Islet cell transplantation as a β-cell replacement therapy is of great interest for treatment of type 1 diabetes mellitus (T1DM), and advances in regenerative medicine are expected to render this option feasible for widespread clinical use in the future. Indeed, there have been many reports of successful in vitro generation of insulin-producing cells (IPCs) from mesenchymal stem cells (MSC), such as bone marrow stem cells1,2 and adipose-derived stem cells (ADSCs).3,4 Some studies reported have demonstrated their beneficial effects in vivo, leading to clinical trials as a new type of transplantation.5,6 We have also reported the generation of IPCs from ADSCs both in vitro and in vivo.7–10 However, some major issues regarding these IPCs remain to be addressed before their clinical use can be considered.6,11 One of these is whether IPCs generated from ADSCs isolated from patients with T1DM for autotransplantation are capable of appropriate insulin production. We intend to obtain 1 cm3 adipose tissue from T1DM patients to isolate and passage ADSCs for differentiation into IPCs using a modified protocol9 (Fig. 1A) and autotransplantation into the mesentery laparoscopically. Therefore, this issue must be resolved for our strategy to be applicable.
The mechanisms of T1DM are still unclear. However, T1DM is caused by absolute insulin exhaustion because of autoimmune insulitis, which destroys pancreatic β cells by cytotoxic CD8+ T cells.12 Therefore, transplanted IPCs may be attacked by recipient autoimmunity. Programmed death-ligand 1 (PDL-1) plays a major role in suppressing acquired immunity in autoimmune diseases,13 pregnancy, and allogeneic transplantation. Islet PDL-1 expression may limit the effects of autoimmunity including autoimmune insulitis.13–15 Although it is unknown whether IPCs transplanted into T1DM patients are subjected to immunological attack by cytotoxic T cells, expression of PDL-1 supports the prospects for clinical application of IPCs.
Here, we characterize ADSCs isolated from the adipose tissue of T1DM patients and compare them with the generated IPCs. Superfluous adipose tissue was obtained during abdominal surgical procedures performed on patients (including T1DM patient) at our institute. Adipose-derived stem cells were isolated from this tissue, and cultured and differentiated into IPCs using our 2-step 3-dimensional xenoantigen-free culture method (Fig. 1A).8,9 We then investigated the differences of cell quality and the functions between generated IPCs and isolated ADSCs, and measured PDL-1 expressions in the isolated ADSCs and generated IPCs from T1DM and nondiabetic patients.
MATERIALS AND METHODS
Adipose tissue was obtained from patients who underwent abdominal surgery at our institute. The study was approved by Tokushima University Hospital (Public title: “The differentiation of IPCs from adipose tissue-derived stromal cells”; UMIN number: UMIN000035546, January 10, 2018). Four patients (1 patient with T1DM and 3 nondiabetic patients) were enrolled in this study. Written informed consent was obtained from all patients.
Adipose-derived stem cells were isolated from adipose tissue of the T1DM patient and 3 nondiabetic patients using a commercial ADSC isolation kit (Funakoshi Co., Ltd, Tokyo, Japan) as described previously.9 Briefly, diced adipose tissue was cultured in a 3-dimensional, hydroxyapatite-coated, polyethylene-polypropylene, nonwoven fabric matrix. After 1 to 2 weeks, fibroblast-like ADSCs had proliferated in the matrix, which were collected by trypsinization and subcultured (Fig. 1A). These ADSCs were used for IPC generation at 95% confluence. As a control, commercially obtained ADSCs (StemPro Human Adipose-Derived Stem Cells, R7788-115) from Invitrogen (Waltham, Mass) were cultured in accordance with the manufacturer's guidelines.
Insulin-producing cells were generated by 3-dimensional culture using our 2-step protocol.7–10 Briefly, 5 × 105 ADSCs were seeded into 96-well ultralow attachment plates as a 3-dimensional culture (Sigma-Aldrich Japan Co., LLC., Tokyo, Japan). In Step 1, the cells were cultured for 7 days in a differentiation cocktail of Dulbecco’s modified Eagle's medium/F12 (Thermo Fisher Scientific Inc., Waltham, Mass), 1% human albumin (Wako, Osaka, Japan), 1% B27 supplement (Thermo Fisher Scientific), 1% N2 supplement (Thermo Fisher Scientific), 50 ng/mL activin A (Sigma-Aldrich), 10 nM exendin-4 (Sigma-Aldrich), and 0.1 mg/mL RCP μ-piece (Fujifilm, Tokyo, Japan). The cells were then cultured for 14 days in differentiation cocktail, plus 50 ng/mL human hepatocyte growth factor (Funakoshi Co., Ltd), 10 mM nicotinamide (Sigma-Aldrich), and histone deacetylase (HDAC) inhibitor (Sigma-Aldrich) (Fig. 1A).
Surface markers of ADSCs were analyzed using a FACSVerse and BD FACSuite software version 10 (BD Biosciences, Tokyo, Japan), as described previously.9,16,17 Briefly, ADSCs were collected by trypsinization and washed twice with phosphate-buffered saline. Five microliters of antibodies against CD31 (PE conjugated; e-Bioscience, San Diego, Calif), CD34 (FITC-conjugated; e-Bioscience), CD45 (PerCP/Cy5.5 conjugated; BioLegend, San Diego, Calif), CD90f (APC conjugated; e-Bioscience), CD105 (APC conjugated; e-Bioscience), or CD146 (APC conjugated; Bio Legend) were added to the cell pellets after addition of 1300 μL FACS buffer. The cells were incubated at room temperature (RT) for 30 minutes in the dark, washed with FACS buffer, and then analyzed by the FACSVerse.
We assessed the morphology of IPCs after fixation using a cell-fixing kit (Funakoshi Co., Ltd), as described previously.8–10 Briefly, the cells were placed in gels and fixed with 10% paraformaldehyde overnight. Then, 4-μm-thick paraffin-embedded sections were prepared, stained with dithizone,18 and examined by light microscopy.
Cells were pelleted by centrifugation, placed on a coverslip, and incubated with a primary antibody against insulin (aa287-299, LS-B129; LSBio, Seattle, Wash) at a dilution of 1:100 in PBS for 1 hour at RT. They were then incubated with biotinylated secondary antibody and treated with a streptavidin-biotin-horseradish peroxidase complex. Positive staining was visualized with diaminobenzidine after counterstaining with Mayer hematoxylin.
For antigen retrieval, 4-μm-thick paraffin embedded sections were boiled in 0.01 mol/L sodium citrate (pH 6.0) for 30 minutes. Blocking was performed with 5% goat serum/phosphate-buffered saline for 60 minutes, and then the sections were stained with an anti-programmed cell death ligand-1 antibody (17952-1-AP; Proteintech, Rosemont, Ill) at a dilution of 1:100 in blocking solution overnight at 4°C. After washing, the sections were stained with a secondary antibody (EnVision + Dual Link System-HRP, Dako, Tokyo, Japan) for 60 minutes at RT. Then, the sections were counterstained with hematoxylin-eosin and mounted using Mount-Quick (Daido Sangyo, Tokyo, Japan).
Samples were prepared as described previously.8,10 Briefly, samples were fixed with 2% paraformaldehyde and 2% glutaraldehyde in 0.1 M phosphate buffer (PB) (pH 7.4) at 4°C overnight, washed 3 times with 0.1 M PB for 30 minutes each wash, and postfixed with 2% osmium tetroxide in 0.1 M PB at 4°C for 2 hours. The sections were then dehydrated in 50% and 70% ethanol solutions for 10 minutes each at 4°C, 90% ethanol for 10 minutes at RT, and then 3 times in 100% ethanol for 10 minutes each at RT. The samples were treated with propylene oxide twice for 30 minutes each, placed in a 70:30 mixture of propylene oxide and resin (Quetol-812; Nisshin EM Co., Tokyo, Japan) for 1 hour, and then transferred to fresh 100% resin, followed by polymerization at 60°C for 48 hours. The polymerized resin was sectioned at 70 nm with a diamond knife using an ultramicrotome (Ultracut UTC; Leica, Vienna, Austria). The sections were mounted on copper grids and stained with 2% uranyl acetate at RT for 15 minutes, followed by secondary staining with a lead stain solution (Sigma-Aldrich Co., Tokyo, Japan) at RT for 3 minutes. The grids were observed under a transmission electron microscope (JEM-1400Plus; JEOL Ltd., Tokyo, Japan) at an acceleration voltage of 100 kV.
Glucose-Stimulated Insulin Secretion
Glucose-stimulated insulin secretion of IPCs was calculated as described previously.8–10,19–21 The IPCs were cultured in RPMI-1640 medium containing 2.2 mM glucose for 1 hour, medium containing 22 mM glucose for 1 hour, and then the original medium for an additional 1 hour. The insulin concentration in the culture supernatant was analyzed by an enzyme-linked immunosorbent assay (Akrin-011H, Shibayagi, Shibukawa, Japan) using a microplate reader at a wavelength of 450 nm. Total DNA was extracted to estimate the cell count. The stimulation index (SI) was then calculated by dividing the amount of insulin secreted during the high glucose incubation by the insulin secretion during the low glucose incubation. Three independent experiments were performed.
Eight-week-old nu-nu nude mice were purchased from CIEA, JAC Inc. (Tokyo, Japan) and bred at the Tokushima University animal facility. The mice were intraperitoneally injected with 200 mg/kg streptozotocin (STZ) to induce diabetes.8–10 Nonfasting glucose was measured in blood collected from the tail vein using an Accu-chek Aviva (Rossi DC Japan Inc., Tokyo, Japan). Diabetes was diagnosed when a single value of greater than 400 mg/dL or 2 consecutive values of greater than 350 mg/dL were obtained.8–10 Ninety-six IPCs clusters (containing 1.7 × 106 functional cells8,9) were carefully handpicked and suspended in 0.1 ml Hank’s balanced salt solution (Thermo Fisher Scientific). Then, the IPCs were transplanted into the mesentery of each mouse using a 20G needle as reported previously.10 Under-kidney capsule transplantation is usually used in in vivo islet and IPC studies, but we had previously found that intramesentery IPC transplantation has more rapid normalization effects on the hyperglycemic state compared with under-kidney capsule transplantation.8,10 Angiogenesis of transplanted IPCs is not similar to that of intraperitoneal injection according to the pathological evaluations (data of this mechanism submitted). The study was approved by the Animal Care and Use Committee of the Tokushima University (T29-29, June 14, 2017) and was performed in accordance with the NIH Guide for the Care and Use of Laboratory Animals.
Data are presented as mean and standard deviation or median and range for quantitative variables and number (percentage) for qualitative variables. Univariate analysis was performed using one-way analysis of variation, paired and unpaired t tests, Scheffe test, or the log-rank test, as appropriate. P less than 0.05 was considered to represent statistical significance. All P values are 2-sided. Analyses were performed using JMP 14 (SAS Institute Inc., Tokyo, Japan).
Surface Markers of Freshly Isolated and Commercially Available ADSCs
We characterized the isolated ADSCs by identifying cell surface markers. Flow cytometric analysis revealed that the ADSCs isolated from the T1DM patient were CD31−CD34−CD45−CD90+CD105+CD146− (Fig. 1B), and both freshly isolated ADSCs, which were isolated from patient fat tissue, and commercially available ADSCs demonstrated the reported changes in stromal and stem cell-associated markers with freezing/thawing cycles.9,22,23 This pattern of surface marker expression was similar in subcutaneous and visceral fat, and there were no major differences between ADSCs from the T1DM patient and commercially available ADSCs. However, there were fewer CD105+ cells among the commercially provided ADSCs of both subcutaneous and visceral origins.
Morphological Investigations for ADSCs and IPCs
The isolated ADSCs were relatively small, spindle-shaped cells, which were consistent with the typical fibroblastoid morphology of ADSCs.16 After 21 days, the generated IPCs appeared as clusters that were strongly stained with dithizone (Fig. 1C).
Generated IPCs From T1DM Patient Fat Tissue Strongly Express Insulin
The IPCs generated from the T1DM patient fat tissue were viable and strongly expressed insulin in their cytoplasm, according to immunofluorescence conducted using an anti-insulin antibody and 4′,6-diamidino-2-phenylindole (Fig. 1D).
SI of IPCs From T1DM Patient’s Fat Tissue Was the Same Level as IPCs From Nondiabetic Patients’ Fat Tissue and From Commercially Available ADSCs
The glucose SI reflecting the amount of insulin secreted autonomously by the IPCs in high and low glucose (22 and 2.2 mM, respectively) at day 21 was similar for adipose tissues from the T1DM patient and patients without DM, and the commercially obtained ADSCs (Fig. 1E). Interestingly, the SI of IPCs generated from visceral fat tissue was significantly lower than that of IPCs generated from subcutaneous fat tissue of the nondiabetic patients (P < 0.001; Tukey test) and commercially obtained ADSCs (P < 0.05; Tukey test).
Generated IPCs From T1DM Patient Subcutaneous Fat Tissue Contain Granules Resembling Insulin Secretion Granules of Naive Pancreatic β-Cells
Electron microscopic analysis revealed that the generated IPCs contained many dense cystic microstructures that morphologically resembled insulin secretion granules (Fig. 2A). Moreover, these dense cystic microstructures were surrounded many small particles (Fig. 2B).
Reversal of Diabetes in Nude Mice Was Achieved With Transplantation of IPCs Generated From Subcutaneous Fat of T1DM Donor and as Effective as IPCs Generated From Nondiabetic Donors
Insulin-producing cells were transplanted into the mesentery of STZ-induced diabetic nude mice as reported previously.10 The control sham-operated group was injected with 0.1 mL normal saline into the mesentery. As a result, the nonfasting blood glucose concentrations of all IPC-transplanted nude mice decreased gradually and were normal from 9 days after transplantation, including the mouse transplanted with IPCs generated from the T1DM patient subcutaneous adipose tissue. The normal blood glucose status persisted until 30 days after transplantation (Fig. 3A). These results showed that IPCs derived from type 1 diabetes mellitus patient fat tissue at the same cell numbers of IPCs generated from nondiabetic patient fat tissues converted STZ-induced diabetic nude mice to normoglycemia. When normoglycemia was defined as a blood glucose concentration of less than 150 mg/dL, the normoglycemia rate of STZ-induced diabetic nude mice was significantly higher than that of the sham-operated group (P = 0.0069; log-rank test. Fig. 3B).
Generated IPCs From T1DM Patient Fat Tissue and Commercially Available ADSCs Strongly Express PDL-1
We next investigated the expression of PDL-1 in ADSCs and IPCs generated from commercially provided ADSCs and ADSCs isolated from the T1DM patient adipose tissue. Microscopy revealed that PDL-1 expression was low in ADSCs from the T1DM patient subcutaneous adipose tissue and commercially provided cells, but IPCs generated from both of these sources strongly expressed PDL-1 (Fig. 4).
We have focused on MSCs, especially ADSCs as a feasible cell source for β-cell replacement thrapy.24,25 Adipose-derived stem cells are easy to procure, less invasive procedure compared with other MSCs, less ethical problems (often under local anesthesia) compared with ES cells and are reported that their multipotency is superior to other MSCs.26 Indeed, we succeeded in the establishment of effective IPC generation in vitro and demonstrated their in vivo functions.8–10 These IPCs are monoclonal (insulin positive, glucagon negative, and pancreatic peptide negative).7 Moreover, these generated IPCs had a DNA quantity of 88.4% and same level of cell viability by ATP assay compared with original ADSCs.9 If the stable generation of own effective IPCs can be realized by an easy, safe and low cost procedure, IPCs autotransplantation would be extremely beneficial for T1DM patients. However, there are some major issues to be resolved before clinical application.8–10
The first issue is whether functional IPCs can be generated from a T1DM patient’s own fat tissue. The present study demonstrated not only that functional IPCs could be generated from ADSCs freshly isolated from human adipose tissue but also that these IPCs, including those obtained from the T1DM patient, were capable of secreting enough insulin to normalize the blood glucose concentrations of diabetic nude mice. Freshly isolated ADSCs from the T1DM patient and commercially provided ADSCs had similar cell surface marker expression, reflecting both stromal and stem cell characteristics. Moreover, the generated IPCs demonstrated the ability to secrete insulin autonomously in vitro, even when derived from T1DM patient adipose tissue. Indeed, these IPCs performed similarly to those derived from nondiabetic patients. However, the insulin responsiveness was significantly lower when the IPCs were derived from visceral fat tissue. The number of T1DM patients was small. However, this is because the number of T1DM patients who receive surgery is small. Thus, we will try to enroll more T1DM patients for further investigations.
The second issue is whether generated IPCs from a T1DM patient's own fat tissue are attacked by autoimmunity. It is considered that transplanted IPCs are not rejected by allogeneic immunity, but transplanted IPCs still have a risk of autoimmune attack, even though the cells are from self-tissue. If generated IPCs are attacked by cytotoxic T cells in an autoimmune manner, they would be gradually destroyed similarly to naive islets.12 To avoid this situation, studies have reported cytoprotection methods or microcapsule of cells.27,28 However, these methods do not guarantee longevity during slow autoimmunological destruction. A realistic solution is transplantation of a large volume of cells or multitransplantation such as clinical islet transplantation.29 According to our estimation, 3.0 × 106 functional IPCs/patient kg are required for conversion of the hyperglycemic state with a single transplantation (data not shown), and we plan to use a large animal transplantation model to prove this. Moreover, we have used some reagents with cytoprotective effects on pancreatic β cells.30,31 However, the other possibility still remains. The interpretation of the PDL-1 expression level in islets of T1DM patients remains controversial. However, some studies have shown that islet PDL-1 expression might limit the effects of autoimmunity, including autoimmune insulitis.13–15 It has been reported that PDL-1 expression in β cells minimizes the effects of autoimmunity in a mouse model of T1DM (the NOD mouse) and T1DM patients.13–15 Although it is still unknown whether IPCs transplanted into T1DM patients are subjected to immunological attack by cytotoxic T cells, leading to severe insulitis, this strong expression of PDL-1 suggests that transplanted IPCs may avoid immunological surveillance and survive long-term. If generated IPCs avoid autoimmune attack by neogenesis of IPCs, insulin-producing cell autotransplantation may be a new alternative treatment to cure T1DM.
In summary, reversal of diabetes in a murine model was achieved by transplantation of IPCs generated from subcutaneous fat of a T1DM donor and was as effective as IPCs generated from nondiabetic donors. Moreover, the strong expression of PDL-1 in the generated IPCs suggests the potential for immune circumvention. However, further studies are required to determine the applicability of these approaches. Such studies may also provide insights into why β cells are subjected to autoimmunity in T1DM patients.
The authors thank Mark Cleasby, PhD, and Micthell Arico from Edanz Group (https://en-author-services.edanzgroup.com/ac) for editing a draft of this article.
1. Xin Y, Jiang X, Wang Y, et al. Insulin-producing cells differentiated from human bone marrow mesenchymal stem cells in vitro ameliorate streptozotocin-induced diabetic hyperglycemia. PLoS One
2. Oh JE, Choi OK, Park HS, et al. Direct differentiation of bone marrow mononucleated cells into insulin producing cells using pancreatic β-cell-derived components. Sci Rep
3. Lin G, Wang G, Liu G, et al. Treatment of type 1 diabetes with adipose tissue-derived stem cells expressing pancreatic duodenal homeobox 1. Stem Cells Dev
4. Chandra V, Swetha G, Muthyala S, et al. Islet-like cell aggregates generated from human adipose tissue derived stem cells ameliorate experimental diabetes in mice. PLoS One
5. Päth G, Perakakis N, Mantzoros CS, et al. Stem cells in the treatment of diabetes mellitus—focus on mesenchymal stem cells. Metabolism
6. Dadheech N, Shapiro AMJ. Human induced pluripotent stem cells in the curative treatment of diabetes and potential impediments ahead. In: Turksen K, ed. Cell Biology and Translational Medicine, Volume 5. Advances in Experimental Medicine and Biology
vol. 1144. Cham, Switzerland: Springer; 2018:25–35.
7. Ikemoto T, Feng R, Shimada M, et al. A new 2-step acceleration protocol using a histone deacetylase inhibitor to generate insulin-producing cells from adipose-derived mesenchymal stem cells. Pancreas
8. Ikemoto T, Feng R, Iwahashi SI, et al. In vitro and in vivo effects of insulin-producing cells generated by xeno-antigen free 3D culture with RCP piece. Sci Rep
9. Wada Y, Ikemoto T, Morine Y, et al. The differences in the characteristics of insulin-producing cells using human adipose-tissue derived mesenchymal stem cells from subcutaneous and visceral tissues. Sci Rep
10. Ohta S, Ikemoto T, Wada Y, et al. A change in the zinc ion concentration reflects the maturation of insulin-producing cells generated from adipose-derived mesenchymal stem cells. Sci Rep
11. Sneddon JB, Tang Q, Stock P, et al. Stem cell therapies for treating diabetes: progress and remaining challenges. Cell Stem Cell
12. Atkinson MA, Eisenbarth GS, Michels AW. Type 1 diabetes. Lancet
13. Keir ME, Liang SC, Guleria I, et al. Tissue expression of PD-L1 mediates peripheral T cell tolerance. J Exp Med
14. Osum KC, Burrack AL, Martinov T, et al. Interferon-gamma drives programmed death-ligand 1 expression on islet β cells to limit T cell function during autoimmune diabetes. Sci Rep
15. Wang CJ, Chou FC, Chu CH, et al. Protective role of programmed death 1 ligand 1 (PD-L1) in nonobese diabetic mice: the paradox in transgenic models. Diabetes
16. Silva KR, Côrtes I, Liechocki S, et al. Characterization of stromal vascular fraction and adipose stem cells from subcutaneous, preperitoneal and visceral morbidly obese human adipose tissue depots. PLoS One
17. Song X, Hong C, Zheng Q, et al. Differentiation potential of rabbit CD90-positive cells sorted from adipose-derived stem cells in vitro. In Vitro Cell Dev Biol Anim
18. Ikemoto T, Noguchi H, Fujita Y, et al. New stepwise cooling system for short-term porcine islet preservation. Pancreas
19. Hannan NR, Fordham RP, Syed YA, et al. Generation of multipotent foregut stem cells from human pluripotent stem cells. Stem Cell Reports
20. Maehr R, Chen S, Snitow M, et al. Generation of pluripotent stem cells from patients with type 1 diabetes. Proc Natl Acad Sci U S A
21. Sakata N, Egawa S, Sumi S, et al. Optimization of glucose level to determine the stimulation index of isolated rat islets. Pancreas
22. Yoshimura K, Shigeura T, Matsumoto D, et al. Characterization of freshly isolated and cultured cells derived from the fatty and fluid portions of liposuction aspirates. J Cell Physiol
23. Mitchell JB, McIntosh K, Zvonic S, et al. Immunophenotype of human adipose-derived cells: temporal changes in stromal-associated and stem cell-associated markers. Stem Cells
24. Yamada S, Shimada M, Utsunomiya T, et al. Trophic effect of adipose tissue-derived stem cells on porcine islet cells. J Surg Res
25. Saito Y, Shimada M, Utsunomiya T, et al. The protective effect of adipose-derived stem cells against liver injury by trophic molecules. J Surg Res
26. Chen G, Shi X, Sun C, et al. VEGF-mediated proliferation of human adipose tissue-derived stem cells. PLoS One
27. Strand BL, Coron AE, Skjak-Braek G. Current and future perspectives on alginate encapsulated pancreatic islet. Stem Cells Transl Med
28. Tomei AA, Villa C, Ricordi C. Development of an encapsulated stem cell-based therapy for diabetes. Expert Opin Biol Ther
29. Ikemoto T, Noguchi H, Shimoda M, et al. Islet cell transplantation for the treatment of type 1 diabetes in the USA. J Hepatobiliary Pancreat Surg
30. Wada Y, Takata A, Ikemoto T, et al. The protective effect of epigallocatechin 3-gallate on mouse pancreatic islets via the Nrf2 pathway. Surg Today
31. Syed I, Rubin de Celis MF, Mohan JF, et al. PAHSAs attenuate immune responses and promote β cell survival in autoimmune diabetic mice. J Clin Invest