IGF-I mediates regeneration of endocrine pancreas by increasing beta cell replication through cell cycle protein modulation in mice
Recovery from diabetes requires restoration of beta cell mass. Igf1 expression in beta cells of transgenic mice regenerates the endocrine pancreas during type 1 diabetes. However, the IGF-I-mediated mechanism(s) restoring beta cell mass are not fully understood. Here, we examined the contribution of pre-existing beta cell proliferation and transdifferentiation of progenitor cells from bone marrow in IGF-I-induced islet regeneration.
Streptozotocin (STZ)-treated Igf1-expressing transgenic mice transplanted with green fluorescent protein (GFP)-expressing bone marrow cells were used. Bone marrow cell transdifferentiation and beta cell replication were measured by GFP/insulin and by the antigen identified by monoclonal antibody Ki67/insulin immunostaining of pancreatic sections respectively. Key cell cycle proteins were measured by western blot, quantitative RT-PCR and immunohistochemistry.
Despite elevated IGF-I production, recruitment and differentiation of bone marrow cells to beta cells was not increased either in healthy or STZ-treated transgenic mice. In contrast, after STZ treatment, IGF-I overproduction decreased beta cell apoptosis and increased beta cell replication by modulating key cell cycle proteins. Decreased nuclear levels of cyclin-dependent kinase inhibitor 1B (p27) and increased nuclear localisation of cyclin-dependent kinase (CDK)-4 were consistent with increased beta cell proliferation. However, islet expression of cyclin D1 increased only after STZ treatment. In contrast, higher levels of cyclin-dependent kinase inhibitor 1A (p21) were detected in islets from non-STZ-treated transgenic mice.
These findings indicate that IGF-I modulates cell cycle proteins and increases replication of pre-existing beta cells after damage. Therefore, our study suggests that local production of IGF-I may be a safe approach to regenerate endocrine pancreas to reverse diabetes.
KeywordsBeta cells Bone marrow-derived cells Cell cycle Islet regeneration Replication
protein kinase B
forkhead box O1
green fluorescent protein
antigen identified by monoclonal antibody Ki67
cyclin-dependent kinase inhibitor 1A
cyclin-dependent kinase inhibitor 1B
Pancreatic beta cells are responsible for producing the insulin required to maintain glucose homeostasis. Both type 1 and 2 diabetes result from reduced beta cell mass. The mechanisms that lead to endocrine pancreas regeneration and maintenance are under extensive investigation. Regeneration of beta cell mass may arise from replication of pre-existing beta cells [1, 2, 3] or from transdifferentiation of progenitor cells from pancreatic ducts, acini or even from bone marrow (BM) . Several growth factors have been identified as being able to upregulate beta cell mass [5, 6]. We have shown that Igf1 expression in beta cells of diabetic transgenic mice (1) regenerates the endocrine pancreas, probably by increasing beta cell replication and neogenesis, (2) counteracts hyperglycaemia and (3) protects islets from lymphocytic infiltration and beta cell death by apoptosis [7, 8]. Therefore, IGF-I may be a key factor for restoring pancreatic beta cell mass during diabetes. Nevertheless, the IGF-I-mediated mechanism(s) responsible for endocrine pancreas regeneration in vivo are not fully understood.
IGF-I has the ability to recruit circulating BM cells to sites of skeletal muscle damage . Likewise, IGF-I overproduction increases stem cell number and growth in ischaemic cardiac tissue, leading to an increase in myocyte turnover . BM cells have also been shown to differentiate into beta cells in adult mice . Injection of BM cells into diabetic recipients results in pancreatic regeneration and reduction of hyperglycaemia . In contrast, other studies found little or no evidence of BM cell differentiation into beta cells in healthy pancreas or after induced damage in mice [12, 13, 14] or in non-diabetic humans . Experimental conditions, such as the BM transplantation protocol, the method used for identification of cell fate and the model of pancreatic damage may account for these differences. Furthermore, it cannot be ruled out that the presence of specific (growth) factors may increase the recruitment and differentiation of BM cells into beta cells in islets of diabetic pancreases. Therefore, IGF-I production may contribute to endocrine pancreas regeneration through recruitment and differentiation of BM cells into functional beta cells.
Recent studies have implicated the insulin and IGF-I signalling pathways in the regulation of beta cell proliferation [16, 17]. In the adult pancreas, pre-existing beta cell proliferation occurs at a low rate as the main mechanism of beta cell maintenance . Moreover, beta cell growth matches changes in systemic insulin demand, which increases during common physiological states such as pregnancy or ageing, and also during obesity and insulin resistance . Genetic lineage tracing in adult mice has shown that pre-existing beta cells are a major source of new beta cells, both in normal conditions and after partial pancreatectomy . Furthermore, recent studies have revealed that adult beta cells exhibit equal proliferation potential [2, 20]. Similarly, we have observed that streptozotocin (STZ)-treated Igf1-expressing transgenic mice presented increased BrdU incorporation in islets, suggesting increased beta cell replication . In contrast, mice lacking Irs2 and Igf1 receptors show marked reduction of beta cell mass and develop diabetes . Beta cell overproduction of protein kinase B (AKT) leads to beta cell proliferation and islet hyperplasia [21, 22]. AKT can also phosphorylate the transcriptional regulator forkhead box O1 (FOXO1), leading to its nuclear exclusion. FOXO1 regulates the expression of several cell cycle components and decreased nuclear levels of FOXO1 have been shown to increase beta cell replication [23, 24, 25]. Beta cell proliferation is modulated by the interaction of a diverse set of protein components (cyclins, cyclin-dependent kinases [CDKs] and cyclin kinase inhibitors) that comprise the cell cycle molecular machinery . FOXO1 regulates levels of the cyclin kinase inhibitor protein 1B (p27), while phosphorylated (P)-AKT can phosphorylate p27 and induce its nuclear exclusion [27, 28]. Therefore, IGF-I production in islets may modulate cell cycle machinery to increase beta cell replication and endocrine pancreas regeneration during diabetes.
Here, to determine the mechanisms by which IGF-I can lead to recovery of beta cell mass during diabetes, we examined the contribution of IGF-I to BM cell transdifferentiation and to proliferation of pre-existing beta cells in the pancreas of transgenic mice. While a contribution of BM cell transdifferentiation was found to be unlikely, we found that IGF-I-mediated regeneration of beta cell mass occurred predominantly by increased beta cell replication through modulation of key cell cycle proteins.
C57Bl6/SJL transgenic mice expressing Rip-I/Igf1 or the β-actin/Gfp chimeric genes [7, 29] were used. Male 6-week-old green fluorescent protein (GFP)-transgenic mice were BM donors. Recipients were 2-month-old male and female wild-type and Igf1 transgenic mice. For diabetes induction, mice received five intraperitoneal injections, on consecutive days, of STZ (30 or 40 mg/kg) dissolved in 0.1 mol/l citrate buffer (pH 4.5). Diabetes was assessed by measuring blood glucose levels using an analyser (Glucometer Elite; Bayer, Leverkusen, Germany). In all experiments, non-STZ-treated mice were sex- and age-matched with STZ-treated groups. Animal care and experimental procedures were approved by the Ethics Committee in Animal and Human Experimentation of the Universitat Autònoma de Barcelona.
Bone marrow transplantation
Bone marrow was flushed from the medullary cavities of tibiae and femurs with Iscove’s modified Dulbecco’s medium using a 25-gauge needle. Recipient mice were irradiated using a myeloablative regimen (10 Gy fractionated in two doses, 4 h apart) with x-ray equipment (MG324; Philips, Hamburg, Germany) set at 300 kV, 10 mA, delivered at a dose rate of 1.03 Gy/min. After 2 h, recipient mice received 1 × 107 unfractionated BM cells by tail vein injection. The level of haemopoietic chimerism was determined in peripheral blood nucleated cells by flow cytometry (Epics XL; Coulter Electronics, Hialeah, FL, USA).
Immunohistochemical and morphometrical analysis
Pancreases were fixed for 24 h in formalin, embedded in paraffin and sectioned. Immunohistochemical detection of insulin, glucagon, IGF-I, GFP, CD45, GLUT-2, CDK4, p27, Ki67 and cyclin D1 was performed as indicated in Electronic supplementary material (ESM). Morphometrical analysis is also described in ESM.
Gene expression analysis
For quantitative RT-PCR analysis, total RNA was extracted from isolated islets (ESM) using isolation reagent (Tripure; Roche Molecular Biochemicals, Mannheim, Germany) and Rneasy Mini Kit (Qiagen, Hilden, Germany). Total RNA (1 µg) was reverse-transcribed for 1 h at 37°C using a kit (Omniscript Reverse transcriptase; Qiagen). Quantitative PCR was performed in SmartCycler II (Cepheid, Sunnyvale, CA, USA) using the Quantitect SYBR green kit (Qiagen). Primer sequences are shown in the ESM.
Western blot analysis
Isolated islets were homogenised in protein lysis buffer. Proteins (20–150 µg) were separated by 10% SDS-PAGE (wt/vol.), transferred to polyvinylidene difluoride membranes and probed overnight at 4°C with primary antibodies against IGF-I (R&D Systems, Minneapolis, MN, USA), phospho-AKT, AKT, phospho-FOXO1, FOXO1 (Cell Signalling, Danvers, MA, USA), CDK4 (Santa Cruz, Palo Alto, CA, USA) and β-actin (Abcam, Cambridge, UK). Detection was performed using horseradish peroxidase-labelled anti-goat IgG or horseradish peroxidase-labelled anti-rabbit IgG (DAKO, Glostrup, Denmark) and western blotting detection reagent (ECL Plus; Amersham, Freiburg, Germany).
All values are expressed as the means ± SEM. Differences between groups were compared by Student’s t test. A p value of less than 0.05 was considered statistically significant.
Bone marrow cell transplantation in transgenic mice overexpressing
Igf1 in beta cells IGF-I can participate in the recruitment of BM cells into the tissue where it is produced . Therefore, the production of IGF-I by transgenic islets may increase BM cell recruitment and/or differentiation after STZ treatment. Lethally irradiated C57Bl6/SJL wild-type and Igf1-expressing transgenic mice were transplanted with 1 × 107 BM cells from donor Gfp-expressing transgenic mice . ESM Fig. 1a summarises the experimental design. To examine the haemopoietic engraftment of recipient mice the proportion of GFP-positive cells in peripheral blood was determined by flow cytometry 28 days after cell transplantation. A high percentage of chimerism (>85%) was observed in all animals 1 month after the transplant (ESM Table 1) and was maintained 4 months after transplant (data not shown). Furthermore, immunohistochemical analysis revealed that recipient lymph node cells were replaced by donor GFP-producing cells (ESM Fig. 1b). In addition, high levels of IGF-I were detected in islets from transgenic mice 4 months after transplant, indicating that expression of the transgene was not altered (ESM Fig. 1c).
Counteraction of diabetes in BM-transplanted Igf1-expressing transgenic mice
Three months after STZ treatment and 4 months after the transplant, all groups of BM transplanted mice were killed. Double insulin and glucagon immunostaining of non-STZ-treated wild-type and Igf1 transgenic pancreas showed islets with normal distribution of insulin-expressing cells in the core and glucagon-expressing cells in the periphery (Fig. 1d). Diabetic wild-type mice showed islet destruction and lack of insulin-expressing cells. In contrast, STZ-treated transgenic mice showed large islets with altered distribution of alpha and beta cells, since alpha cells no longer formed a mantle around the beta cell core. These results suggested that beta cell destruction and/or regeneration may lead to disorganisation of islet cell distribution in STZ-treated transplanted transgenic mice (Fig. 1d).
Contribution of BM cells to endocrine pancreas
IGF-I induces replication of beta cells
IGF-I modulates the expression of beta cell cycle proteins
Transgenic mice expressing IGF-I in islets regenerate the endocrine pancreas after STZ-induced damage, which suggests that this growth factor may be key in restoring beta cell mass when expressed in the pancreas during diabetes [7, 8]. Defining the mechanisms by which IGF-I regulates the proliferation of pancreatic beta cells is an essential prerequisite for the design of new strategies to regenerate this cell population. IGF-I-mediated effects in pancreas regeneration may be direct, by acting on and inducing replication of remaining beta cells, and/or indirect, by acting in cell types other than beta cells, which in turn may differentiate into beta cells.
IGF-I has been shown to increase recruitment of BM cells to sites of skeletal and cardiac muscle damage to enhance regeneration [9, 35]. We observed that healthy wild-type and Igf1 transgenic-transplanted mice showed GFP-positive cells in the interstitium of the endocrine pancreas. However, after analysis of more than 10,000 beta cells, only one cell was found expressing both GFP and insulin (less than 0.01%), indicating that no significant differentiation of BM cells into beta cells occurred in transgenic healthy animals. Thus, IGF-I production in islets did not increase BM cell transdifferentiation. In contrast, it has been reported that about 1.7% to 3% of BM cells transdifferentiate into beta cells at 4 to 6 weeks after transplantation . However, other studies have also reported little or no BM cell differentiation to beta cells [12, 13, 14, 36, 37].
The ability of BM cells to differentiate into other cell types may be dependent on tissue-specific damage [9, 38]. Several reports indicate that STZ treatment of BM-transplanted mice did not increase BM cell transdifferentiation into beta cells and mice developed hyperglycaemia [12, 13, 14]. Similarly, we found a strong reduction in the beta cell mass and no GFP-positive/insulin-positive cells in diabetic wild-type mice. In all these studies, BM cells were transplanted prior to STZ treatment, indicating that STZ-induced damage did not mobilise BM cells to the pancreas. In contrast, injection of c-Kit-positive BM cells into already diabetic mice reduced hyperglycaemia . Amelioration of glycaemia was not due to BM cell transdifferentiation into beta cells, but rather to the BM cell capacity to initiate pancreatic tissue regeneration . Since only one GFP-positive/insulin-positive cell was detected, we conclude that BM cells do not contribute to endocrine pancreas regeneration in diabetic Igf1 transgenic mice and suggest that BM cell differentiation into beta cells in vivo is a rare event. Our results clearly contrast with studies performed in injured skeletal and cardiac muscles, in which IGF-I increased BM cell recruitment [9, 10, 35].
When transplanted mice were treated with low doses of STZ (5 × 30 mg/kg), wild-type mice developed hyperglycaemia, while transgenic mice were normoglycaemic. The second administration of a higher dose of STZ (5 × 40 mg/kg) to the transgenic mice led to a progressive increase in glycaemia, which subsequently and gradually decreased. This suggested that IGF-I production partially protected islets from destruction. Thus, 10 days after STZ treatment, an increase in apoptotic beta cells was detected in Igf1-expressing islets, which was lower than that in wild-type islets. Alternatively and/or concomitantly, an IGF-I-mediated increase in beta cell replication may have contributed to prevention of hyperglycaemia in STZ-treated transplanted transgenic mice. This was consistent with higher Ki67-positive/insulin-positive cell number in islets of transgenic mice even at 3 months after STZ treatment. Thus, our results suggest that IGF-I-mediated proliferation of pre-existing beta cells and decreased apoptosis could increase beta cell mass and reduce hyperglycaemia.
Increased P-AKT levels in transgenic islets indicated activation of the IGF-I receptor signalling pathway. Similarly, active AKT overproduction in beta cells leads to higher proliferation rate and islet hyperplasia [21, 22]. Consistent with AKT activation, Igf1-expressing islets exhibited increased FOXO1 phosphorylation. In addition to downregulating p27 expression by direct phosphorylation of FOXO1, AKT also promotes nuclear exclusion of p27 by direct phosphorylation . Consistent with increased P-AKT and P-FOXO1, a decrease in the expression and nuclear immunolocalisation of p27 in Igf1-expressing islets was also observed. The protein p27 inactivates CDK4 and blocks beta cell proliferation, and both CDK4 null mice and transgenic mice overexpressing p27 in beta cells show greatly reduced beta cell mass and diabetes [32, 39]. Nuclear localisation of CDK4 was observed in beta cells of transgenic islets, both before and after STZ treatment. However, non-STZ-treated Igf1 transgenic mice did not present uncontrolled beta cell proliferation, suggesting that the increase in CDK4 induced by IGF-I is not sufficient to enhance replication. In agreement with these results, we did not detect a significant increase in Ki67-positive beta cells in the pancreas of either 2-month-old or 6-month-old Igf1-expressing transgenic mice compared with wild-type mice.
CDK4 partners with D-cyclins for activation . It has been reported that cyclin D2 is expressed at higher levels than cyclin D1 and is crucial for controlling postnatal beta cell expansion . Igf1-expressing islets showed similar cyclin D1 and D2 levels to wild-type, suggesting that levels of these cyclins were probably too low to increase CDK4 activity. However, after STZ treatment an increase in the expression of cyclin D1 in transgenic islets was observed, which probably increased beta cell proliferation. It has been suggested that cyclin D1 plays a critical accessory role to cyclin D2 to partner with CDK4 and promote beta cell duplication . Similarly, transgenic mice overexpressing Igf1 in skeletal muscle showed regeneration and increased levels of cyclin D1 only after tissue damage . In this regard, overexpression of cyclin D1 increased beta cell replication in vitro  and in vivo . Furthermore, increased beta cell replication by glucagon-like peptide-1 involves transcriptional induction of cyclin D1 . Parallel to increased cyclin D1, a normalisation of p21 expression in STZ-treated Igf1 transgenic islets compared with wild-type islets was observed. It has been postulated that p21 may inhibit beta cell growth by acting as a molecular ‘brake’ on beta cell cycle progression . In islets from non-STZ-treated transgenic mice, we found increased expression of p21, which was not detected after STZ treatment. This finding suggests that IGF-I may increase beta cell replication predominantly under stress conditions that lead to beta cell destruction. These results are also consistent with the fact that no over-regeneration has ever been detected in STZ-treated Igf1-expressing transgenic mice after normalisation of glycaemia [7, 8]. In addition, normoglycaemic regenerated STZ-treated transgenic mice show normal glucose tolerance (data not shown; ), indicating that these mice maintain physiological glucose homeostasis. These findings also agree with the fact that Igf1-expressing transgenic mice did not develop tumours.
In summary, this study demonstrates that IGF-I-mediated beta cell mass regeneration during diabetes was not dependent on BM cell transdifferentiation. In contrast, we show that Igf1 expression increased beta cell replication and decreased beta cell apoptosis. This study agrees with genetic lineage tracing studies demonstrating that beta cell self-duplication is the major source of new beta cells during adult life in mice [1, 47]. Furthermore, adult beta cells exhibit equal proliferation potential and expand from within a vast and seemingly uniform pool of mature beta cells [2, 20]. Since local IGF-I production is safe and does not lead to over-regeneration, this study supports the notion that Igf1 gene transfer to the pancreas in vivo to induce beta cell replication and prevent beta cell apoptosis may be key for reversal of type 1 and 2 diabetes. However, rodents have a much higher capacity for beta cell replication than humans . Although it is not clear to what extent beta cell replication contributes to the maintenance of beta cell mass in adult humans, a 100-fold increase in the frequency of beta cell proliferation has been reported in an 89-year-old man with recent-onset type 1 diabetes . Furthermore, IGF-I production prevents islet lymphocytic infiltration and decreases beta cell apoptosis [7, 8], which could overcome the increased vulnerability to apoptosis of replicating beta cells described in type 1 and 2 diabetes . Recently, progress has been achieved in genetic manipulation of mouse and diabetic dog pancreas in vivo [49, 50]. Thus, new gene therapy approaches for diabetes focused on endocrine pancreas regeneration may be envisaged in the future.
We thank M. Watford and C. J. Mann for helpful discussions. J. Agudo and V. Jimenez were recipients of predoctoral fellowships from Ministerio de Educación, Cultura y Deporte, Spain. E. Ayuso and A. Salavert were recipients of predoctoral fellowships from Direcció General de Recerca, Generalitat de Catalunya. This work was supported by grants from Plan Nacional I + D + I (SAF2005–02381), Instituto de Salud Carlos III (CIBER de Diabetes y Enfermedades Metabólicas Asociadas), Spain, and from the European Community (BetaCellTherapy, FP6-2004-512145).
Duality of interest
The authors declare that there is no duality of interest associated with this manuscript.
- 43.Cozar-Castellano I, Takane KK, Bottino R, Balamurugan AN, Stewart AF (2004) Induction of beta-cell proliferation and retinoblastoma protein phosphorylation in rat and human islets using adenovirus-mediated transfer of cyclin-dependent kinase-4 and cyclin D1. Diabetes 53:149–159PubMedCrossRefGoogle Scholar
- 46.Cozar-Castellano I, Weinstock M, Haught M, Velazquez-Garcia S, Sipula D, Stewart AF (2006) Evaluation of beta-cell replication in mice transgenic for hepatocyte growth factor and placental lactogen: comprehensive characterization of the G1/S regulatory proteins reveals unique involvement of p21cip. Diabetes 55:70–77PubMedCrossRefGoogle Scholar