Insulin-Producing Cells Differentiated from Human Bone Marrow Mesenchymal Stem Cells In Vitro Ameliorate Streptozotocin-Induced Diabetic Hyperglycemia

Abstract

Background: The two major obstacles in the successful transplantation of islets for diabetes treatment are inadequate supply of insulin-producing tissue and immune rejection. Induction of the differentiation of human bone marrow-derived mesenchymal stem cells (hMSCs) into insulin-producing cells (IPCs) for autologous transplantation may alleviate those limitations.

Methods: hMSCs were isolated and induced to differentiate into IPCs through a three-stage differentiation protocol in a defined media with high glucose, nicotinamide, and exendin-4. The physiological characteristics and functions of IPCs were then evaluated. Next, about 3 × 10(6) differentiated cells were transplanted into the renal sub-capsular space of streptozotocin (STZ)-induced diabetic nude mice. Graft survival and function were assessed by immunohistochemistry, TUNEL staining and measurements of blood glucose levels in the mice.

Results: The differentiated IPCs were characterized by Dithizone (DTZ) positive staining, expression of pancreatic β-cell markers, and human insulin secretion in response to glucose stimulation. Moreover, 43% of the IPCs showed L-type Ca2+ channel activity and similar changes in intracellular Ca2+ in response to glucose stimulation as that seen in pancreatic β-cells in the process of glucose-stimulated insulin secretion. Transplantation of functional IPCs into the renal subcapsular space of STZ-induced diabetic nude mice ameliorated the hyperglycemia. Immunofluorescence staining revealed that transplanted IPCs sustainably expressed insulin, c-peptide, and PDX-1 without apparent apoptosis in vivo.

Conclusions: IPCs derived from hMSCs in vitro can ameliorate STZ-induced diabetic hyperglycemia, which indicates that these hMSCs may be a promising approach to overcome the limitations of islet transplantation.

Fig 1. The morphological and phenotypical characteristics of cultured human bone marrow mesenchymal stem cells (hMSCs).

Human bone marrow samples were obtained from healthy volunteers by lumbar puncture. Bone mononuclear cells were isolated from human bone marrow samples by density gradient centrifugation in a Percoll solution (1.073 g/ml) and cultured in L-DMEM with 10% fetal bovine serum. Eight to twelve days later, individual colonies were selected and subcultured. The morphology of cultured cells was observed and the 5^th passages (P5) cells were used to identified the phenotypical characteristics. A: The morphological features of cultured hMSCs at 5th days, the 3rd and 6th passages. B: Cell cycle analysis revealed that most of P5 hMSCs were in the G0/G1 phase (quiescent phase, 86.7 ± 2.8%) with a small population of cells in the S+G2/M phase (active proliferative phase, 14.4 ± 2.8%). C: Flow cytometry analysis disclosed that P5 hMSCs were positive for CD44, CD73, CD166 and CD105, but negative for CD34, CD31, and CD45. D: Immunofluorescence staining showed almost all of P5 hMSCs expressed the antigens of CD73, CD44, CD105, CD166, CD29 and c-kit. E: Under transmission electron microscope, hMSCs exhibited abundant microvilli on cell surface (upper), the irregular shape of nucleus with plentiful euchromatin (middle) and abundant organelles and glycogen granules in cytoplasm (bottom). To detect multi-lineage differentiation potential, P5 hMSCs were cultured in adipogenic, osteogenic, or chondrogenic medium for 2–4 weeks. F: After the induction, lipid droplets in cytoplasm and Oil-Red-positive adipogenic cells were observed. G: The osteogenic differentiation of hMSCs was demonstrated by mineral deposits formed and positive Von Kossa staining (left). The chondrogenic differentiation was reflected by positive Alcian blue staining (right). H: Karyotype analysis revealed the normal karyotype of P5 hMSCs. Scale bars: 50 μm for A, D, F, G and 1 μm for E.

Fig 2. HMSCs differentiated into insulin-producing cells (IPCs).

hMSCs at passage 5 (P5) were induced to differentiate into IPCs by a 30-day protocol. Briefly, hMSCs were induced in high glucose (23.3 mmol/L)-DMEM with 5% FBS for 15 days (referred to as HD-hMSCs); then cultured in L-DMEM medium containing 5% FBS and 20 μmol/L nicotinamide for 7 days; at last, Exendix-4 at 10 μmol/L was added into the medium for another 7-day incubation. After the 30-day induction, cells were fixed or harvested for characterization. A: During the 15 days of induction, small cell aggregates were formed, and during continued culture, more islet-like clusters and round epithelial-like cells appeared. B: The islet-like clusters (middle) and round epithelial-like cells (right) of differentiated hMSCs were positive for Dithizone staining (red color), while the undifferentiated hMSCs were negative. C and D: Immunofluorescence staining showed that islet-like clusters expressed both insulin (C) and c-peptide (D). E: Western blot assay further confirmed the expression of insulin and c-peptide proteins in the induced cells, the fetal pancreas from 6-month-old aborted fetus was used as control. F: The insulin gene expression was evaluated by real-time PCR in hMSCs and IPCs, for which the fetal pancreas was used as positive control. G: Flow cytometry analysis demonstrated that there was 14.5 ± 2.2% of insulin-positive cells and 5.5 ± 1.1% of c-peptide positive cells in the induced cells from P5 hMSCs. All in vitro data were obtained from at least 3 independent experiments. Data are presented as means ± SD (* P<0.05 compared to hMSCs; ^& P<0.05 compared between hMSCs and fetal pancreas). Scale bars: 50 μm for A-D.

Fig 3. Gene expression and insulin-releasing functions of IPCs in vitro.

hMSCs were induced and treated under the same condition described in Fig 2. A: Immunofluorescence staining for PDX-1 (green) and nuclear staining with DAPI (blue) on hMSCs, HD-MSCs and IPCs. The PDX-1 staining in hMSCs was negative, while HD-MSCs and IPCs expressed PDX-1 in nuclei (green). B: Western blot showed that PDX-1 was activated in the induced cells (HD-hMSCs) and the expression of PDX-1 was significantly elevated in IPCs. C: The cell aggregates/clusters from HD-hMSCs and IPCs groups were picked, digested and re-cultured to get the monolayer cells, and the co-staining for PDX-1 (green) and c-peptide (red) was performed on the monolayer cells. hMSCs were negative for both PDX-1 and c-peptide, and HD-hMSCs only expressed the PDX-1, while the monolayer cells from IPCs co-expressed PDX-1 and c-peptide. D: Expression levels of Pdx1, Ngn3, Pax4, Nkx6.1, Glut2 and glucagon genes by quantitative real-time PCR test. The IPCs expressed Pdx1, Ngn3, Pax4, Nkx6.1, Glut2 and Glucagon, similar to the expressing of fetal pancreas. E: The human insulin secreted by IPCs in response to glucose stimulation was examined by ELISA assay. Insulin released amount was calculated by insulin secreted in the culture medium for 2 h after 5.6mM and 23.3mM glucose stimulation. Compared to hMSCs, IPCs show the insulin secretion in a glucose dose-dependent manner. Data are presented as means ± SD from the 3 independent experiments (*P<0.05 compared to hMSCs; ^&P<0.05 compared between 5.6mM and 23.3mM glucose groups). Scale bars: 50μm for A and C.

Fig 4. Measurement of L-type Ca²⁺ channels during differentiation process and intracellular Ca²⁺ concentration in IPCs in response to glucose stimulation.

The whole-cell patch-clamp experiments were performed on hMSCs, HD-MSCs and IPCs to determine the amount and functional change of L-type Ca²⁺ channel during the differentiation process. Nifedipine at 10 μM suppressed the inward currents. Nifedipine-sensitive inward Ca²⁺ currents (I_Ca.L) were recorded by 300-ms voltage step between +10 and +60 mV from a holding potential of –50 mV (to inactivate INa). A: The I_Ca.L was recorded among the hMSCs, HD-MSCs and IPCs. B: The I-V relationship of I_Ca.L displayed the maximum current peak appeared at 10 mV in hMSCs, HD-MSCs and IPCs (left), and the mean peak current amplitudes in IPCs is higher than that in hMSCs or HD-MSCs (right, *P<0.05 compared to hMSCs). C: The changes of intracellular Ca²⁺ concentration in IPCs in response to glucose stimulation were measured by labeling with Fluo-3/AM under laser confocal scanning microscopy. The results were expressed as respective fluorescence intensity of 10 of IPCs (left) and average fluorescence intensity of 30 of IPCs (right). After stimulated with 30mmol/L glucose at 240 s, the intracellular Ca²⁺ concentration rapidly increased in IPCs due to significant Ca²⁺ influx when extracellular Ca²⁺ existing. When Ca²⁺ channels were blocked by 25 mg/ml of Verapamil at 480 s and then stimulated with 3.5 mM glucose at 720 s, intracellular Ca²⁺ rapidly increased again in IPCs due to the release of Ca²⁺ from calcium stores. D: The images of intracellular Ca²⁺ before (left) and after 2 times of glucose stimulation (middle and right) in IPCs, reflected by green fluorescence. Scale bars: 50μm for D.

Fig 5. Proliferation and apoptosis of IPCs.

hMSCs were induced and treated under the same condition described in Fig 2. A: Immunofluorescence staining for PCNA (red) and insulin (green) on hMSCs and IPCs (400×). The PCNA, not insulin was strongly expressed in undifferentiated hMSCs. While a portion of IPCs expressed both PCNA and insulin. B: The IPCs did not show significant apoptosis and necrosis compared with hMSCs examined by Annexin V-EGFP/PI apoptosis kit. Scale bars: 50μm for A.

Fig 6. Transplantation of IPCs into STZ-treated diabetic nude mice in vivo.

IPCs or hMSCs were transplanted into renal capsule of STZ-induced type 1 diabetic male BALB/C nude mice. Blood glucose was measured every 3 days for 21 days. Twenty-one days after transplantation, mice were sacrificed by heart perfusion of 4% phosphate-buffered formalin and the kidneys were isolated, fixed in 10% formalin, embedded and sectioned for histological studies. A: Blood glucose of the diabetic mice implanted with IPCs were normalized and remained euglycemic throughout the observation period of 21 days, whereas those receiving no cells or hMSCs remained hyperglycemic (*P<0.05 compared to non-transplanted or hMSCs groups, n = 5). B: Gross appearance (left) and hematoxylin and eosin staining showed the presence of transplanted IPCs under the renal capsule (arrows). C: TUNEL staining revealed that there was no obvious apoptosis in the transplanted hMSCs or IPCs under the renal capsule. D: Immunofluorescence staining indicated that IPCs expressed insulin, PDX-1 and c-peptide under the renal capsule. Scale bars: 50μm for B-D.