Differential Effects of p27 in Regulation of β-Cell Mass During Development, Neonatal Period, and Adult Life

p27^Kip1-deficient mice are in C57/B6 background and have been described previously (13). Experiments were performed in 3- to 4-month-old female mice. RIP7-rtTA mice express the reverse-tetracycline transactivator under the control of the rat insulin II gene and are in C57Bl/6 (B6)/CBA mice (14). These mice were kindly provided by Gerhard Christofori and Shimon Efrat. The tetO^p27Kip mice have been previously described and contain p27 under the regulation of tetracycline-responsive element. RIP7-rtTA/tetO^p27Kip mice (DT) were obtained by crossing hemizygous RIP7-rtTA with hemizygous tetO^p27Kip mice. Control and experimental animals were on comparably mixed background. All procedures were performed in accordance with the Washington University Animal Studies Committee.

The studies performed to determine the role of p27 in β-cells from adult mice were performed in 8-week-old male mice. Mice on doxycycline treatment were given 2 mg/ml doxycycline in a 5% sucrose solution instead of drinking water for 18 weeks. To overexpress p27 during development, slow-release doxycycline pellets (60-day release; 0.7 mg doxycycline per day) (Innovative Research of America, Sarasota, FL) were implanted subcutaneously using a trochar as described by the manufacturer. Females were administered doxycycline pellets before mating so that the transgenic offspring were exposed to the drug in utero. Induction of p27 in the early neonatal period was achieved by adding doxycycline to the drinking water after birth. To induce p27 in the neonatal period only, doxycycline in the drinking water was administered to neonates from females that did not have the doxycycline pellet implanted. The same proportion of males and females were included in each experimental group.

Islet isolation was accomplished by collagenase digestion as described previously (15). After isolation, islets were handpicked and lysed in lysis buffer. The following antibodies were used: p27 (Cell Signaling, Beverly, MA) and α-tubulin (Sigma, St. Louis, MO). Protein obtained from islets from RIP7-rtTA/tetO^p27Kip mice (DT) and single transgenic (ST) mice, (50 μg) (∼100 islets) on chronic doxycycline treatment were used for each experiment. Membranes containing islet lysates were incubated for 24 h with primary antibodies at the dilutions recommended by the manufacturer. Immunoblotting experiments were performed at least three times in duplicate.

Pancreatic tissue was fixed overnight in 3.7% formalin solution and embedded in paraffin using standard techniques. Immunohistochemistry for insulin (Sigma), bromodeoxyuridine (BrdU) (Amersham, Piscataway, NJ), Ki67 (Novocastra, Burlingame, CA), and p27 (Cell Signaling) was performed in 5-μm sections. Assessment of apoptosis was performed by immunofluorescence using cleaved caspase 3 antibodies (Cell Signaling) using the dilution recommended by the manufacturer.

β-Cell mass was calculated by point-counting morphometry from five insulin-stained sections (4 μm) separated by 200 μm using the BQ Classic98 MR software package (Bioquant, Nashville, TN) as described previously (16). Proliferation was assessed in insulin- and BrdU-stained sections from p27-deficient mice injected with BrdU for 6 h as previously described (15). Insulin- and Ki67-stained (Novocastra) sections were also used to assess proliferative rate when indicated. At least 2,000 insulin-stained cells were counted for each animal. To calculate the mean size of the individual β-cells, the β-cell area was divided by the number of β-cell nuclei in the covered area. The β-cell size measurements were performed in at least 100 islets from insulin-stained sections.

Fasting blood samples were obtained from the tail vein after overnight or 6 h fasting as indicated in the corresponding figure legend. Glucose was measured on whole blood using AccuChek II glucometer (Roche Diagnostics, Indianapolis, IN). Plasma insulin levels were determined using a rat insulin enzyme-linked immunosorbent assay kit (Linco, St. Louis, MO). Glucose tolerance tests were performed in 18-h fasted animals by injecting glucose (2 mg/g i.p.) as described previously (15). Insulin tolerance tests were done in 4-h fasted mice followed by glucose measurements at 30, 60, and 120 min after intraperitoneal insulin injection using 0.75 units/kg.

All values are expressed as means ± SE. Paired Student’s t test was used for all other comparisons. Differences were considered statistically significant at a P value <0.05.

To begin to elucidate the role of endogenous p27^kip in glucose homeostasis, we used p27-deficient mice (p27^−/−) (13). As previously described, p27^−/− mice were heavier compared with heterozygous or wild-type controls (Fig. 1A). Blood glucose levels after 6 h fasting were significantly lower in 3-month-old p27^−/− mice (Fig. 1B). The decreased glucose level in p27^−/− mice was associated with hyperinsulinemia in 6-h fasted mice (Fig. 1C). To determine glucose homeostasis after a glucose challenge, we performed an intraperitoneal glucose tolerance test (Fig. 1C). No difference in glucose levels was observed after an overnight fast (Fig. 1D). Blood glucose in p27^−/− mice was significantly reduced at 30 and 60 min after glucose injections (P < 0.05; Fig. 1D). The results of these experiments suggest that p27 deficiency results in increased weight and improved glucose tolerance.

Knowing that insulin tolerance tests in p27^−/− mice were unaltered (data not shown), we reasoned that the hyperinsulinemia observed could result from alterations in β-cell mass. Islet mass was augmented by twofold in p27^−/− mice compared with wild-type mice (Fig. 2A). The size of individual β-cells was not altered in p27^−/− mice (data not shown). To determine whether the alteration in β-cell mass results from increased proliferation, we assessed the rate of BrdU incorporation in p27-deficient mice and controls. The proliferative rate observed in p27^−/− mice was increased compared with that of wild-type mice (P < 0.05; Fig. 2B). These results indicate that p27 may function as a negative regulator of β-cell proliferation. However, it is unclear whether the effect of p27 occurs during development or postnatally.

Overexpression of p27 using the rat insulin promoter induces diabetes (12). To determine the importance of p27 in regulation of β-cell proliferation during the different stages required to achieve adult β-cell mass, we used a doxycycline-inducible system. These mice were generated by crossing mice expressing the tetracycline reverse transactivator under the control of the rat insulin promoter (RIP-rTTA) with mice expressing p27 under the control of the tetracycline operator (tetO^p27) (DT). The control group included mice containing one of the transgenes (ST). The mice with one of the transgenes (RIP-rTTA or tetO^p27) were combined for the analysis since no difference in glucose homeostasis or weight was observed between the groups. The metabolic phenotype of the ST was similar to the wild-type mice (data not shown). To induce p27 in β-cells from adult mice, 8-week-old mice were given doxycycline in their drinking water for 18 weeks. Immunoblotting for p27 in islets from mice chronically receiving doxycycline showed a two-band pattern representing the phosphorylated and nonphosphorylated forms of p27 as described (12) (Fig. 3A). p27 expression was increased by 2.8-fold in islets from DT mice chronically receiving doxycycline (Fig. 3B). These results were confirmed by immunostaining for p27 in pancreatic sections from mice after 18 weeks on doxycycline (data not shown).

To begin to elucidate the metabolic phenotype of these mice, we first measured glucose and insulin levels at different time points during doxycycline treatment. Similar weight gain was observed in all groups after doxycycline treatment (data not shown). Glucose levels in DT mice during the doxycycline treatment were not different than those obtained from ST mice (Fig. 3C). Serum concentration of insulin was similar in DT and ST mice during the experiment (Fig. 3D). Complementary assessment of carbohydrate metabolism was performed by glucose tolerance tests. Glucose tolerance tests before and 6, 12, and 16 weeks after (data not shown for 6 and 12 weeks) doxycycline treatment were similar among the groups (Fig. 3E and F). Taken together, these experiments suggest that overexpression of p27 in β-cells from adult mice has no effect on carbohydrate metabolism.

Significant reductions on β-cell mass should be achieved before detecting glucose intolerance. To determine whether the lack of metabolic changes were associated with alterations in islet morphology, we measured β-cell mass and proliferation (Fig. 4). Islet mass showed no difference in the DT mice compared with ST mice (Fig. 4A). The size of individual β-cells, estimated by measurement of β-cell area, was not affected (data not shown). The examination of the β-cell proliferation by BrdU staining in DT mice after 18 weeks of doxycycline treatment was no different than that of ST mice (Fig. 4B). These results indicate that overexpression of p27 in the adult β-cell does not alter β-cell mass and proliferation.

Based on the lack of phenotype found in our experiments and the severe diabetes observed by overexpressing p27 in β-cells using the insulin promoter (12), we hypothesized that p27 could play a more important role during development and/or the early postnatal period. To test this, we assessed the effect of p27 overexpression by doxycycline treatment during pregnancy and after birth. This was achieved by implanting doxycycline pellets subcutaneously in nonpregnant females before mating, followed by doxycycline in the water after the pups were born. Interestingly, 3-week-old DT mice showed significant hyperglycemia and hypoinsulinemia after 6 h of fasting (Fig. 5A and B). Glucose tolerance tests in 4-week-old mice demonstrated that DT mice were hyperglycemic at 30 and 60 min after glucose injections compared with ST mice (P < 0.05; Fig. 5C). The hyperglycemic phenotype was more severe at 8 weeks of doxycycline treatment (P < 0.05; Fig. 5C). Insulin sensitivity assessed by insulin tolerance tests was not different between DT mice and controls (data not shown). As shown in Fig. 5E and F, the hyperglycemic phenotype observed in DT mice resulted from reduced β-cell mass and decreased proliferation. Assessment of apoptosis by cleaved caspase 3 staining demonstrated that apoptotic rate was similar in DT and control mice (data not shown). These studies suggest the phenotype of mice overexpressing p27 during β-cell development and early postnatal life is a more severe phenotype than that observed after overexpression of β-cells from adult mice. These data also indicate that p27 plays a more important role during proliferative conditions such as those observed in development and the early postnatal period.

To determine the role of p27 in β-cell proliferation during the early postnatal period and the contribution of this stage to the hyperglycemic phenotype described above, we administered doxycycline to newborn mice. After 6 h of fasting, 3-week-old DT mice exhibited higher glucose levels compared with ST mice (P < 0.05; Fig. 6A). Glucose tolerance tests performed in 3-week-old mice showed that glucose levels in DT mice were significantly higher at 30 and 60 min after glucose injections (Fig. 6B). After 8 weeks of age, DT mice exhibited severe hyperglycemia at 30, 60, and 120 min after glucose injections, suggesting that the severity of the phenotype is progressive (Fig. 6C). The results of these experiments suggest that p27 is an important regulator of β-cell cycle during the proliferative wave in the early postnatal period.

To determine whether the metabolic and/or the β-cell mass phenotype was reversible, we followed mice after discontinuation of doxycycline treatment. For these studies, DT and ST mice were treated with doxycycline during development and postnatally for 4 weeks. At this time, doxycycline treatment was discontinued in half of the DT mice and all the experimental groups followed for 4 more weeks. As shown previously, DT mice treated with doxycycline during β-cell development and 4 weeks after birth exhibited fasting hyperglycemia and glucose intolerance assessed by intraperitoneal glucose tolerance tests (Fig. 7A and B). Compared with ST mice, glucose values after 6 h fasting were higher in DT mice with doxycycline and DT mice in which doxycycline was discontinued for 2 weeks (Fig. 7C). In contrast to the glucose tolerance results observed before discontinuation of doxycycline treatment, DT mice that were not receiving doxycycline for 2 weeks showed significant hyperglycemia only at 30 min after glucose load (Fig. 7D). Three weeks after discontinuation of doxycycline, fasting glucose in DT mice was similar to that of ST mice (Fig. 7E). Glucose tolerance in DT mice, in which doxycycline was discontinued for 3 weeks, showed almost complete recovery of hyperglycemia (Fig. 7F). Fasting glucose and glucose tolerance in DT mice that were off doxycycline for 4 weeks were not different compared with that of ST control mice (Fig. 7G and H). These data suggest that the abnormalities in carbohydrate metabolism induced by overexpression of p27 during β-cell development and postnatally are reversible.

To determine whether the correction of glucose intolerance after removal of doxycycline treatment resulted from recovery of β-cell mass, we performed islet morphometry in the mice used in Fig. 7. Mice treated with doxycycline exhibited reduced β-cell mass, confirming the results showed in Fig. 5E. More importantly, DT mice in which doxycycline treatment was discontinued for 4 weeks exhibited β-cell mass, which was comparable to that of controls (Fig. 7A). The assessment of β-cell proliferation showed that the proliferative rate in mice that received doxycycline throughout the experiment was lower than that of control ST mice (Fig. 8). Our findings imply that the proliferative rate in DT mice, which did not received doxycycline for the last 4 weeks of the experiment, was intermediate but not statistically different than that of the ST controls or DT mice treated with doxycycline (Fig. 8). These experiments suggest that the rapid improvement of the metabolic phenotype results from restoration of β-cell mass.

The result of the current studies assessed the role of p27 in the regulation of β-cell mass by using p27-null mice and doxycycline-inducible mice overexpressing p27 in β-cells. The present work shows that p27 deficiency results in increased β-cell proliferation and mass and improved glucose tolerance. Doxycycline-induced overexpression of p27 in β-cells from adult mice showed normal glucose tolerance and comparable β-cell mass and proliferation. In contrast, overexpression of p27 during β-cell development and the first 4 weeks of postnatal life resulted in glucose intolerance and decreased β-cell mass and proliferation. This same phenotype was also observed by overexpressing p27 only after birth, suggesting that p27 is critical in determining adult β-cell mass. Taken together, these studies suggest that p27 is a major determinant for β-cell cycle and mass and that the major effect of this CKI occurs during β-cell development and the early wave of proliferation that occurs during the first month of life. This work also implies that p27 could be a critical regulatory molecule under conditions that require β-cell expansion, such as adaptation to insulin resistance.

Recent data from several laboratories suggest that modulation of the cyclin D–cdk4 complex activity is a major step that controls cell cycle progression, proliferation, and maintenance of β-cell mass (1–4). Moreover, overexpression of cdk4 by adenoviral transfer in human islets increases proliferation in vitro (5). The cell cycle is negatively regulated by two families of CKIs. The INK4 CKIs, p16^INK4a, p15^INK5b, and p19^INK4d, have been shown to play an important role in regulation of β-cell proliferation (1,17,18). The Cip/Kip CKIs primarily inhibit Cdk2 but also have a role in assembling cdk4/cyclin D complex. Of the CKIs, p27^Kip1 is thought to be the primary modulator of the proliferative status in most cell types, where it functions to induce and maintain the quiescent state (13,19,20). Recent studies showed that p27 deficiency prevented the development of overt hyperglycemia in both Irs2^−/− and Lepr^−/− mice by increased islet mass and serum insulin concentration (12). Our studies extend these data to the β-cell phenotype of p27-deficient mice in physiological conditions. In contrast to a previous report (21), our data showed that these mice have improved glucose tolerance. This is most likely the result of different genetic backgrounds of the mice and a lower glucose dose (1 mg/g) used for the glucose tolerance test. Our findings imply that the improved glucose tolerance results from increased β-cell mass and proliferation. Since these mice lack p27 during development, it is unclear if the major effect of p27 in regulation of β-cell mass and proliferation occurs early during pancreas development or if there is an effect of this CKI in driving cell cycle entry in mature β-cells. This could explain the fact that a 30% increase in proliferation in p27-null mice results in a twofold change in β-cell mass. In contrast to the striking changes in β-cell mass in mice containing a mutated cdk4 that is resistant to inhibition by p16, the phenotype of p27-deficient mice is rather mild. These experiments suggest that p27 deficiency does not result in unrestricted cell cycle progression and as shown by Franklin et al. (17) and Karnik et al. (18); deficiency in other CKIs are necessary to induce islet hyperplasia.

Based on the observations by Uchida et al. (12) and our results obtained in the p27-deficient mice, it was reasonable to expect that overexpressing p27 in β-cells from adult mice would result in diabetes. Surprisingly, overexpression of p27 in 8-week-old mice for 18 weeks had no effect on glucose tolerance or β-cell mass. The lack of a phenotype when p27 was expressed in adult mice is most likely explained by the low rate of proliferation and minimal turnover of adult β-cells as reported recently (22). Taking into consideration that the reported half-life of adult β-cells is 40 days (23), we can assume that 18 weeks would be enough time to see a phenotype, although it is possible that overexpression of p27 for a longer period of time would lead to a detectable phenotype. Our experiments expressing p27 during embryonic stages and early postnatal life suggest that p27 is critical to achieve a final functional β-cell mass in adult mice and that p27 acts as a key brake for progression through cell cycle induced by mitogenic signals. These data also confirm that the proliferative wave during the early postnatal period is crucial for achieving a final β-cell mass. Finally, the restoration of glucose tolerance and β-cell mass after discontinuing doxycycline is another example of the regenerative potential of β-cells. Interestingly, the proliferative rate measured after achieving normoglycemia in mice in which doxycycline was discontinued for 4 weeks was intermediate between that of DT mice receiving doxycycline and ST control mice. A potential explanation for the intermediate proliferative rates observed after discontinuation of doxycycline is that these measurements were performed after achieving normoglycemia and normal β-cell mass. It is possible that higher proliferative rates would have been observed at earlier times during the recovery process. Less likely, it is interesting to speculate that neogenesis could have contributed to the recovery of β-cell mass.

The results of these genetic studies suggest that the use of inducible systems of CKIs could generate information that cannot be obtained by traditional transgenesis and suggest that CKIs can have diverse effects during the different stages necessary to achieve adult β-cell mass. Another important conclusion of these experiments is that the major effect of p27 in regulation of β-cell cycle in physiological conditions occurs only during embryogenesis and early postneonatal period. Our results imply that p27 is a major regulator of cell cycle in conditions that require expansion of β-cell mass and that this CKI could contribute to the abnormal adaptation of β-cells observed in type 2 diabetes. The lack of unrestricted proliferation in p27-deficient mice and the importance of this CKI in improving the adaptation to insulin resistance in genetic models make this molecule an attractive target for pharmacological modulation.

This work was supported in part by a Junior Faculty Award and Career Development Award from the American Diabetes Association (to E.B.-M.).

The authors acknowledge the support of the Washington University Diabetes Research and Training Center for insulin measurements and histology and the Washington University Digestive Diseases Research Core Center grant morphology core.