Growth hormone receptor regulates {beta} cell hyperplasia and glucose-stimulated insulin secretion in obese mice

J Clin Invest. doi:10.1172/JCI45027.
Copyright © 2011, The American Society for Clinical Investigation. Yingjie Wu1, Chengyu Liu2, Hui Sun1, Archana Vijayakumar1, Pejman Raeisi Giglou1, Ruifang Qiao1, Joshua Oppenheimer1, Shoshana Yakar1 and Derek LeRoith1

1Division of Endocrinology, Diabetes and Bone Disease, Department of Medicine, and Department of Oncological Sciences, Mount Sinai School of Medicine, New York, New York, USA.
2Transgenic Core Facility, National Heart, Lung, and Blood Institute, NIH, Bethesda, Maryland, USA.

Address correspondence to: Derek LeRoith or Shoshana Yakar, Endocrinology/Diabetes and Bone Disease, The Mount Sinai School of Medicine, One Gustave L. Levy Pl, Box 1055, Annenberg Building Room 23-66B, New York, New York 10029, USA. Phone: 212.241.6306; Fax: 212.241.4218; E-mail: derek.leroith@mssm.edu (D. LeRoith); shoshana.yakar@mssm.edu (S. Yakar).

Published May 9, 2011
Received for publication September 7, 2010, and accepted in revised form March 16, 2011.

Insulin, growth hormone (GH), and insulin-like growth factor–1 (IGF-1) play key roles in the regulation of ß cell growth and function. Although ß cells express the GH receptor, the direct effects of GH on ß cells remain largely unknown. Here we have employed a rat insulin II promoter–driven (RIP-driven) Cre recombinase to disrupt the GH receptor in ß cells (ßGHRKO). ßGHRKO mice fed a standard chow diet exhibited impaired glucose-stimulated insulin secretion but had no changes in ß cell mass. When challenged with a high-fat diet, ßGHRKO mice showed evidence of a ß cell secretory defect, with further deterioration of glucose homeostasis indicated by their altered glucose tolerance and blunted glucose-stimulated insulin secretion. Interestingly, ßGHRKO mice were impaired in ß cell hyperplasia in response to a high-fat diet, with decreased ß cell proliferation and overall reduced ß cell mass. Therefore, GH receptor plays critical roles in glucose-stimulated insulin secretion and ß cell compensation in response to a high-fat diet.

ß Cell mass changes according to insulin demand, and loss of ß cell hyperplasia in the face of insulin resistance is fundamental to the pathogenesis of type 2 diabetes (1). Although ß cell hypertrophy and neogenesis contribute to enlargement ?f ß cell mass, growing evidence indicates that ß cell hyperplasia is mainly controlled by ß cell proliferation as demonstrated in the cyclin D2–knockout mouse (2). Previous studies have implicated the insulin receptor (IR), but not the IGF-1 receptor (IGF-1R), in the maintenance of ß cell mass and adaptation to high-fat diet (HFD) feeding (3). ß Cell–specific IR-knockout (ßIRKO), IGF-1 receptor–knockout (ßIGFRKO), and double-knockout mutants exhibited normal growth and development of ß cells (4), as did mice with knockout of IR substrate–1 (IRS-1) (5). However, ßIRKO, but not ßIGFRKO, mice exhibited an age-dependent decrease in ß cell mass and eventually developed diabetes (6), suggesting that the IR plays a role in the maintenance of adult ß cell mass. When challenged with HFD, ßIRKO mice showed poor islet compensatory growth as compared with ßIGFRKO and control mice (3), suggesting a major role for the IR in compensatory increases in ß cell mass.

The role of the growth hormone receptor (GHR) in ß cells is not well understood. Previous reports have shown that growth hormone (GH) stimulates insulin gene expression, biosynthesis, and release in ß cells of rodents and humans. The first physiological evidence for a role of GHR in ß cells came from GHR- (7) and prolactin receptor–null (PRLR-null) (8) mice, which exhibited reduced ß cell mass, impaired glucose tolerance, and increased insulin sensitivity. However, these mice displayed compromised growth and significant changes in body adiposity; thus, a direct causal effect could not be established. To resolve these systemic effects, Lee et al. specifically ablated the downstream mediator of GH, STAT5a/b, in ß cells using the Cre transgene (9). Pdx-Cre-Stat5fl/fl mice developed functional islets, which suggested that STAT5 is not essential for ß cell proliferation or function. Nonetheless, with age Pdx-Cre-Stat5fl/fl mice developed mild glucose intolerance, probably due to increased body adiposity. In contrast, RIP-Cre-Stat5fl/fl mice exhibited mild obesity, hyperglycemia, and glucose intolerance, and it was suggested that these were secondary to partial Stat5 gene ablation in the hypothalamus. Similar findings were demonstrated by a dominant negative form of STAT5 under the RIP promoter (10), while expression of constitutively active STAT5 counteracted these effects (10). Altogether, STAT5 may have negligible effects on ß cell mass or function under normal conditions, but during obesity it may play a role in preserving ß cell mass and function. Nonetheless, STAT5 can also be activated by other stimuli, such as IL-2, IL-3, the type 1 interferon receptor, and leptin (11); thus, these studies merely allude to the role of GHR in ß cells. Furthermore, the effects of GH on lipid accumulation in ß cells and its adverse effects of lipids during obesity may not be mediated solely via STAT5. Thus, we took a direct approach, inactivating GHR in ß cells to unequivocally dissect the role of GHR in determining ß cell mass and function during normal and pathophysiological conditions.

Generation of ß cell–specific GHR KO (ßGHRKO) mice. Specific Ghr inactivation in ß cells was achieved using the Cre/loxP system (Supplemental Figure 1; supplemental material available online with this article; doi: 10.1172/JCI45027DS1). Specificity of Ghr gene recombination was validated by PCR using genomic DNA extracted from islets dissected by laser capture microdissection (Figure 1A) and other tissues by standard protocols. In control mice, homozygous for the floxed Ghr allele, and RIP-Cre transgenic mice, the recombinant allele was undetectable in all tissues (Figure 1B), while in ß cell–specific GHR–knockout (ßGHRKO) mice, the Ghr recombinant allele (null allele) was detected only in islets and whole pancreas, but not in exocrine pancreas (Exo), hypothalamus, spleen, or liver, indicating ß cell–specific Cre-mediated excision of the Ghr gene (Figure 1B).

Figure 1 Assessment of Ghr excision and expression in ßGHRKO mice. (A) Isolation of the islets from pancreas by laser capture microdissection. Original magnification, ×10. (B) PCR analysis of GHR excision in different tissues of ßGHRKO mice using primers S6 and R2 (see Supplemental Figure 1A). LL, mice homozygous for the floxed Ghr allele; LLc, ßGHRKO mice; WWc, RIP-Cre transgenic mice; I, islet; Exo, exocrine; P, pancreas; H, hypothalamus; S, spleen; L, liver. “L” arrowhead indicates the floxed allele, “W” indicates the wild-type allele, and “–” indicates the recombinant allele. (C) Immunofluorescence analysis of GHR protein (red) and insulin (green) in pancreatic sections. (D) Immunofluorecence analysis of Cre (red) and insulin (green) in pancreatic sections. (E) Gene expression of Ghr and the Cre transgene in islets analyzed by real-time PCR. *P < 0.05. (F) Phosphorylation of STAT5 in liver upon intravenous injection of GH to obese control and ßGHRKO mice, as assessed by Western immunoblotting. (G) Immunofluorecence analysis of phosphorylated STAT5 protein in pancreatic sections from obese mice (insulin [INS], red; pSTAT5, green; and nucleus, blue). Scale bars: 200 µm (C, D, and G).

Immunostaining of whole pancreas with anti-GHR antibody was positive in control mice, while no staining of GHR was detected in islets of ßGHRKO mice (Figure 1C). Furthermore, immunostaining with anti-Cre recombinase antibody (Figure 1D) was positive in islets of ßGHRKO mice but not in controls. Expression of the Ghr gene according to real-time PCR (Figure 1E) decreased by 95% in ßGHRKO as compared with controls. Additionally, we could not find evidence for Ghr gene recombination in the hypothalamus, as no differences were found in the immunostaining pattern of GHR, CRE, NPY, POMC, or AGRP (Supplemental Figure 2, A and B), nor in their expression levels assessed by real-time PCR (Supplemental Figure 2, C–E).

Last, we tested STAT5 phosphorylation and nuclear localization in liver and pancreas, respectively. GH injected (12.5 µg/100 g body weight) intravenously stimulated STAT5 phosphorylation in liver of both control and ßGHRKO mice (Figure 1F). However, while in control islets STAT5 translocated into the nucleus following GH injection, in ßGHRKO mice we hardly detected nuclear STAT5 staining (Figure 1G). Together, our results show that GHR action in ßGHRKO mice is ablated in ß cells but is intact in other tissues.

Ghr inactivation in ß cells does not affect islet size or insulin content in mice fed chow. Male (Supplemental Figure 3A) and female (Supplemental Figure 3B) mice were followed from 2 to 15 weeks of age. Body weight and body composition (assessed by MRI) of ßGHRKO mice were indistinguishable from those of controls throughout the study. Serum IGF-1 and insulin levels tested at 8 and 16 weeks were similar between the groups (Supplemental Figure 3, C and D), indicating that Ghr inactivation in ß cells does not affect growth and development.

Overexpression of GH in mice resulted in increased ß cell mass (12), while Ghr-null mice (7) as well as GH antagonist (13) transgenic mice exhibited significant reductions in islet size and number. In vivo experiments with rats bearing GH-secreting tumors (14) also showed ß cell hyperplasia, as seen in patients with acromegaly (15). We therefore sought to determine whether ß cell ablation of GHR affected islet size. Sections throughout the pancreatic head, body, and tail revealed no differences in H&E staining (Supplemental Figure 3, E and F) or in insulin staining (Supplemental Figure 3G) between controls and ßGHRKO mice at 16 weeks. This was in accordance with similar levels of insulin in serum of control and ßGHRKO mice (Supplemental Figure 3D) and islet insulin content measured in islet extracts by RIA (data not show), suggesting that loss of Ghr in ß cells does not affect islet development.

The metabolic consequences of GHR ablation in ß cells were determined by intraperitoneal glucose tolerance and insulin tolerance tests (GTT and ITT) in males at 16 weeks, which revealed no differences between ßGHRKO and control mice (Supplemental Figure 3, H and I). These data are in agreement with data obtained from mice expressing a dominant negative form of STAT5b (10), or mice with ß cell–specific ablation of STAT5b (9) fed chow. However, we found that the first phase of insulin secretion, observed shortly following glucose bolus, was blunted in ßGHRKO mice (Supplemental Figure 3J), suggesting that when mice are challenged with high glucose, maximal insulin secretion is low, while insulin secretion at a more physiologic (meal) glucose level is normal. In contrast, arginine-stimulated insulin release from the islets was normal in ßGHRKO mice (Supplemental Figure 3K), indicating that GHR in ß cell is involved in glucose-stimulated first-phase insulin secretion.

ßGHRKO mice exhibit impaired ß cell hyperplasia when fed HFD. ß Cell hyperplasia is compromised in ßIRKO mice when fed HFD (4). To investigate whether ßGHRKO mice manifest ß cell hyperplasia, we fed mice with HFD from weaning. Both control and ßGHRKO mice fed HFD exhibited a similarly marked increase in body weight and body adiposity (Figure 2A). GTT after 24 weeks of HFD showed that both obese control and ßGHRKO mice became glucose intolerant (Figure 2B), but obese ßGHRKO mice showed significantly higher glucose levels. ITT indicated that obese ßGHRKO mice responded to exogenous insulin in the same manner as controls (data not show). Obese ßGHRKO mice (40 weeks on HFD) exhibited severely blunted first-phase glucose-stimulated insulin secretion (GSIS) (Figure 2C) but showed normal arginine-stimulated insulin secretion (Figure 2D). While ß cell mass in obese control mice markedly increased in response to high-fat feeding, obese ßGHRKO mice showed no increase in ß cell mass, resulting in an about a 2.5-fold decrease in mean ß cell mass (Figure 2, E and F), suggesting an impaired ability to adapt to the higher insulin demand in response to a HFD. Interestingly, islet insulin content did not differ between obese control and obese ßGHRKO mice (Supplemental Figure 4), suggesting an insulin secretion defect in ßGHRKO mice. Ex vivo GSIS in isolated islets revealed increased insulin secretion in response to elevations in glucose concentrations in both groups (Figure 2G). However, when corrected to islet DNA, islets isolated from obese ßGHRKO mice secreted significantly less insulin than those isolated from obese controls (Figure 2G), while arginine-stimulated insulin secretion was not affected (Figure 2H). To verify the integrity and activity of the KATP channels in islets, we studied GSIS in the presence of KCl or glibenclamide. We found that islets from obese controls and ßGHRKO significantly increased insulin secretion in response to both stimuli, suggesting that both channels are intact.

Figure 2 HFD-induced obesity causes impaired glucose tolerance and insulin secretion in ßGHRKO mice. (A) Body weight of male control and ßGHRKO mice fed regular chow (RC) or HFD. (B) GTT in male mice after 24 weeks on HFD. ßGHRKO mice did not exhibit first-phase GSIS (C), but displayed normal first-phase arginine-stimulated (Arg-stimulated) insulin secretion (D). Numbers in parentheses indicate sample size. (E) HFD-induced islet hyperplasia is shown by immunostaining using anti-insulin antibody. Scale bars: 200 µm. (F) ß Cell mass was quantified in insulin-stained pancreas sections. (G) GSIS in isolated islets from obese mice in response to different glucose (Glu) concentrations and in the presence of KCl or glibenclamide (Glib). (H) Arginine-simulated insulin secretion in isolated islets from obese mice. *P < 0.05.

Obese ßGHRKO mice exhibit decreased ß cell proliferation. Islets from obese ßGHRKO mice exhibited significant reductions in cellular proliferation, as evident by significant decrease in Ki67- (Figure 3, A and B) or PCNA-positive (data not shown) cells. This correlated with a decreased percentage of cyclin D2–positive cells in islets of obese ßGHRKO mice (Figure 3, C and D). Cyclin D2 gene expression, assessed by real-time PCR, was also decreased in isolated islets of obese ßGHRKO mice (Figure 3E). Additionally, insulin, glucagon, Glut2, caspase-3, P16, or P27 gene expression revealed no differences between islets from obese controls or ßGHRKO mice (data not shown). We speculate that the defect in ß cell hyperplasia in obese ßGHRKO mice is due to decreased proliferation, which may partially be cyclin D2 mediated. Interestingly, however, data from mice expressing a dominant negative STAT5b (dnSTAT5b) specifically in ß cells (10) show that pancreatic insulin content and the relative ß cell mass significantly increased in response to high-fat feeding. Nonetheless, similar to our findings, ß cell replication was lower and associated with decreased islet expression of cyclin D2 in obese dnSTAT5b mice (10). This may suggest that ß cell compensatory response to HFD is not mediated by STAT5b. In support of this, studies with rat ß cells have shown that activation of GHR results in increased PKC activity, leading to elevations in Ca2+ and concomitant increases in diacylglycerol (DAG) that together play a role in transmitting the mitogenic effects of GH into a proliferative response (16). Other studies with rat ß cells have shown that upon stimulation with GH, there was a rapid stimulation of JAK2 and Src protein phosphorylation and a rise in intracellular Ca2+, which is required for cell replication (17).

Figure 3 Reduced ß cell proliferation in obese ßGHRKO mice. (A) Immunofluorecence analysis of Ki67 in pancreatic sections. Ki67 is stained green, GLUT2 red, and the nucleus blue. (B) Quantification of Ki67-positive cells per 1,000 cells based on 20–25 islets in each group. (C) Immunohistochemical staining for cyclin D2 in pancreas sections from obese control and ßGHRKO mice. Top and bottom rows each show representative images of the two groups. (D) Quantification of nuclear cyclin D2 staining, with determination of strong nuclear staining (Brown+++) and weak nuclear staining (Brown+). (E–I) Gene expression of cyclin D2 (Cnnd2), Gck, Irs2, Kir6.2, and Sur1 in isolated islets from obese control and ß-GHRKO mice. Scale bars: 200 µm (A and C). Numbers in parentheses indicate sample size. *P < 0.05.

To begin to understand the molecular mechanism involved in impaired insulin secretion, we performed gene expression studies in isolated islets. The expression levels of glucokinase (Gck), which plays a key role in glucose metabolism, were reduced in islets of obese ßGHRKO mice (Figure 3F). Previous studies showed that IRS-2 is a crucial regulator of ß cell survival and function (18). Thus, ß cell–specific Irs2 gene inactivation resulted in reduced ß cell mass and GSIS (19). Accordingly, we found that the expression level of the Irs2 was significantly reduced in islets of obese ßGHRKO mice (Figure 3G). Immunostaining with anti-GCK and -IRS2 antibodies showed significant reductions in fluorescent signal in ßGHRKO islets (Supplemental Figure 5). Our results are consistent with the phenotype of Gck+/– mice, which showed insufficient ß cell hyperplasia when fed HFD (20). Overexpression of IRS-2 in ß cells partially rescued the diabetic phenotype of HFD-fed Gck+/– mice, indicating that IRS-2 is the downstream mediator of glucose signaling and ß cell replication. Last, expression levels of the SUR/KIR KATP channels, which are crucial for the regulation of glucose-induced insulin secretion and are the target for the sulfonylureas, were similar in obese control and ßGHRKO mice (Figure 3, H and I).

In conclusion, deletion of GHR specifically in pancreatic islet ß cells was associated with a lack of compensatory hyperplasia in response to HFD-induced obesity. Furthermore, GHR is apparently important for GSIS from the ß cells. While the exact mechanism(s) involved are undefined, results from the present study suggest that GHR signaling for these important processes is not solely dependent on STAT5.

Animals. All mice were on the C57BL/6 genetic background. All experimental procedures were approved by the Institutional Animal Care and Use Committee of Mount Sinai School of Medicine (New York, New York, USA). Generation of Ghr-floxed mice and the Ghr ß cell–specific knockout (ßGHRKO) mice is detailed in Supplemental Methods. HFD, 60% (wt/wt) fat content, was obtained from Research Diets Inc.

Intraperitoneal glucose and insulin tolerance tests. GTTs and ITTs were performed as detailed previously (21).

Serum insulin and IGF-1. Serum insulin and IGF-1 levels were determined using RIA (Millipore).

Glucose- and arginine-stimulated insulin secretion. Glucose (3 g/kg) or l-arginine (15 mM/l in 0.2 M PBS) was injected intraperitoneally to overnight fasted mice. Blood glucose was measured at the indicated time points. Ex vivo GSIS and arginine-stimulated insulin secretion were done using a modified protocol detailed in Supplemental Methods.

Histology and immunohistochemistry. Histology and immunohistochemistry were performed in paraffin-embedded tissue sections. Antibody description is provided in Supplemental Methods. ß Cell mass was calculated from 3–5 pancreatic sections per mouse stained with anti-insulin antibody. Relative ß cell area was calculated using NIH ImageJ software, and ß cell mass was calculated as ß cell area multiplied by pancreatic weight of 6 mice per group.

Gene expression. RNA was isolated using TRIzol (Invitrogen). RNA integrity was verified using Bioanalyzer (Agilent Technologies). cDNA was generated using oligo(dT) primers (Invitrogen). Real-time PCR was performed with the QuantiTect SYBR Green PCR kit (QIAGEN) in ABI PRISM 7900HT detection systems (Applied Biosystems). Sequences of primers used for real-time PCR are presented in Supplemental Table 1.

Statistics. All data are expressed as mean ± SEM and were analyzed using unpaired 2-tailed Student’s t test. A P value less than 0.05 was considered significant.

View Supplemental data

We thank Chunxin Wang and Xiaoli Chen for helpful discussion; Pedro L. Herrera and Domenico Accili at Columbia University for providing the RIP-Cre mice; and Andrew Huang, Amy Wu, and Laya Rajan for their help in quantification analysis.

Conflict of interest: The authors have declared that no conflict of interest exists.

Citation for this article: J Clin Invest doi:10.1172/JCI45027.

Weir GC, Bonner-Weir S. A dominant role for glucose in beta cell compensation of insulin resistance. J Clin Invest. 2007;117(1):81–83. Georgia S, et al. Cyclin D2 is essential for the compensatory beta-cell hyperplastic response to insulin resistance in rodents. Diabetes. 2010;59(4):987–996. Okada T, et al. Insulin receptors in beta-cells are critical for islet compensatory growth response to insulin resistance. Proc Natl Acad Sci U S A. 2007;104(21):8977–8982. Ueki K, et al. Total insulin and IGF-I resistance in pancreatic beta cells causes overt diabetes. Nat Genet. 2006;38(5):583–588. Kulkarni RN. Receptors for insulin and insulin-like growth factor-1 and insulin receptor substrate-1 mediate pathways that regulate islet function. Biochem Soc Trans. 2002;30(2):317–322. Otani K, et al. Reduced beta-cell mass and altered glucose sensing impair insulin-secretory function in betaIRKO mice. Am J Physiol Endocrinol Metab. 2004;286(1):E41–E49. Liu JL, et al. Disruption of growth hormone receptor gene causes diminished pancreatic islet size and increased insulin sensitivity in mice. Am J Physiol Endocrinol Metab. 2004;287(3):E405–E413. Freemark M, et al. Targeted deletion of the PRL receptor: effects on islet development, insulin production, and glucose tolerance. Endocrinology. 2002;143(4):1378–1385. Lee JY, Gavrilova O, Davani B, Na R, Robinson GW, Hennighausen L. The transcription factors Stat5a/b are not required for islet development but modulate pancreatic beta-cell physiology upon aging. Biochim Biophys Acta. 2007;1773(9):1455–1461. Jackerott M, et al. STAT5 activity in pancreatic beta-cells influences the severity of diabetes in animal models of type 1 and 2 diabetes. Diabetes. 2006;55(10):2705–2712. Hennighausen L, Robinson GW. Interpretation of cytokine signaling through the transcription factors STAT5A and STAT5B. Genes Dev. 2008;22(6):711–721. Parsons JA, Bartke A, Sorenson RL. Number and size of islets of Langerhans in pregnant, human growth hormone-expressing transgenic, and pituitary dwarf mice: effect of lactogenic hormones. Endocrinology. 1995;136(5):2013–2021. Chen NY, et al. Effects of streptozotocin treatment in growth hormone (GH) and GH antagonist transgenic mice. Endocrinology. 1995;136(2):660–667. Parsons JA, Hartfel MA, Hegre OD, McEvoy RC. Effect of MtTW15 mammosomatotropic tumors on pancreatic islet hormones. Diabetes. 1983;32(1):67–74. Hellman B, Angervall L. The frequency distribution of the number and volume of the islets of Langerhans in man. 3. Studies in diabetes of early onset, insuloma and acromegaly. Acta Pathol Microbiol Scand. 1961;53:230–236. Sjoholm A, et al. Rapid Ca2+ influx and diacylglycerol synthesis in growth hormone-mediated islet beta-cell mitogenesis. J Biol Chem. 2000;275(28):21033–21040. Berlanga JJ, Fresno Vara JA, Martin-Perez J, Garcia-Ruiz JP. Prolactin receptor is associated with c-src kinase in rat liver. Mol Endocrinol. 1995;9(11):1461–1467. Withers DJ, et al. Disruption of IRS-2 causes type 2 diabetes in mice. Nature. 1998;391(6670):900–904. Cantley J, et al. Pancreatic deletion of insulin receptor substrate 2 reduces beta and alpha cell mass and impairs glucose homeostasis in mice. Diabetologia. 2007;50(6):1248–1256. Terauchi Y, et al. Glucokinase and IRS-2 are required for compensatory beta cell hyperplasia in response to high-fat diet-induced insulin resistance. J Clin Invest. 2007;117(1):246–257. Wu Y, Sun H, Yakar S, LeRoith D. Elevated levels of insulin-like growth factor (IGF)-I in serum rescue the severe growth retardation of IGF-I null mice. Endocrinology. 2009;150(9):4395–4403.

 

Mitochondrial dysfunction in patients with primary congenital insulin resistance

J Clin Invest. doi:10.1172/JCI46405.
Copyright © 2011, The American Society for Clinical Investigation. Alison Sleigh1, Philippa Raymond-Barker2, Kerrie Thackray3, David Porter4, Mensud Hatunic3, Alessandra Vottero5, Christine Burren6, Catherine Mitchell7, Martin McIntyre8, Soren Brage9, T. Adrian Carpenter1, Peter R. Murgatroyd2, Kevin M. Brindle10, Graham J. Kemp11, Stephen O’Rahilly3, Robert K. Semple3 and David B. Savage3

1Wolfson Brain Imaging Centre, University of Cambridge, Cambridge, United Kingdom.
2Wellcome Trust Clinical Research Facility (WTCRF), Addenbrooke’s Hospital, Cambridge, United Kingdom.
3Metabolic Research Laboratories, Institute of Metabolic Science, University of Cambridge, Cambridge, United Kingdom.
4Siemens AG Healthcare Sector, Erlangen, Germany.
5Department of Pediatrics, University of Parma, Parma, Italy.
6University Hospitals Bristol National Health Service Trust, Bristol, United Kingdom.
7Hillingdon Hospital, Hillingdon, United Kingdom.
8Royal Alexandra Hospital, Paisley, United Kingdom.
9Medical Research Council Epidemiology Unit, Institute of Metabolic Science, University of Cambridge, Cambridge, United Kingdom.
10Cancer Research UK Cambridge Research Institute, Li Ka Shing Centre, Cambridge, United Kingdom.
11Department of Musculoskeletal Biology and Magnetic Resonance and Image Analysis Research Centre, University of Liverpool, Liverpool, United Kingdom.

Address correspondence to: D.B. Savage, Metabolic Research Laboratories, Institute of Metabolic Science, University of Cambridge, Addenbrooke’s Hospital, Hills Road, Cambridge CB2 0QQ, United Kingdom. Phone: 44.1223.767.923; Fax: 44.1223.330.598; E-mail: dbs23@medschl.cam.ac.uk.

Authorship note: Robert K. Semple and David B. Savage contributed equally to this work.

Published May 9, 2011
Received for publication January 14, 2011, and accepted in revised form March 23, 2011.

Mitochondrial dysfunction is associated with insulin resistance and type 2 diabetes. It has thus been suggested that primary and/or genetic abnormalities in mitochondrial function may lead to accumulation of toxic lipid species in muscle and elsewhere, impairing insulin action on glucose metabolism. Alternatively, however, defects in insulin signaling may be primary events that result in mitochondrial dysfunction, or there may be a bidirectional relationship between these phenomena. To investigate this, we examined mitochondrial function in patients with genetic defects in insulin receptor (INSR) signaling. We found that phosphocreatine recovery after exercise, a measure of skeletal muscle mitochondrial function in vivo, was significantly slowed in patients with INSR mutations compared with that in healthy age-, fitness-, and BMI-matched controls. These findings suggest that defective insulin signaling may promote mitochondrial dysfunction. Furthermore, consistent with previous studies of mouse models of mitochondrial dysfunction, basal and sleeping metabolic rates were both significantly increased in genetically insulin-resistant patients, perhaps because mitochondrial dysfunction necessitates increased nutrient oxidation in order to maintain cellular energy levels.

Insulin resistance underpins the tight association between the type 2 diabetes and obesity pandemics. Yet despite enormous scientific endeavor, understanding of the molecular pathogenesis of insulin resistance remains incomplete. One of the most characteristic and consistent metabolic features of the obese insulin-resistant state is lipid accumulation in sites other than white adipose tissue, so-called ectopic fat (1). Triglyceride accumulation in the liver and skeletal muscle is strongly associated with insulin resistance in these tissues, and while triglyceride is not itself thought to cause insulin resistance, more reactive lipid species, such as diacylglycerol and ceramide, are believed to impair insulin action (1).

Ectopic fat accumulation presumably reflects a cellular mismatch between the sum of lipid delivery and synthesis and the sum of lipid oxidation and disposal. A combination of recent human and rodent studies have convincingly documented mitochondrial abnormalities in insulin-resistant and diabetic states, highlighting the potential importance of impaired mitochondrial fat oxidation in the accumulation of ectopic fat and giving rise to the suggestion that mitochondrial dysfunction may be the primary defect in prevalent obesity-related insulin resistance (reviewed in ref. 2). These studies include ex vivo morphological and biochemical analyses of tissue samples and/or isolated mitochondria (3–6) as well as noninvasive in vivo magnetic resonance spectroscopy (MRS) measures (7–13). However, although the associations appear robust, the direction of causality remains uncertain (2, 14).

Where human genetic variants are identified that directly produce 1 out of 2 associated phenomena, investigating whether those same genetic variants are also associated with the second phenomenon offers a powerful means of investigating causality in the association. Such Mendelian randomization is now increasingly used in large human epidemiological genetic studies (15); however, the principle is also applicable to rare monogenic disorders. A classic early example of this was the association observed between rare mutations in the LDLR gene, which produce high LDL cholesterol, and premature atherosclerotic disease (16), establishing that high levels of LDL cholesterol are causally linked to atherosclerosis.

Studying patients with mitochondrial genetic defects producing primary mitochondrial dysfunction seems at first sight to offer the potential to conduct a similar Mendelian randomization experiment to test whether mitochondrial dysfunction in humans can produce insulin resistance, but this is likely to be confounded by deleterious effects of generalized mitochondrial dysfunction on ß cell and muscle function (17, 18). However, patients with congenital, severe insulin resistance due to mutations in the insulin receptor (INSR) gene are well described and, though uncommon, afford instead the reciprocal opportunity to test whether primary insulin resistance can produce mitochondrial dysfunction, potentially accounting for some but not necessarily all of their observed association in commoner forms of insulin resistance.

Characteristics of the participants. Seven patients from 4 unrelated families with dominant-negative heterozygous missense mutations (P1236A, A1135E, M1138K, and A1121P) affecting highly conserved residues within the tyrosine kinase domain of the insulin receptor were selected for mitochondrial function studies. Such mutations are well known to produce severe insulin resistance with an autosomal dominant pattern of inheritance (19). All patients had acanthosis nigricans, a cutaneous marker of severe insulin resistance, and the typical biochemical profile of patients with primary insulin receptor defects (“insulin receptoropathies”), including severe hyperinsulinemia and normal triglyceride and HDL cholesterol levels, usually with preserved or elevated serum adiponectin levels (Supplemental Table 1; supplemental material available online with this article; doi: 10.1172/JCI46405DS1). Glycated hemoglobin levels were normal in 4 INSR participants and only minimally elevated in the other 3 (Supplemental Table 1). Intramyocellular lipid levels were similar to those of matched controls in 3 out of 4 INSR mutation carriers (Supplemental Table 1), which is in keeping with the similarly normal hepatic fat levels previously documented in this disorder (20).

As well as being age-, gender-, and BMI-matched (Table 1), control volunteers (n = 12) were deliberately selected for being sedentary, as none of the patients with INSR mutations undertook regular exercise. Fasting biochemical parameters were all within normal limits in the healthy controls (Table 1). All patients and controls wore an Actiheart monitor (CamNtech) and underwent a standard graded exercise calibration test, from which we derived an estimate of maximal oxygen consumption (VO2 max), measured in ml/kg of fat-free mass/min. Results from these estimates were similar in the 2 groups (Table 1).

Table 1 Characteristics of the healthy volunteers and patients with severe insulin resistance due to loss-of-function INSR mutations

Assessment of oxidative phosphorylation function in vivo. The theory of measuring chemical exchange rates using nuclear magnetic resonance magnetization transfer techniques was first introduced by Forsen and Hoffman in 1963 (21). Its application to measure ATP synthesis rates in muscle in vivo has been widely used despite concerns that the measurement will contain a nonoxidative, glycolytic component that could be as large as 80% at rest and that resting ATP turnover may not be the most relevant measure of mitochondrial function (22–27). An alternative approach is to assess the kinetics of replenishment of the phosphocreatine (PCr) pool after exercise, which relies purely on oxidative ATP synthesis (28). This is most conveniently quantified as the PCr recovery half time (t1/2), defined as the time taken for PCr to recover by half the amount it was depleted, which is inversely proportional to functional “mitochondrial capacity.” A recent study in which rats were treated with the mitochondrial complex 1 inhibitor diphenyleneiodonium highlighted reservations about measuring ATP turnover using magnetization (saturation) transfer techniques by documenting a significant slowing of PCr recovery but no change in the rate of ATP synthesis determined from saturation transfer (ST) measurements (26). Given these concerns, we determined both resting ST and postexercise PCr recovery in this study.

All participants completed the MRS scans, apart from 2 INSR subjects who did not undertake the ST measurement due to claustrophobia. The rate of ATP synthesis measured with the ST technique (ST VATP) was similar in both groups (Figure 1, A and B, and Table 1), whereas the rate of PCr recovery after exercise was significantly slower in patients with severe insulin resistance due to INSR mutations than that in healthy controls (Figure 1, C and D, and Table 1). This difference remained significant after correcting for minor differences in VO2 max (Table 1). There was no correlation between ST VATP and the postexercise PCr recovery half time (t1/2) (Spearman’s rho, r = –0.235, P = 0.363).

Figure 1 31P MRS measurements of mitochondrial function. (A) Representative ST spectra, with saturation of the ?-ATP resonance (right, bottom) and corresponding control spectrum (right, top). The 2 spectra are superimposed (left) to show the difference (?) in the Pi resonance. (B) ST VATP in both the controls (white bars; n = 12) and in patients with INSR mutations (black bars; n = 5). (C) Mean fractional PCr recovery curves for controls (gray squares; n = 12) and for patients with INSR mutations (black circles; n = 7). Five spectra were averaged to give a time resolution of 10 seconds for clarity in this figure. The monoexponential fit of the mean recovery rate constant is shown for controls (gray line) and INSR patients (black line). (D) Half time for PCr recovery (t1/2) as measured from the recovery rate after exercise for both controls (white bars; n = 12) and for patients with INSR mutations (black bars; n = 7). In B–D, data are mean ± SEM.

The significantly slowed rate of PCr recovery after exercise in the INSR patients shows that insulin resistance due to a well-defined primary defect in insulin signaling is associated with evidence of mitochondrial dysfunction in vivo. This suggests that the association between mitochondrial dysfunction and insulin resistance reported in prevalent forms of insulin resistance of unknown etiology cannot be assumed to imply that mitochondrial dysfunction causes insulin resistance. Our data are consistent with the mitochondrial dysfunction reported in insulin-deficient patients with type 1 diabetes (29) and several recent murine studies reporting mitochondrial dysfunction in mice with either primary genetic defects in the insulin signaling cascade (30–32) or high-fat feeding–induced insulin resistance, in which the insulin resistance was shown to precede mitochondrial dysfunction (33).

Metabolic rate measurements. Changes in body weight reflect a mismatch between energy intake and energy expenditure. Although human genetic studies increasingly suggest that changes in energy intake are the primary driver of weight changes in most cases in humans, changes in metabolic rate can also lead to weight gain or weight loss. For example, thyrotoxicosis promotes mitochondrial uncoupling and ultimately weight loss (34). As well as being a major determinant of activity-associated energy expenditure, skeletal muscle is also a significant contributor (~20%–30%) to resting energy expenditure (35). In order to determine whether insulin resistance and/or mitochondrial dysfunction might alter metabolic rate, we evaluated basal metabolic rate (BMR) and sleeping metabolic rate (SMR) in our patients with INSR mutations and in the healthy controls. Surprisingly, we found that both the BMR and SMR were significantly increased in patients with INSR mutations when corrected for fat-free mass (Figure 2). While both these independent measures of metabolic rate reflect whole body rather than just skeletal muscle energy metabolism, they suggest that reduced mitochondrial function need not lead to a reduction in resting energy expenditure.

Figure 2 Energy expenditure in patients with loss-of-function INSR mutations. BMR and SMR in healthy controls (white bars; n = 11) and in patients with INSR mutations (black bars; n = 7). Results are expressed per kilogram of fat-free mass (FFM). Data are mean ± SEM.

These human observations contradict the intuitive expectation that defects in mitochondrial energy metabolism will lead to reduced energy expenditure. Nevertheless, the data are consistent with observations in mice with primary defects in mitochondrial function, in which the resultant increase in cellular AMP levels activate AMP kinase, producing a lean insulin-sensitive phenotype (36). The primary determinant of the rate of oxidative phosphorylation is thought to be the homeostatic need to defend the cellular energy charge, so one of a number of possible explanations for these observations is that mitochondrial dysfunction necessitates increased nutrient oxidation in order to maintain cellular energy levels.

Germline dominant-negative defects in the insulin receptor gene may confidently be assumed to produce congenital, severe insulin resistance, and co-inheritance of a primary mitochondrial defect in all the unrelated patients studied is vanishingly unlikely. Thus the association of genetic defects in insulin receptor function with impaired oxidative phosphorylation in vivo may be taken to establish that primary insulin resistance can produce secondary mitochondrial dysfunction. Such an observation in one group with a rare monogenic form of insulin resistance is not necessarily directly transposable to the situation in prevalent obesity-related insulin resistance, which most likely encompasses a heterogeneous group of postreceptor defects. Nevertheless there is currently no evidence that our findings are not generalizable to prevalent insulin resistance, and, even viewed most conservatively, they demonstrate that the association between prevalent insulin resistance and mitochondrial dysfunction must not be assumed to be solely accounted for by a unidirectional effect of primary mitochondrial dysfunction on insulin sensitivity.

Participants. Each participant provided written informed consent, and all studies were conducted in accordance with the principles of the Declaration of Helsinki. Clinical studies were approved by the National Health Service Research Ethics Committee, United Kingdom, and were conducted in the WTCRF.

Patients with severe insulin resistance due to loss-of-function mutations in the INSR gene were identified as part of a long-standing program of research into genetic and acquired forms of severe insulin resistance. Patients with features of Donohue syndrome or Rabson-Mendenhall syndrome were not included, as these patients tend to manifest severe metabolic disturbances, including poorly controlled hyperglycemia, a known cause of mitochondrial dysfunction (37). Instead we recruited patients with normal or near-normal glycated hemoglobin (HBA1C) levels. Healthy age-, gender-, and BMI-matched control volunteers were recruited by advertisement. All were sedentary non-smokers without medical disorders likely to affect energy metabolism and without a family history of diabetes.

Experimental protocol and magnetic resonance studies. See the Supplemental Methods for details regarding these studies.

Biochemical assays. Insulin, leptin, and adiponectin were measured as previously described (20).

Statistics. All statistics were performed in SPSS PASW Statistics 18 (SPSS Inc.). Quantitative data are presented as mean ± SEM. The 2-tailed independent-sample t test was used to compare means between groups, with significance classed as P < 0.05.

View Supplemental data

We thank all the participants and J. Harris, L. McGrath, and the staff of the WTCRF for assistance with clinical studies. This work was supported by grants from the Wellcome Trust (to S. O’Rahilly, R.K. Semple, and D.B. Savage), the UK National Institute for Health Research Cambridge Biomedical Research Centre, the UK Medical Research Council Centre for Obesity and Related Metabolic Diseases, and the Clinical Research Infrastructure Grant.

Conflict of interest: The authors have declared that no conflict of interest exists.

Citation for this article: J Clin Invest doi:10.1172/JCI46405.

Savage DB, Petersen KF, Shulman GI. Disordered lipid metabolism and the pathogenesis of insulin resistance. Physiol Rev. 2007;87(2):507–520. Patti ME, Corvera S. The role of mitochondria in the pathogenesis of type 2 diabetes. Endocr Rev. 2010;31(3):364–395. Kelley DE, He J, Menshikova EV, Ritov VB. Dysfunction of mitochondria in human skeletal muscle in type 2 diabetes. Diabetes. 2002;51(10):2944–2950. Morino K, et al. Reduced mitochondrial density and increased IRS-1 serine phosphorylation in muscle of insulin-resistant offspring of type 2 diabetic parents. J Clin Invest. 2005;115(12):3587–3593. Ritov VB, et al. Deficiency of electron transport chain in human skeletal muscle mitochondria in type 2 diabetes mellitus and obesity. Am J Physiol Endocrinol Metab. 2010;298(1):E49–E58. Ritov VB, Menshikova EV, He J, Ferrell RE, Goodpaster BH, Kelley DE. Deficiency of subsarcolemmal mitochondria in obesity and type 2 diabetes. Diabetes. 2005;54(1):8–14. Petersen KF, et al. Mitochondrial dysfunction in the elderly: possible role in insulin resistance. Science. 2003;300(5622):1140–1142. Petersen KF, Dufour S, Befroy D, Garcia R, Shulman GI. Impaired mitochondrial activity in the insulin-resistant offspring of patients with type 2 diabetes. N Engl J Med. 2004;350(7):664–671. Szendroedi J, et al. Muscle mitochondrial ATP synthesis and glucose transport/phosphorylation in type 2 diabetes. PLoS Med. 2007;4(5):e154. Schrauwen P, Schrauwen-Hinderling V, Hoeks J, Hesselink MK. Mitochondrial dysfunction and lipotoxicity. Biochim Biophys Acta. 2010;1801(3):266–271. Schrauwen-Hinderling VB, et al. Impaired in vivo mitochondrial function but similar intramyocellular lipid content in patients with type 2 diabetes mellitus and BMI-matched control subjects. Diabetologia. 2007;50(1):113–120. Schrauwen-Hinderling VB, Roden M, Kooi ME, Hesselink MK, Schrauwen P. Muscular mitochondrial dysfunction and type 2 diabetes mellitus. Curr Opin Clin Nutr Metab Care. 2007;10(6):698–703. Phielix E, et al. Lower intrinsic ADP-stimulated mitochondrial respiration underlies in vivo mitochondrial dysfunction in muscle of male type 2 diabetic patients. Diabetes. 2008;57(11):2943–2949. Stump CS, Short KR, Bigelow ML, Schimke JM, Nair KS. Effect of insulin on human skeletal muscle mitochondrial ATP production, protein synthesis, and mRNA transcripts. Proc Natl Acad Sci U S A. 2003;100(13):7996–8001. Davey Smith G, Leary S, Ness A, Lawlor DA. Challenges and novel approaches in the epidemiological study of early life influences on later disease. Adv Exp Med Biol. 2009;646:1–14. Brown MS, Goldstein JL. A receptor-mediated pathway for cholesterol homeostasis. Science. 1986;232(4746):34–47. Murphy R, Turnbull DM, Walker M, Hattersley AT. Clinical features, diagnosis and management of maternally inherited diabetes and deafness (MIDD) associated with the 3243A>G mitochondrial point mutation. Diabet Med. 2008;25(4):383–399. Szendroedi J, et al. Impaired mitochondrial function and insulin resistance of skeletal muscle in mitochondrial diabetes. Diabetes Care. 2009;32(4):677–679. Semple RK, Savage DB, Halsall DJ, O’Rahilly S. Syndromes of severe insulin resistance and/or lypodystrophy. In: Weiss RE, Refetoff S, eds.Genetic Diagnosis of Endocrine Disorders . Burlington, Massachusetts, USA: Elsevier; 2010:39–44. Semple RK, et al. Postreceptor insulin resistance contributes to human dyslipidemia and hepatic steatosis. J Clin Invest. 2009;119(2):315–322. Forsen S, Hoffman RA. Study of moderately rapid chemical exchange reactions by means of nuclear magnetic double resonance. J Chem Phys. 1963;39:2892–2901. Brindle KM, Blackledge MJ, Challiss RA, Radda GK. 31P NMR magnetization-transfer measurements of ATP turnover during steady-state isometric muscle contraction in the rat hind limb in vivo. Biochemistry. 1989;28(11):4887–4893. Brindle KM, Radda GK. 31P-NMR saturation transfer measurements of exchange between Pi and ATP in the reactions catalysed by glyceraldehyde-3-phosphate dehydrogenase and phosphoglycerate kinase in vitro. Biochim Biophys Acta. 1987;928(1):45–55. Kemp GJ. The interpretation of abnormal 31P magnetic resonance saturation transfer measurements of Pi/ATP exchange in insulin-resistant skeletal muscle. Am J Physiol Endocrinol Metab. 2008;294(3):E640–E642. Campbell-Burk SL, Jones KA, Shulman RG. 31P NMR saturation-transfer measurements in Saccharomyces cerevisiae: characterization of phosphate exchange reactions by iodoacetate and antimycin A inhibition. Biochemistry. 1987;26(23):7483–7492. van den Broek NM, Ciapaite J, Nicolay K, Prompers JJ. Comparison of in vivo postexercise phosphocreatine recovery and resting ATP synthesis flux for the assessment of skeletal muscle mitochondrial function. Am J Physiol Cell Physiol. 2010;299(5):C1136–C1143. Kingsley-Hickman PB, et al. 31P NMR studies of ATP synthesis and hydrolysis kinetics in the intact myocardium. Biochemistry. 1987;26(23):7501–7510. Taylor DJ, Bore PJ, Styles P, Gadian DG, Radda GK. Bioenergetics of intact human muscle. A 31P nuclear magnetic resonance study. Mol Biol Med. 1983;1(1):77–94. Karakelides H, et al. Effect of insulin deprivation on muscle mitochondrial ATP production and gene transcript levels in type 1 diabetic subjects. Diabetes. 2007;56(11):2683–2689. Cheng Z, et al. Foxo1 integrates insulin signaling with mitochondrial function in the liver. Nat Med. 2009;15(11):1307–1311. Cheng Z, Tseng Y, White MF. Insulin signaling meets mitochondria in metabolism. Trends Endocrinol Metab. 2010;21(10):589–598. Boudina S, et al. Contribution of impaired myocardial insulin signaling to mitochondrial dysfunction and oxidative stress in the heart. Circulation. 2009;119(9):1272–1283. Bonnard C, et al. Mitochondrial dysfunction results from oxidative stress in the skeletal muscle of diet-induced insulin-resistant mice. J Clin Invest. 2008;118(2):789–800. Lebon V, et al. Effect of triiodothyronine on mitochondrial energy coupling in human skeletal muscle. J Clin Invest. 2001;108(5):733–737. Zurlo F, Larson K, Bogardus C, Ravussin E. Skeletal muscle metabolism is a major determinant of resting energy expenditure. J Clin Invest. 1990;86(5):1423–1427. Pospisilik JA, et al. Targeted deletion of AIF decreases mitochondrial oxidative phosphorylation and protects from obesity and diabetes. Cell. 2007;131(3):476–491. Savage DB, Semple RK, Chatterjee VK, Wales JK, Ross RJ, O’Rahilly S. A clinical approach to severe insulin resistance. Endocr Dev. 2007;11:122–132.

View the original article here