Transient Suppression of TGFβ Receptor Signaling Facilitates Human Islet Transplantation

Abstract

Although islet transplantation is an effective treatment for severe diabetes, its broad application is greatly limited due to a shortage of donor islets. Suppression of TGFβ receptor signaling in β-cells has been shown to increase β-cell proliferation in mice, but has not been rigorously examined in humans. Here, treatment of human islets with a TGFβ receptor I inhibitor, SB-431542 (SB), significantly improved C-peptide secretion by β-cells, and significantly increased β-cell number by increasing β-cell proliferation. In addition, SB increased cell-cycle activators and decreased cell-cycle suppressors in human β-cells. Transplantation of SB-treated human islets into diabetic immune-deficient mice resulted in significant improvement in blood glucose control, significantly higher serum and graft insulin content, and significantly greater increases in β-cell proliferation in the graft, compared with controls. Thus, our data suggest that transient suppression of TGFβ receptor signaling may improve the outcome of human islet transplantation, seemingly through increasing β-cell number and function.

Transplantation of islets or β-cells is an effective therapy for a deficiency of functioning β-cells in type 1 and some type 2 diabetes patients (1, 2). Nevertheless, the shortage of donor pancreases potentially restrains the application of successful human islet transplantation. Although great efforts have been made to generate functional β-cells from non-β-cell sources, no reliable clinical application has been so far available. Although postnatal β-cell growth predominantly results from β-cell replication (3–5), the β-cell replication rate under normal conditions is fairly low, and gets progressively lower with increasing age (6–10). So far, manipulations to increase the numbers of human β-cells for transplantation have not been satisfactory. An attractive option is to enhance β-cell proliferation through manipulation of intracellular signaling pathways in β-cells (11).

TGFβ receptor signaling has recently been demonstrated to affect pancreatic β-cell proliferation, in addition to its critical participation in the regulation of pancreas development and β-cell function in mice (12–17). TGFβ receptor signaling is initiated by the binding of TGFβ ligands to the TGFβ type II receptor (TGFbRII) to recruit and catalyze phosphorylation of TGF type II receptor (TGFbRI), which subsequently phosphorylates the transcription factors small mothers against decapentaplegic (SMAD)2/SMAD3 (14–17). Phosphorylated SMAD2/SMAD3 generates a complex with SMAD4 and then translocates into the nucleus to modulate gene expression and regulate cellular functions, eg, cell cycle controls (14–17). SMAD7 is a potent inhibitor for TGFβ receptor signaling (16). TGFβ receptor signaling may suppress β-cell proliferation through SMAD-mediated modulation of cell-cycle regulators (15, 16, 18). However, most evidence supporting a critical role for TGFβ receptor signaling in regulating the proliferation of β-cells has been obtained in mice, not yet in human.

SB-431542 (SB) is a specific TGFbRI inhibitor with no effect on receptors for bone morphogenetic proteins, ERK, c-Jun N-terminal protein kinases, or p38 MAPK pathways (19). Although the effects of SB on TGFβ receptor signaling suppression are prolonged, they are not permanent and appear to be reversible (19). We have previously used SB to successfully inhibit TGFβ receptor signaling in mouse β-cells (16). Here, we hypothesize that SB may similarly inhibit TGFβ receptor signaling in human β-cells, and we expect this inhibition of TGFβ receptor signaling to result in increased β-cell proliferation, and subsequently, β-cell number.

Materials and Methods

Mouse manipulation

All mouse experiments were approved by the Animal Research and Care Committee at the Children's Hospital of Pittsburgh and the Institutional Animal Care and Use Committee from the University of Pittsburgh. Female NOD/SCID mice of 12 weeks of age were purchased from The Jackson Laboratory. Administration of a single dose of alloxan (ALX) (Sigma-Aldrich) at 50-mg/kg body weight induced sustained hyperglycemia in NOD/SCID mice, as has been described before (5). Fasting blood glucose and ip glucose tolerance tests (IPGTTs) were performed as described before (14). Islet transplantation under the mouse kidney capsule and quantification of graft vessel density have been described before (20). Measurement of serum or graft insulin levels was performed with a human insulin ELISA kit (ALPCO), and C-peptide from the collected media 24 hours after treatment with SB was measured using a human C-peptide ELISA assay (ALPCO) (20, 21).

Human islet culture and treatment

Human islets were isolated from healthy, nondiabetic organ donors by the University of Chicago Transplant Center. Human islets were kept in CMRL 1066-based media containing 25 mmol/L glucose supplemented with 15% fetal calf serum (Invitrogen) for 2–7 days before the experiments. Five independent human islet batches from 3 male donors and 2 female donors, ages ranging from 32 to 55, were used in this study. Each experiment used islets from the same batch to compare different groups. The final data are from a summary of 5 experiments using these 5 batches accordingly. SB (Sigma-Aldrich) was dissolved in dimethylsulfoxide (DMSO) and applied to cultured human islets with a final concentration of 20 μmol/L. The control human islets were treated with DMSO of the same volume. Twenty-four hours after treatment with SB, the human islets were either analyzed, or transplanted under the kidney capsule of the diabetic NOD/SCID mice that had received ALX injection 7 days before.

3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay

To assay human islet cell growth, 5 islets were seeded onto a 96-well plate, subjected to a Cell Viability kit (MTT) (Roche), and quantified by absorbance value (OD) at 570 nm in a microtiter plate reader (Promega). Reported values were the mean from 5 repeats.

Immunostaining and Western blotting

Deidentified cadaveric nondiabetic human pancreas was obtained from the Pathology Department of Children's Hospital of Pittsburgh. Immunostaining and Western blotting were performed as described before (5, 14, 16, 20). Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) staining was performed with In Situ Cell Death Detection kit (Roche). Primary antibodies are listed in Table 1. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a protein loading control in Western blotting. Secondary antibodies are cyanine, indocarbocyanine, or horseradish peroxidase-conjugated rat, rabbit, or guinea pig-specific (Jackson ImmunoResearch). Nuclear staining was performed with Hoechst 33342 (HO) (1:1000; Becton-Dickinson Biosciences). Densitometry of Western blottings was quantified with NIH ImageJ software.

Table 1.

Antibody Table

Peptide/Protein Target	Name of Antibody	Manufacturer, Catalog Number, and/or Name of Individual Providing the Antibody	Species Raised in; Monoclonal or Polyclonal	Dilution Used
Insulin	Insulin	DAKO, A0564	Guinea pig; polyclonal	1:500
Glucagon	GCG	Cell Signaling, 2760s	Rabbit; polyclonal	1:200
Somatostatin	SOM	DAKO, A0566	Rabbit; polyclonal	1:300
Pancreatic polypeptide	PP	Abcam, AB113694	Rabbit; polyclonal	1:200
casp3	casp3	Cell Signaling, 9662s	Rabbit; polyclonal	1:200
Ki-67	Ki-67	DAKO, M7249	Rat; polyclonal	1:200
SMAD2	SMAD2	Cell Signaling, 3101s	Rabbit; polyclonal	1:1000
pSMAD2	pSMAD2	Cell Signaling, 5339s	Rabbit; polyclonal	1:1000
SMAD3	SMAD3	Cell Signaling, 9523s	Rabbit; polyclonal	1:1000
pSMAD3	pSMAD3	Cell Signaling, 9520s	Rabbit; polyclonal	1:1000
SMAD7	SMAD7	Santa Cruz Biotechnology, Inc, SC-11392	Rabbit; polyclonal	1:500
p27	p27	Cell Signaling, 9520s	Rabbit; polyclonal	1:2000
CyclinD1	CyclinD1	Cell Signaling, 2978s	Rabbit; polyclonal	1:1000
CyclinD2	CyclinD2	Cell Signaling, 2924s	Rabbit; polyclonal	1:1000
CDK4	CDK4	Cell Signaling, 12790s	Rabbit; polyclonal	1:1000
Cytochrome C	Cytochrome C	Cell Signaling, 4280s	Rabbit; polyclonal	1:1000
Caspase 9	Caspase 9	Cell Signaling, 9508s	Mouse; monoclonal	1:1000
Beclin-1	Beclin-1	Cell Signaling, 3738s	Rabbit; polyclonal	1:1000
GAPDH	GAPDH	Cell Signaling, 2118s	Rabbit; polyclonal	1:1000
TGFbRI	TGFbRI	Cell Signaling, 3712s	Rabbit; polyclonal	1:1000

Data analysis

Quantifications were performed as described before (14, 16). All values are depicted as mean ± SE. Five repeats were analyzed in each condition. All data were statistically analyzed using one-way ANOVA with a Bonferroni correction, followed by Fisher's exact test. Significance was considered when P < .05.

Results

Treatment of human islets with SB significantly increases β-cell proliferation

Because SB has been shown to inhibit TGFβ receptor signaling in mouse β-cells, we hypothesized that SB may similarly inhibit TGFβ receptor signaling in human β-cells to affect β-cell proliferation. First, TGFbRI staining in nondiabetic human pancreas showed that TGFbRI was exclusively expressed by β-cells in the human islets with no TGFbRI staining in non-β islet endocrine cells (Figure 1, A and B). Next, we treated cultured human islets with SB for 24 hours and examined cell proliferation by Ki-67 staining. The control human islets received DMSO (Control). We found that the percentage of Ki-67+ β-cells in human islets significantly increased after SB treatment (Figure 2, A and B), suggesting that SB significantly increases β-cell proliferation in human islets.

Figure 1.

TGFbRII is exclusively detected in β-cells in human islets. A and B, Representative immunostaining images for TGFbRI (red), insulin (green), and HO (blue) (A) and TGFbRI (red), glucagon/somatostatin/pancreatic polypeptide cocktail (GCG/SOM/PP) (green), and HO in human pancreas (B). Scale bars, 50 μm.

Figure 2.

Treatment of human islets with SB significantly increases β-cell number and function. A–F, Human islets were treated with SB for 24 hours, followed by staining for Ki-67, or TUNEL, or cleaved casp3. The control human islets received DMSO, the solvent for SB (control, CTL). A and B, Percentage of Ki-67+ β-cells by quantification (A) and by representative confocal images (B). C and D, Percentage of TUNEL+ β-cells by quantification (C) and by representative confocal images (D). E and F, Percentage of casp3+ β-cells by quantification (E) and by representative confocal images (F). A positive cell is shown in higher magnification in an inset. G, MTT assay. H, Secreted C-peptide in the culture media. *, P < .05. NS, nonsignificant. n = 5. Scale bars, 10 μm.

Treatment of human islets with SB does not increase β-cell apoptosis

In order to find out whether treatment of human islets with SB may increase β-cell apoptosis, we analyzed apoptotic β-cells by TUNEL assay, and by cleaved Caspase 3 (casp3). We found that the percentage of TUNEL+ β-cells (Figure 2, C and D) or casp3+ β-cells (Figure 2, E and F) in human islets was not altered after SB treatment. These data suggest that SB does not increase β-cell apoptosis in human islets.

SB treatment increases human islet cell number and C-peptide secretion

Because entering an activated cell cycle does not necessarily guarantee that the cell will physically divide, we then analyzed the β-cell number after SB treatment in an MTT assay. We detected a significant increase in cell number in SB-treated human islets, compared with controls, 3 days after treatment (Figure 2G). Moreover, the SB treatment also significantly increased C-peptide secretion (Figure 2H).

Molecular mechanisms underlying SB-induced increases in human β-cell proliferation

In order to understand the underlying mechanisms for increasing cell number after SB treatment, we analyzed human islets by Western blotting after SB treatment. We found that SB significantly inhibited the phosphorylation of TGFbRI, SMAD2, and SMAD3 (Figure 3, A and C), consistent with previous reports. Moreover, SB activated SMAD7 (Figure 3D), significantly decreased cell-cycle suppressor p27 (Figure 3E), and significantly increased cell-cycle activators CyclinD1 (Figure 3F), CyclinD2 (Figure 3G), and cyclin-dependent kinase 4 (CDK4) (Figure 3H), consistent with the positive effect of SB on cell proliferation. However, SB did not alter the levels of key apoptotic genes Cytochrome C, Caspase 9, and a key autophagy-associated gene Beclin-1 (Figure 3, I–K), consistent with the effect of SB treatment on cell apoptosis.

Figure 3.

Molecular mechanisms underlying SB-induced increases in human β-cell proliferation. Representative Western blottings and quantification for phosphorylated TGFbRI (pTGFbRI) and TGFbRI (A), phosphorylated SMAD2 (pSMAD2) and SMAD2 (B), phosphorylated SMAD3 (pSMAD3) and SMAD3 (C), SMAD7 (D), p27 (E), CyclinD1 (F), CyclinD2 (G), CDK4 (H), Cytochrome C (I), Caspase 9 (J), and Beclin-1 (K). GAPDH is a loading control in D–K. SB, SB-treated human islets; CTL, control human islets treated with DMSO. *, P < .05. NS, nonsignificant. n = 5.

Pretreatment of human islets with SB improves the outcome of human islet transplantation through augmented β-cell proliferation and function

To extrapolate these findings toward clinical applications, we evaluated the effects of pretreating human islets with SB-431542 on the outcome of islet transplantation. A subtherapeutic number (150) of human islets were pretreated with SB or DMSO as a control for 24 hours. These islets were then transplanted under the kidney capsule of diabetic NOD/SCID mice (10 each) that had received an ALX injection 7 days before. Three days after transplantation, 5 mice in each group were killed for analysis of β-cell proliferation in the graft, and the other 5 mice in each group were followed for 28 days and then analyzed for IPGTT and graft insulin. During the process, fasting blood glucose and serum insulin were examined (Figure 4A). We found that, beginning 7 days after human islet transplantation, fasting glucose levels in mice that received SB-treated human islets decreased significantly, compared with control mice (Figure 4B). At 28 days after transplantation, the glucose tolerance in mice that received SB-treated human islets was significantly improved, compared with control mice (Figure 4C). At both day 3 and day 28 after islet transplantation, significantly higher serum insulin was detected in mice that received SB-treated human islets, compared with control mice (Figure 4D). At day 28 after islet transplantation, significantly higher graft insulin content was detected in mice that received SB-treated human islets, compared with control mice (Figure 4E). In addition, significantly higher graft vessel density was found in SB-treated human islets, compared with control mice (Figure 4, F and G), and significantly higher numbers of Ki-67+ β-cells were detected in the graft 3 days after transplantation (Figure 4, H and I). Together, these data suggest that pretreatment of human islets with SB may improve the outcome of human islet transplantation, apparently through augmentation of β-cell proliferation and function.

Figure 4.

Pretreatment of human islets with SB improves the outcome of human islet transplantation through augment of β-cell proliferation and function. A, A subtherapeutic number (150) of human islets were pretreated with SB or DMSO as a control for 24 hours. These islets were then transplanted under the kidney capsule of diabetic NOD/SCID mice (10 each) that had received ALX injection 7 days before. Three days after transplantation, 5 mice in each group were killed for analysis of β-cell proliferation in the graft, and the other 5 mice in each group were followed for 28 days and then analyzed for IPGTT and graft insulin. During the process, fasting blood glucose and serum insulin were examined. B, Fasting glucose. C, IPGTT at 28 days after transplantation. D, Serum insulin at day 3 and day 28. E, Graft insulin at day 28 after islet transplantation. F and G, Graft vessel density based on CD31, by quantification (F) and representative images (G). H and I, Percentage of Ki-67+ β-cells in the graft 3 days after transplantation, by quantification (H) and by representative images (I). Yellow arrows pointed to Ki-67+ β-cells. *, P < .05. n = 5. Scale bars, 50 μm. 150 islets, mice transplanted with 150 human islets; 150 SB-islets, mice transplanted with 150 SB-treated human islets.

Discussion

Increasing graft β-cell number by promoting self-replication is believed to be the most practical approach for improving the outcome of islet transplantation. Genetic manipulation of human β-cells may create severe safety concerns of generating insulinomas or other neoplasms. Other approaches have applied cotransplantation of endothelial or other cell types that express high levels of angiogenic factors, eg, vascular endothelial growth factor A, or have overexpressed such factors in β-cells. These manipulations have been shown to not only improve graft revascularization, but also to increase β-cell proliferation (22–25). However, these approaches also have considerable shortcomings. β-cells are highly differentiated cells; overexpression of angiogenic factors in β-cells at high levels, driven by either the cytomegalovirus promoter or the insulin promoter, may aberrantly induce expression of a great number of proteins not normally present in β-cells. The potential effects of these ectopic proteins on normal β-cell function are uncertain. Also, increased protein processing in β-cells may create endoplasmic reticulum stress. On the other hand, the addition of other cell types (eg, endothelial cells or duct cells) to the islet grafts may well generate new antigens to be recognized by the host immune system after allograft transplantation. Therefore, an ideal way to increase the absolute number of β-cells in grafts may be through transient pharmacological treatment over a short pretransplant period, thus avoiding any genetic alteration of the β-cells, and avoiding a long culture period that may alter β-cell phenotype. Also, if successful, such an approach may avoid the potential systemic toxicity of treating islet transplant patients with β-cell-augmenting drugs.

Here, we used such a short pretransplant manipulation of human islets. By analyzing β-cell proliferation and apoptosis, we demonstrate that inhibition of TGFβ receptor signaling using the TGFbRI inhibitor SB increases β-cell proliferation, without inducing cell apoptosis. Although autophagy-induced cell death is unlikely to occur in this model, we did measure the Beclin-1 levels, a key autophagy related protein. Because the Beclin-1 levels remained unchanged with SB treatment, we conclude that SB does not induce appreciable cell death through autophagy. Moreover, examination of the absolute cell number confirmed an increase in cell number. However, we could not exclude that this increased cell number was partially contributable to proliferation of non-β islet cells. Our examination of Ki-67+ cells in vitro and in vivo did not support this possibility, because most Ki-67+ inside the islets appeared to be insulin+ β-cells by confocal microscopy. In addition, we found TGFβ receptors to be expressed on β-cells in human islets. We have previously shown that M2 macrophages produced TGFβ1 and epidermal growth factor (EGF) to induce Smad7 expression and to increase β-cell proliferation in a mouse pancreatic duct ligation model (14, 16). Here, SB-induced suppression of TGFβ-receptor also induced SMAD7. The induction of SMAD7 after either activation (in vivo pancreatic duct ligation model) or suppression (in vitro SB) of TGFβ-receptor signaling may depend on the presence of EGF, because EGF caused a nuclear exclusion of phosphorylated SMAD2 (16).

In order to translate our findings to clinical human islet transplantation, we devised a simple and clinically usable protocol. Human islets were treated with SB for 24 hours, and then the islets were transplanted into diabetic mice. This approach avoids the need for systemic treatment with a TGFβ signaling inhibitor, which may well be toxic. We used a subtherapeutic number of human islets for transplantation to allow distinction between the function of control islets and islets treated with SB. Although the inhibitory effects of SB on TGFbRI are reversible, SB may be active inside β-cells for a prolonged time, even after the islets have been transplanted into the mice, and this sustained presence inside the cells appears to not be harmful. In addition, the reversible effects of SB on TGFbRI would allow for the proliferating β-cells to later revert to a normal status.

Although the in vitro effects of SB to increase human β-cell numbers appear to be modest, the effects of SB on human islet transplant function were surprisingly large. The increases in serum and graft insulin content, as well as the improvements in blood glucose control by pretreatment of islets with SB, were significantly more enhanced compared with controls. Of note, these in vivo effects appear to result from not only increased β-cell proliferation, but also from an improved β-cell function (17, 21) due to increased insulin secretion and improved graft vascularization found in the current study.

In summary, here we provide an effective and clinically usable protocol to improve human islet transplantation through pharmacologically increasing functional β-cells and β-cell function using transient TGFβ receptor signaling suppression.