Retraction: 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.

Obesity is a worldwide epidemic, which stems from excess high-calorie intake and an increasingly sedentary lifestyle. Obesity is mainly caused by alterations in the metabolic profile of an individual that results in increased fat accumulation, hyperinsulinemia, hyperglycemia and hyperlipidemia (1, 2). The metabolic abnormalities observed during obesity lead to the development of insulin resistance (IR) and type 2 diabetes (T2D) over time; however, under certain circumstances the development of IR and T2D is independent of obesity (3). Although defects in a plethora of signaling pathways have been discovered in IR, little is known about the factors that induce IR. Wang et al. (4) have recently described the identification and characterization of NYGGF4 or phosphotyrosine interaction domain-containing protein 1 (Pid1), a novel gene that is primarily expressed in adipose tissue, heart, and skeletal muscle and was up-regulated in sc adipose tissue of obese subjects. Functional analysis indicated that ectopic expression of Pid1 in 3T3L1 preadipocytes enhances the proliferation of 3T3L1 cells by regulating the S-phase of the cell cycle, without affecting adipocyte differentiation of 3T3L1 cells in vitro (4). Studies by Zhang et al (5) further reported that Pid1 overexpression in differentiated 3T3L1 cells impairs insulin-stimulated glucose uptake. The inhibitory effects of Pid1 overexpression on glucose uptake in 3T3L1 cells was shown to be due to decreased Glut-4 translocation to the cell membrane and reduced tyrosine phosphorylation of insulin receptor substrate 1 (IRS1) and serine phosphorylation of Akt, independent of the Ras/MAPK insulin-signaling pathway (5). Consistent with this observation, shRNA-mediated knock down of Pid1 in 3T3L1 adipocytes increases both basal and insulin-stimulated glucose uptake. Moreover, pretreatment of Pid1 knockdown cells with a phosphatidylinositol 3-kinase inhibitor, blocked the enhanced insulin-stimulated glucose uptake observed, suggesting a role for Pid1 in PI3K-mediated insulin signaling (6). Overexpression of Pid1 was further shown to have no affect on the levels of secreted adipokines, including TNF-α, IL-6, adiponectin, and resistin, in conditioned medium collected from Pid1-overexpressing 3T3L1 cells (5). However, it is important to highlight that TNF-α and free fatty acids significantly increase Pid1 expression, whereas adiponectin and resistin decrease Pid1 expression in 3T3L1 cells (7). It has also been reported that Pid1 transcript levels were up-regulated during differentiation of human preadipocytes and further enhanced by TNF-α addition in a time- and dose-dependent manner (7). Collectively, these data suggest that the increased levels of Pid1 induce IR in adipocytes.

Studies have also focused on determining the role of Pid1 in the development of IR in myoblast cells in culture. Overexpression of Pid1 in L6 myotubes reduces IRS1/PI3K/Akt signaling by inhibiting IRS1 phosphorylation during insulin stimulation (8). These studies clearly demonstrate that Pid1 overexpression in L6 myotubes preferentially interferes with insulin-stimulated glucose uptake. Consistent with these results, knockdown of Pid1 in C2C12 myotubes also enhances both basal and insulin-stimulated glucose uptake, Glut-4 protein translocation to cell membrane, and IRS1 and Akt phosphorylation (9). In order to identify genetic factors that specifically induce IR in skeletal muscle during obesity, we performed a microarray on the RNA isolated from human myoblasts cultured from IR and insulin-sensitive (IS) subjects. Microarray analysis showed that Pid1 expression is significantly up-regulated in myoblasts established from IR subjects. Using mouse models of obesity and IR, we have shown that expression of Pid1 is significantly higher in adipose, skeletal muscle, and liver tissues in a time-dependent manner during high-fat diet (HFD) feeding in mice and, moreover, correlates with the progression of IR in these mice. PID1 overexpression in human myoblasts inhibited insulin signaling, whereas ablation of PID1 expression improved insulin signaling in human myoblasts. Moreover, we now show that the myokine myostatin (Mstn), a growth factor previously associated with obesity and T2D in humans, up-regulates PID1 expression through the NF-κB signaling pathway in skeletal muscle, thus identifying a novel mechanism through which PID1 is up-regulated during obesity and IR.

Materials and Methods

Human subjects

All human primary myoblasts used in this study were derived from overweight, but otherwise healthy, Singaporean Chinese ethnic males (aged 21–40 years with body mass index 18.5–30 kg/m²) that were actively recruited to participate in a cross-sectional metabolic phenotype study code named Singapore Adult Metabolism Study (SAMS-1). Individuals were deemed IS or insulin resistant (IR) based on Insulin sensitivity index (ISI) and Homeostatic Model of Assessment of IR (Homeostatic Model of Assessment of IR [HOMA-IR]) calculations derived from hyperinsulinemic euglycemic glucose clamp analysis. Insulin sensitivity was measured using the hyperinsulinemic-euglycemic glucose clamp technique after of 10–12 hours of overnight fasting. Two plastic iv cannulas were inserted, one into the antecubital vein for infusion of 20% dextrose solution and insulin, and the second into a contralateral dorsal hand vein for blood sampling. After baseline blood samples were taken, insulin was infused at a fixed rate of 40 mU/m² body surface area/min for the duration of the clamp (120 minutes) in order to achieve plateau (steady-state) insulin levels, about 100 mU/mL above basal concentration to suppress hepatic glucose output. The plasma glucose concentration was clamped at euglycemic levels over the whole duration by monitoring the glucose concentration serially at 5-minute intervals using a glucose oxidase-based glucose analyzer (YSI 2300 STATPLUS, YSI Incorporated, Life Sciences, Yellow Spring Instruments) and adjusting the infusion rate of the hypertonic dextrose solution. The insulin sensitivity index (ISI) was calculated using the mean glucose disposal rate during the final 30 minutes of the insulin infusion (M-value) per kg of lean mass (obtained from dual energy x-ray absorptiometry) and the steady-state insulin concentration (I) during this same period. HOMA-IR was calculated as [fasting insulin (μU/mL) × fasting glucose (mmol/L)] ÷ 22.5. Insulin resistant (IR) and IS individuals had an ISI of <4.0 and >4.0 and an HOMA-IR of >2.5 and <2.5 respectively.

Collection of muscle biopsy and human primary myoblasts cultures

Subjects were instructed to avoid vigorous exercise for at least 3 days before the skeletal muscle biopsy. Percutaneous biopsy of the vastus lateralis muscle was performed under local anesthesia (1% lidocaine) using aseptic technique by a physician using a 16G full-core microbiopsy needle system (BioPince) in a subset of individuals who provided informed consent for this procedure. Human primary myoblasts cultures were established from approximately 10 mg muscle biopsy collected in PBS. Briefly, muscle tissue was minced and then digested with 0.2% collagenase type 1A (Sigma-Aldrich; catalog no. C5894) for 90 minutes at room temperature with shaking at 750 rpm, after which the cultures were pelleted down and passed through the 100-μm filter to remove undigested tissue. The cells were resuspended in growth medium (20% FBS; Invitrogen, 10% HS; Invitrogen, 1% PS; Invitrogen and 1% chick embryo extract; catalog no. C3999 from US Biologicals). Cultures were enriched for myoblasts by preplating on uncoated plates for 3 hours. After 3 hours of attachment, the supernatant having cells were preplated on to matrigel (Becton Dickinson)-coated plates and myoblast cultures were maintained at 37°C and 5% CO₂.

Animals

C57BL/6J (wild type) male mice (8 weeks of age) were obtained from National University of Singapore-Centre for Animal Resources. Mstn^−/− mice were a gift from Professor See-Jin Lee (Johns Hopkins University) and were housed and maintained at Nanyang Technological University Animal house in Singapore. All experimental procedures were approved by the Institutional Animal Care and Use Committee (IACUC) in Singapore. For HFD experiments, wild type (WT, n = 8) mice were fed with either HFD (58V8, Test Diet) or chow diet (CD) (58Y2, Test Diet) for 19 weeks. Following 4, 9, 12, 15, and 19 weeks of HFD regimen mice were euthanized with CO₂ and tissues were harvested for further analysis. For glucose treatment studies, wild-type mice were injected with either equal volumes of saline (n = 8) or with 2 mg/g body weight (BW) of glucose sc for 12 weeks, with an additional 25% glucose water provided as a further source of glucose. For recombinant human (rh) Mstn protein (rhMstn) injection studies 6-week-old male wild-type mice were randomly grouped into 2 groups. One group of mice (n = 8) was injected 3 times per week with an identical volume of saline, and the second group (n = 8) received 5 μg/g BW of rhMstn for a period of 12 weeks. The rhMstn protein was expressed and purified from Escherichia coli as described previously (10). Normal physiologic changes were monitored for all mice during the trial. Fasting blood glucose, glucose tolerance, and insulin tolerance testing were performed on all trial mice and were analyzed as previously published (11, 12).

Cell culture and treatments

The established mouse C2C12 myoblast cell line were purchased from American Type Culture Collection (ATCC) and was maintained as previously described (13). C2C12 cells were grown in proliferation medium (10% FBS, 1% P/S, and DMEM; Invitrogen) and differentiated in differentiation medium (DMEM, 2% HS, and 1% P/S; Invitrogen) for 96 hours to form fully differentiated myotubes. Human primary myoblasts (hMb15 and hMb5) (14, 15) were a kind gift of Drs. Vincent Mouly and Gillian Butler-Browne and were maintained as previously described (16, 17). Chinese hamster ovary (CHO) cells overexpressing Mstn or CHO-control cells were a kind gift from Dr. Se-Jin Lee (Johns Hopkins University). The CHO-Mstn and CHO-control conditioned medium was obtained from either CHO-Mstn or CHO-control cells. Treatment with either CHO-Mstn or CHO-control was performed as previously described (10). To induce IR in vitro, hMb15 myoblasts were treated for 24 hours with either 0.25 mM sodium palmitate (catalog no. P9767; Sigma), as described previously (18), or with 10 ng/mL TNF-α (catalog no. 170-076-10; Miltenyi Biotec GmbH).

Plasmids

Human PID1 pLKO.1 lentiviral shRNA vectors (catalog no. RHS4533) were purchased from Open-Biosystems with clone identification nos. TRCN0000123236, TRCN0000123234, and RCN0000123235. The murine Mstn overexpression (pCMV6-Mstn) vector (catalog no. MG227629) and pCMV6-Entry empty vector control (catalog no, MR227629) were purchased from Origene. The pMDLg/pRRE (ID 12251), pRSV-Rev (ID 12253), pMD2.G (ID 12259), and pLJM1 (ID 19319) plasmids were obtained from Addgene. The pLightSwitch-Prom empty (catalog no. S790005) and pLightSwitch-PID1 promoter-reporter vector (catalog no. S708635) was purchased from the SwitchGear Genomics Store,. The pGL3P (catalog E1761) firefly luciferase control vector was purchased from Promega Corp. Mstn-specific shRNA plasmids (catalog no. TG512805) were obtained from Origene.

Antibodies and immunoblotting (IB)

Mouse anti-Pid1 (catalog no. ab107978) and rabbit anti-Mfn1 (catalog no. ab104274) antibodies were purchased from Abcam. Human anti-PID1 (catalog no. HPA036103) and mouse anti-Tubulin (T9026) were purchased from Sigma-Aldrich. Rabbit anti-Akt1/2/3 (catalog no. sc-8312), rabbit anti-p-Akt1/2/3 (Ser-473)-R (catalog no. sc-7985), rabbit anti-Drp-1 (catalog no. sc-32898), and mouse anti-Mfn2 (catalog no. sc-100560) antibodies were purchased from Santa Cruz Biotechnology, Inc. Mouse anti-Glut-4 (catalog no. 2299), rabbit anti-IRS1 (catalog no. 2382), and mouse anti-p-IRS-1/p-Tyr-100 (catalog no. 9411) were purchased from Cell Signaling Technology, Inc. All IB experiments were performed as previously described (19). To assess for membrane localization of Glut-4, membrane fractions were isolated using the MEM-PER Plus Membrane protein extraction kit (catalog no. 89842) from Thermo Scientific as per the manufacturer guidelines.

Real-time quantitative PCR (qPCR) and primer sequences used

qPCR was performed on total RNA isolated from cells and tissues as described previously (10). The primers used in the current manuscript are listed below. Mouse Pid1 isoform 1 (mIso1) forward, 5′-agg act cat agt ggc tgc aag-3′ and reverse, 5′-aca tct tct cga gcc agt gtg t-3′; mouse Pid1 isoform 2 (mIso2) forward, 5′-gtg gca gcc ggc tac tga-3′ and reverse, 5′-agg act gct ggt ata ggt tcc-3′; human PID1 isoform 1 (hIso1) forward, 5′-cct tcc act gtg agg aca-3′ and reverse, 5′-gct ggg aga ggt tcc ttt tt-3′; human PID1 isoform 2 (hIso2) forward, 5′-ctg gca acc acc tga tgt ta-3′ and reverse, 5′-atg gag cca aac ttg gaa tg-3′; human PID1 isoform 3 (hIso3) forward, 5′-gga gcg cct gca ggt tac-3′ and reverse, 5′-atg gag cca aac ttg gaa tg-3′; human PID1 isoform 4 (hIso4) forward, 5′-ggt aac ctt gca gcc act gt-3′ and reverse, 5′-cgc ctg cag cac ttt cag-3′; human Drp1 forward, 5′-TGG GCG CCG ACA TCA-3′ and reverse, 5′-GCT CTG CGT TCC CAC TAC GA-3′; human Mfn1 forward, 5′-GGC ATC TGT GGC CGA GTT-3′ and reverse, 5′-ATT ATG CTA AGT CTC CGC TCC AA-3′; and human Mfn2 forward, 5′-GCT CGG AGG CAC ATG AAA GT-3′ and reverse, 5′-ATC ACG GTG CTC TTC CCA TT-3′. The sequences of the gene-specific primers used to validate the microarray results are available upon request. The quantification of mtDNA:nuDNA ratio was calculated as described previously (20).

Microarray analysis on IS and IR myoblasts

RNA samples from independent proliferating human primary myoblast cultures derived from 2 insulin resistant (IR) and 2 IS Chinese individuals were collected in TRIzol reagent, and RNA isolation was performed as per the manufacturers protocol (Invitrogen Life Technologies) and purified using the RNeasy Mini kit (Qiagen). Microarray analysis was performed by Agilent Technologies (Singapore) using the Agilent SurePrint G3 Human GE 8 × 60K array. Genes that were up- or down-regulated by ≥1.5 and P < .05 were considered as significantly altered. The heat map was established using GeneSpring GX 7.3 software (Silicon Genetics) with default settings: 1) All samples interpretation, 2) Pearson correlation for similarity, and 3) Average linkage. The heat map for the validated gene expression was generated using Orange 2.0b data mining, fruitful and fun software (http://orange.biolab.si).

Generation of the enhanced green fluorescent protein (EGFP)-PID1 construct and overexpression of PID1 in hMb15 myotubes

The human PID1 isoform 4 expression construct (EGFP-PID1) was generated by PCR amplification of the coding region of PID1 isoform 4 and its subsequent cloning into the EGFP-tagged pLJM1 viral vector backbone. The following PCR primers were used for PID1 isoform 4 amplification, 5′-AGC GCT ATG TGG CAG CCG GCC ACG-3′ and reverse 5′-ACC GGT GCC ATC ATC GGA TTC CAA TTC-3′. Human myoblasts (hMb15) total RNA was used as the RT template. The EGFP-PID1 construct was packaged with third-generation lentiviral vectors consisting of pMDLg/pRRE (ID 12251) and pRSV-Rev (ID 12253) packaging vectors and the pMD2.G envelope vector (ID 12259), all obtained from Addgene. The generation of EGFP-PID1 viral particles and subsequent infection of hMb15 myoblasts was performed as previously described (21), after which total RNA and protein were harvested for analysis of human PID1 levels.

Lentiviral-mediated knockdown of Pid1 in hMb15 myotubes

To transfect human primary myoblasts with PID1-specific shRNA constructs, a lentiviral-mediated transduction approach was undertaken. Human PID1 shRNA vectors (EGFP-shPID1) or the pLKO.1 empty vector (EGFP-shCon) were packaged with third-generation lentiviral vectors consisting of pMDLg/pRRE (ID 12251) and pRSV-Rev (ID 12253) packaging vectors and the pMD2.G envelope vector (ID 12259), all obtained from Addgene. The generation of PID1-specific shRNA containing viral particles and subsequent infection of hMb15 myoblasts was performed as previously described (21), after which total RNA and protein were harvested for analysis of human PID1 levels.

Glucose uptake

Glucose uptake was assessed in hMb15 human myoblasts using a fluorescent glucose uptake cell-based assay kit (Cayman Chemical Company; catalog no. 600471), as per the manufacturers protocol. Briefly, 10 000 cells per well were plated in a 96-well plate and were subjected to 2 successive incubations with serum-free MEM-α for 15 minutes each at 37°C/5% CO₂. After this, 2 mL of fresh serum-free MEM-α medium was added and the cells were subjected to 16 hours of starvation followed by stimulation with or without increasing concentrations (0.01 μM, 0.1 μM, and 1 μM) of insulin (Sigma; catalog no. I6634) for 15 minutes at 37°C/5% CO₂. Following insulin stimulation the cells were washed 3 times with 250 μL of PBS after which 150 μg/mL of 2-deoxy-2-[(7-nitro-2,1,3-benzoxadiazol-4-yl) amino]-D-glucose (2-NBDG) fluorescently labeled glucose was added in serum-free MEM-α medium followed by incubation for 10 minutes at 37°C/5% CO₂. Glucose uptake stopped by adding the glucose uptake inhibitor Apigenin at a dilution of 1:1000. The plate was then centrifuged at 400 × g for 5 minutes. The solution was then aspirated and the cells were washed once with 100 μL of PBS buffer followed by centrifugation at 400 × g for 5 minutes. After centrifugation, 100 μL of PBS buffer was added to each well and fluorescence was measured with an excitation of 490 nm and emission range of 510–570 nm. The cell lysates were collected and protein concentration was measured using Bradford's reagent. Blank (no 2-NBDG) readings were subtracted from all samples and all data was normalized to total protein per well.

Small interfering RNA (siRNA)-mediated knockdown of Mstn

Four different Mstn-specific siRNA oligos siRNA 1 5′-AGA UAC AAA UCC UCA GUA AUU-3′, siRNA 2 5′-GGA UUG GAU UAU CGC UCC UAA UU-3′, siRNA 3 5′-UUA UCA CGC UAC AAC GGA AUU-3′, siRNA 4 5′-GAU UGG AUU AUC GCU CCU AAA UU-3′, and negative control siRNA 5′-UUC UCC GAA CGU GUC ACG UUU-3′ (Neg siRNA) were purchased from Sigma. The oligos were chosen from the previous publication by Furalyov et al. (22). The siRNA (620 ng/2 mL) were transfected with Lipofectamine-LF2000 (LF2000 reagent; Invitrogen) as per the manufacturer's guidelines onto cells cultured in 6-well plates. After validating the results, we used a 1:1 ratio mixture of siRNA 2 and siRNA 4 for all relevant experiments.

Transient transfections and luciferase assays

To study the effect of Mstn overexpression on Pid1 expression, C2C12 myoblasts were transfected with either empty reporter vector (pCMV6) or Mstn overexpression vector (pCMV6-Mstn) in 6-well plates using Lipofectamine 2000 (LF2000 reagent; Invitrogen). Similarly, to study PID1 promoter activity, C2C12 myoblasts were cotransfected with either empty vector control (pLightSwitch-Prom) or human PID1 promoter-reporter luciferase construct (pLightSwitch-PID1) together with the pGL3P firefly luciferase control vector in 6-well plates using Lipofectamine 2000 (LF2000 reagent; Invitrogen). Transient transfections were performed as described below. Briefly, 3 μg DNA and 8 μL Lipofectamine 2000 was diluted in 100 μL of prewarmed DMEM. The DNA+DMEM mix and Lipofectamine 2000+DMEM mix were then mixed together by drop-wise addition and incubated at room temperature for 30 minutes. A constant volume (200 μL) of the transfection mix was then added into 2 mL of DMEM/well and incubated for 48 hours at 37°C/5%CO₂. For human PID1 promoter studies, C2C12 cells were treated with CHO-Mstn, CHO-control and/or 20 μM/mL⁻¹ of the NF-κB-specific inhibitor BAY-117085 (catalog no. B5681; Sigma) following transfection, as per the previously published protocol (23). Assessment of promoter-reporter luciferase activity was performed as previously reported (21). Relative luciferase activity in all of the extracted protein samples was measured in triplicates and normalized against Firefly luciferase activity from the pGL3P control vector.

Chromatin immunoprecipitation (ChIP)

ChIP was performed using hMb15 myoblasts transfected with either the empty vector control (pLightSwitch-Prom) or the human PID1 promoter-reporter construct (pLightSwitch-PID1) and treated with or without CHO-Mstn in the absence or presence of BAY-117085. ChIP was performed as described previously (21). The resulting products were visualized by agarose gel electrophoresis. The sequences of the primers used to detect the PID1 promoter-specific NF-κB elements are forward, 5′-CGG GGC TAC GGT TTC CTT ATT-3′ and reverse, 5′-TCC TAG TGA GGG GTG AAT GG-3′. The following antibodies were used to pull down the DNA/protein complexes: rabbit anti-NF-κB (p65) catalog no. sc-372 from Santa Cruz and rabbit anti-IgG (catalog no. X0903; DakoCytomation).

Statistical analysis

All data represent mean ± SEM and P values were calculated using either Student's t test or 2-way ANOVA.

Results

Pid1 is up-regulated in insulin-resistant human skeletal muscle myoblasts

To identify novel genes associated with the early onset of T2D in humans, we performed microarray analysis to compare the gene expression changes between human primary myoblast cultures derived from overweight IR and IS Chinese individuals. These IR and IS human primary myoblasts were further characterized in vitro and were shown to be IR or IS, respectively, based on the ability of insulin to stimulate both phosphorylation of Akt and glucose uptake (Bonala Sabeera Bonala, Sudarsanareddy Lokireddy, Craig McFarlane, Radiance Lim, Hilary Chua, Sreekanth Patnam, Melvin Khee Shing Leow, Chin Meng Khoo,Yung Seng Lee, E Shyong Tai, Peter D. Gluckman, Mridula Sharma, and Ravi Kambadur, personal communication). The results of the microarray are summarized in a heat map (Figure 1A, left panel) and significantly (P < .05) up-regulated and down-regulated genes are listed in Table 1 and Table 2, respectively. Importantly, genes that have been previously reported to be involved in IR and T2D in rodent, mouse, and human models, such as Casitas B-lineage lymphoma protooncogene b (Cblb), Smad family protein 1/2/3 (Smad1/2/3), sterol-regulatory element-binding transcription factor 1 (SREBP1), interleukin 6 (IL6), myozenin 3 (MyoZ3), TNFα-induced protein 6 (TNFAIP6), IRS1, IGF-binding protein 2/3/4/5/6 (IGFBP2/3/4/5/6), low density lipoprotein receptor-related protein 6 (LRP6), and growth and differentiation factors 1/5/11 (GDF1/5/11) (24–38) were similarly differentially regulated between IR and IS human myoblasts. Sixty significantly differentially regulated genes were further validated through quantitative real-time PCR (qPCR), and the results are summarized in a heat map (Figure 1A, right panel). Microarray analysis and subsequent validation through qPCR revealed that PID1 (phosphotyrosine interaction domain-containing protein-1) was up-regulated by approximately 9-fold in IR myoblasts compared with IS myoblasts (Table 1 and Figure 1A, right panel). We further validated the expression of PID1 in several IS and IR human primary myoblast cultures derived from different individuals (n = 7 for each), and we confirmed a significant increase in PID1 mRNA and protein levels in IR myoblasts, when compared with IS myoblasts (Figure 1, B and C).

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 (***).

Table 1.

List of Genes Up-Regulated in IR Myoblasts Compared with IS Myoblasts

Genbank Accession No.	Gene Symbol	Description	Fold Regulation
			IS	IR	(P< .05)
NM_005858	hAKAP8	Homo sapiens A kinase (PRKA) anchor protein 8 (AKAP8)	1.33	4.30	.00086
NM_170662	hCBLB	H. sapiens Cas-Br-M (murine) ecotropic retroviral transforming sequence b	1.00	9.99	.00200
NM_147190	HCERS5	H. sapiens ceramide synthase 5	1.33	7.51	.00150
NM_004864	hGDF15	H. sapiens growth differentiation factor 15	2.19	4.57	.00091
NM_00100157	HGDF6	H. sapiens growth differentiation factor 6	1.07	9.30	.00186
NM_182828	HGDF7	H. sapiens growth differentiation factor 7	1.01	9.86	.00197
NM_002099	hGYPA	H. sapiens glycophorin A	1.91	3.44	.00069
NM_005346	HHSPA1B	H. sapiens heat shock 70-kDa protein 1B	2.55	6.91	.00138
NM_000596	HIGFBP1	H. sapiens IGF-binding protein 1	1.14	8.77	.00175
NM_00125385	HIGFBP7	H. sapiens IGF-binding protein 7	1.05	9.51	.00190
NM_00100753	HIGFBP1	H. sapiens IGF-binding protein-like 1	1.45	6.91	.00138
NM_000600	hIL6	H. sapiens IL-6 (interferon, β 2)	1.37	7.30	.00146
NM_005586	hMDFI	H. sapiens MyoD family inhibitor (MDFI)	1.21	8.27	.00165
NM_133371	HMYOZ3	H. sapiens myozenin 3	3.86	8.59	.00052
NM_001100818	hPID1	H. sapiens PID1	1.00	9.98	.00200
NM_003942	hRPS6KA4	H. sapiens ribosomal protein S6 kinase A4	1.08	9.27	.00185
NM_004755	hRPS6KA5	H. sapiens ribosomal protein S6 kinase A5	1.01	9.88	.00198
NM_014496	hRPS6KA6	H. sapiens ribosomal protein S6 kinase A6	1.12	8.93	.00179
NM_003161	hRPS6KB1	H. sapiens ribosomal protein S6 kinase B1	1.14	8.77	.00175
NM_012234	HRYBP	H. sapiens RING1 and YY1 binding protein	1.20	8.36	.00167
NM_005900	HSMAD1	H. sapiens SMAD family member 1	1.00	9.99	.00200
NM_005901	HSMAD2	H. sapiens SMAD family member 2	1.15	8.69	.00174
NM_001145102	HSMAD3	H. sapiens SMAD family member3 (SMAD3)	1.08	9.28	.00186
NM_001005291	hSREBF1	H. sapiens sterol regulatory element binding transcription factor 1	1.12	8.97	.00179
NM_007115	hTNFAIP6	H. sapiens TNF α-induced protein 6	3.02	5.31	.00106
NM_003279	HTNNC2	H. sapiens troponin C type 2 (fast)	1.18	8.48	.00170
NM_014408	HTRAPPC3	H. sapiens trafficking protein particle complex 3	1.05	9.54	.00191
NM_016146	HTRAPPC4	H. sapiens trafficking protein particle complex 4	1.01	10.50	.00210
NM_001606	HABCA2	H. sapiens ATP-binding cassette, subfamilyA (ABC1)	2.80	6.30	.00126

Gene symbols, gene name description, and accession numbers together with fold up-regulation and P values in IS and IR myoblasts are given. CREB, cAMP response element-binding protein; SMAD, Sma and Mad related protein. PID, Phosphotyrosine interaction domain-containing protein; IGF, insulin growth factor; TNF, tumor necrosis factor.

Table 2.

List of Genes Down-Regulated in IR Myoblasts Compared with IS Myoblasts

GenBank Accession No.	Gene Symbol	Description	Fold Regulation
			IS	IR	(P < .05)
NM_005544	hIRS1	Homo sapiens IRS1	3.30	0.33	.00047
NM_138714	hNFAT5	H. sapiens nuclear factor of activated T-cells 5, tonicity-responsive	5.18	0.52	.00052
NM_033139	hCALD1	H. sapiens caldesmon 1	2.02	0.20	.00658
NM_017812	hCHCHD3	H. sapiens coiled-coil-helix-coiled-coil-helix domain containing 3	3.15	0.32	.01252
NM_001927	hDES	H. sapiens desmin	5.99	0.60	.01764
NM_000597	hIGFBP2	H. sapiens IGF-binding protein 2	2.09	0.21	.01102
NM_001013398	hIGFBP3	H. sapiens IGF-binding protein 3	6.08	0.61	.02713
NM_001552	hIGFBP4	H. sapiens IGF-binding protein 4	3.34	0.03	.05921
NM_000599	hIGFBP5	H. sapiens IGF-binding protein 5	4.44	0.44	.01799
NM_002178	hIGFBP6	H. sapiens IGF-binding protein 6	11.11	0.25	.09813
NM_003637	hITGA10	H. sapiens integrin, α 10	2.91	0.29	.00127
NM_001001431	hTNNT2	H. sapiens troponin T type 2	1.61	0.16	.00909
NM_001042782	hTNNT3	H. sapiens troponin T type 3	3.15	0.31	.00662
NM_213674	hTPM2	H. sapiens tropomyosin2 (β)	1.91	0.19	.00506
NM_001043353	hTPM3	H. sapiens tropomyosin 3	1.84	0.18	.00685
NM_014402	hUQCRQ	H. sapiens ubiquinol-cytochrome c reductase, complex III subunit VII	1.51	0.15	.01483
NM_002334	hLRP4	H. sapiens low density lipoprotein receptor-related protein 4	1.48	0.15	.00591
NM_004490	hGRB14	H. sapiens growth factor receptor-bound protein 14	1.70	0.17	.16309
NM_130435	hPTPRE	H. sapiens protein tyrosine phosphatase, receptor type, E	1.74	0.17	.03514
NM_005811	hGDF11	H. sapiens growth differentiation factor 11	1.77	0.18	.00010
NM_000557	hGDF5	H. sapiens growth differentiation factor 5	19.37	1.94	.00848
NM_000331	hSAA1	H. sapiens serum amyloid A1	8.48	0.18	.00044
NM_001104	hACTN3	H. sapiens actinin, α 3	8.27	0.21	.00500
NM_004380	hCREBBP	H. sapiens CREB binding protein	2.59	0.86	.00100
NM_001043353	hTPM3	H. sapiens tropomyosin 3	9.98	0.52	.00010
NM_003280	hTNNC1	H. sapiens troponin C type 1	9.27	0.79	.00325
NM_001126133	hTNNT1	H. sapiens troponin T type 1	9.88	0.61	.01000
NM_001492	hGDF1	H. sapiens growth differentiation factor 1	8.93	0.61	.00020
NM_003281	hTNNI1	H. sapiens troponin I type 1	8.77	0.14	.00840
NM_004374	hCOX6C	H. sapiens cytochrome c oxidase subunit VIc	8.36	0.20	.00070

Gene symbols, gene name description, and accession numbers together with fold down-regulation and P values in IS and IR myoblasts are given. CREB, cAMP response element-binding protein. IGF, Insulin growth factor

Identification and tissue distribution of Pid1 isoforms in humans and mice

An in silico search for the protein sequence of Pid1 revealed 4 different isoforms of human PID1 and 2 different isoforms of mouse Pid1 protein (Figure 2A). We noted that the identified isoforms of Pid1 protein differed in amino acid length at the N terminus. The UniProt IDs and amino acid lengths of the different human and mouse Pid1 isoforms are given in Table 3. Subsequent expression analysis revealed that whereas isoform 1 (hIso1) and isoform 3 (hIso3) of human PID1 were expressed at undetectable or very low levels, human PID1 isoform 2 (hIso2) and isoform 4 (hIso4) were readily detectable in both hMb15 myoblast and myotube cultures, with hIso4 the predominantly expressed (Figure 2B). Furthermore, expression analysis also revealed relatively higher expression of both hIso2 and hIso4 in hMb15 myotube cultures when compared with myoblasts (Figure 2B). We next assessed the expression of hIso2 and hIso4 in IS and IR human myoblasts, and results revealed that both hIso2 and hIso4 were significantly up-regulated in IR myoblasts, when compared with IS myoblasts (Figure 2C). Next we analyzed the expression of mouse Pid1 isoform 1 (mIso1) and isoform 2 (mIso2) in mouse myoblast cultures. Both C2C12 myoblasts and primary myoblast cultures, derived from C57BL/6J (wild type) mice, displayed expression of Pid1 mIso1 and mIso2 (Figure 2D). Similar to hIso2 and hIso4, qPCR analysis demonstrated that mIso1 and mIso2 were expressed higher in myotubes, when compared with myoblasts (Figure 2D). In addition, we further characterized the tissue distribution of mouse Pid1 isoforms between different mouse tissues. Subsequent qPCR analysis revealed that mIso1 and mIso2 were highly expressed in spleen, liver, and lung and moderately expressed in brain, white adipose tissue (inguinal and epididymal fat) and brown adipose tissue (Figure 2E). However, the expression of both mIso1 and mIso2 was relatively lower in heart, intestine, kidney, skin, testes, and skeletal muscle (Figure 2E). Further analysis of skeletal muscle fiber type-specific expression revealed that mIso1 and mIso2 were significantly higher in predominantly slow-twitch soleus muscle, when compared with fast or mixed muscle fiber types (Figure 2F).

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.

Table 3.

Human and Mouse PID1/Pid1 Isoforms

UniProt ID	Protein Name	Amino Acid Length	Species
Q7Z2X4-1	PID1 Isoform-1	250	Human
Q7Z2X4-2	PID1 Isoform-2	248	Human
Q7Z2X4-3	PID1 Isoform-3	168	Human
Q7Z2X4-4	PID1 Isoform-4	217	Human
Q3UBG2	Pid1 Isoform-1	217	Mouse
Q3UBG2-2	Pid1 Isoform-2	203	Mouse

Human and mouse PID1/Pid1 isoforms, with their respective amino acid lengths and UniProt IDs are given.

Pid1 is up-regulated in peripheral tissues during diet-induced obesity in mice

In order to assess whether induction of IR alters the expression of Pid1, WT mice were fed with HFD, and qPCR analysis was performed to assess mIso1 and mIso2 mRNA expression in adipose, skeletal muscle, and liver tissues. HFD feeding resulted in the progressive development of IR across time, as measured by both glucose tolerance and insulin tolerance tests (Supplemental Figure 1, A and B, published on The Endocrine Society's Journals Online web site at http://mend.endojournals.org). Concomitant with the development of IR, significantly increased body weight and white adipose tissue weight was observed across time in HFD-fed mice, when compared with CD-fed controls, despite comparable food intake between CD- and HFD-fed mice throughout the diet regimen (Supplemental Figure 1, C–E). As shown in Figure 3A, mIso1 and mIso2 mRNA expression in adipose tissues was increased by 9 weeks of HFD feeding and was maintained throughout the 19-week HFD feeding regimen (Figure 3A). However, in contrast, we only observed significant increase in mIso1 and mIso2 mRNA expression in skeletal muscle from 15 weeks of HFD feeding onward (Figure 3B). In liver tissue we observed a different expression pattern for mIso1 and mIso2 mRNA expression, when compared with that observed in either adipose or skeletal muscle tissue. Specifically, the expression of mIso1 and mIso2 was similarly increased at both 9 and 12 weeks of HFD feeding, when compared with CD-fed controls. However, although still higher than CD-fed controls, the expression of mIso1 and mIso2 sharply declined from 15 weeks of HFD feeding onward (Figure 3C). Consistent with qPCR analysis, we also observed a similar pattern of Pid1 protein levels in adipose, skeletal muscle, and liver tissues across time in mice fed with HFD, when compared with CD-fed mice (Figure 3D). These data suggest that Pid1 expression is temporally regulated during HFD-induced IR and that Pid1 expression is up-regulated early in adipose and liver tissues and later in skeletal muscle in response to HFD feeding in mice. In addition to HFD feeding, we also assessed the changes in Pid1 expression during high-glucose injection in mice. WT mice were subsequently injected with high glucose (2 mg/g BW) for 12 weeks and, as expected, high-glucose feeding resulted in an increase in adipose tissue mass and the development of IR (data not shown). The expression of mIso1 and mIso2, as well as protein levels of Pid1, was subsequently measured in adipose, skeletal muscle, and liver tissues. Consistent with HFD feeding, the mRNA expression of mIso1 and mIso2 and protein levels of Pid1 were significantly up-regulated in adipose (Figure 3, E and H), skeletal muscle (Figure 3, F and H), and liver tissue (Figure 3, G and H) during high glucose-induced obesity and IR.

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 (**).

Myostatin up-regulates Pid1 expression in peripheral tissues

The levels of the myokine myostatin (Mstn) are up-regulated during obesity, IR, and T2D in mouse and humans models (39–41). In agreement with this, absence of or blockade of Mstn leads to reduced adipose tissue content, increased peripheral tissue fatty acid oxidation, and improved insulin sensitivity in mice, even when subjected to chronic HFD feeding (11). Because the expression of both Pid1 and Mstn are increased during obesity and IR, we speculated that Mstn might play a role in promoting the increased expression of Pid1 during obesity and IR. Treatment with CHO-Mstn significantly up-regulated hIso2 and hIso4 mRNA expression levels in hMb15 myotubes (Figure 4A). In addition, mIso1 and mIso2 expression was significantly up-regulated in C2C12 myotubes following treatment with CHO-Mstn (Figure 4B). Moreover, transfection of an Mstn overexpressing construct (pCMV6-Mstn) also resulted in increased mRNA expression of mIso1 and mIso2 (Figure 4C) and elevated Pid1 protein levels (Figure 4D) in 96-hour differentiated C2C12 myoblasts. To further confirm the in vitro results, we analyzed the mRNA expression of mIso1 and mIso2 as well as Pid1 protein levels in mice injected with rhMstn (5 μg/g BW) for 12 weeks. Subsequent qPCR analysis revealed a significant increase in the mRNA expression of both mIso1 and mIso2 in adipose, skeletal muscle, and liver tissue collected from rhMstn protein-injected mice, when compared with saline-injected controls (Figure 4, E–G). In agreement with this, increased levels of Pid1 protein were also detected in adipose, skeletal muscle, and liver tissue collected from rhMstn protein-injected mice, when compared with saline-injected controls (Figure 4H). To further validate Mstn regulation of Pid1, we knocked down Mstn expression in C2C12 myoblasts using Mstn-specific siRNA and assessed the mRNA expression of mIso1 and mIso2 and protein levels of Pid1. Subsequent qPCR confirmed knockdown of Mstn upon transfection of siRNA-2 and siRNA-4 (Figure 4I), because such C2C12 myoblasts were transfected with equal amounts of both siRNA-2 and siRNA-4 for further analysis. Consistent with Mstn regulation of Pid1, we observed a significant reduction in the mRNA expression of mIso1 and mIso2 (Figure 4J), together with reduced protein levels of Pid1 (Figure 4K) in 96-hour differentiated Mstn-specific siRNA-transfected C2C12 myoblasts. In agreement with this, reduced mIso1 and mIso2 expression was noted in gastrocnemius (Gas) muscle collected from Mstn^−/− mice, when compared with WT (Figure 4L). Moreover, reduced Pid1 protein levels were observed in adipose, skeletal muscle, and liver tissue isolated from Mstn^−/− mice, when compared with WT mice (Figure 4M). Taken together, these data clearly demonstrate that Mstn positively regulates Pid1 expression in peripheral tissues.

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.

Myostatin signals through NF-κB to induce Pid1 gene expression in skeletal muscle

To further investigate the mechanism of Mstn activation of Pid1, we performed in silico analysis of the upstream sequence of the human and mouse PID1/Pid1 gene promoter. Subsequent sequence analysis identified putative Sp1, GATA, and sterol regulatory element (STRE) binding sites within both the human and mouse PID1/Pid1 upstream sequences. Furthermore, 2 putative NF-κB binding sites (−0.514-kb and −0.573-kb) in the proximal human PID1 promoter and 1 putative NF-κB binding site (−4229-kb) in the mouse Pid1 promoter were identified (Figure. 5A). Our laboratory has recently shown that Mstn can signal via NF-κB to activate TNF-α expression (23). Therefore, to confirm that Mstn induces PID1 expression via NF-κB, C2C12 myoblasts were transfected with a human 0.978kb PID1 promoter-reporter construct (pLightSwitch-PID1) and subjected to treatment with CHO-Mstn. Treatment of PID1-promoter reporter-transfected C2C12 myoblasts with CHO-Mstn resulted in a dose-dependent increase in PID1 luciferase reporter activity, when compared with CHO-control treated myoblasts (Figure 5B). In agreement with this, siRNA-mediated knockdown of Mstn in PID1-transfected C2C12 myoblasts resulted in reduced PID1 luciferase reporter activity, when compared with CHO-control-treated myoblasts (Figure 5C). In order to confirm that NF-κB plays a role in Mstn regulation of PID1 expression, PID1-transfected C2C12 myoblasts were treated with the NF-κB-specific inhibitor BAY-117082 and subjected to treatment with CHO-Mstn. As shown in Figure 5D, Mstn failed to induce PID1 promoter-reporter activity in the presence of NF-κB-specific inhibitor BAY-117082. To validate that NF-κB binds to the putative NF-κB binding sites in the human PID1 promoter, we performed ChIP analysis. The results revealed binding of NF-κB to the putative NF-κB sites within the human PID1 promoter, which was enhanced via addition of CHO-Mstn and blocked through BAY-117082 treatment (Figure 5E).

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.

Overexpression of PID1 in human myoblasts leads to impaired insulin signaling

Lentiviral-mediated transduction of an EGFP-tagged PID1-overexpressing construct (EGFP-PID1) into hMb15 myoblasts cultures (Figure. 6A) resulted in a dramatic increase in PID1 expression (Figure 6B). Consistent with a role in promoting IR, we observed reduced insulin-stimulated glucose uptake in PID1-overexpressing hMb15 myoblasts, when compared with controls. Moreover the level of glucose uptake in PID1-overexpressing hMb15 myoblasts remained comparable to basal levels despite treatment with increasing concentrations of insulin (Figure 6C). In agreement with this, Western blot analysis revealed a significant reduction in the levels of the insulin signaling intermediates pIRS1, pAkt, and the levels of plasma membrane-associated Glut-4 in PID1-overexpressing hMb15 myoblasts (EGFP-PID1), when compared with control (EGFP-Con) (Figure 6D). Similarly, when we overexpressed PID1 in a second independent human primary myoblast clone (hMb5) we also observed significant reduction in pIRS1, pAkt, and plasma membrane-associated Glut-4 levels, when compared with control (EGFP-Con) (Figure 6E), suggesting a consistent effect of PID1 on insulin signaling in human myoblasts. Quantification of mtDNA content in PID1 overexpressing hMb15 myoblasts (EGFP-PID1) revealed a significant reduction in the mtDNA:nuDNA copy number, when compared with control (EGFP-Con) (see Figure 8A). Consistent with this, we also observed a significant reduction in the levels of mitochondrial fusion genes, mitofusin1 (Mfn1) and mitofusin2 (Mfn2), as assessed through qPCR (see Figure 8B) and Western blot (see Figure 8C) analysis, in PID1-overexpressing myoblasts (EGFP-PID1). In contrast to the expression pattern noted for mitochondrial fusion genes, we observed that the levels of the mitochondrial fission gene Drp1 were up-regulated in PID1-overexpressing hMb15 myoblasts (EGFP-PID1), compared with control cells (EGFP-Con) (see Figure 8, B and C).

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 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.

Knockdown of PID1 in human myoblasts enhances insulin signaling

Lentiviral-mediated transduction of an EGFP-tagged PID1-specific shRNA constructs into hMb15 myoblasts cultures (Figure 7A) resulted in an approximately 80% reduction in PID1 mRNA levels (Figure 7B, left panel) and about 70% reduction in PID1 protein levels (Figure 7B, right panel). Subsequent analysis of glucose uptake revealed that knockdown of PID1 resulted in improved dose-dependent insulin-stimulated glucose uptake in hMb15 myoblasts (Figure 7C). Furthermore, Western blot analysis revealed increased levels of pIRS1, pAkt, and plasma membrane-associated Glut-4 in response to insulin stimulation in PID1 knockdown hMb15 myoblasts (EGFP-shPID1), when compared with control shRNA (EGFP-shCon) expressing myoblasts (Figure 7D). Independently we knocked down PID1 in hMb5 myoblasts and, much like what we observed in hMb15 myoblasts, the levels of pIRS1, pAkt, and plasma membrane-associated Glut-4 were enhanced in response to insulin stimulation, when compared with control shRNA (EGFP-shCon)-expressing myoblasts (Figure 7E). As expected, mtDNA:nuDNA copy number was significantly higher in PID1 knockdown hMb15 myoblasts (EGFP-shPID1), when compared with control (EGFP-shCon) myoblasts (Figure 8D). Furthermore, we also observed a significant increase in the levels of mitochondrial fusion genes Mfn1 and Mfn2 (Figure 8, E and F) in PID1 knockdown myoblasts (EGFP-shPID1). However, PID1 knockdown did not alter the levels of the mitochondrial fission gene Drp1 in hMb15 myoblasts (EGFP-shPID1) (Figure 8, E and F).

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.

Inhibition of PID1 improves insulin signaling during palmitate-, TNF-α-, and Mstn-induced IR in myoblasts

In order to further prove that induction of insulin resistance is associated with increased PID1 expression, we established in vitro models of IR and monitored PID1 gene expression. The saturated fatty acid, palmitate, is known to induce IR in C2C12 myotubes through increased ceramide accumulation (18). In order to understand the relationship between palmitate loading and changes in PID1 gene expression, hMb15 myoblasts were treated with 0.25 mM palmitate, and subsequent qPCR analysis revealed surprising reduction in PID1 expression upon lipid (palmitate) loading (Figure 9A). However, as expected, lipid loading resulted in reduced pIRS1 and pAkt levels in hMb15 myoblasts (Figure 9D). It is important to mention that, during localization studies in the presence or absence of palmitate, we observed increased localization PID1 protein toward the cell membrane (data not shown), suggesting that localization of PID1, as opposed to increased expression of PID1, may play an important role in inhibiting pIRS1 and pAkt signaling in response to lipid loading. Nevertheless we observed a significant increase in PID1 mRNA expression in both TNF-α (Figure 9B)- and CHO-Mstn (Figure 9C)-treated hMb15 myoblasts, concomitant with reduced pIRS1 and pAkt levels (Figure 9, E and F, respectively). Moreover, we noted that treatments with palmitate (Figure 9D)-, TNF-α (Figure 9E)-, or CHO-Mstn (Figure 9F)-mediated reduction in pIRS1 and pAkt levels was blocked in the absence of PID1, suggesting, that PID1 is a critical mediator of palmitate-, CHO-Mstn-, and TNF-α-mediated IR. Consistent with this, knockdown of PID1 in the presence of palmitate (Figure 9D and Supplemental Figure 2A), TNF-α (Figure 9E and Supplemental Figure 2B) or Mstn (Figure 9F and Supplemental Figure 2C) resulted in improved pIRS1 and pAkt levels, suggesting that inhibition of PID1 is able to rescue palmitate-, TNF-α-, and Mstn-mediated IR in human myoblasts.

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.

Discussion

Pid1 is a novel phosphotyrosine-binding domain-containing protein that was initially identified as an obesity-related gene in sc fat tissue from obese individuals. In addition to adipose tissue, Pid1 is also expressed in skeletal muscle; however, whereas the metabolic function of Pid1 in adipocytes is well characterized through either overexpressing or knockdown experiments, the impact and importance of Pid1 in muscle metabolism is less understood. Recent studies have documented that Pid1 has a potential role in inhibiting insulin-stimulated glucose uptake in rodent myotubes. Despite this, the physiologic role of Pid1 during obesity and IR in vivo as well as the function of Pid1 in human skeletal muscle remain to be characterized.

During our molecular screen we noted a significant up-regulation of PID1 expression in insulin resistant (IR) human primary myoblasts derived from insulin resistant individuals. The individuals participating in this study were all overweight (nonobese) with either normal insulin sensitivity or had significant IR (non-T2D) as determined by hyperinsulinemic euglycemic glucose clamp. Hence these results suggest that increased expression of PID1 may not only be associated with obesity but may also be associated with the development of early preobesity-associated IR in human skeletal muscle. In silico characterization of PID1 revealed the presence of 4 different human PID1 isoforms and 2 different mouse Pid1 isoforms in the UniProt database. The isoforms of human (PID1) and mouse (Pid1) were derived from the same gene due to alternative splice variation. Molecular analysis of human PID1 isoform expression in human hMb15 myoblasts and myotubes revealed that PID1 hIso4 is predominantly expressed, hIso2 is moderately expressed, and hIso1 and hIso3 are expressed at low levels in muscle. Taken together, these data suggest that human PID1 hIso2 and hIso4 may have more prominent skeletal muscle-specific functions, when compared with hIso1 and hIso3. Importantly, the expressions of hIso2 and hIso4 were both significantly up-regulated in IR human myoblasts, when compared with IS human myoblasts, further suggesting that both hIso2 and hIso4 may be associated with the development of IR in human skeletal muscle. Analysis of mouse Pid1 mIso1 and mIso2 in C2C12 and mouse primary myoblast cultures derived from WT mice revealed that both mouse Pid1 mIso1 and mIso2 were highly expressed. In addition, expression analysis of mouse Pid1 isoform tissues distribution suggested that Pid1 is ubiquitously expressed. However, a relatively higher level of Pid1 expression was seen in liver, spleen, and lungs and is moderately expressed in heart, white adipose tissue, and brown adipose tissue. Relatively low levels of expression are seen in skeletal muscle, kidney, and intestine. However, the tissues distribution pattern of Pid1 isoform in mouse differs to that observed and reported by Wang et al. (4) in human tissues. This could be explained by the species specificity of Pid1 gene regulation in different tissues (4). Muscle fiber type-specific expression of mouse Pid1 isoforms suggested that the predominantly slow twitch oxidative fibers such as soleus muscle have relatively higher expression of Pid1 mIso1 and mIso2 when compared with fast twitch and mixed fiber type-containing muscles. It is well characterized that overexpression of Pid1 reduces mitochondrial number and affects mitochondria morphology (7). In fact, overexpression of Pid1 led to reduced mitochondrial number and impaired ATP production in 3T3L1 cells (42). In addition overexpression of Pid1 led to increased levels of reactive oxygen species, increased expression of peroxisomal proliferator-activated receptor coactivator-1α and peroxisomal proliferator-activated receptor coactivator -1β, and elevated levels of Mfn1 protein, suggesting that Pid1 overexpression in adipocytes impairs mitochondrial function and oxidative metabolism (7). In agreement with this, we also found reduced mtDNA copy number and reduced levels of mitophagy-associated proteins in Pid1-overexpressing myoblasts. Taken together, these data suggest that Pid1 might play a role in maintaining normal mitochondrial biogenesis and oxidative metabolism in slow twitch muscles, which have high mitochondrial number and generate energy through oxidative phosphorylation. However, further work will need to be performed to confirm this.

Previous in vitro studies have clearly demonstrated that overexpression of Pid1 is associated with obesity and the development of IR. However, the physiologic role of Pid1 during high calorie intake (HFD and high glucose) in vivo is not known. In the current manuscript, molecular studies revealed that Pid1 expression was up-regulated by 9 weeks in adipose tissues of mice fed with HFD and was maintained throughout the 19-week diet regimen. The early up-regulation of Pid1 in adipose tissues of HFD-fed mice may be due to enhanced activation of Pid1 by circulatory free fatty acids and proinflammatory cytokines (TNF-α) released from adipocytes, as reported earlier (43). Furthermore, Pid1 has been previously shown to enhance adipocyte proliferation and hypertrophy (4); therefore, the early increase in Pid1 expression observed during HFD feeding may play a role in regulating the rapid expansion of adipocyte size and number that is normally observed during HFD feeding. It is widely accepted that abdominal fat deposition has a major contribution to the development of fatty liver (44). Consistent with this, we observed significantly increased fat accumulation at 9, 12, 15, and 19 weeks concomitant with comparable up-regulation in Pid1 expression in the liver, similar to that observed in adipose tissues at 9, 12, 15, and 19 weeks. However, it is important to mention that the relative expression of Pid1 tended to decrease in the liver during the latter stages of HFD feeding regimen (15 and 19 weeks). We speculate that this reduction in Pid1 may be due to a compensatory mechanism to improve metabolic defects and IR in the liver during HFD feeding at acute stages of obesity; however, further work is currently underway to study this phenomenon in more detail. In contrast to adipose and liver, Pid1 expression was up-regulated in skeletal muscle during the latter stages (15 and 19 weeks) of HFD regimen. It is noteworthy to mention that palmitate loading, which results in development of IR in myoblasts, also did not significantly up-regulate PID1 expression; however, unpublished preliminary data from our laboratory suggests that palmitate treatment affect the cellular localization of PID1, with a greater abundance of PID1 localized toward the cell membrane in response to lipid loading (data not shown). This phenomenon is consistent with previously published studies in which PID1 protein was shown to be colocalized with LRP4 protein at the perinuclear region in H4 and Nuero2A cells. Therefore, we speculate that localization of PID1 in response to HFD and lipid loading may be more relevant in the development of IR rather than PID1 expression levels in skeletal muscle.

To understand and validate the functions of PID1 in human skeletal muscle further, we overexpressed PID1 in hMb15 and hMb5 human myoblasts. Consistent with previous reports (45), overexpression of PID1 drastically reduced the efficiency of insulin-stimulated glucose uptake in hMb15 myoblasts with an associated reduction in pIRS1, pAkt, and membrane associated Glut-4 protein levels. Furthermore, we also knocked down PID1 expression in hMb15 and hMb5 myoblasts and observed enhanced basal and insulin-stimulated glucose uptake and increased pIRS1, pAkt, and membrane associated Glut-4 protein levels. These results are highly consistent with previous reports highlighting Pid1 as a downstream regulator of the insulin-signaling pathway, resulting in inhibition of glucose uptake in rodent muscle cells and adipocytes (8, 9, 46). Previous reports have shown that several adipokines/cytokines are able to up-regulate Pid1 expression in adipocytes (7). However, in muscle cells the upstream activators of Pid1 are poorly understood. Here we show, for the first time, that the myokine Mstn, inducer of IR, is one of the potent activators of Pid1 in skeletal muscle. Pid1 expression was increased in response to Mstn treatment of C2C12 and hMb15 myotubes in vitro and in response to rhMstn injection into mice. Moreover, absence of or knockdown of Mstn resulted in decreased Pid1 levels. Through in silico analysis we identified the presence of NF-κB-binding sites in the human and mouse PID1 promoter, and subsequent PID1 promoter-reporter studies and ChIP analysis collectively revealed that Mstn regulates PID1 through a mechanism involving NF-κB (Figure 10). Furthermore, in addition to myokines, we have also shown that the cytokine TNF-α can also up-regulate PID1 expression, which is consistent with previously published work (47). More importantly, we now show that inhibition of PID1 can improve both Mstn- or TNF-α-induced reduction of the insulin-signaling intermediates pIRS1 and pAkt.

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.

In summary, here we have shown, for the first time, that PID1 is a marker of early obesity and IR in humans. Expression of Pid1 during high-fat feeding suggested a potential role for Pid1 to regulate systemic obesity and IR. Moreover, activation of Pid1 by Mstn via NF-κB signaling pathway in muscle cells suggests that Pid1 expression can be regulated by myokines (Figure 10). Thus, Pid1 is a novel regulator of early obesity and IR in skeletal muscle.