Islet-Sparing Effects of Protein Tyrosine Phosphatase-1b Deficiency Delays Onset of Diabetes in IRS2 Knockout Mice

Irs2^+/− mice on a mixed genetic background (C57BL/6 × 129Sv) were intercrossed with Ptp1b^−/− mice on a similar genetic mixed background (also C57BL/6 × 129Sv) to yield Irs2^+/−::Ptp1b^+/− mice, which were then intercrossed to yield a subsequent generation of doubly heterozygotic mice. Double heterozygotes were then intercrossed to yield Irs2^+/− mice and Irs2^+/−::Ptp1b^−/− mice, which were then intercrossed to yield the genotypes of this study: wild-type mice and Ptp1b^−/−, Irs2^−/−, and Irs2^−/−::Ptp1b^−/− mice. Irs1^+/− mice were intercrossed with Ptp1b^−/− mice in a similar fashion. Mice were maintained at a nonbarrier animal facility in the Joslin Diabetes Center and fed Mouse Diet 5020 9F (9% fat calculated by weight, 21.6% fat calculated by kilocalories) (PMI Nutrition International, Richmond, IN). Genotyping of animals was done by PCR, as described previously (5,19). Male mice were used to characterize glucose homeostasis phenotypes and pancreatic pathology. Random-fed glucose and insulin measurements were performed as previously described. Intraperitoneal glucose tolerance tests were performed on mice fasted for 15–16 h with 2 g d-glucose per kilogram body weight as described previously (23). Intraperitoneal insulin tolerance tests were performed on 6-h fasted mice with 0.75 units (6-week-old mice) or 1.5 units/kg (15-week-old mice), similar to previously described methods (23). Area under the curve (AUC) analysis was calculated with trapezoid rule algorithms for individual glucose tolerance tests and reported as grams per deciliter per minute.

All results are reported as mean ± SE for equivalent groups. To minimize the effect of statistical variance, statistical comparisons were only performed between Irs2^−/− mice and Irs2^−/−::Ptp1b^−/− mice (and not other possible combinations) and were compared with independent t tests (unpaired and two-tailed) reported as P values.

Immunohistochemical localization of antigens and double-label immunohistochemistry were performed similarly to previously described methods (23). Pancreas samples were dissected from fed mice and fixed with Bouin’s solution overnight. Five-micrometer longitudinal sections of paraffin blocks were rehydrated with xylene followed by decreasing concentrations of ethanol, microwaved in 0.01 mol/l sodium citrate (pH 6.0) for 20 min and permeabilized with 1% Triton X-100 in PBS before primary antisera incubation. Guinea pig anti-insulin, rabbit anti-glucagon antibodies (Zymed Laboratories, South San Francisco, CA) and goat anti-Pdx1 (Santa Cruz Biotech, Santa Cruz, CA) were used as primary antibodies. Secondary antibodies were labeled with fluoroscein isothiocyanate or rhodamine (Jackson ImmunoResearch Laboratories, West Grove, PA). Adult β-cell area was measured by acquiring images at ×10 from 14–20 adjacent nonoverlapping images of insulin sections from 3-month-old male mice, at least six animals per genotype with a Zeiss Axiovert microscope (Carl Zeiss MicroImaging, Thornwood, NY). Images were analyzed for area with Improvision Open Lab software density slice software (Improvision Scientific Imaging, Lexington, MA). Results of β-cell quantification are expressed as the percentage of the total surveyed area containing cells positive for insulin. Islet density and size were calculated from captured insulin-stained images. Results of Irs2^−/− mice and Irs2^−/−::Ptp1b^−/− mice were compared with independent t tests (unpaired two tailed) and reported as P values.

We intercrossed Irs2^+/− mice with Ptp1b1^−/− mice to generate compound heterozygotes (Irs2^+/−::Ptp1b^+/−), which were backcrossed to produce the nine expected genotypes at the expected Mendelian frequencies (data not shown). Male wild-type, Irs2^−/−, Ptp1b^−/−, and Irs2^−/−::Ptp1b^−/− mice were selected for this investigation to focus on the effect of Ptp1b disruption upon glucose tolerance and diabetes in Irs2^−/− mice (4–6,18,19). We fed all of the mice a regular diet containing 9% fat (calculated by weight) to reduce dietary fat-induced differences in weight that occur between wild-type mice and Ptp1b^−/− mice (18,19). Consistent with our previous findings, a weight difference was barely perceptible between wild-type and Ptp1b^−/− mice at 6 weeks of age, whereas Ptp1b^−/− mice weighed 20% less than wild type at 15 weeks (Fig. 1). By contrast, Irs2^−/−::Ptp1b^−/− mice weighed 20% less than Irs2^−/− mice at 6 weeks, whereas the difference was almost imperceptible by 15 weeks of age, when Irs2^−/− mice develop severe diabetes (Fig. 1).

As previously shown, Irs2^−/− mice displayed fasting and random-fed hyperglycemia between 6 and 15 weeks of age (Table 1). Compared with wild-type mice, Ptp1b^−/− mice displayed normal fasting blood glucose at 6 weeks and slightly lower values at 15 weeks, whereas no differences were detected during random feeding (Table 1). Ptp1b disruption in Irs2^−/− mice reduced fasting and random-fed glucose at 6 and 15 weeks (Table 1). However at 15 weeks Irs2^−/−::Ptp1b^−/− mice displayed fasting hyperglycemia (Table 1). Still, some Irs2^−/−::Ptp1b^−/− mice survived for as long as 9 months, whereas Irs2^−/− mice were severely diabetic and invariably died near 15 weeks (data not shown). Thus, disruption of Ptp1b improved glucose tolerance, delayed the onset of diabetes, and prolonged survival in the Irs2^−/− mice.

Next, we assessed glucose tolerance during intraperitoneal glucose injections at 6 and 15 weeks of age. Compared with wild-type mice at 6 weeks, Ptp1b^−/− mice displayed nearly identical glucose tolerance, with indistinguishable AUCs (Fig. 2A and C). Ptp1b^−/− mice displayed greatly improved glucose tolerance by 15 weeks, consistent with previous findings (18,19). Irs2^−/− mice were severely glucose intolerant at 6 and 15 weeks (Fig. 2). Although Ptp1b disruption consistently improved glucose tolerance of the Irs2^−/− mice, Irs2^−/−::Ptp1b^−/− mice still displayed significant glucose intolerance at 6 and 15 weeks (Fig. 2).

To estimate peripheral insulin sensitivity, fasted mice were tested by intraperitoneal injection of insulin (0.75 units/kg body wt insulin at 6 weeks and 1.5 units/kg at 15 weeks), and glucose levels were measured over the next 60 min. Unlike previous insulin tolerance test results of Ptp1b^−/− mice on a similar diet at 10–14 weeks, our Ptp1b^−/− mice displayed a normal response to injected insulin at both ages compared with wild-type controls (Fig. 3) (18,19). The divergence in results may reflect differences in genetic background or testing conditions (in this study we used 1.5 units/kg insulin instead of 0.75 units/kg). Compared with wild-type mice, insulin weakly reduced blood glucose in Irs2^−/− mice at 6 weeks of age and had no effect at 15 weeks (Fig. 3). Ptp1b deficiency normalized insulin sensitivity in young Irs2^−/−::Ptp1b^−/− mice (Fig. 3A); however, by 15 weeks, peripheral insulin sensitivity was diminished in Irs2^−/−::Ptp1b^−/− mice (Fig. 3B). Thus, in the absence of Irs2, deletion of Ptp1b increased peripheral insulin sensitivity in young mice but had minimal effects as the mice aged.

Irs2 mediates peripheral insulin action and plays an essential role in pancreatic β-cell function and survival. Without Irs2, β-cell mass decreases until the mice die between 14 and 17 weeks (5). Genetic manipulation to rescue β-cell function in Irs2^−/− mice, by upregulation of Irs2 or Pdx1 in β-cells or by suppression of Foxo1, promotes islet growth, normalizes glucose tolerance, and prevents diabetes (23–25). Because Irs2^−/−::Ptp1b^−/− mice survive beyond 16 weeks, we investigated their insulin levels and β-cell mass. Compared with wild-type mice, fasting and fed serum insulin levels in Ptp1b^−/− mice were indistinguishable at 6 and 15 weeks, consistent with similar peripheral insulin sensitivity indicated by the glucose and insulin tolerance tests, but different from previous results where Ptp1b^−/− mice had decreased fed serum insulin values (18,19). This discrepancy could be due to the larger experimental study groups in our previous report or to differences in genetic background. Fasting insulin was elevated 10-fold in both Irs2^−/− and Irs2^−/−::Ptp1b^−/− mice, indicating compensation for peripheral insulin resistance (Table 1). These results further demonstrate that loss of Ptp1b failed to normalize completely peripheral insulin sensitivity in the Irs2^−/−::Ptp1b^−/− mice (Table 1). Random-fed serum insulin levels were quite variable, and while the average serum insulin was higher in Irs2^−/−::Ptp1b^−/− mice than in Irs2^−/− mice at both 6 weeks and 15 weeks, these differences did not reach statistical significance.

Next we investigated pancreatic histopathology to establish the effect of Ptp1b disruption on islet size at 15 weeks. Previous studies show that islet mass is inversely related to peripheral insulin sensitivity (1). We found that pancreatic cross-sectional β-cell area was reduced in Ptp1b^−/− mice, consistent with their increased peripheral insulin receptor signaling from Ptp1b disruption (Fig. 4B and C). As previously shown, β-cells were barely detected in Irs2^−/− islets, as β-cell area and islet size in pancreatic sections were decreased (even though Irs2^−/− mice are severely insulin resistant) (Fig. 4). In contrast to findings in wild-type mice, Ptp1b deletion increased insulin and Pdx1 immunostaining as well as β-cell area and islet size in Irs2^−/−::Ptp1b^−/− mice. These findings suggest that in the absence of Irs2, Ptp1b deficiency improved β-cell function, possibly through increased insulin receptor or Irs1 signaling (Fig. 4). Although Ptp1b deletion promoted compensatory islet growth and function in Irs2^−/−::Ptp1b^−/− mice, islet function was still insufficient and diabetes developed at 15 weeks.

Our results illustrate that superimposed Ptp1b disruption in Irs2^−/− mice improves glucose tolerance by partially increasing peripheral insulin sensitivity in young mice and promoting some compensatory β-cell function. These compensatory effects are insufficient to completely restore islet function, however, and diabetes eventually occurs in Irs2^−/−::Ptp1b^−/− mice.

Ptp1b normally inhibits insulin signaling by dephosphorylating the insulin receptor as it passes by the endoplasmic reticulum during recycling (15). Without Ptp1b, the rate of receptor inactivation is reduced, which increases insulin sensitivity. In this study Ptp1b disruption in wild-type mice fed a low-fat diet (9%) resulted in mild fasting hypoglycemia, greatly improved glucose tolerance, and slight increases in peripheral insulin sensitivity by 15 weeks of age (Figs. 1–3). In this study, we found that superimposed Ptp1b disruption in Irs2^−/− mice almost completely normalized peripheral insulin sensitivity at 6 weeks of age (Fig. 3A). However, 6-week-old Irs2^−/− and Irs2^−/−::Ptp1b^−/− mice displayed fasting serum insulin values that were equally hyperinsulinemic, revealing persistent insulin resistance that is not entirely corrected by the absence of Ptp1b (Table 1). Supporting this idea, by 15 weeks of age Ptp1b disruption had little effect to rescue peripheral insulin resistance of Irs2^−/−::Ptp1b^−/− mice (Fig. 3B). Thus, we conclude that Ptp1b disruption partially improves peripheral insulin sensitivity in Irs2^−/−::Ptp1b^−/− mice.

Both Irs2 and Ptp1b appear to be critical regulators of body weight. Irs2 mediates antiobesity signals that are observed in healthy female Irs2^−/− mice, possibly caused by impaired hypothalamic leptin signaling (26). In contrast, Ptp1b appears to antagonize metabolic signals that promote normal body weight: Ptp1b^−/− mice are resistant to obesity and can partially rescue loss of leptin signaling (16–19). Supporting previous findings, we found that Ptp1b disruption modestly reduced body weight in wild-type mice by 15 weeks (19). Ptp1b disruption had a profound effect in the Irs2^−/− background, reducing body weight in 6-week-old Irs2^−/−::Ptp1b^−/− mice 20% below Irs2^−/− mice (Fig. 1). However, as Irs2^−/− mice developed frank diabetes, this difference was lost and Irs2^−/−::Ptp1b^−/− mice and Irs2^−/− mice had equivalent body weights (Fig. 1). Future studies will be required to determine the reason for the greater decrease in body weights of young Irs2^−/−::Ptp1b^−/− mice compared with Ptp1b^−/− mice.

Ptp1b deletion had differential effects on islet formation in wild-type and Irs2^−/− mice. β-Cell area decreased in Ptp1b^−/− mice consistent with improved glucose tolerance, fasting hypoglycemia, and slightly improved peripheral insulin sensitivity. This inverse relation between insulin sensitivity and insulin secretion is well established in rodents and humans (1). However, the relation between peripheral insulin action and islet mass is entirely disrupted in Irs2^−/− mice because β-cell area decreases as peripheral insulin sensitivity decreases. Relative to Irs2^−/− mice, peripheral insulin action improves in the Irs2^−/−::Ptp1b^−/− mice and β-cell area increased as the islet-trophic signals are partially restored. However, this compensatory β-cell function is eventually insufficient to meet insulin requirements, and diabetes occurs in Irs2^−/−::Ptp1b^−/− mice.

Although Irs2^−/− mice are severely resistant to insulin, several genetic maniupulations can prevent diabetes in Irs2^−/− mice by promoting islet expansion and restoring β-cell function. Transgenic expression of Irs2 itself in pancreatic and β-cells prevents diabetes in Irs2^−/− mice (25). Furthermore, changes that mimic the effects of Irs2 signaling restore β-cell function in Irs2^−/− mice. For example, genetic downregulation of the forkhead transcription factor Foxo1 restores β-cell function and prevents diabetes in Irs2^−/− mice (24), as does upregulation of Pdx1 (23). These results also suggest that β-cell failure, and not peripheral insulin resistance, is the major cause of diabetes in Irs2^−/− mice. Indeed, we found that islet Pdx1 immunostaining was increased in pancreatic sections from Irs2^−/−::Ptp1b^−/− mice compared with Irs2^−/− mice. Genetic background also appears to play an important role to modify islet function in Irs2^−/− mice as Irs2^−/− mice on different genetic backgrounds with higher basal Pdx1 expression (CBA and C57BL/6) have milder defects in islet function (27,28).

Previous work suggests that the type 1 IGF receptor (IGF1R) plays an important role in β-cell growth and survival; however, the exact location where IGF1R action is required is not clear because β-cell-specific Igf1r disruption does not reduce β-cell mass (29). In contrast, β-cell-specific disruption of the insulin receptor results in decreased β-cell mass as the mice age (30). Although Ptp1b has never been demonstrated to be expressed in β-cells, our RT-PCR analysis suggests that Ptp1b mRNA is expressed in isolated islets (data not shown). The disruption of Ptp1b could therefore increase insulin receptor and/or Igf1 receptor islet signaling, although β-cell expansion via these pathways is still not sufficient to fully restore β-cell mass of Irs2^−/−::Ptp1b^−/− mice. In the absence of Irs2, Irs1 signals presumably would be used to mediate insulin receptor signaling. Previous experiments show that Irs1 promotes partial β-cell function in the Irs2^−/− mice, but it is not sufficient for long-term β-cell function (7). We cannot exclude the possibility that in the absence of Ptp1b, other Irs proteins in islets may also contribute to β-cell survival or proliferation. Thus, glucose tolerance is improved in the Irs2^−/−::Ptp1b^−/− mice and diabetes is delayed relative to Irs2^−/− mice, but compensatory β-cell expansion is not sufficient to completely prevent diabetes. Nevertheless, in human patients with type 2 diabetes PTP1b might be expected to promote both peripheral insulin sensitivity and β-cell function, as the disease does not result from a complete lack of IRS2 (31).

Even though Ptp1b^−/− mice display improved glucose tolerance, they do not develop tumors. In contrast, PTEN^+/− mice develop tumors from hyperactivity of the PI 3-kinase cascade (32,33). The different outcome might be related to the specificity of Ptp1b to act upon the insulin receptor and not the IGF1R receptor in the whole organism context. Supporting this idea, Ptp1b^−/− mice do not display an overgrowth phenotype. Notably, however, in fibroblasts Ptp1b may downregulate IGF1R signaling (34). Indeed, some distinct developmental roles of Irs2 that might be mediated by the IGF1R are not restored in Irs2^−/−::Ptp1b^−/− mice, including ovarian function and female fertility (data not shown). To further test the role of Ptp1b in IGF→Irs protein-related somatic growth signaling, we have also created Irs1 and Ptp1b double knockout mice. Preliminary data suggest that Ptp1b deficiency does not modify the growth phenotype of Irs1 knockout mice: Irs1^−/−::Ptp1b^−/− mice were ∼60% normal size at various ages, similar in size to Irs1^−/− mice (data not shown). These data support the hypothesis that Ptp1b has tissue-specific roles to dephosphorylate and downregulate tyrosine kinase-dependent signaling in vivo and that Ptp1b disruption does not promote general somatic growth. This finding diminishes concerns that Ptp1b inhibitors might promote tumor growth. However, an alternative explanation could be that enhanced Irs2 signaling cannot completely compensate for Irs1 deficiency. The specificity of PTP1b toward the insulin receptor and the lack of oncogenic effects from Ptp1b disruption in vivo provide further evidence that pharmacological PTP1b antagonists could be safe and effective diabetes treatments, if residual IRS2-like signals persist in β-cells of human patients with type 2 diabetes.

This work was supported by National Institutes of Health Grants to M.F.W. (DK55326, DK43808, and DK38712), to B.G.N. (CA 49152 and DK60838), and to B.B.K. (DK60839) and by Howard Hughes Medical Institute (HHMI) funds to M.F.W. J.A.K. was supported by a Lawson Wilkins Pediatric Endocrine Society Fellowship, a National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) training grant, a Juvenile Diabetes Research Foundation Fellowship, and a Charles H. Hood Foundation Child Health Research Grant.