. Bars represent mean fold change in mouse Pid1 gene expression ± SEM normalized to GAPDH from triplicate experiments. P < .01 (**) and P < .001 (***). H, IB analysis of Pid1 protein levels in adipose, muscle (Gas), and liver tissues isolated from WT mice (C57BL/6J) injected with either saline or 2 mg/g BW glucose (n = 4 per group) for 12 weeks. The levels of tubulin were assessed to ensure equal loading of protein. The graphs displays densitometry analysis of PID1 protein levels normalized to tubulin and expressed as arbitrary units (A.U). The data represent the mean ± SEM of triplicate experiments. P < .05 (*) and P < .01 (**).

Figure 4.

Myostatin Up-Regulates Pid1 Expression in Peripheral Tissues. A, qPCR analysis of human PID1 hIso2 and hIso4 mRNA expression in hMb15 myotubes treated with either CHO-control or CHO-Mstn (1:4 dilution). Bars represent mean fold change in human PID1 gene expression ± SEM normalized to GAPDH from triplicate experiments. P < .01 (**). B, qPCR analysis of mouse Pid1 isoform 1 (mIso1) and isoform 2 (mIso2) mRNA expression in C2C12 myotubes treated with either CHO-control or CHO-Mstn (1:4 dilution). Bars represent mean fold change in mouse Pid1 gene expression ± SEM normalized to GAPDH from triplicate experiments. P < .05 (*). C, qPCR analysis of mouse Pid1 isoform 1 (mIso1) and isoform 2 (mIso2) mRNA expression in 96-hour differentiated C2C12 myoblasts transfected with either an empty vector control (pCMV6) or Mstn-overexpressing construct (pCMV6-Mstn). Bars represent mean fold change in mouse Pid1 gene expression ± SEM normalized to GAPDH from triplicate experiments. P < .001 (***). D, IB analysis of Mstn and Pid1 protein levels 96 hours differentiated C2C12 myoblasts transfected with either an empty vector control (pCMV6) or Mstn-overexpressing construct (pCMV6-Mstn). The levels of tubulin were assessed to ensure equal loading of protein. qPCR analysis of mouse Pid1 isoform 1 (mIso1) and isoform 2 (mIso2) mRNA expression in adipose (E), muscle (F), and liver (G) tissues isolated from WT mice (C57BL/6J) injected with either saline or rhMstn (5 μg/g BW). Bars represent mean fold change in mouse Pid1 gene expression ± SEM normalized to GAPDH from triplicate experiments. P < .001 (***). H, IB analysis of Pid1 protein levels in adipose, skeletal muscle, and liver tissues isolated from either saline or rhMstn (5 μg/g BW)-injected WT mice (C57BL/6J). The levels of tubulin were assessed to ensure equal loading of protein. I, IB analysis of Mstn protein levels in C2C12 myoblasts transfected with either nonsilencing control siRNA (Neg siRNA) or 4 different Mstn-specific siRNAs (Mstn siRNA-1, Mstn siRNA-2, Mstn siRNA-3, Mstn siRNA-4). The levels of tubulin were assessed to ensure equal loading of protein. J, qPCR analysis of mouse Pid1 isoform 1 (mIso1) and isoform 2 (mIso2) mRNA expression in 96-hour differentiated C2C12 myoblasts transfected with transfected with either nonsilencing control siRNA (Neg siRNA) or Mstn-specific siRNA (Mstn siRNA) on C2C12 myoblasts. Bars represent mean fold change in mouse Pid1 gene expression ± SEM normalized to GAPDH from triplicate experiments. P < .05 (*). K, IB analysis of Mstn and Pid1 protein levels in 96-hour differentiated C2C12 myoblasts transfected with (+) or without (−) Mstn-specific siRNA (Mstn siRNA). The levels of tubulin were assessed to ensure equal loading of protein. L, qPCR analysis of mouse Pid1 isoform 1 (mIso1) and isoform 2 (mIso2) mRNA expression in muscle (Gas) isolated from WT and Mstn^−/− mice (n = 3 per group). Bars represent mean fold change in Pid1 gene expression ± SEM normalized to GAPDH from triplicate experiments. P < .05 (*). M, IB analysis of Pid1 protein levels in adipose, muscle (Gas), and liver tissues isolated from either WT or Mstn^−/− mice (n = 2 per group). The levels of tubulin were assessed to ensure equal loading of protein.

Figure 5.

Myostatin Signals via NF-κB to Up-Regulate Pid1 Gene Expression. A, Schematic representation of human PID1 (−1.3 kb) and mouse Pid1 (−4.5-kb) upstream elements. Locations of putative Sp1, GATA, STRE, and NF-κB binding elements, as identified through in silico analysis, are indicated. Arrows denote transcription start sites in both human and mouse PID1/Pid1 promoters. B (upper panel), Schematic representation of the 0.978-kb proximal human PID1 promoter-reporter construct (PID1 promoter) used for subsequent analysis. The locations of the putative NF-κB binding elements are indicated. B (lower panel), Assessment of PID1 promoter-reporter luciferase activity in C2C12 myoblasts cotransfected with either empty vector control (pLightSwitch-Prom) or PID1 promoter (pLightSwitch-PID1), and treated with CHO-control (−) or a 1:4 (+) or 1:2 (++) dilution of CHO-Mstn. Luciferase activity was normalized to firefly luciferase activity and expressed as relative luciferase activity. Bars represent relative luciferase activity ± SEM and are representative of triplicate experiments. P < .01 (**). C, Assessment of PID1 promoter-reporter luciferase activity in C2C12 myoblasts transfected with either pLightSwitch-Prom or pLightSwitch-PID1 promoter and further treated with the absence (−) or presence (+) of Mstn-specific siRNA (Mstn siRNA). Luciferase activity was normalized to firefly luciferase activity and expressed as relative luciferase activity. Bars represent relative luciferase activity ± SEM and are representative of triplicate experiments. P < .01 (**). D, Assessment of PID1 promoter-reporter luciferase activity in C2C12 myoblasts transfected with either pLightSwitch-Pro or pLightSwitch-PID1 and treated with (+) or without (−) a 1:4 dilution of CHO-Mstn in the absence (−) or presence (+) of the NF-κB-specific inhibitor BAY-117085. Luciferase activity was normalized to firefly luciferase activity and expressed as relative luciferase activity. Bars represent relative luciferase activity ± SEM and are representative of triplicate experiments. P < .01 (**) and P < .001 (***). E, ChIP analysis of NF-κB interaction with NF-κB binding elements in the human PID1 promoter and further treated with CHO-control (−) or CHO-Mstn (+) in the absence (−) or presence (+) of BAY-117085. PCR amplicons for input Genomic DNA Input-NF-κB (PID1), negative control antibody (anti-IgG), and anti-NF-κB (p65) are indicated.

Figure 6.

PID1 Overexpression in hMb15 Myoblasts Induces IR. A, Immunofluorescence images showing EGFP-positive hMb15 myotubes transduced with either empty vector (EGFP-Con) or an EGFP-tagged-PID1-overexpressing construct (EGFP-PID1). An image of untransduced negative control (Negative) hMb15 myotubes is also shown. Scale bars represent 100 μm. B, IB analysis of PID1 protein levels in hMb15 myotubes following lentiviral-mediated transduction of either empty vector (EGFP-Con) or an EGFP-tagged-PID1-overexpressing construct (EGFP-PID1). The levels of tubulin were assessed to ensure equal loading of protein. C, Assessment of 2-NBDG glucose uptake in hMb15 myoblasts after lentiviral-mediated transduction of either empty vector (EGFP-Con) or a EGFP-tagged-PID1-overexpressing construct (EGFP-PID1) during basal conditions (0 μM) and following stimulation with increasing concentrations of insulin (0.01 μM, 0.1 μM, and 1 μM). The bars represent fold increase ± SEM and are representative of triplicate experiments. P < .01 (**). D, IB analysis of pIRS1, IRS1, pAkt, Akt, and plasma membrane (PM) Glut-4 protein levels in hMb15 myoblasts following lentiviral-mediated transduction of either empty vector (EGFP-Con) or a EGFP-tagged-PID1-overexpressing construct (EGFP-PID1) and treatment without (−) or with increasing concentrations of insulin (0.01 μM, 0.1 μM, and 1 μM). The levels of tubulin were assessed to ensure equal loading of protein. E, IB analysis of pIRS1, IRS1, pAkt, Akt, and plasma membrane (PM) Glut-4 protein levels in hMb5 myoblasts following lentiviral-mediated transduction of either empty vector (EGFP-Con) or a EGFP-tagged-PID1-overexpressing construct (EGFP-PID1) and treatment without (−) or with increasing concentrations of insulin (0.01 μM, 0.1 μM, and 1 μM). The levels of tubulin were assessed to ensure equal loading of protein.

Figure 7.

PID1 Knockdown in hMb15 Myoblasts Improves IS. A, Immunofluorescence images showing EGFP-positive hMb15 myotubes transduced with either an EGFP-tagged nonsilencing control shRNA vector (EGFP-shCon) or an EGFP-tagged PID1-specific shRNA construct (EGFP-shPID1). Scale bars represent 100 μm. B (left panel), qPCR analysis of PID1 mRNA expression levels in hMb15 myotubes following lentiviral-mediated transduction of either an EGFP-shCon or EGFP-shPID1. Bars represent mean fold change in PID1 gene expression ± SEM normalized to GAPDH. P < .01 (**). B (right panel), IB analysis of PID1 protein levels in hMb15 myotubes following lentiviral-mediated transduction of either EGFP-shCon or shEGFP-shPID1. The levels of tubulin were assessed to ensure equal loading of protein. C, Assessment of 2-NBDG glucose uptake in hMb15 myoblasts after lentiviral-mediated transduction of either an EGFP-shCon or EGFP-shPID1. during basal conditions (0 μM) and following stimulation with increasing concentrations of insulin (0.01 μM, 0.1 μM, and 1 μM). The bars represent fold increase ± SEM and are representative of triplicate experiments. P < .05 (*) and P < .01 (**). D, IB analysis of pIRS1, IRS1, pAkt, Akt, and plasma membrane (PM) Glut-4 protein levels in hMb15 myoblasts following lentiviral-mediated transduction of either EGFP-shCon or EGFP-shPID1 and treatment without (−) or with increasing concentrations of insulin (0.01 μM, 0.1 μM, and 1 μM). The levels of tubulin were assessed to ensure equal loading of protein. E, IB analysis of pIRS1, IRS1, pAkt, Akt, and plasma membrane (PM) Glut-4 protein levels in hMb5 myoblasts following lentiviral-mediated transduction of either EGFP-shCon or EGFP-shPID1 and treatment without (−) or with increasing concentrations of insulin (0.01 μM, 0.1 μM, and 1 μM). The levels of tubulin were assessed to ensure equal loading of protein.

Figure 8.

PID1 Regulates Mitochondrial Biogenesis and Function in Myoblasts. A, qPCR analysis of mtDNA:nuDNA ratio in hMb15 myoblasts following lentiviral-mediated transduction of either an EGFP-Con or EGFP-PID1. P < .01 (**). B, qPCR analysis of Mfn1,Mfn2, and Drp1 mRNA expression levels in hMb15 myotubes following lentiviral-mediated transduction of either an EGFP-Con or EGFP-PID1. Bars represent mean fold change in Mfn1, Mfn2, and Drp1gene expression ± SEM normalized to GAPDH. P < .01 (**). C, IB analysis of Mfn1, Mfn2, and Drp1 protein levels in hMb15 myoblasts following lentiviral-mediated transduction of either EGFP-Con or EGFP-PID1. The levels of tubulin were assessed to ensure equal loading of protein. D, qPCR analysis of mtDNA:nuDNA ratio in human myoblasts following lentiviral-mediated transduction of either an EGFP-shCon or EGFP-shPID1. P < .01 (**). E, qPCR analysis of Mfn1,Mfn2, and Drp1 mRNA expression levels in hMb15 myotubes following lentiviral-mediated transduction of either an EGFP-shCon or EGFP-shPID1. Bars represent mean fold change in Mfn1, Mfn2, and Drp1gene expression ± SEM normalized to GAPDH. P < .01 (**). F, IB analysis of Mfn1, Mfn2, and Drp1 protein levels in hMb15 myoblasts following lentiviral-mediated transduction of either EGFP-shCon or EGFP-shPID1. The levels of tubulin were assessed to ensure equal loading of protein.

Figure 9.

Palmitate, TNF-α, and Mstn Signal via PID1 to Induce IR in Human Myoblasts. A, qPCR analysis of PID1 mRNA expression levels in hMb15 myotubes treated without (−) or with (+) 0.25 mM palmitate (PA) for 24 hours. Bars represent mean relative mRNA levels of PID1 normalized to GAPDH. P < .05 (*). B, qPCR analysis of PID1 mRNA expression levels in hMb15 myotubes without (−) or with (+) CHO-Mstn for 24 hours. Bars represent mean relative mRNA levels of PID1 normalized to GAPDH. P < .001 (***). C, qPCR analysis of PID1 mRNA expression levels in hMb15 myotubes without (−) or with (+) 10 ng/mL TNF-α for 24 hours. Bars represent mean relative mRNA levels of PID1 normalized to GAPDH. P < .001 (***). D (upper panel), qPCR analysis of PID1 mRNA expression levels in hMb15 myotubes following lentiviral-mediated transduction of either an EGFP-shCon or EGFP-shPID1 and treatment without (−) or with (+) 0.25 mM palmitate (PA) for 24 hours. D (lower panel), IB analysis of pIRS1, IRS1, pAkt, and Akt protein levels in hMb15 myoblasts following lentiviral-mediated transduction of either EGFP-shCon or EGFP-shPID1 and treatment without (−) or with (+) 0.25 mM palmitate (PA) for 24 hours. E (upper panel), qPCR analysis of PID1 mRNA expression levels in hMb15 myotubes following lentiviral-mediated transduction of either an EGFP-shCon or EGFP-shPID1 and treatment without (−) or with (+) 10 ng/mL TNF-α for 24 hours. E (lower panel), IB analysis of pIRS1, IRS1, pAkt, and Akt protein levels in hMb15 myoblasts following lentiviral-mediated transduction of either EGFP-shCon or EGFP-shPID1 and treatment without (−) or with (+) 10 ng/mL TNF-α for 24 hours. F (upper panel), qPCR analysis of PID1 mRNA expression levels in hMb15 myotubes following lentiviral-mediated transduction of either an EGFP-shCon or EGFP-shPID1 and treatment without (−) or with (+) CHO-Mstn for 24 hours. F (lower panel), IB analysis of pIRS1, IRS1, pAkt, and Akt protein levels in hMb15 myoblasts following lentiviral-mediated transduction of either EGFP-shCon or EGFP-shPID1 and treatment without (−) or with (+) CHO-Mstn. The levels of tubulin were assessed to ensure equal loading of protein.

Figure 10.

Molecular Signaling Mechanisms of Pid1 Gene Regulation and Risks Associated in Skeletal Muscle. Excess Mstn activates NF-κB transcription factor translocation from cytoplasm into the nucleus. In the nucleus NF-κB recognizes NF-κB binding sites present on the proximal Pid1 promoter and initiates Pid1 gene transcription. The up-regulated Pid1 interferes with insulin signaling and reduces muscle glucose uptake efficiency. The development of obesity, T2D, and IR are associated with increased Pid1 activity.