Inhibition of TGF-β Signaling Promotes Human Pancreatic β-Cell Replication

Replication of β-cells is the primary mechanism for maintenance and expansion of β-cell mass in response to changing insulin demands (1–4), and failure of such adaptive expansion can result in diabetes (5,6). Induction of p16^INK4a expression, a product of the Ink4a/Arf locus, is typical in adult tissues (7,8) and contributes to reduced self-renewal across mammalian tissues, including β-cells (9–11). This reflects in the reduced ability of the endocrine pancreas for adaptive expansion and regeneration (12,13). The Ink4a/Arf locus has also been linked with type 2 diabetes in genome-wide association studies (14). Thus, manipulation of mechanisms that regulate the Ink4a/Arf locus could serve a tool for promoting adult β-cell replication.

Several studies, including ours, have shown that distinct repressive complexes of polycomb group proteins regulate β-cell replication during aging by epigenetically targeting the Ink4a/Arf locus (15,16). Polycomb repressive complex 2 (PRC2) contains histone H3K27 methyltransferase Ezh2 and marks the target chromatin with lysine 27 trimethylation of histone H3 (H3K27me3). This directs the recruitment of polycomb repressive complex 1 (PRC1; contains Bmi1 and ubiquitin ligase-Ring1B), leading to the ubiquitination of lysine 119 in histone H2 (H2AK119) (11,17). Together, these epigenetic modifications result in the repression of the Ink4a/Arf locus in juvenile β-cells to maintain replicative potential. The loss of polycomb complex binding in adult β-cells results in the recruitment of histone methyltransferase Mll1, which mediates trimethylation of lysine 4 in histone H3 (H43K4me3), an activating histone modification (15,16). In β-cells, Mll1 exists in a complex with histone H3K27me3 demethylase JmjD3, which removes the repressive H3K27me3 modifications (18). Together, the recruitment of the Mll1-JmjD3 complex leads to the induction of p16^INK4a expression (15,16,18), resulting in reduced β-cell replication. Our recent work shows that a combination of polycomb replenishment with loss of Mll1 complex binding can drive replication in adult β-cells (18).

Although recent work has unraveled cellular signals, such as platelet-derived growth factor that regulate the repression of p16^INK4a expression and β-cell self-renewal in young adults (19), the signals that mediate the induction of Ink4a/Arf locus in late adult life are not known. Here, we report that transforming growth factor-β (TGF-β) signaling induces Ink4a expression leading to replicative decline in β-cells through the recruitment of Smad3 as a part of the Mll1 complex. Smad3 is a downstream effector of the canonical TGF-β signaling and translocates to the nucleus upon its phosphorylation as a consequence of ligand binding to the TGF-β receptor (20). We demonstrate that inhibition of TGF-β signaling using small molecules can reduce the levels of p16^INK4a in a conserved fashion across cell types. Finally, we show that small molecule inhibitors of TGF-β signaling can be used to induce β-cell replication by lowering the cellular levels of p16^INK4a. Together, these studies provide a therapeutically relevant example of how cellular signals modulate β-cell replication via targeting epigenetic pathways.

For the experiments involving mouse islets and mouse in vivo regeneration studies, the animals were maintained by mating wild-type males and females on a C57BL/6J background. Male NOD Cg-Prkdc^scidIl2rg^tm1Wjl/SzJ (NOD-scid IL-2Rg^null [NSG]) (denoted NOD-scid-γ or NSG) mice (8–10 weeks old) were used for human islet transplant experiments. All animal experiments were performed in accordance with National Institutes of Health policies on the use laboratory animals and approved by the Animal Research Committee of the Office for the Protection of Research Subjects at the University of California, Los Angeles, and Joslin Diabetes Center Institutional Animal Care and Use Committee protocols.

For experiments on the effect of TGF-β inhibitor on mouse β-cell replication in vivo, wild-type C57BL/6J mice (age 9 months) were divided into two groups (n = 9 per group) and injected intraperitoneally with a single dose of TGF-β inhibitor SB431542 (5 mg/kg body weight [BW]) or vehicle control (50% [v/v] DMSO), ensuring that the volume injected was kept up to 100 µL. Previous long-term studies have shown that the concentration of DMSO used as vehicle in these studies has no adverse effects on mouse health (21). Pancreata from these mice were harvested and processed for histology and analyzed for proliferation after 1 week (4). For human islet graft experiments, male NSG mice (n = 4) were used for experiments and grafted with human islets (1,000 islet equivalents) under the kidney capsule as previously described (22,23). One week after transplantation (when the islets developed vascular anastomosis), mice were injected intraperitoneally with vehicle or TGF-β inhibitor (5 mg/kg/BW, 50% [v/v] DMSO) twice weekly for 3 weeks. Four weeks after transplantation, mice were intraperitoneally injected with BrdU (Sigma-Aldrich, St. Louis, MO) (100 mg/kg/BW) for 3 days and killed 6 h after receiving the final BrdU injection. Blood was collected before and 2 and 4 weeks after islets were grafted for the measurement of human insulin (Human Insulin ELISA Kit; Mercodia) and human C-peptide (C-Peptide ELISA Kit; Mercodia). In vivo glucose-stimulated insulin secretion (3 g/kg/BW, 20% dextrose) was performed at 3 weeks post-transplantation after overnight fasting. The graft-containing kidneys and pancreata were embedded in paraffin and prepared for histological analysis (24).

Mouse embryonic fibroblasts (MEFs) were maintained on dishes coated with 0.2% porcine gelatin in DMEM-based medium containing 10% FBS at 37°C in a 5% CO₂ environment. Mouse islets were isolated using the Liberase/DNase enzyme mix (Roche Diagnostics) as previously described (25). In brief, the pancreas was inflated with the enzyme mix through the common bile duct and digested at 37°C. Islets were enriched by gradient and handpicked. Human islet samples from healthy adult organ donors without diabetes who died of acute trauma or anoxia were obtained from the Integrated Islet Distribution Program, as previously described (4). From 100 to 200 islets were cultured in a 35 mm × 10 mm easy-grip tissue culture dish (Falcon) in DMEM (CellGro) containing 10% FBS, supplemented with 11 mmol/L d-glucose (mouse islets) or 5 mmol/L d-glucose (human islets). For experiments involving the effect of TGF-β on cellular aging and proliferation, early-passage MEFs (passage 2 [P2]) in culture were supplemented with 5.0 ng/mL TGF-β (PeproTech) for 2 days (for proliferation assay) and for the indicated number for days (for growth curve), with daily change of culture medium. For experiments with TGF-β inhibitors, islets from old mice (8–12 months old) and late-passage MEFs (passage 9–12 [P9–12]) were cultured with one of the following TGF-β inhibitors, SB431542 (15 μmol/L; Sigma-Aldrich), SB505124 (5 μmol/L; Sigma-Aldrich), or SD208 (10 μmol/L; Sigma-Aldrich), or control small molecule phosphatidylinositide 3 kinase inhibitor LY294002 (10 μmol/L; Sigma-Aldrich) for 3 days for replication analysis and the indicated number of days for cell growth curves, with daily change of culture medium. For cell growth measurements, the number of cells was measured at indicated times and after indicated treatments by using a hemocytometer and plotted as a function of time. Human islets from cadaveric donors were cultured for 4 days with SB431542 (40 μmol/L; Sigma-Aldrich) or vehicle control (DMSO) and then harvested for replication analysis. Western blots with the indicated antibodies were performed, as previously described (25).

A standard immunofluorescence staining protocol was used to detect various protein markers in mouse pancreatic islet sections (4). Cultured islets were harvested, fixed in 4% paraformaldehyde, embedded in histogel matrix, and processed according to the above protocol. MEFs were fixed as monolayers in culture (on gelatin-coated plastic ware) for 20–30 min at room-temperature with the BD Cytofix/Cytoperm solution (BD Pharmingen) and washed thoroughly three times with 1× BD Perm/Wash buffer, followed by permeabilization at room temperature in the same buffer. Primary antibodies were diluted in the blocking solution (3% BSA in Tris-buffered saline with Tween) at the following dilutions: guinea pig anti-insulin (diluted 1:400; Dako), mouse anti-Ki67 (diluted 1:50; BD Pharmingen), mouse anti-BrdU (diluted 1:2; GE Healthcare/Amersham RPN-202); anti-phosphorylated histone H3 (diluted 1:200; Millipore pHH3), rabbit anti-Pdx1 (diluted 1:200; Millipore), and mouse anti-p16 (diluted 1:400; Santa Cruz sc-1661). Donkey- and goat-derived secondary antibodies conjugated to FITC or Cy3 were diluted 1:500 (Jackson ImmunoResearch Laboratories).

Slides were viewed using a Leica DM6000 microscope, and images were acquired using Openlab software (Improvision). The proliferation index was calculated using Ki67, BrdU, or pHH3 staining as a measure of cell proliferation by costaining the islet sections for Ki67 (or BrdU or pHH3) and insulin or Pdx1 for β-cells and by costaining for Ki67 with the nuclear marker DAPI for MEFs. β-Cell area was evaluated by point-counting morphometry on immunofluorescence-stained pancreatic sections. Cell death was detected by TUNEL assay (ApopTag S7100; Millipore). At least 2,000–3,000 β-cell nuclei were counted per animal.

As previously published, 2.5 mg anti-Mll1 antibody (Millipore 05-765) or control IgG was used for each immunoprecipitation on MEF lysates. Briefly, antibodies were bound to Protein G Sepharose (Millipore) for 2–3 h and incubated with MEF extracts prepared overnight at 4°C on a rotator. The immunoprecipitates were analyzed by Western blotting with antibodies against JmjD3 (Millipore; 07-1434), Smad3 antibody (Bethyl Laboratories; A300-109A), and control Dnmt3a (Imgenex; IMG-268A). Samples for Western blotting were lysed in cell dissociation buffer (Invitrogen) supplemented with protease inhibitors cocktail (Calbiochem), as described. The following primary antibodies were used: mouse anti-p16 (Santa Cruz; sc-1661), mouse anti-Smad3 (BD Biosciences; 610842), rabbit anti-phosphorylated (p)Smad3 (07-1389), and mouse anti–β-tubulin (Sigma-Aldrich; T8328). Each Western blot is a representative from three independent experiments using different sample preparations.

Chromatin immunoprecipitation (ChIP) was performed as previously reported (26), with minor modifications. The islets (150–200 islets per group, made into single-cell suspension) or MEFs (1 × 10⁵ cells) were treated with 1% paraformaldehyde at room temperature for 20 min to crosslink the DNA with bound proteins. Formaldehyde was quenched by adding glycine to a final concentration of 125 mmol/L, followed by cell lysis in 100 μL lysis buffer consisting of 50 mmol/L Tris HCl (pH 8.0), 10 mmol/L EDTA, 1% SDS, supplemented with 1× complete proteinase inhibitors (Calbiochem), and sonicated to yield DNA fragments with an average size of 500 bp using Bioruptor (Diagenode). A total of 2.5 μg antibody was bound to 20 μL Protein-A/G Dynabeads (Invitrogen), depending on the antibody isotype, for 2 h at 4°C with agitation. The following antibodies were used: anti–3me-H3K27 (Millipore; 07-449), anti–H3K4 me3 (Cell Signaling; 9751S), anti-JmjD3 (Millipore; 07-1534), anti-MLL1 (Millipore; 05-765), and anti-Smad3 (BD Biosciences; 610842), or normal mouse IgG as a control. One-fourth fraction of the sonicated chromatin was incubated overnight at 4°C with the antibody-bead complexes in total volume of 200 μL in RIPA buffer containing 10 mmol/L Tris HCl (pH 8.0), 140 mmol/L NaCl, 1 mmol/L EDTA, 1% Triton X-100, 0.1% SDS, and 0.1% Na-deoxycholate supplemented with 1× complete proteinase inhibitors. After washing four times with radioimmunoprecipitation assay buffer and twice with Tris-EDTA buffer, chromatin was eluted, followed by reverse crosslinking at 68°C for 4 h with vigorous agitation in the presence of Proteinase K (Sigma-Aldrich). The DNA fragments were then purified using phenol-chloroform extraction and ethanol precipitation. The resulting DNA was quantified and served as a template for the real-time PCR reactions, performed using ABI7900HT (Applied Biosystems), with 1× Fast SYBR Green Mix (Applied Biosystems). The DNA enrichment after ChIP was estimated as percentage bound-to-input ratio determined by real-time PCR. The primers used to amplify the Ink4a/Arf locus have been described before (11,16).

All data are expressed as mean ± SE. Mean and SEM values were calculated from at least triplicates of a representative experiment. The statistical significance of differences was measured by unpaired Student t test for experiments with two groups and a continuous outcome or by one-way ANOVA for repeat measures. A P value of <0.05 indicated statistical significance.

To understand the mechanisms that mediate the recruitment of the Mll1 complex to the Ink4a/Arf locus during replicative decline, we first used the MEFs model. P2 MEFs are characterized by low p16^INK4a and high proliferation, whereas P9–12 MEFs accumulate p16^INK4a, resulting in reduced proliferation (11,18). Coimmunoprecipitation of Mll1 complex components, followed by mass spectrometry, suggested that the Mll1 complex interacted with the TGF-β signaling pathway effectors (data not shown). Therefore, we assessed the interaction between the Mll1 complex and TGF-β pathway using anti-Mll1 antibody immunoprecipitate from P9 MEFs. Western blotting using the Smad3 antibody revealed that Smad3 interacts with Mll1 (Fig. 1A and B). Using ChIPs, we compared the binding of Smad3 to the Ink4a/Arf locus in P2 and P9 MEFs. Smad3 was recruited to the Ink4a/Arf locus in the P9 MEFs, suggesting a correlation between Smad3 binding and induction of p16^INK4a (Fig. 1C). This was further confirmed by treatment of P2 MEFs with TGF-β, which resulted in recruitment of Smad3 to the Ink4a/Arf locus, along with Mll1 and JmjD3 (Fig. 1D and Supplementary Fig. 1A), resulting in increased H3K4me3 (activating histone modification) and reduced H3K27me3 (repressive modification) levels (Supplementary Fig. 1B and C). This resulted in a significant reduction in the proliferation of P2 MEFs treated with TGF-β, due to the accumulation of p16^INK4a (Fig. 1E–G and Supplementary Fig. 1D), which was associated with increased levels of pSmad3 (Supplementary Fig. 1E). Taken together, these observations suggested a novel role for TGF-β in the transcriptional regulation of Ink4a/Arf locus.

To confirm if TGF-β signaling is associated with reduced proliferative potential through induction of p16^INK4a expression, we treated P9 MEFs with SB431542 (27), a well-characterized inhibitor of TGF-β signaling that blocks phosphorylation of Smad3 via the ALK5 receptor. Treatment of P9 MEFs with SB431542 reduced the association of Smad3 (Fig. 2A), Mll1, and JmjD3 to the Ink4a/Arf locus (Fig. 2B), resulting in reduced H3K4me3 and increased H3K27me3 levels (Supplementary Fig. 2A and B). This led to reduced levels of p16^INK4a and increased proliferation of P9 MEFs treated with SB431542 (Fig. 2C–E and Supplementary Fig. 2C). This suggested that indeed the decline in proliferative potential due to p16^INK4a accumulation in P9 MEFs is mediated by TGF-β signaling. To test if the effect of TGF-β inhibitor on induction of proliferation was reversible, we cultured P9 MEFs in the presence of SB431542, followed by inhibitor withdrawal after 4 days of treatment. Withdrawal of SB431542 resulted in a decline in the proliferation compared with P9 MEFs that were continuously maintained in the presence of inhibitor (Supplementary Fig. 2D), suggesting a transient effect of TGF-β inhibitor on increasing proliferation.

On the basis of the observations in MEFs, we examined the effect of TGF-β signaling on regulation of the Ink4a/Arf locus in islets. ChIP analysis showed increased recruitment of Smad3 to the Ink4a/Arf locus in islets from 9-month-old mice compared with islets from younger juvenile mice (Fig. 3A). Treatment of pancreatic islets isolated from 9- to 12-month-old mice with SB431542 resulted in reduced binding of Smad3, Mll1, and JmjD3 complex to the Ink4a/Arf locus (Fig. 3B and Supplementary Fig. 3A), concomitant with increased H3K27me3 and reduced H3K4me3 levels at the Ink4a/Arf locus (Supplementary Fig. 3B and C). We then cultured the islets from adult mice (aged 9–12 months) with TGF-β inhibitor or vehicle control. The β-cells (marked by Pdx1) from islets treated with SB431542 showed increased replication (Fig. 3C and D) due to reduced levels of p16^INK4a in these islets (Fig. 3E). The specific effect of inhibition of TGF-β signaling on β-cell replication was further confirmed by treatment of islets from aged mice with two other known TGF-β signaling inhibitors, namely SB505124 and SD208, which showed a similar effect on increased replication, whereas LY294002, a phosphatidylinositide 3 kinase inhibitor used as a control, did not result in any significant increase in replication (Supplementary Fig. 3D and E). Thus, treatment with small molecule inhibitors of TGF-β signaling can promote β-cell replication in islets from adult mice that typically exhibit resistance to induction of replication and regeneration (13).

To establish whether treatment with TGF-β inhibitor can stimulate β-cell regeneration in adult mice in vivo, we injected 9-month-old mice intraperitoneally with one dose of SB431542 or vehicle control (n = 9 per group). The pancreas was harvested after 1 week, and coimmunostaining of pancreatic sections for Ki67 and insulin showed increased β-cell replication of SB431542 in treated mice compared with control mice (Fig. 4A and B). The increase in β-cell proliferation was accompanied by a reduction in p16^INK4a levels in islets (Fig. 4C). We previously showed that polycomb regulation of p16^INK4a-dependent changes that limits β-cell replication is conserved in human islets (18). We then examined whether β-cell replication can be induced in human islets by inhibition of the TGF-β pathway. Treatment of human islets (from donors aged 40–60) with SB431542 resulted in increased β-cell replication compared with vehicle control (Fig. 4D). The increased replication in the inhibitor-treated group correlated with reduced binding of Mll1 and Smad3 at the Ink4a/Arf locus (Supplementary Fig. 4A and B). Thus, inhibition of TGF-β signaling improves β-cell replication in both mice and humans.

To further evaluate the stimulatory effect of TGF-β inhibitor on human β-cell proliferation, we used a model in which 1,000 human islet equivalents were transplanted under the kidney capsule of male NSG mice (28). One week after transplantation, mice receiving human islet grafts were treated with the TGF-β inhibitor (5 mg/kg/BW, 50% [v/v] DMSO) or vehicle control (DMSO), and sham-operated mice did not receive any treatment. Graft survival and functionality was examined by insulin secretory responses to glucose, using an ultrasensitive human insulin ELISA assay, in mice that received human islets (Supplementary Fig. 5A). Detection of human insulin and C-peptide levels in the circulation 2 and 4 weeks after transplantation indicated the presence of viable human islet grafts (Supplementary Fig. 5B and C). Immunohistochemical analysis of human islet graft–bearing kidney sections showed a ∼10-fold (P < 0.05) increase in BrdU incorporation in the TGF-β inhibitor–treated group, indicating a transition from G1 to S-phase of the cell cycle (Fig. 5A–G and Supplementary Fig. 5D). The enhanced mitosis was confirmed by an approximately fourfold (P < 0.05) increase in Ki67⁺ β-cells in the treated group (Fig. 5H and Supplementary Fig. 5E), and an augmentation in pHH3⁺ β-cells suggesting progression into the G2/M phases (P = 0.07) (Fig. 5I). In untreated mice, the rate of human β-cell replication was virtually undetectable, consistent with previous findings under similar conditions (22,23). Analysis of p16^INK4a expression in these graft samples by immunohistochemistry showed reduced p16^INK4a levels upon treatment with TGF-β inhibitor, consistent with our earlier observations in mice (Fig. 5J). Furthermore, examination of endogenous mouse pancreas of both groups demonstrated significantly lower β-cell apoptosis (P = 0.005) in the TGF-β inhibitor–treated group compared with control mice (Supplementary Fig. 6A). Despite the increased β-cell replication upon TGF-β inhibitor treatment of mouse islets and human islet grafts, and decreased β-cell apoptosis in mouse islets, the β-cell area, and islet density were not different between groups (Supplementary Fig. 6B and C). Thus, treatment with TGF-β inhibitor can be effectively used to promote replication of human β-cells.

In parallel, we evaluated endogenous islets of the NSG mice that received the TGF-β inhibitor. Consistent with our previous data in independent studies (Fig. 4B), the group that received TGF-β inhibitors exhibited a significant increase in three different proliferation markers, namely, BrdU (0.27% ± 0.06 vs. 0.06% ± 0.02, P = 0.028), Ki67 (0.2% ± 0.03 vs. 0.09% ± 0.01, P = 0.047), and pHH3 (0.12% ± 0.02 vs. 0.03% ± 0.02, P = 0.039) (Fig. 6A–C) in the pancreatic β-cells. Similar to human islet grafts without treatment, β-cell replication in untreated and sham-operated mice was minimal. In addition, the pancreatic β-cell proliferation was accompanied by a reduction in p16^INK4a levels in the islets (Fig. 6D). Taken together, our studies suggest that TGF-β signaling induces age-related accumulation of p16^INK4a, leading to replicative decline and that treatment with TGF-β inhibitor can effectively induce replication in both mouse and human β-cells.

Previous work on TGF-β signaling has largely focused on pancreatic development and islet homeostasis and function (29–34). Here, we delineate the mechanism by which TGF-β signaling controls replicative senescence across cell types, including mouse and human β-cells, by regulating the Ink4a/Arf locus. More importantly, we have used small molecule inhibitors of the TGF-β pathway to stimulate mouse and human β-cell replication in vitro and in vivo. TGF-β signaling is known to inhibit cell proliferation, and disruptions in TGF-β signaling are associated with overcoming senescence (35–37), as reviewed by Padua and Massagué (38). It is important to note that we observed a two- to threefold increase in β-cell replication with TGF-β inhibitors, which is comparable to the rates of replication observed during physiological β-cell mass expansion.

TGF-β1, -2, and -3 ligands are endogenously expressed within the human pancreas, not just in islets, but also in ducts as well as in the acini (3). This suggests that TGF-β ligands may work in both autocrine and paracrine fashion in the pancreas. TGF-β1 levels have been reported to increase with age in circulation as well as within several tissues (27,28). TGF-β signaling has been shown to be a critical signal in pancreatic development (39). Loss of TGF-β type II receptor shows an increase in endocrine precursors and proliferating endocrine cells, suggesting that TGF-β signaling modulates the growth and differentiation of endocrine progenitors (40).

Disruptions in the canonical TGF-β signaling in the exocrine tissue result in islet hypertrophy (31), thus suggesting a role for TGF-β in regulating β-cell proliferation and islet homeostasis. Transient loss of Smad7 in the islets, using an inducible loss of function model, resulted in islet enlargement (32). More recently, genetic loss of Smad2 and -3 has been reported to result in a more robust proliferative response upon β-cell loss (29). Another study showed loss of Smad2 induced islet hyperplasia and impaired insulin secretion (30). Studies using combined pan-pancreatic knockdown of TGF-β receptors I and II have also implicated TGF-β signaling in the regulation of β-cell proliferation (33). However, most of such work examining the TGF-β signaling on β-cell proliferation has been performed in the context of adaptive expansion in permissive age windows. Our work is the first to identify a key role for TGF-β signaling in age-dependent proliferative decline in pancreatic β-cells and to also establish a mechanism of cell cycle control via the INK4a locus.

TGF-β has been reported to stimulate insulin secretion (41), insulin gene transcription, and islet function (34). Further, TGF-β has been recently reported to have a protective effect on islet grafts in the murine models (40). Also, another recent study shows that inhibition of TGF-β signaling can reverse β-cell dedifferentiation in diabetes, suggesting an effect on function (42). These studies raise the possibility that treatment with TGF-β inhibitor may promote islet function and islet graft survival. In our studies using human islet grafts in humanized mice, however, the control and inhibitor-treated mice both showed comparable graft survival and human insulin and C-peptide secretion profiles. Considering Ink4a repression by polycomb group proteins is a well-conserved mechanism across cell types, additional work that addresses targeting of the small molecule inhibitors of the TGF-β signaling to the islet β-cells is warranted. Our study for the first time shows the connection between TGF-β signaling and p16^Ink4a-dependent replicative senescence in β-cells, thus providing a mechanistic insight into the age-dependent decline of β-cell replication. The experiments presented here establish a novel link between TGF-β signaling and the epigenetic regulation of β-cell replicative potential. These data provide an example of a pharmacological agent that can promote the replication of human β-cells with translational therapeutic potential to restore functional β-cell mass in patients with diabetes.

Acknowledgments. The authors are grateful to Dr. Shuen-Ing Tschen (University of California, Los Angeles) for discussions and to Emily Snyder and James Maksymetz (University of California, Los Angeles) for technical support.

Funding. This work was supported by a start-up grant from the Larry L. Hillblom Foundation, a transition award from JDRF, and a Diabetes Research Center (University of California, Los Angeles-University of California, San Diego) Pilot and Feasibility Award to S.D.; an advanced postdoctoral fellowship from JDRF to E.D.; grants from National Institute of Diabetes and Digestive and Kidney Diseases to R.N.K. and A.B.; and grants from JDRF and the Helmsley Charitable Trust to A.B.

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

Author Contributions. S.D. performed the experiments and analyses on the role of TGF-β signaling on Ink4a/Arf locus in MEFs and β-cells and the effect of TGF-β inhibitor on mouse β-cell proliferation. E.D. performed the experiments and analyses on the effect of TGF-β inhibitor on human islet transplants in humanized mouse model. S.D. and A.B. and conceived and planned the experiments and interpreted data on the role of TGF-β signaling in MEFs and mouse β-cells. E.D. and R.N.K. contributed to the discussion on design and interpreted the data on human islet transplants in the humanized mouse model. S.D., E.D., R.N.K., and A.B. wrote the manuscript. A.B. is the guarantor of this work and, as such, had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.