Age-Dependent Decline in β-Cell Proliferation Restricts the Capacity of β-Cell Regeneration in Mice

We obtained 6-week-old (young) and 7- to 8-month-old (old) male C57BL/6 mice from The Jackson Laboratory. For the high-fat diet experiment, 12 young or 12 old mice were fed with either normal diet (4.4% of total calories derived from fat, 3.9 kcal/g; Harlan Teklad) or high-fat diet (55% of total calories derived from fat, 4.8 kcal/g; Harlan Teklad) for 8 weeks. For exendin-4 experiments, six young and six old mice were injected intraperitoneally with either exendin-4 (10 nmol/kg; Sigma) or PBS for 7 days. For the STZ (Sigma-Aldrich) experiment, six young and six old mice were injected with a single dose of 90 mg/kg freshly prepared STZ in citrate buffer (pH 4.5) and killed at 7 days. Bmi1^−/− mice were obtained from Maarten van Lohuizen of the Netherlands Cancer Institute. For Bmi1^−/− mice, targeted disruption of the Bmi1 allele has been described before (29). The animals were maintained by mating Bmi1^+/− males and females on a C57BL/6J background. Mice were fed ad libitum on standard diet and keep under a 12-h light/dark circle. All animal protocols were approved by the chancellor's animal research committee at the University of California, Los Angeles.

Fasting blood glucose was measured after overnight fasting. An insulin tolerance test was performed after a 6-h fast. Blood glucose was measured before intraperitoneal injection of insulin (0.75 mU/g body wt) and then 20, 40, and 60 min after injection. No anesthesia was used during the experiment. Glucose tolerance testing was performed after overnight fasting. Blood glucose levels (mg/dl) were measured before intraperitoneal injection of glucose (2 mg dextrose/g body wt) and then 15, 30, 60, and 120 min after injection.

Pancreatic tissue was processed as previously described (11). In brief, the pancreas was dissected and fixed in 4% formaldehyde before being embedded in paraffin. Then, 5-μm sections were deparaffinized and rehydrated, followed by antigen retrieval using antigen unmasking buffer (Vector Labs), and then permeabilized in 0.4% Triton X-100/Tris-buffered saline. Slides were blocked with 3% IgG-free BSA (Jackson ImmunoResearch Laboratories) and treated with anti-mouse insulin (Dako), anti-ki67 (BD Pharmingen), anti–proliferating cell nuclear antigen (PCNA; Lab Vision), anti-Bmi1 (Millipore), anti-p16^Ink4a (Santa Cruz), anti–cyclin D2 (Santa Cruz), or anti-p27 (Santa Cruz) antibody followed by fluorescein isothiocyanate– or Cy3-conjugated secondary antibodies (The Jackson Laboratory). Terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) assay was detected using an in situ cell death detection kit (Roche) according to manufacturer's instructions. Slides were mounted with Vectashield with 4′6-diamidino-2-phenylindole (Vector Labs), and images were obtained with a Leica DM6000 microscope using Openlab software (Improvision). The immunofluorescence data presented are representatives of at least five animals per group in each case.

β-Cell mass was measured as previously described (11). In brief, five to eight sections from each pancreas were stained with anti-mouse insulin antibody (Dako) and scanned by a Leica DM6000 microscope. Montage images were made by ImageJ software. The cross-sectional areas of pancreas and β-cells were determined by ImagePro software. β-Cell mass per pancreas was estimated as the product of the relative cross-sectional area of β-cells per total tissue and the weight of the pancreas and was calculated by examining pancreata from at least three animals for each genotype.

Liberase propidium iodide–purified enzyme blend for rodent islet isolation (Roche) was infused at 3.5 mg/ml into the pancreas via the bile duct. Inflated pancreata were then removed and incubated in Liberase propidium iodide for 18 min at 37°C. Islets were dissociated from the exocrine tissues by shaking vigorously several times, followed by Histopaque (Sigma-Aldrich) gradient. Islets were handpicked under a dissecting microscope and lysed by tissue extraction buffer (Invitrogen). Lysates with equal amounts of protein were resolved by SDS-PAGE, followed by transferring to polyvinylidene fluoride membrane for immunoblotting. The membranes were probed with specific antibodies against p16^Ink4a (Santa Cruz), Bmi1 (Millipore), and β-tubulin (Sigma-Aldrich). The data presented are representative of at least three experiments.

We performed chromatin immunoprecipitation (ChIP) analysis using a Millipore ChIP kit (no. 17-295) according to the manufacturer's instructions, with minor modifications. The islets (150–200 islets per group) were treated with 2% paraformaldehyde at room temperature for 20 min to cross-link the DNA with bound proteins. After washing, the islets were resuspended in SDS lysis buffer with protease inhibitors and sonicated to shear the chromatin. The chromatin was then precleared and incubated with 2–5 mg of anti–acetyl histone H3 lysine 9 (H3K9; no. 07-532; Millipore), or normal mouse IgG as a control, overnight at 4°C with agitation. After immunoprecipitation, the chromatin was harvested, the cross-links were reversed, and the DNA was purified and precipitated. The resulting DNA was quantified and served as a template for the real-time PCR, performed using a LightCycler FastStartPLUS DNA SYBR Master kit (Roche) and Light Cycler PCR equipment (Roche). The DNA enrichment after ChIP was estimated as the percentage bound-to-input ratio, determined by real-time PCR. The primers used to amplify the Ink4a/Arf locus are as follows: primer set 1: forward 5′GAGTACAGCAGCGGGAGCAT-3′, reverse 5′-GAACTTCACCAAGAAAACCCTCTCT-3′; primer set 2: forward 5′-GTCCGATCCTTTAGCGCTGTT-3′, reverse 5′-AGCCCGGACTACAGAAGAGATG-3′; primer set 3: 5′CCGGAGCCACCCATTAAACTA-3′, reverse 5′-CAAGACTTCTCAAAAATAAGACACTGAAA-3′; primer set 4: forward 5′-CCCAACACCCACTTGAGGAA-3′, reverse 5′-CAGAGGTCACAGGCATCGAA-3′; and primer set 5: (negative control, HoxC13 exon 2) forward 5′-CATTTTTCACTGATTTCCTAAGCA-3′, reverse 5′-CAATGATGTCACCCCTCCTC-3′.

Min6 cells were maintained in Dulbecco's modified Eagle's medium with 10% fetal bovine serum and transfected using Lipofectamine 2000 (Invitrogen) following the manufacturer's instructions with 1 mg of Bmi1 construct (in pcDNA3 myc-HisA, resulting in expression of NH₂-terminal myc-tagged Bmi1) and the control enhanced green fluorescent protein construct (in pcDNA3), allowing expression from the cytomegalovirus promoter in either case.

All data were expressed as the means ± SE. Mean and SE values were calculated from at least triplicates of a representative experiment. The statistical significance of differences was measured by unpaired Student's t test and confirmed by one-way ANOVA for repeat measures. P < 0.05 indicated statistical significance. The P values indicated in the graphs are from Student's t test.

To ascertain whether age plays a role in the adaptive β-cell mass expansion associated with insulin resistance, young (6-week-old) and old (7- to 8-month-old) mice were fed a high-fat diet. Insulin tolerance testing performed after 8 weeks of high-fat diet showed that both young and old mice displayed decreased insulin sensitivity compared with mice on a normal diet (Fig. 1,A). No statistically significant difference in body weight between groups was noted at the beginning of the study. The body weight of both young and old mice showed a 50% increase after 8 weeks of high-fat diet (data not shown). Glucose tolerance tests showed that blood glucose levels of old mice after 8 weeks of high-fat diet remained >400 mg/dl by the end of the 120-min testing period. By contrast, young mice after 8 weeks of high-fat diet reached baseline by the end of the 120-min testing period (Fig. 1,B). Measurement of fasting blood glucose revealed that older mice displayed fasting blood glucose levels consistently >200 mg/dl after 8 weeks of high-fat diet, indicating that these mice were unable to maintain glucose homeostasis (Fig. 1 C). In contrast, young mice fed a high-fat diet did not show fasting blood glucose levels that were significantly different from young mice fed a normal diet.

The inability of old mice, unlike young mice, after 8 weeks of high-fat diet to maintain glucose homeostasis led us to analyze whether the adaptive expansion of β-cell mass associated with insulin resistance differed in the different age-groups. Young mice after 8 weeks of high-fat diet showed a fivefold increase in β-cell mass compared with the normal-diet group. In contrast, old mice fed high-fat diet did not show any changes in β-cell mass compared with control old mice (Fig. 1 D). Moreover, pancreas weight did not show any significant increase after 8 weeks of high-fat diet in both young and old mice (data not shown). Thus, young mice, unlike old mice, have the capacity to expand β-cell mass to adapt to insulin resistance.

To analyze whether increased proliferation could account for the expansion of β-cell mass in young mice fed a high-fat diet, the number of Ki67- and PCNA-positive β-cells was quantified in mice fed high-fat diet and those fed normal diet. A 7.5-fold increase of Ki67-positive β-cells and a 2.3-fold increase of PCNA-positive β-cells was observed in the young mice fed high-fat diet compared with the control group. In contrast, the number of Ki67- and PCNA-positive β-cells did not change in old mice after a high-fat diet compared with those fed normal diet (Fig. 2,A and B). No TUNEL staining (as a measure of apoptosis) was observed in all groups (data not shown). Because decreased proliferative capacity has been correlated with increased expression of p16^Ink4a with age (15), we determined the levels of p16^Ink4a in isolated islets from mice of increasing age. As shown inFig. 2,C, p16^Ink4a levels increase with age in isolated islets. Immunohistochemistry revealed that β-cells of 1-month-old mice showed little expression of p16^Ink4a, in contrast to high levels of p16^Ink4a in 6-month-old mice (Fig. 2,D). To determine whether p16 levels played a role in the adaptive expansion of β-cell mass in response to high-fat diet, we isolated islets from young and old mice fed normal diet or high-fat diet and measured p16^ink4a levels by Western blot. In islets isolated from young mice, low levels of p16^ink4a were apparent, which decreased after 8 weeks of high-fat diet feeding. In contrast, high levels of p16^Ink4a in islets from old mice did not change with high-fat diet (Fig. 2 E). We next examined whether levels of Bmi1, a polycomb group protein that is known to regulate the Ink4a locus, were also affected. High Bmi1 levels were apparent in islets isolated from young mice, and increased levels of Bmi1 were apparent after high-fat diet. In contrast, islets isolated from old mice did not show appreciable levels of Bmi1.

Glucoincretin hormones GLP-1 and its long-acting peptide analogue exendin-4 have been suggested to induce β-cell mass expansion by increasing β-cell proliferation (30). A 1-week daily injection of exendin-4 resulted in a twofold increase in β-cell mass (30). To further investigate the effect of aging on β-cell mass expansion and β-cell proliferation, young and old mice were injected intraperitoneally with exendin-4 daily for 7 days, and β-cell mass was analyzed. After exendin-4 treatment, β-cell mass in young mice showed a twofold increment compared with control group injected with PBS. In contrast, old mice treated with exendin-4 did not show any increases in their β-cell mass (Fig. 3 A). These data suggest that as adaptive expansion associated with insulin resistance, the capacity of β-cell mass to expand in response to glucoincretin hormones challenge is severely diminished with aging.

To analyze whether increased proliferation could account for the expansion of β-cell mass in young mice treated with exendin-4, the number of Ki67- and PCNA-positive β-cells was quantified. A 4-fold increase of Ki67-positive β-cells and a 5.5-fold increase of PCNA-positive β-cells was observed in the young mice treated with exendin-4 compared with the control group. In contrast, the number of Ki67- and PCNA-positive β-cells did not change in old mice after treatment with exendin-4 compared with untreated old mice (Fig. 3,B and C). We next examined whether the failure of old mice to expand their β-cell mass on exendin-4 treatment is dependent on levels of p16^Ink4a. A marked decline of p16^Ink4a expression levels was observed in islets of young mice (Fig. 3,D and E), which is concomitant with increased Bmi1 levels (Fig. 3,H and I). In contrast, exendin-4 treatment did not lead to a decrease of p16^Ink4a levels in the β-cells in the old mice (Fig. 3,F and G). Furthermore, levels of Bmi1 (Fig. 3 J and K) did not show much difference between the exendin-4–treated mice and the PBS control group.

We first assessed whether p16^Ink4a was a specific cell cycle target of Bmi1 by analyzing expression levels of p16^Ink4a along with other cell cycle regulators involved in β-cell proliferation. Immunohistochemistry analysis of pancreatic sections from 14-day-old Bmi1^−/− mice and wild-type littermates showed p16^Ink4a was upregulated in islets of Bmi1^−/− mice (Fig. 4). Other cell cycle targets cyclin D2 and p27 were unaffected by the absence of Bmi1 (Fig. 4). To directly test whether Bmi1 regulates p16^Ink4a, we measured the levels of p16^Ink4a in islets isolated from 4-week-old wild-type and Bmi1^−/− mice. Immunoblots showed increased levels of p16^Ink4a in islets isolated from Bmi1^−/− mice (Fig. 5,A). To further confirm whether Bmi1 represses p16^Ink4a, we generated an expression vector that allowed expression of Bmi1 fused with an NH₂-terminal myc tag. The expression of myc-tagged Bmi1 in the Min6 cells resulted in reduced levels of p16^Ink4a protein, thus confirming the repressive effect of Bmi1 on p16^Ink4a expression (Fig. 5,B). We next assessed whether Bmi1 regulated p16^Ink4a expression by affecting the chromatin structure of the Ink4a/Arf locus that encodes p16^Ink4a. We determined whether histone acetylation at the Ink4a/Arf locus, which results in the loosening of chromatin and lends itself to transcription, was regulated by Bmi1. We examined the levels of H3K9 acetylation associated with the Ink4a/Arf locus in islets isolated from 6-week-old Bmi1^−/− mice and their wild-type littermates. In the absence of Bmi1, the levels of H3K9 acetylation were increased, suggesting that Bmi1 mediated repression of p16^Ink4a via modulation of chromatin structure (Fig. 5 C).

The absence of Bmi1 in young mice, which results in premature increase in p16^Ink4aexpression in the β-cells, allowed us to test whether upregulation of p16^Ink4a in young Bmi1^−/− mice would affect β-cell expansion. We treated 6-week-old Bmi1^−/− mice and their wild-type littermates daily with exendin-4 for 7 days and quantified β-cell proliferation and β-cell mass. After exendin-4 treatment, a twofold increase of β-cell mass was observed in the wild-type mice compared with the PBS-injected control group (Fig. 5,D). In contrast, no difference in β-cell mass was found between exendin-4– and PBS-treated Bmi1^−/− mice. To ascertain whether β-cell proliferation was affected in Bmi1^−/− mice treated with exendin-4, we quantified β-cell proliferation in wild-type and Bmi1^−/− mice after exendin-4 treatment. We observed a threefold increase in β-cell proliferation after exendin-4 treatment in wild-type mice, whereas no increase was observed in Bmi1^−/− mice (Fig. 5 E). These data further underscore the importance of the age-dependent rise of p16^Ink4a levels in regulating age-dependent decline in β-cell proliferation and β-cell mass expansion.

To assess the effect of aging on β-cell regeneration capacity, we adapted a standard regeneration model that uses a single injection of STZ to lead to a loss of ∼50% of the β-cells (31). We used this model for two reasons:1) an increase in β-cell proliferation is observed 7 days after STZ injection that correlates with regeneration of β-cell mass, and 2) the residual β-cell mass is sufficient to maintain blood glucose levels, thus potentially minimizing the role of hyperglycemia in β-cell proliferation and apoptosis. Young and old mice were subject to a single dose of 90 mg/kg STZ to induce partial loss of β-cells, and Ki67- and PCNA-positive β-cells were evaluated 7 days later. Both young and old mice that had received single-dose STZ administration displayed hyperglycemia (Fig. 6,A). A number of proliferating β-cells were observed in young mice after STZ administration. In contrast, old mice injected with STZ had limited numbers of Ki67- or PCNA-positive β-cells (Fig. 6 B, C, and E). Quantification of Ki67-positive β-cells in young mice after STZ administration revealed a fourfold increase in proliferating β-cells, whereas no difference was observed in older mice after STZ administration. Quantification of PCNA-positive β-cells showed a fivefold increase in proliferating β-cells in young mice after STZ administration, whereas no difference was observed in older mice. These data indicate that young mice, unlike old mice, have the capacity to proliferate and regenerate β-cell mass after β-cell destruction.

The experiments presented here were designed to evaluate whether age plays a role in the proliferative and regenerative capacity of the pancreatic β-cell. We used three distinct models of endogenous β-cell renewal and regeneration: adaptive expansion associated with insulin resistance, exendin-4–induced β-cell expansion, and the regeneration of β-cells after toxin-mediated ablation. In all three models, young mice had the capacity to expand or regenerate β-cells, whereas the β-cell mass of old mice did not show similar plasticity. We also show that the age-dependent decline in the capacity for β-cell mass expansion coincides with reduced β-cell proliferation. Our data suggests that age-dependent decline in β-cell proliferation curtails the ability of the endocrine pancreas to expand or regenerate β-cell mass to respond to changes in metabolic demands, growth factors, or injury.

Our data raise an important issue in assessing therapeutic strategies in diabetes research that are largely carried out in young mice, when the β-cell proliferative and regeneration capabilities are still retained. It could be argued that older mice with more limited regeneration capacity would serve as a better model for assessing therapies for humans. For instance, the age of NOD mice in most of the studies could contribute to the success in the reversal of diabetes in NOD mice but failures in humans (32,33). It has been suggested that incretin hormone therapy might serve to foster β-cell regeneration in humans with either type 1 or type 2 diabetes, but it is likely that GLP-1–induced increased β-cell mass may not be achievable in typical middle-aged patients with the capacity for β-cell expansion so predictably low in these human subjects (34).

The plasticity of β-cell mass was correlated with β-cell proliferation and the regulation of the Ink4/Arf locus. The product of the Ink4a/Arf (Cdkn2a) locus, p16^Ink4a (a negative regulator of CDK4 D-type cyclins), is thought to be involved in aging, and in genome-wide studies it has linked with type 2 diabetes (35,36). Moreover, the age-induced increase in p16^Ink4a has been shown to limit the regenerative capacity of β-cells with aging (15). Several studies suggest a role for polycomb group proteins in the regulation of expression from the Ink4a/Arf locus, and Bmi1, a key component of the polycomb complex, has been widely implicated in this process (37,38). We have correlated β-cell proliferation with changes in expression of Bmi1 and p16^Ink4a in β-cells that serve as both biological markers and effectors of aging. Our data suggests that age-related changes within the β-cell may be a significant contributing factor to reduced tissue homeostasis and regeneration in older individuals. Whereas the simplest model is that the Bmi1/p16^Ink4a pathway operates in the β-cell to regulate β-cell proliferation via CDK4 cyclin D2, it is possible that this pathway acts in putative stem cells that lead to regulation of β-cell mass. Activation of the Bmi1 in distinct cell types within the pancreas will shed light on the compartment that regulates β-cell mass.

Our data indicate that age-related changes that lead to β-cell senescence may be a significant contributing factor to reduced tissue homeostasis and regeneration in older mice. In fact, these senescence-inducing mechanisms presumably act to prevent the development of malignancy in the aging organism, but this protection comes at the cost of lost proliferative potential (39). However, we need to be cautious in interpreting experimental data from short-lived rodents to long-lived humans, where cellular senescence might be regulated by alternative mechanisms. We suggest diminished proliferative capacity of β-cells is the primary cause of why β-cells in old mice expand less successfully. It is also worth pointing out that the β-cell population is heterogeneous, and aging may change the composition of the β-cell population (34). Other distinct processes that may operate in humans to account for diminished regenerative capacity with aging include altered stem/progenitor cell function, an increased tendency for apoptosis, alterations in growth factor profiles, and important changes in immune responses. It will be crucial in the future to better characterize these underlying mechanisms to develop new therapeutic strategies for diabetes because the progression of diabetes is likely to alter how the pancreas copes normally with tissue homeostasis. Understanding the mechanisms that regulate declining capacity for β-cell expansion with aging is not only important from the perspective of regeneration, but it also should be taken into account when selecting donor islets to be used in cellular therapies.

In summary, we have shown here that the ability to expand β-cell mass in mice declines with age and is correlated with reduced β-cell replication. Furthermore, we have correlated the age-dependent decline in the regenerative capacity of β-cells with the Bmi1/Ink4/Arf pathway. The latter pathway might therefore represent a molecular target in efforts to foster β-cell regeneration in the treatment of diabetes.

This study was supported by National Institutes of Health Grant R01 DK-068763 and the Larry Hillblom and Juvenile Diabetes Research Foundations (to A.B.).

No potential conflicts of interest relevant to this article were reported.

We thank Dr. Peter Butler for advice and helpful discussion. We thank Maarten van Lohuizen, Netherlands Cancer Institute, for the Bmi1^−/− mice.