Dysregulation of insulin receptor substrate 2 in β cells and brain causes obesity and diabetes

The molecular link between obesity and β cell failure that causes diabetes is difficult to establish. Here we show that a conditional knockout of insulin receptor substrate 2 (Irs2) in mouse pancreas β cells and parts of the brain — including the hypothalamus —increased appetite, lean and fat body mass, linear growth, and insulin resistance that progressed to diabetes. Diabetes resolved when the mice were between 6 and 10 months of age: functional β cells expressing Irs2 repopulated the pancreas, restoring sufficient β cell function to compensate for insulin resistance in the obese mice. Thus, Irs2 signaling promotes regeneration of adult β cells and central control of nutrient homeostasis, which can prevent obesity and diabetes in mice.

Type 2 diabetes affects 18.2 million people in the United States and is increasingly common among young people (1, 2). Most disease models point to obesity as an important cause of type 2 diabetes, because of its association with skeletal muscle insulin resistance and pancreatic β cell failure (3, 4). However, dysregulated peripheral and central signaling — owing to many different causes — might initiate and sustain the cascade of events that progress to obesity and diabetes. Most if not all insulin signals are produced or modulated through tyrosine phosphorylation of insulin receptor substrate 1 (IRS1); IRS2 or its homologs; or other scaffold proteins including Src-homology-2 (SH2) domains including SH2 domain containing Shc, Casitas B-lineage lymphoma proto-oncogene (c-CBL), adapter protein with a PH and SH2 domain (APS) and SH2 and PH domain containing signaling mediator 1 (SH2B), Grb-2 associated binder-I (GAB1), GAB2, or dreadlocks (DOCK, an SH3-SH2 adaptor protein) isoforms (5). Although the role of each of these substrates merits attention, cell-based experiments and work with transgenic mice suggest that many insulin responses are directly integrated with counterregulatory hormones and proinflammatory cytokines through tyrosine and serine phosphorylation of IRS1 and IRS2 (5). Dysregulation of these IRS proteins by proinflammatory cytokines or genetic deletion impairs glucose tolerance as a result of peripheral insulin resistance; however, systemic Irs2^–/– mice also develop life-threatening diabetes at 2–3 months of age because Irs2 is required for β cell growth, function, and survival (6–9). The progression of Irs2^–/– mice toward diabetes is retarded or prevented by modifying elements of the insulin/IGF-signaling cascade that promote compensatory β cell function — including downregulation of protein tyrosine phosphatase Ptp1b or the transcription factor Foxo1; or upregulation of Akt (also called protein kinase B, or PKB) or pancreatic duodenal homeobox 1 (Pdx1) (9–14). Transgenic upregulation of Irs2 in pancreatic β cells also prevents diabetes in Irs2^–/– mice, obese mice, and streptozotocin-induced diabetic mice by promoting sufficient and sustained compensatory insulin secretion (9). Thus, the Irs2 branch of the insulin/IGF-signaling cascade is a common link between peripheral insulin action and secretion.

A molecular basis for the close association between obesity and peripheral insulin resistance is difficult to establish. Dysregulated signaling — rather than antidotal consumption of high-calorie diets — might contribute to the early development of obesity that progresses to diabetes (15–17). Insulin, leptin, and adiponectin are important peripheral signals that inform the brain of short- and long-term nutrient availability (17–19). Pharmacologic inhibition of insulin signaling in the hypothalamus increases food intake, and conditional KO of the insulin receptor in the brain causes obesity in mice on high-fat diets (20–23). Leptin secreted from adipocytes promotes satiety and energy utilization, at least in part by promoting α melanocyte–stimulating hormone (αMSH) production in the hypothalamus (15). Mutations that disrupt neuronal leptin signaling or melanocortin signaling increase food intake, body weight, and peripheral insulin resistance in mice and people that progresses to diabetes if β cell function also deteriorates (24–26). Adiponectin, another adipocyte-derived hormone, enhances hepatic and muscle insulin action and promotes energy expenditure through signaling in the hypothalamus (18, 27); however, adiponectin is reduced in obese people and rodents (28, 29).

Previous work suggests that Irs2 signaling plays an important role in the CNS for brain growth, female fertility, and nutrient homeostasis (30). Since IRS2 is highly expressed in the hypothalamus, its signaling cascade may be responsible for integrating central control of nutrient homeostasis and appetite regulation with peripheral insulin action and β cell function (9). Female Irs2^–/– mice — which develop diabetes more slowly than do male mice — are hyperphagic and obese until severe diabetes causes weight loss (30). To test the role of selective Irs2 dysregulation in obesity and diabetes, we flanked the Irs2 gene with loxP recombination sites (floxed Irs2, or fIrs2) and crossed these mice with transgenic mice expressing Cre recombinase under control of the rat insulin-2 promoter (TgN[Ins2Cre]25Mgn mice, or cr² mice). Cre recombinase is expressed strongly in β cells and weakly in certain brain regions of these transgenic cr² mice (31, 32). Thus, our strategy strongly deleted fIrs2 alleles from β cells and weakly deleted them from brain and certain neurons of the hypothalamus. Analysis of the mice suggests that dysregulated Irs2 signaling could be a common link between obesity and β cell failure that progresses to diabetes; and that Irs2 signaling is an important element for β cell regeneration.

Conditional disruption of floxed Irs2 allele by cr². To determine whether partial dysregulation of Irs2 signaling links obesity and diabetes, we intercrossed mice containing homozygous fIrs2 alleles with cr² mice (Figure 1A). PCR analysis of islets from the compound fIrs2:cr² mice (4 weeks old) confirmed that fIrs2 alleles were almost completely deleted, which was consistent with the reported efficacy of cr² in β cells (Figure 1B). Since islets are composed mainly of β cells, Irs2 mRNA levels decreased at least 90% (Figure 1C). fIrs2 alleles were partially deleted in the brain, including regions of the hypothalamus (Figure 1, B and D), which is consistent with the reported expression of cr² in some neurons (31, 32). At 8 weeks, Irs2 protein levels were reduced at least 30% in fIrs2:cr² hypothalamic extracts (Figure 1E). By contrast, fIrs2 was intact and protein levels were normal in other tested tissues, including muscle, liver, and fat (Figure 1B and data not shown).

Figure 1

Creation of fIrs2 mice and cr²-mediated recombination/deletion of fIrs2. (A) A schematic diagram of the 4 steps in the generation of the fIrs2 containing embryonic stem cells (see Methods for details). Homol recomb, homologous recombination. (B) Recombination/deletion of fIrs2 was revealed in various mouse tissues at 4 weeks of age by PCR analysis of genomic DNA. L, liver; K, kidney; H, heart; S, spleen; B, brain, F, fat; M, muscle. The intact fIrs2 allele (fIrs2⁺) yields a 750-bp fragment, whereas the recombined/deleted allele (fIrs2^_) yields a 250-bp fragment. (C) Real-time PCR analysis of Irs2 expression in isolated islets, normalized by cyclophilin expression. (D) PCR analysis of hypothalamic DNA to reveal recombination event. (E) Immunoprecipitation and Western blot analysis of hypothalamus with antibodies against Irs2. (F) Brain sections were immunostained with antibodies against Irs2 (red) and Cre (green), and the images were merged in Openlab Image software (version 3.14; Improvision). Similar results were obtained with multiple sections from 2 to 8 mice. Asterisks indicate third ventricle. Magnification, ×20.

To verify that fIrs2 was disrupted in hypothalamic neurons, we immunostained sections of the arcuate nucleus with antibodies against Irs2 (green stain) and Cre recombinase (red stain); low background was confirmed in systemic Irs2^–/– brain sections that lack both proteins (Figure 1F). Irs2 was detected in the arcuate nucleus from WT or cr² mice, and Cre recombinase was detected in cr² mice, where it was expressed in some (yellow staining) but not all Irs2-positive neurons (Figure 1F). However, coexpression of Irs2 and Cre was largely abolished in fIrs2:cr² sections, which confirmed that fIrs2 was deleted wherever Cre was expressed (Figure 1F); similar results were found in paraventricular nucleus (data not shown).

Growth of the fIrs2:cr² mice. The fIrs2:cr² mice reached a normal body weight at 4 weeks of age but then grew more rapidly and were 30% heavier than controls when the experiment was terminated at 32 weeks (Figure 2A). At 8 weeks, fIrs2:cr² mice were 8 grams heavier than the controls (fIrs2:cr², 38.2 ± 0.6 g, P < 0.01 compared with fIrs2, 29.8 ± 0.5 g; cr², 29.4 ± 0.8 g; and WT, 29.1 ± 0.6 g); and they consumed 60% (P < 0.05) more chow (9% fat) and 65% (P = 0.001) more water than did the control mice during a 24-hour interval (Figure 2, B and C). Unlike systemic Irs2^–/–mice, which have small brains (33), the fIrs2:cr² mice had a normal brain size (fIrs2:cr², 0.41 ± 0.01 g; fIrs2, 0.40 ± 0.01 g; cr², 0.40 ± 0.02 g; and WT, 0.40 ± 0.02 g).

Figure 2

Growth and nutrient homeostasis of the fIrs2:cr² mice. (A) Male littermates fed regular or low-fat chow were weighed weekly from postnatal day 21 until 32 weeks of age; each point represents the average ± SE of at least 6 mice. Omitted error bars are smaller than the symbol. (B) Food intake was determined with 8-week-old male mice over 24 hours using a Comprehensive Lab Animal Monitoring System; average ± SE for 4 animals per genotype is reported. (C) Cumulative water intake was determined with 8-week-old male mice over 24 hours using a Comprehensive Lab Animal Monitoring System; average ± SE for 3 animals per genotype is reported. (D) Average lean and fat body mass ± SE was determined on three 8-week-old male of the indicated genotypes using a dual energy X-ray absorptiometry (DEXA) scanner (GE Medical Systems Lunar) according to the manufacturer’s instructions. (E) Average ± SE plasma insulin and leptin levels were determined on at least fifteen 8-week-old random fed male mice of each genotype using a rat insulin or mouse leptin ELISA Kit. *P < 0.05; **P < 0.01.

The fIrs2:cr² mice were 10% longer — snout to anus — than the controls (fIrs2:cr², 107.9 ± 0.5 mm, P < 0.01 compared with fIrs2, 98 ± 2 mm; cr², 96.5 ± 0.4 mm; WT, 98 ± 1 mm). On the normal 9% fat diet, 8-week-old fIrs2:cr² mice had 2-fold more body fat and increased lean body mass (Figure 2D). Chemical analysis confirmed that fat mass was increased disproportionately in fIrs2:cr² mice (fIrs2:cr², 18% ± 3% fat, n = 5; fIrs2, 11% ± 1% fat, n = 4; P = 0.005). Consistent with the increased adiposity, serum leptin and insulin levels were elevated 2-fold in the fIrs2:cr² mice compared with controls (Figure 2E). Because of their larger size, fIrs2:cr² mice produced more heat than control mice during a 24-hour period (fIrs2:cr², 14.1 ± 0.1 kcal; WT, 11.0 ± 1.5 kcal). Although caloric restriction achieved by feeding the mice a low-fat (5%) diet reduced the weight of all the mice, the fIrs2:cr² mice were consistently heavier (Figure 2A).

Glucose homeostasis in the fIrs2:cr² mice. Diabetes developed in the fIrs2:cr² mice between 8 and 10 weeks of age, as fasting glucose levels exceeded those of the control mice (fIrs2:cr², 110 ± 9 mg/dl, P < 0.01 compared with fIrs2, 63 ± 2 mg/dl; cr², 60 ± 4 mg/dl; WT, 62 ± 6 mg/dl). Moreover, random-fed glucose was consistently 2-fold higher in the fIrs2:cr² mice than in the controls (Figure 2A). Glucose disposal rates measured in the fIrs2:cr² mice — during a hyperinsulinemic/euglycemic clamp test — were reduced 50%, which revealed peripheral insulin resistance, an important hallmark of type 2 diabetes (Figure 2B). By contrast, hepatic glucose output measured before and during experimental hyperinsulinemia was similar in all of the mice examined (Figure 3C).

Figure 3

Glucose homeostasis in the fIrs2:cr² mice. (A) Average ± SE random-fed blood glucose levels were determined using a Glucometer Elite in at least fifteen 8-week-old male mice of the indicated genotypes. (B and C) Glucose-disposal rates and hepatic glucose output before and during hyperinsulinemia (average ± SE) were determined on 6_10 fasted and conscious 10-week-old male mice of the indicated genotypes. (D) Pancreas insulin content was determined using rat insulin ELISA Kit following acid-ethanol extraction. Results are average ± SE for five 8-week-old mice per genotype. (E) After a 16-hour overnight fast, blood glucose (average ± SE) was determined on at least fifteen 8-week-old male mice of the indicated genotypes following intraperitoneal injection of 2 g D-glucose/kg body weight. (F) The integrated glucose tolerance (GTT, t = 0 Ø 120 min) for the 8-week old male mice fed the ordinary diet (data in E) was compared with that for 16-week-old male mice fed a low-fat diet (LFD, 5%); these ages were selected for comparison to match the body weights. The glucose tolerance curves were integrated using the trapezoid rule implemented in Sigmaplot 8.0 (Systat Software Inc.). *P < 0.05; **P < 0.01.

The insulin content of the fIrs2:cr² pancreas was reduced more than 2-fold compared with that in the control mice (Figure 3D). This result is consistent with a positive role for IRS2 in β cell function, especially during compensation for peripheral insulin resistance (9). Intraperitoneal glucose injections following a 16-hour overnight fast confirmed glucose intolerance and diabetes in fIrs2:cr² mice (Figure 3E). Although glucose tolerance improved when the fIrs2:cr² mice were placed on the low-fat diet, fIrs2:cr² mice remained hyperinsulinemic (fIrs2:cr², 2.5 ± 0.2 ng/ml, P < 0.01 compared with fIrs2, 0.7 ± 0.1 ng/ml; cr², 0.6 ± 0.1 ng/ml; WT, 0.5 ± 0.1 ng/ml) and relatively glucose intolerant in comparison with control mice on the low-fat diet (Figure 3F).

β cell growth and function in fIrs2:cr² mice. The growth, function, and survival of β cells are essential for normal glucose homeostasis, especially during compensation for peripheral insulin resistance. At 4 weeks of age when the fIrs2:cr² mice had a normal body weight, relative β cell content — estimated by the average β cell area in pancreas sections — and β cell size were indistinguishable from those of control mice (Table 1). The expression of several genes that promote β cell function and growth — hepatocyte nuclear factor 3 β (Hnf3β), Hnf4α, and Pdx1 — were decreased in fIrs2:cr² islets; however, glucose transporter 2 (Glut2) mRNA levels were nearly normal (Table 1). DNA synthesis — estimated by in situ BrdU incorporation—was not significantly different in fIrs2:cr² islets (Table 1). Nevertheless, β cell content in the fIrs2:cr² mice did not increase during the next 4 weeks, contrary to what was observed in 8-week-old control mice. These characteristics were consistent with the reduced insulin content of the fIrs2:cr² pancreas (Figure 3D).

Table 1

Islet characteristics

Based upon our experience with systemic Irs2^–/– mice (9), we expected hyperglycemia to progress and eventually kill the fIrs2:cr² mice; however, diabetes always resolved spontaneously between 6 and 8 months of age (Figure 4A). At 6 months, relative β cell content of fIrs2:cr² mice was normal in comparison with WT mice; and by 10 months, the relative β cell mass increased in all of the mice to compensate for peripheral insulin resistance that always emerges as the mice age and gain weight (Table 1). The nondiabetic fIrs2:cr² mice remained heavier than the control mice (fIrs2:cr², 68 ± 2 g; fIrs2, 51 ± 2 g; cr², 53 ± 1 g); and owing to a larger pancreas at 10 months, the total β cell mass in fIrs2:cr² mice was slightly greater than that in controls (Table 1). At 10 months, pancreas sections from fIrs2:cr² mice contained more islets, but a disproportionate number of the islet sections contained fewer than 10 β cells (Table 1). Thus, persistent insulin resistance was balanced by the growth of functional β cells that maintained compensatory hyperinsulinemia in mice at 10 months of age (Figure 4B); and during the next year, these mice did not develop diabetes.

Figure 4

Pancreas β cell function in aged fIrs2:cr² and fIrs2:cr²:Irs1^+/_ mice.(A) Random-fed blood glucose levels were determined at the indicated ages. Averages ± SE were determined from at least 5 male mice per genotype. (B) Random insulin levels were determined in male mice of the indicated genotypes and ages. Averages ± SE were determined from at least 5 mice per genotype. (C) Cre-mediated recombination/deletion of fIrs2 alleles was determined by real-time PCR analysis in single islets isolated from male fIrs2:cr² mice at the indicated ages; or determined in hypothalamus from 10-month-old fIrs2 or fIrs2:cr² mice. (D) Real-time PCR analysis of cr² expression in isolated fIrs2:cr² islets at the indicated ages; mRNA levels were normalized against cyclophilin expression, and the averages ± SE were determined for 2 male mice at each age. (E) Body weight of male littermates of the indicated genotypes that were fed regular chow and weighed weekly from postnatal day 21 until 16 weeks of age; each point represents the average ± SE of at least 6 mice. (F) Three representative pancreatic sections obtained from 6-month-old mice of the indicated genotypes immunostained with antibodies against insulin (green) or glucagon (red).

What reversed the progression to diabetes? We isolated individual islets from fIrs2:cr² mice at 4, 6, and 10 months of age and used PCR to identify the presence or absence of the fIrs2 allele. At 4 months, most islets lacked fIrs2, whereas a few islets displayed a mixture of deleted and intact fIrs2 (Figure 4C). At 6 months, islets contained either deleted fIrs2, intact fIrs2, or a mixture of these alleles; and by 10 months, deleted fIrs2 was never detected (Figure 4C). Thus, fIrs2^–/– β cells were progressively replaced by β cells containing fIrs2. The progressive decline of Cre recombinase in islets between 1 and 12 months of age suggests that regenerated β cells arise from cells that weakly express Cre recombinase (Figure 4D); however, we cannot establish the source of these new β cells.

The role of IRS1 in β cell regeneration. Peripheral insulin action is mediated by both Irs1 and Irs2 (5); however, systemic Irs1^–/– mice never develop diabetes, as Irs2 promotes β cell growth and function to compensate for peripheral insulin resistance (34). Previous work shows that Irs1 haploinsufficiency strongly inhibits residual growth, function, and survival of β cells in systemic Irs2^–/– mice, as β cells are nearly undetected when systemic Irs2^–/–:Irs1^+/– mice die at 30 days of age (34). To determine whether Irs1 contributes to the regeneration of β cells in fIrs2:cr² mice, we intercrossed fIrs2:cr² mice with systemic Irs1^+/– mice to produce compound [fIrs2:cr²]:Irs1^+/– mice. Unlike systemic Irs1^+/– mice or fIrs2:cr² mice, the [fIrs2:cr²]:Irs1^+/– mice developed diabetes at 4 weeks (Figure 4A). The early-onset diabetes progressed rapidly to severe hyperglycemia as serum insulin levels fell during the next several months (Figure 4B). As a result of progressive and severe hypoinsulinemia, the [fIrs2:cr²]:Irs1^+/– mice gained less weight than the fIrs2:cr² mice during the study period (Figure 4E). When the [fIrs2:cr²]:Irs1^+/– mice were sacrificed at 6 months, relative β cell content was reduced 8-fold in comparison with the insulin-resistant but normal-sized Irs1^+/– mice (Figure 4F and Table 1). Although β cells containing insulin were detected in the [fIrs2:cr²]:Irs1^+/– pancreas at 6 months, the small islets were disproportionately populated by α cells (Figure 4F). The striking discrepancy between the β cell mass of fIrs2:cr²:Irs1^+/–mice compared with fIrs2:cr² mice was unexpected, as β cells retaining fIrs2 alleles should grow in both cases; unfortunately, it was impossible to isolate intact islets from 6 month old [fIrs2:cr²]:Irs1^+/– mice for genotyping.

Hypothalamic function in fIrs2:cr² mice. The fIrs2:cr² mice closely resemble agouti mice or mice deficient in pro-opiomelanocortin (Pomc^–/–) or melanocortin 4 receptor (Mc4r^–/–) (15), which suggests that Irs2 mediates some effects of insulin upon the melanocortin pathway in the hypothalamus. To evaluate insulin signaling, we prepared hypothalamic extracts from WT or fIrs2:cr² mice 15 minutes after intravenous insulin or saline injections. Irs2 protein levels and tyrosine phosphorylation were reduced 30% in fIrs2:cr² extracts in comparison with WT, whereas levels of signal transducer and activator of transcription 3 (Stat3) protein — measured as a control — were equal in both mice (Figure 5A). Moreover, less p85α/β was detected in Irs2 immunoprecipitates, and insulin-stimulated Akt phosphorylation was decreased in fIrs2:cr² hypothalamic extracts (Figure 5A). These results confirmed that insulin signaling was reduced in fIrs2:cr² hypothalamus. By comparison, leptin-stimulated Stat3 or Akt phosphorylation were not significantly changed in fIrs2:cr² mice (data not shown).

Figure 5

Analysis of hypothalamic function in the fIrs2:cr² mice. (A) Immunoblots of hypothalamus lysates. After an overnight fasting, anesthetized 8-week-old mice were injected intravenously with 50 μl of saline or saline containing 5 units of insulin. Fifteen minutes later, the hypothalamus was isolated and lysed. Lysates were immunoprecipitated with anti_Irs2 antibody, resolved by SDS-PAGE, and then immunoblotted with antibodies against Irs2, phosphotyrosine (pY20), or p85. Lysates were also blotted independently with antibodies against Akt or phospho-Akt [Akt(Pi)], or Stat3. Data are representative of at least three independent experiments. (B) The relative expression of various mRNAs_Npy (1421690_s_at), Agrp (1419127_at), Cart (1422825_at) or Pomc1 (1455858_x_at; 1433800_a_at), Stat3 (1426587_a_at; 1426587_a_at; 1460700_at), Socs3 (1455899_x_at; 1456212_x_at) or PcSK2 (1448312_at; 1444147_at) was determined in hypothalamic extracts using Affymetrix GeneChip Mouse Expression Array 430A. The fold change is shown with the 90% confidence interval for expression in 3 fIrs2:cr² mice compared with the average control values from two WT, 2 cr² and 2 fIrs2 mice. Where more than one probe occurs on the Genechip, the average is reported. Cntr, control. (C) Brain sections from WT or fIrs2:cr² mice were immunostained with antibodies against αMSH (green) or Irs2 (red). Yellow appears in the digitally merged images where αMSH and Irs2 are coexpressed. Magnification, ×20; *3rd ventricle. (D) Body weight and blood glucose levels of fIrs2:cr², [fIrs2:cr²]:rip13^Irs2, and WT male mice at 8 weeks; average ± SE was determined from 5 animals of each genotype. No tg, no transgene.

Food intake and insulin secretion increases αMSH secretion from POMC neurons in the hypothalamus, which stimulate MC4R signaling that promotes satiety (35). Our results were consistent with characteristics of reduced MC4R signaling, including excess appetite: Pomc1 mRNA levels — measured on Affymetrix GeneChip Mouse Expression Array 430A — were reduced almost 2-fold (90% confidence) in the hypothalamus of 16-hour–starved fIrs2:cr² mice (Figure 5B). Other neuropeptide mRNAs — neuropeptide Y (Npy), agouti-related protein (Agrp), and cocaine and amphetamine regulated transcript (Cart) — were not changed; and the mRNA levels encoding Stat3, suppressor of cytokine signaling 3 (Socs3), and pro-protein convertase-2 — which processes αMSH — were normal (Figure 5B).

Since αMSH is a product of the POMC transcript in the arcuate nucleus, we immunostained hypothalamic sections from 8-week-old WT or fIrs2:cr² mice with antibodies against Irs2 or αMSH (Figure 5C). The size of the hypothalamus appeared normal. In WT mice, Irs2 and αMSH were coexpressed — as indicated by the yellow staining — in many but not all of the neurons near the 3rd ventricle; however, yellow staining was absent in the fIrs2:cr² sections, which suggested that many αMSH neurons no longer expressed Irs2. These results show that some POMC neurons in the arcuate nucleus of fIrs2:cr² mice were uncoupled from Irs2 signaling.

To confirm that weight gained by the fIrs2:cr² mice during the first 8 weeks of life was due to dysregulated Irs2 signaling in the CNS and not β cell failure, we restored β cell function by introducing transgenic Irs2 into the β cells of fIrs2:cr² mice. This was accomplished through an intercross with rip13^Irs2 mice, which express recombinant Irs2 in the β cells but not in the hypothalamus (9). Hyperglycemia did not occur in the [fIrs2:cr²]:rip13^Irs2 mice; however, the mice continued to gain excess weight (Figure 5D). Thus, partially dysregulated Irs2 signaling in the brain, rather than in β cells, appears to be responsible for hyperphagia and obesity in the fIrs2:cr² mice.

Our results show that partial dysregulation of Irs2 signaling in β cells and brain — including the hypothalamus — may explain the close association between obesity, peripheral insulin resistance, and β cell failure that characterizes type 2 diabetes. Whether dysregulation of IRS2 signaling contributes to type 2 diabetes and obesity in humans is unknown. However, allelic variations at the human IRS2 locus appear to be linked to obesity, glucose intolerance, and polycystic ovarian syndrome (36–38). Moreover, IRS1 and IRS2 functions are strongly inhibited by many heterologous signals, especially proinflammatory cytokines produced during acute trauma or chronic metabolic stress (5). Moreover, IRS2 expression is strongly induced by the cAMP response element binding protein (CREB), which is activated by cAMP agonists, including glucagon-like peptide 1 (GLP1) in β cells (10). Dysregulation of GLP1 signaling has been reported in type 2 diabetes, which might impair IRS2 expression in β cells and the brain (39, 40).

Male fIrs2:cr² mice develop diabetes between 2 and 3 months of age, at least in part because β cells fail to compensate for peripheral insulin resistance. Unlike systemic Irs2^–/– mice that develop life-threatening diabetes earlier in life owing to complete loss of β cells, the fIrs2:cr² mice survive with diabetes until β cell regeneration repopulates the islets with functional β cells that cure diabetes. In fIrs2:cr² mice, the new β cells must arise from cells that never expressed Cre. The progressive decline of Cre recombinase in islets is consistent with this mechanism. Inactivation of cr² by hyperglycemia — as shown for the insulin-2 gene itself—might explain the progressive loss of Cre that prevented deletion of fIrs2 alleles in the new β cells (41); however, more work is need to understand the exact mechanism involved.

New β cells are thought to arise by division of β cells themselves; by differentiation of islet precursors residing in pancreatic duct epithelium; or from bone marrow (42). Direct lineage tracing in adult mice suggests that terminally differentiated β cells retain a significant proliferative capacity in vivo and are an important — and possibly the only — source of new β cells in adult mice (42). If β cells that never express Cre recombinase are the sole source of islet regeneration in the fIrs2:cr² mouse, the regenerative capacity must be significant. Intact fIrs2 is barely detected in pooled islets at 2 months of age, which is consistent with previous reports showing that cr² mediates nearly complete deletion of floxed genes in β cells (32, 43, 44). At least a portion of the weak PCR signal revealing intact fIrs2 in these islets arises from non–β cells. Thus, β cells in fIrs2:cr² islets that retain Irs2 expression represent a distinct minority. Regardless of the source of regenerated β cells, IRS2 signaling must play an important role in the process.

By contrast, systemic deletion of one Irs1 allele from fIrs2:cr² mice caused early-onset diabetes and impaired β cell regeneration. This result is unexpected, because Irs1 plays a limited role in islet growth (34). The absence of 1 Irs1 allele should not inhibit regeneration if the new β cells arise by replication of a few β cells that contain intact fIrs2. It remains possible that Irs1-dependent precursors contributed to the expansion of β cell mass in the fIrs2:cr² mice. Alternatively, Irs1 signaling might sustain the function of Irs2-deficieint β cells long enough to permit the few Irs2-containing β cells an opportunity to undergo division and repopulate the islets.

Diabetes is averted by the expression of transgenic Irs2 in the fIrs2:cr² β cells, which suggests that strategies to promote Irs2 expression may prevent or cure type 2 diabetes. Irs2 mRNA and Irs2 protein is strongly upregulated in β cells by cAMP Ø CREB signaling (10). A relevant example of this mechanism is the increased expression of IRS2 during activation of the GLP1 receptor by GLP1 itself or the long-acting homolog exendin-4 (9, 10). Improved insulin secretion, increased insulin sensitivity, and better appetite regulation reported in diabetic rodents and people treated with GLP1 or exendin-4 can be explained by increased expression of IRS2 (5, 45). Since Irs2 function is inhibited by serine phosphorylation and ubiquitin-mediated degradation that is mediated by proinflammatory cytokines (46), increased IRS2 expression together with effective management of chronic inflammation might prevent the progression of insulin resistance to diabetes.

The mild disruption of fIrs2 by the cr² in the hypothalamus displays the features of dysregulated melanocortin signaling — increased appetite, lean and fat body mass, linear growth, and peripheral insulin resistance (15). Although obesity is usually thought to contribute directly to peripheral insulin resistance, lean fIrs2:cr² mice were also relatively insulin resistant, hyperinsulinemic, and glucose intolerant compared with the control mice. The results from the hyperinsulinemic clamp suggest that the insulin resistance in the fIrs2:cr² mice resides in extra hepatic tissues — most likely the skeletal muscle — as hepatic glucose output is regulated normally by insulin, whereas glucose disposal is impaired. Previous work shows that dysregulated melanocortin signaling causes peripheral insulin resistance even before obesity develops, possibly because activation of MC4Rs in the paraventricular nucleus promotes peripheral insulin action (47, 48). The reduced levels of Pomc1 mRNA in fIrs2:cr² hypothalamus and the loss of Irs2 from neurons that express αMSH reveal a potential mechanism for these effects. Thus, dysregulated Irs2 signaling in the brain, especially the hypothalamus, could contribute to central and peripheral insulin resistance that predisposes individuals to obesity and diabetes (25).

The deletion of floxed insulin or Igf1 receptors by cr² expression does not alter food intake or body weight (43, 49). If more careful analysis reveals the deletion of these receptor genes in the hypothalamus — as reported here for fIrs2 and elsewhere for floxed Stat3 (32) — then loss of insulin or Igf1 receptor signaling might not promote obesity. This potential result would be similar to that obtained using the phenotype of neuron-specific insulin receptor KO (NIRKO) mice; only when placed on a diet containing 60% fat do NIRKO mice develop obesity.

Our work shows that Irs2 signaling is important for β cell growth, function, and survival. Moreover, our experiments show that β cell regeneration in middle age can cure diabetes in mice and that Irs2 signaling provides a significant advantage for this regeneration. Now we must determine whether the new β cells arise from the division of a few preexisting β cells that escaped fIrs2 deletion or from non–β cell precursors that never express Cre recombinase when they differentiate at middle age. Independent of the mechanism, strategies to enhance the Irs2 branch of the insulin/IGF-signaling pathway in β cells and hypothalamus could be a rational approach to prevent obesity and diabetes.

This work was supported by NIH grant DK55236. X. Lin was a Juvenile Diabetes Foundation (JDF) postdoctoral fellow while this work was conducted.

See the related Commentary beginning on page 886.

Nonstandard abbreviations used: cr², TgN[Ins2Cre]25Mgn; fIrs2, floxed Irs2; GLP1, glucagon-like peptide 1; Irs, insulin receptor substrate; Mc4r, melanocortin 4 receptor; αMSH, α melanocyte–stimulating hormone; Pdx1, pancreatic duodenal homeobox 1; Pomc, pro-opiomelanocortin; Stat3, signal transducer and activator of transcription 3.Conflict of interest: The authors have declared that no conflict of interest exists.

1. Zimmet, P, Alberti, KG, Shaw, J. Global and societal implications of the diabetes epidemic. Nature. 2001. 414:782-787.
   View this article via: PubMed CrossRef Google Scholar
2. Diamond, J. The double puzzle of diabetes. Nature. 2003. 423:599-602.
   View this article via: PubMed CrossRef Google Scholar
3. DeFronzo, RA. Lilly lecture 1987. The triumvirate: beta-cell, muscle, liver. A collusion responsible for NIDDM [review]. Diabetes. 1988. 37:667-687.
   View this article via: PubMed Google Scholar
4. Shulman, GI. Cellular mechanisms of insulin resistance. J. Clin. Invest. 2000. 106:171-176.
   View this article via: JCI PubMed CrossRef Google Scholar
5. White, MF. Insulin signaling in health and disease. Science. 2003. 302:1710-1711.
   View this article via: PubMed CrossRef Google Scholar
6. Withers, DJ, et al. Disruption of IRS-2 causes type 2 diabetes in mice. Nature. 1998. 391:900-904.
   View this article via: PubMed CrossRef Google Scholar
7. Kubota, N, et al. Disruption of insulin receptor substrate 2 causes type 2 diabetes because of liver insulin resistance and lack of compensatory beta-cell hyperplasia. Diabetes. 2000. 49:1880-1889.
   View this article via: PubMed CrossRef Google Scholar
8. Previs, SF, Withers, DJ, Ren, JM, White, MF, Shulman, GI. Contrasting effects of IRS-1 vs IRS-2 gene disruption on carbohydrate and lipid metabolism in vivo. J. Biol. Chem. 2000. 275:38990-38994.
   View this article via: PubMed CrossRef Google Scholar
9. Hennige, AM, et al. Upregulation of insulin receptor substrate-2 in pancreatic β cells prevents diabetes. J. Clin. Invest. 2003. 112:1521-1532. doi:10.1172/JCI200318581.
   View this article via: JCI PubMed Google Scholar
10. Jhala, US, et al. cAMP promotes pancreatic beta-cell survival via CREB-mediated induction of IRS2. Genes Dev. 2003. 17:1575-1580.
    View this article via: PubMed CrossRef Google Scholar
11. Kushner, JA, et al. Islet-sparing effects of protein tyrosine phosphatase-1b deficiency delays onset of diabetes in IRS2 knockout mice. Diabetes. 2004. 53:61-66.
    View this article via: PubMed CrossRef Google Scholar
12. Kitamura, T, et al. The forkhead transcription factor Foxo1 links insulin signaling to Pdx1 regulation of pancreatic β cell growth. J. Clin. Invest. 2002. 110:1839-1847. doi:10.1172/JCI200216857.
    View this article via: JCI PubMed Google Scholar
13. Kushner, JA, et al. Pdx1 restores β cell function in Irs2 knockout mice. J. Clin. Invest. 2002. 109:1193-1201. doi:10.1172/JCI200214439.
    View this article via: JCI PubMed Google Scholar
14. Suzuki, R, et al. Pdx1 expression in Irs2-deficient mouse beta-cells is regulated in a strain-dependent manner. J. Biol. Chem. 2003. 278:43691-43698.
    View this article via: PubMed CrossRef Google Scholar
15. Butler, AA, Cone, RD. Knockout studies defining different roles for melanocortin receptors in energy homeostasis. Ann. N. Y. Acad. Sci. 2003. 994:240-245.
    View this article via: PubMed Google Scholar
16. Farooqi, IS, et al. Clinical spectrum of obesity and mutations in the melanocortin 4 receptor gene. N. Engl. J. Med. 2003. 348:1085-1095.
    View this article via: PubMed CrossRef Google Scholar
17. Schwartz, MW, et al. Central nervous system control of food intake. Nature. 2000. 404:661-671.
    View this article via: PubMed Google Scholar
18. Qi, Y, et al. Adiponectin acts in the brain to decrease body weight. Nat. Med. 2004. 10:524-529.
    View this article via: PubMed CrossRef Google Scholar
19. Myers (Jr), MG. Leptin receptor signaling and the regulation of mammalian physiology. Recent Prog. Horm. Res. 2004. 59:287-304.
    View this article via: PubMed CrossRef Google Scholar
20. Woods, SC, Lotter, EC, McKay, LD, Porte (Jr), D. Chronic intracerebroventricular infusion of insulin reduces food intake and body weight of baboons. Nature. 1979. 282:503-505.
    View this article via: PubMed CrossRef Google Scholar
21. Schwartz, M, Figlewicz, DP, Baskin, DG, Woods, SC, Porte, D. Insulin in the brain: A hormonal regulator of energy balance. Endocr. Rev. 1992. 13:387-414.
    View this article via: PubMed CrossRef Google Scholar
22. Schwartz, MW, Baskin, DG, Kaiyala, KJ, Woods, SC. Model of the regulation of energy balance and adiposity by the central nervous system. Am. J. Clin. Nutr. 1999. 69:584-596.
    View this article via: PubMed Google Scholar
23. Bruning, JC, et al. Role of brain insulin receptor in control of body weight and reproduction. Science. 2000. 289:2122-2125.
    View this article via: PubMed CrossRef Google Scholar
24. Huszar, D, et al. Targeted disruption of the melanocortin-4 receptor results in obesity in mice. Cell. 1997. 88:131-141.
    View this article via: PubMed CrossRef Google Scholar
25. Obici, S, et al. Central melanocortin receptors regulate insulin action. J. Clin. Invest. 2001. 108:1079-1085. doi:10.1172/JCI200112954.
    View this article via: JCI PubMed Google Scholar
26. Challis, BG, et al. A missense mutation disrupting a dibasic prohormone processing site in pro-opiomelanocortin (POMC) increases susceptibility to early-onset obesity through a novel molecular mechanism. Hum. Mol. Genet. 2002. 11:1997-2004.
    View this article via: PubMed CrossRef Google Scholar
27. Yamauchi, T, et al. Cloning of adiponectin receptors that mediate antidiabetic metabolic effects. Nature. 2003. 423:762-769.
    View this article via: PubMed CrossRef Google Scholar
28. Combs, TP, et al. Sexual differentiation, pregnancy, calorie restriction, and aging affect the adipocyte-specific secretory protein adiponectin. Diabetes. 2003. 52:268-276.
    View this article via: PubMed CrossRef Google Scholar
29. Spranger, J, et al. Adiponectin and protection against type 2 diabetes mellitus. Lancet. 2003. 361:226-228.
    View this article via: PubMed CrossRef Google Scholar
30. Burks, DJ, et al. IRS-2 pathways integrate female reproduction and energy homeostasis. Nature. 2000. 407:377-382.
    View this article via: PubMed CrossRef Google Scholar
31. Gannon, M, Shiota, C, Postic, C, Wright, CV, Magnuson, M. Analysis of the Cre-mediated recombination driven by rat insulin promoter in embryonic and adult mouse pancreas. Genesis. 2000. 26:139-142.
    View this article via: PubMed CrossRef Google Scholar
32. Cui, Y, et al. Essential role of STAT3 in body weight and glucose homeostasis. Mol. Cell. Biol. 2004. 24:258-269.
    View this article via: PubMed CrossRef Google Scholar
33. Schubert, M, et al. Insulin receptor substrate-2 deficiency impairs brain growth and promotes tau phosphorylation. J. Neurosci. 2003. 23:7084-7092.
    View this article via: PubMed Google Scholar
34. Withers, DJ, et al. Irs-2 coordinates Igf-1 receptor-mediated beta-cell development and peripheral insulin signalling. Nat. Genet. 1999. 23:32-40.
    View this article via: PubMed Google Scholar
35. Schwartz, MW. Progress in the search for neuronal mechanisms coupling type 2 diabetes to obesity. J. Clin. Invest. 2001. 108:963-964. doi:10.1172/JCI200114127.
    View this article via: JCI PubMed Google Scholar
36. Lautier, C, et al. Complex haplotypes of IRS2 gene are associated with severe obesity and reveal heterogeneity in the effect of Gly1057Asp mutation. Hum. Genet. 2003. 113:34-43.
    View this article via: PubMed Google Scholar
37. Ehrmann, DA, Tang, X, Yoshiuchi, I, Cox, NJ, Bell, GI. Relationship of insulin receptor substrate-1 and -2 genotypes to phenotypic features of polycystic ovary syndrome. J. Clin. Endocrinol. Metab. 2002. 87:4297-4300.
    View this article via: PubMed CrossRef Google Scholar
38. Stefan, N, et al. The Gly1057Asp polymorphism in IRS-2 interacts with obesity to affect beta cell function. Diabetologia. 2004. 47:759-761.
    View this article via: PubMed CrossRef Google Scholar
39. Brubaker, PL, Drucker, DJ. Glucagon-like peptides regulate cell proliferation and apoptosis in the pancreas, gut and central nervous system. Endocrinology. 2004. 145:2653-2659.
    View this article via: PubMed CrossRef Google Scholar
40. During, MJ, et al. Glucagon-like peptide-1 receptor is involved in learning and neuroprotection. Nat. Med. 2003. 9:1173-1179.
    View this article via: PubMed CrossRef Google Scholar
41. Ling, Z, Heimberg, H, Foriers, A, Schuit, F, Pipeleers, D. Differential expression of rat insulin I and II messenger ribonucleic acid after prolonged exposure of islet beta-cells to elevated glucose levels. Endocrinology. 1998. 139:491-495.
    View this article via: PubMed CrossRef Google Scholar
42. Dor, Y, Brown, J, Martinez, OI, Melton, DA. Adult pancreatic beta-cells are formed by self-duplication rather than stem-cell differentiation. Nature. 2004. 429:41-46.
    View this article via: PubMed CrossRef Google Scholar
43. Kulkarni, RN, et al. Tissue-specific knockout of the insulin receptor in pancreatic β cells creates an insulin secretory defect similar to that in type 2 diabetes. Cell. 1999. 96:329-339.
    View this article via: PubMed CrossRef Google Scholar
44. Postic, C, et al. Dual roles for glucokinase in glucose homeostasis as determined by liver and pancreatc beta cell-specific gene knock-outs using Cre recombinase. J. Biol. Chem. 1999. 274:305-315.
    View this article via: PubMed CrossRef Google Scholar
45. Zander, M, Madsbad, S, Madsen, JL, Holst, JJ. Effect of 6-week course of glucagon-like peptide 1 on glycaemic control, insulin sensitivity, and beta-cell function in type 2 diabetes: a parallel-group study. Lancet. 2002. 359:824-830.
    View this article via: PubMed CrossRef Google Scholar
46. Rui, L, Yuan, M, Frantz, D, Shoelson, S, White, MF. SOCS-1 and SOCS-3 block insulin signaling by ubiquitin-mediated degradation of IRS1 and IRS2. J. Biol. Chem. 2002. 277:42394-42398.
    View this article via: PubMed CrossRef Google Scholar
47. Shimomura, I, et al. Decreased IRS-2 and increased SREBP-1c lead to mixed insulin resistance and sensitivity in livers of lipodystrophic and ob/ob mice. Mol. Cell. 2000. 6:77-86.
    View this article via: PubMed CrossRef Google Scholar
48. Obici, S, Feng, Z, Karkanias, G, Baskin, DG, Rossetti, L. Decreasing hypothalamic insulin receptors causes hyperphagia and insulin resistance in rats. Nat. Neurosci. 2002. 5:566-572.
    View this article via: PubMed CrossRef Google Scholar
49. Kulkarni, RN, et al. beta-cell-specific deletion of the Igf1 receptor leads to hyperinsulinemia and glucose intolerance but does not alter beta-cell mass. Nat. Genet. 2002. 31:111-115.
    View this article via: PubMed Google Scholar
50. Niswender, KD, et al. Insulin activation of phosphatidylinositol 3-kinase in the hypothalamic arcuate nucleus: a key mediator of insulin-induced anorexia. Diabetes. 2003. 52:227-231.
    View this article via: PubMed CrossRef Google Scholar
51. Wood, SA, Allen, ND, Rossant, J, Auerbach, A, Nagy, A. Non-injection methods for the production of embryonic stem cell-embryo chimaeras. Nature. 1993. 365:87-89.
    View this article via: PubMed CrossRef Google Scholar
52. Nagy, TR, Clair, AL. Precision and accuracy of dual-energy X-ray absorptiometry for determining in vivo body composition of mice. Obes. Res. 2000. 8:392-398.
    View this article via: PubMed CrossRef Google Scholar
53. Fredrich, RC, et al. Leptin levels reflect body lipid content in mice: evidence for diet-induced resistance to leptin action. Nat. Med. 1995. 1:1311-1314.
    View this article via: PubMed CrossRef Google Scholar
54. Kim, JK, et al. Tissue-specific over-expression of lipoprotein lipase causes tissue-specific insulin resistance. Proc. Natl. Acad. Sci. U. S. A. 2001. 98:7522-7527.
    View this article via: PubMed CrossRef Google Scholar
55. Thevenaz, P, Ruttimann, UE, Unser, M. A pyramid approach to subpixel registration based on intentisy. IEEE Trans. Image Process. 1998. 7:27-41.
    View this article via: CrossRef Google Scholar
56. MIT Biopolymers Laboratory. http://web.mit.edu/biopolymers/www/.
57. Li, C, Wong, WH. Model-based analysis of oligonucleotide arrays: expression index computation and outlier detection. Proc. Natl. Acad. Sci. U. S. A. 2001. 98:31-36.
    View this article via: PubMed CrossRef Google Scholar