The Animal Care and Use Committee of Xi'an Jiaotong University approved all the animal protocols. Sprague-Dawley rats were purchased from the Laboratory Animal Center, Xi'an Jiaotong University, Xi'an, China. Bamei pigs were purchased from Xi'an Zhuque market, Xi'an, China. Rats (Laboratory Animal Center, Xi'an Jiaotong University, Xi'an, China) were raised in a specific-pathogen-free laboratory (temperature 18–26°C, relative humidity 40–70%) and were provided with free access to food and water.

MSCs were acquired from Sprague-Dawley rats (n=8; male, 3 weeks of age, 60–80 g). The rats were sacrificed by the spinal dislocation method before obtaining the bones. Briefly, the femurs and tibiae were removed in a sterile environment and the bone cavity was lightly flushed with phosphate-buffered saline (PBS) using a 21-gauge needle. The flushed samples were centrifuged at 1,000 rpm for 5 min, and the supernatant was removed. The cells were resuspended in complete medium (Cyagen Biosciences, Guangzhou, China), and incubated by the adherence culture method, as previously descrbied (28) to isolate the MSCs. MSCs were identified by surface molecular markers [CD90 (anti-mouse/rat CD90.1(Thy-1.1) PE; 12-0900-81, eBioscience, San Diego, CA USA) and CD34 (anti-mouse CD34 FITC; 11-0341-82, eBioscience)] with the use of a flow cytometer, and their ability to differentiate into osteoblast-like and adipocyte-like cells was assessed by Alizarin Red S and Oil Red O staining (both from Cyagen Biosciences). All experiments were carried out using MSCs at passage 3–6.

SIS isolation and preparation was performed as previously described (29). Briefly, Bamei pigs (male, 6 months of age, 100–120 kg) were sacrificed by a lethal injection of sodium pentobarbital and the SIS was prepared within 4 h. The jejunum was washed with water, and fat, tunica serosa, and tunica muscularis were removed. The jejunum was dipped in 1 g/l peroxyacetic acid (Shaanxi Three Bridge Chemical Co., Ltd., Shaanxi, China) to sterilize and maintained in PBS until use. SIS was identified by hematoxylin and eosin (H&E) staining and scanning electron microscopy (Hitachi, Tokyo, Japan).

To generate the SIS-MSC scaffold, the SIS was cut into 22×22 mm sections and each section was placed on a Nunclon 35-mm petri dish (Thermo Fisher Scientific, Waltham, MA, USA). Approximately 1×10⁶ MSCs were seeded in the dish and cultured for 48 h to form a confluent monolayer.

Islets were isolated from Sprague-Dawley rats (n=10; male, 8 weeks of age, 280–300 g) as previously described (30). The rats were anesthetized by an intraperitoneal injection of pentobarbital before obtaining the pancreas for the islets. The pancreas was digested with 1 mg/ml collagenase P (Roche, Mannheim, Germany) and purified by Histopaque-1077 (Sigma, St. Louis, MO, USA). After washing with RPMI-1640 medium (Gibco, Carlsbad, CA, USA), the islets were distributed into groups of 200 for culture. Islets were identified by dithizone and their viability was estimated by acridine orange/propidium iodide (AO/PI), described below (both from Gibco).

The islets were divided into 3 groups as follows: group A, islets; group B, SIS-islets; and group C, SIS-MSC-islets. In the controls (group A), 200 fresh islets were cultured alone in non-treated 35-mm Petri dishes. For group B, 200 fresh islets were seeded on SIS. For group C, 200 fresh islets were seeded on the SIS-MSC scaffold.

RPMI-1640 containing 10% fetal calf serum was used for all co-culture configurations. All islets were cultured in a carbon dioxide incubator (37°C, 5% CO₂).

The islets from the 3 groups were collected on days 7 and 14. The viability rate of the islets was estimated by AO/PI as follows: viability rate, % = numbers_green/numbers_{green + red} ×100.

Islet function was measured by a glucose-stimulated insulin secretion test. Briefly, the islets were incubated in RPMI-1640 containing 1.67 mmol/l glucose for 2 h, and then incubated in RPMI-1640 containing 16.7 mmol/l glucose for a further 2 h. Insulin secretion was assessed with an enzyme-linked immunosorbent assay (ELISA) kit (Mercodia, Uppsala, Sweden). The insulin release stimulation index (SI) was calculated as follows: SI = insulin concentration (16.7 mmol/l)/insulin concentration (1.67 mmol/l).

In each group, the islets and cultured supernatants were obtained at day 7. The islets were fixed in 10% paraformaldehyde (Beyotime, Shanghai, China) overnight and embedded in paraffin. The paraffin sections were stained with a rabbit insulin monoclonal antibody (1:100; Cell Signaling Technology, Danvers, MA, USA), followed by 3,3′-diaminobenzidine (DAB) immunostaining. Images were acquired using a light microscope (Olympus, Tokyo, Japan). Cytokines [vascular endothelial growth factor A (VEGFA), ciliary neurotrophic factor (CNTF), epidermal growth factor (EGF), hepatocyte growth factor (HGF) and tumor necrosis factor (TNF)] in cultured supernatants were detected using an ELISA kit (Mercodia) in accordance with the manufacturer's instructions.

RNA was extracted from the islets on day 14 using an RNeasy kit (Qiagen, Valencia, CA, USA). Reverse transcription was performed using a FastQuant RT kit with gDNase (Tiangen, Beijing, China) from 2 mg total RNA. Quantitative PCR (qPCR) was performed using a maxima SYBR-Green qPCR master mix (Thermo Fisher Scientific) protocol on the instructions of CFX96 (Bio-Rad, Hercules, CA, USA). The PCR products were amplified using the following primer sets: for glyceraldehyde 3-phosphate dehydrogenase (Gapdh), 5′-TACCCACGGCAAGTTCAACG-3′ and 5′-CACCAGCATCACCCCATTTG-3′; for pancreatic and duodenal homeobox 1 (Pdx1), 5′-GGAACGCTGGAACAGGGAAG-3′ and 5′-CAGTCTCGGTTCCATTCGGG-3′; for insulin 1 (Ins1), 5′-GACCATTGATTCCGTGACAT-3′ and 5′-CACCAGAGCATAGGAGCGAC-3′; and for vascular endothelial growth factor A (Vegfa), 5′-AGGAGTACCCCGATGAGATA-3′ and 5′-ATCTCTCCTATGTGCTGGCT-3′.

Every sample was tested in triplicate and the results were quantified using the Livak 2^−ΔΔCT method, where CT is the cycle threshold. Relative gene expression was determined by the ΔΔCT method relative to the GAPDH gene expression. All data are presented as the means ± SEM. A value of P<0.05 was considered to indicate a statistically significant difference compared to the control group.

The cultured islets were collected on day 14 and treated using a cell smear centrifuge. The smears were fixed in 4% paraformaldehyde and permeabilized with Triton X-100 (Beyotime). After blocking with 10% donkey serum, the smears were incubated serially with a rabbit insulin monoclonal antibody (#3014) or a mouse cluster of differentiation (CD)31 monoclonal antibody (#3528) (1:100; both from Cell Signaling Technology). The smears were then incubated with secondary antibody (donkey anti-rabbit Alexa 488, 711-545-152, 1:500 or donkey anti-mouse Cy3, 715-165-150, 1:500; Jackson Immunoresearch, West Grove, PA, USA). Finally, the smears ware counterstained with 4′,6-diamidino-2-phenylindole (DAPI; Beyotime). Images were obtained using a confocal laser scanning microscope (Leica). Graphical representations are expressed as the average of mean fluorescence intensity (MFI) using arbitrary units (AU) ± SD. A value of P<0.05 was considered to indicate a statistically significant difference compared with the control group, as determined by the Tukey-Kramer post test.

Sprague-Dawley rats (n=30; male, 6–8 weeks of age, 250–300 g) were rendered diabetic by a single intraperitoneal injection of streptozotocin (STZ, 50 mg/kg). The rats were anesthetized by an intraperitoneal injection of pentobarbital before transplantation. The rats were randomly apportioned into 3 groups as follows: group A, islets (control); group B, SIS-islets; and group C, SIS-MSC-islets. In the controls (group A), the rats received 1,000-islet transplantation under the skin. In group B, 1,000 SIS-coated islets were folded to a size of 1×1 cm and fixed on the back under the skin of the rats. In group C, 5×10⁶ MSCs were seeded onto SIS to generate an SIS-MSC scaffold within 48 h, and 1,000 islets were coated and folded as in group B for transplantation. No immunosuppressive protocols were applied in the recipient rats.

To assess islet graft survival and function, blood glucose and insulin concentrations were monitored. Blood was obtained from the tail vein using a syringe. Successful islet function was defined as blood glucose levels <11.1 mmol/l on 2 consecutive days. Graft rejection was defined as blood glucose levels >11.1 mmol/l on 2 consecutive days. The mean survival time of the grafts was recorded.

The statistical significance of the differences was determined using one-way analysis of variance. Statistical analyses were performed using SPSS17.0 software. A P-value <0.05 was considered to indicate a statistically significant difference. Differences among the groups with regard to blood glucose and insulin concentration were analyzed by multivariate analysis.

To characterize the MSCs, bone marrow cells isolated from rats were stained for surface molecular marks (CD90 and CD34) and analyzed by flow cytometry. We found that the cells were positive for CD90 and negative for CD34 (Fig. 1A). In addition, the cells were able to differentiate into osteoblast-like and adipocyte-like cells (Fig. 1B and C). These results confirmed that the isolated cells were MSCs.

To characterize the islets, islets isolated from rats were identified with dithizone and AO/PI staining. We found that the islets were stained with dithizone and AO/PI (Fig. 1D and E). These results indicated that islet isolation was successful.

To characterize SIS, the SIS from Bamei pigs was observed respectively under a light microscope and scanning electron microscope. We found that the SIS was composed of collagen fiber with no cells (Fig. 1F and G), indicating that the isolation was successful.

To examine the effects of the SIS and SIS-MSC scaffold on islets, their viability and function were examined in vitro. We found that the viability was significantly higher in both the SIS group and SIS-MSC group than in the control group (Fig. 2A). The cell viability in the SIS-MSC group appeared to be superior to that of the SIS group. These results suggest that the SIS and SIS-MSC scaffold increase islet viability.

Islet function was determined with a glucose-stimulated insulin secretion test on days 7 and 14. We found that the SI was significantly higher in both the SIS and SIS-MSC groups relative to the control group (Fig. 2B; P<0.05). The SI was significantly higher in the SIS-MSC group compared with the SIS group (P<0.05). These findings suggest that the SIS and SIS-MSC scaffolds enhanced islet function, and that the SIS-MSC scaffold was superior to the SIS scaffold.

To investigate whether SIS-MSC increases insulin secretion, we analyzed the intensity of insulin staining in the islets by immunohistochemistry and immunofluorescence staining. The results of immunohistochemistry revealed that the intensity of insulin was significantly higher in the SIS-MSC group than in either the SIS group or the control group (Fig. 3A). The insulin signal was undetectable and islet morphology became loose in the control group, whereas the insulin signal was detected and islet morphology was compact in the SIS and SIS-MSC groups. Consistently, the results of immunofluorescence staining indicated that the MFI of insulin was markedly higher in the SIS-MSC group than in the SIS or the control groups (Fig. 3B and C). These results revealed that the SIS-MSC scaffold was associated with an increase in insulin levels and may prevent islet destruction.

Subsequently, we examined the gene expression levels of Ins1 and Pdx1 by RT-qPCR. We found that the levels of Ins1 and Pdx1 were significantly higher in the SIS-MSC group than in the SIS and the control groups, and that there was no significant difference in mRNA levels of Ins1 or Pdx1 between the control and SIS groups (Fig. 3D). These results suggest that the SIS-MSC scaffold rather than the SIS scaffold upregulates the gene expression of Pdx1 and Ins1.

CD31 is a marker of the vascular endothelium (31). To investigate whether the SIS-MSC scaffold improves the microcirculation of islets, we performed an immunofluorescence analysis for CD31. Although the islets were positive for CD31 in the 3 groups, the MFI of CD31 was significantly higher in the SIS-MSC group than in the SIS and the control group (Fig. 3B–C). These results suggest that SIS-MSC scaffold boosts islet microcirculation.

We examined the effects of the SIS-MSC scaffold on cytokine secretion using ELISA. The concentrations of VEGFA, CNTF, EGF and HGF in culture media were significantly higher in the SIS-MSC group than in the SIS group or the control group (Fig. 4A–D). Consistently, the results of RT-qPCR revealed that the mRNA levels of Vegfa were significantly higher in the SIS-MSC group compared with the SIS or the control groups (Fig. 4E). By contrast, the concentrations of TNF in the culture media were significantly lower in the SIS-MSC group than in the SIS or the control groups (Fig. 4F). These results suggest that MSCs can secrete growth factors and may decrease inflammation.

To examine the effects of the SIS-MSC scaffold on graft survival and function, we performed islet transplantation in rats and monitored the blood levels of glucose and insulin. While the blood glucose levels were significantly lower in both the SIS and the SIS-MSC groups than in the control group, these levels were markedly lower in the SIS-MSC group than in the SIS group (Fig. 5A). Consistently, the blood insulin levels and graft survival time were significantly higher in the SIS-MSC group relative to the SIS or the control groups (Fig. 5B and C). These findings suggest that the SIS-MSC scaffold improves islet function and prolongs graft survival.