Improvement of islet transplantation in prevascularized muscle sinus tracts

Summary

This study aimed to develop an optimal alternative site for islet transplantation, addressing the limitations of the current clinical standard, the intrahepatic site. In this study, prevascularized muscle sinus tracts (MSTs) in mice were created by implanting and then removing polyamide rods, leveraging a controlled foreign body reaction to generate a highly vascularized, porous collagen matrix. Transplantation of 350 syngeneic islets into these MSTs consistently reversed diabetes in mice, achieving long-term normoglycemia for over 180 days, with efficacy comparable with kidney capsule transplantation. The approach also supported the survival of human islets in immunodeficient mice and demonstrated feasibility in a porcine autotransplantation model. This work establishes the MST as a simple, effective, and clinically promising strategy that enhances islet engraftment through localized prevascularization, potentially improving the therapeutic outcomes of islet transplantation.

Graphical abstract

Highlights

• •Prevascularized muscle sinus tracts can be created via 3-week foreign body reaction
• •Transplantation of 350 syngeneic islets into the site reversed diabetes in mice
• •The approach demonstrated efficacy for human islets and porcine islets
• •Prevascularized muscle sinus tracts can be created in non-human primates

Health sciences; Clinical finding; Clinical procedure as health technology; Biological sciences

Introduction

Type 1 diabetes mellitus (T1D) is a chronic and multifactorial autoimmune disease. In individuals with T1D, infiltration of immune cells into pancreatic islets causes damage or death to the insulin-producing β cells, ultimately resulting in hyperglycemia.¹ Transplantation of isolated pancreatic islets can be an appropriate approach to reverse hyperglycemia for T1D patients because islet transplantation can reduce the risks of pancreatic transplantation and prevent the complications relevant to the daily administration of insulin, such as hypoglycemia and blurred vision.² Since the Edmonton protocol was proposed in 2000, this therapy has demonstrated a more consistent ability to achieve insulin independence.³ In 90% of clinical islet transplantation cases, islet cells are directly transplanted into the liver through the portal vein.⁴ However, there are several notable drawbacks of islet transplantation via portal vein, such as the early graft loss due to instant blood-mediated inflammatory reaction,^5,6,7 as well as the risks of bleeding, embolism,⁸ and localized steatosis. The ability to retrieve the graft is particularly crucial for current research aimed at transplanting islets generated from human pluripotent stem cells,^9,10 but it is challenging when islets are transplanted via the portal vein. These deficiencies indicate that the liver is not the optimal transplantation site for islets.^11,12

We believe that skeletal muscle is a potential site for islet implantation due to its accessibility, capacity for large transplant volumes, and greater oxygen supply. This site has a history of successful use in the autotransplantation of parathyroid glands, demonstrating its ability to support highly metabolically active endocrine tissue over an extended period of time. However, following direct transplantation of islets into muscles, the grafts exhibited necrosis and were replaced by fibrous tissue.^13,14 These reports indicated that muscle was a promising site but should be modified. We hypothesized that the primary obstacle to direct intramuscular islet transplantation is an insufficient initial blood supply, leading to massive early graft loss before revascularization. To overcome this limitation, we sought to create a prevascularized site within the muscle prior to cell delivery.

We derived inspiration from a prevascularized site that was established through the foreign body response (FBR).^15,16 The FBR to an inert implant progressively forms a well-organized, highly vascularized tissue capsule. In this study, we created prevascularized muscle sinus tract (MST) by temporary implantation of polyamide rods (Figure 1A). Specifically, different time periods were set upon polyamide rods implantation to determine the optimal duration of FBR. Subsequent to a 3-week FBR, the isolated islets were transplanted into the sinus tracts following the withdrawal of the implants. It has been demonstrated that the islet grafts enable reversal of diabetes without the need for a permanent encapsulation device or exogenous growth factors. In addition, this approach supported the survival of human islet grafts and was confirmed to be feasible in a large-animal experiment.

Figure 1.

Design and fabrication of the prevascularized muscle sinus tracts

(A) Design of the prevascularized muscle sinus tracts for islet transplantation.

(B) A cylindrical polyamide rod with diameter 1 mm, height 1 cm.

(C) The space in muscles was created, and no active hemorrhage was observed.

(D) Three polyamide rods were implanted in muscles.

(E) The incision was sutured after polyamide rods implantation.

Results

Design and fabrication of the prevascularized transplant site

Three cylindrical polyamide rods with diameter of 1 mm and height of 1 cm were selected as implants to induce FBR in mice (Figure 1B), leading to the accumulation of collagen fibers and significant neovascularization. The surface characteristics of the polyamide rods are listed in Table S1.

The left chest muscles of mice were bluntly separated, and 3 polyamide rods were inserted into the space in the muscles (Figures 1C–1E). Subsequently, a period of time was allowed for FBR, and a collagen fiber tissue with surrounding neovessel structures would be formed around the polyamide rods. Removal of the implants would stop the response, leaving 3 cylinder prevascularized sinus tracts with the same volume as the polyamide rods (24 mm³). This cavity served as an in situ transplant “nest,” providing sufficient volume for graft implantation and survival. Besides, the CCK-8 assay showed that the polyamide rods had no obvious cytotoxicity (Figure S1).

The thin-walled and prevascularized MST was created

To examine the impact of the duration of FBR on the MST, histological analysis was conducted on the cross sections of the MST at different time points (24 h, 1 week, 2 weeks, 3 weeks, and 4 weeks). The schematic illustration is shown in Figure 2A.

Figure 2.

The thin-walled and prevascularized MST was created

(A) A schematic illustration depicting the evolution of the cross section of the MST during FBR.

(B) Masson trichrome staining of the cross sections of the MST at 24h, 1w, 2w, 3w, and 4w. Collagen fibers stained blue, muscles stained red. Scale bars: 1 mm (upper row) and 100 μm (lower row).

(C) Immunohistochemical staining of the unprocessed tissue and MST for macrophage (brown). Scale bars: 100 μm.

(D) Immunofluorescent staining of the unprocessed tissue and MST for CD31 (green). Data are presented as mean ± SD; scale bars: 100 μm (1-way ANOVA followed by the Tukey test). ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

For evaluation of the collagen deposition, inflammatory cells infiltration, and angiogenesis, Masson trichrome staining (Figure 2B), F4/80 immunohistochemical staining (Figure 2C), and CD31 immunofluorescent staining (Figure 2D) were performed on these specimens. At 24 h, minimal presence of collagen fiber and neovascularization was detected. The inflammatory cells aggregated as a result of FBR. During the first week, a substantial deposition of blue collagen fibers encircled the polyamide rods, accompanied by the infiltration of numerous inflammatory cells. Although neovascularization was observable, the vascular density remained low due to the thick collagen fibers. At 2 weeks, a notable reduction in collagen fiber thickness and inflammatory cell count was observed. As the progression of FBR continued, there was a noticeable presence of newly formed blood vessels at 3 weeks. Subsequently, at 4 weeks, collagen fibers were gradually redeposited, albeit with a limited number of inflammatory cells.

Specifically, the deposition of collagen fibers showed a pattern of increasing, decreasing, and increasing again, with the highest amount at week 1 and the lowest at week 3. Inflammatory cells were evenly distributed along the collagen fibers, and the number of nuclei peaked at the first week, suggesting a decrease in inflammation since then. During the FBR, vessel density exhibited a greater increase compared with unprocessed muscle starting from week 2. The peak in vessel density was observed at week 3, attributed to heightened neovascularization and reduced fiber content. Notably, vessel density in the MST was higher than that in the subcutaneous sinus tracts after 3 weeks of FBR (Figure S2).

In addition, we detected extensive differences in tissue inflammatory cytokine expression, in C57BL/6 mice, in response to the polyamide rods at different time points (Figure 3). IL-6 expression was elevated within 1 week, while other indicators were strongly elevated within only 24 h. All cytokines were diminished gradually after the 1-week experiment. These suggested that the FBR could induce temporary and strong cytokine responses, contributing to inflammatory cell recruitment and neovascularization.

Figure 3.

Concentrations of inflammatory cytokines in peri-implant muscle tissue during the foreign body reaction

(A–G) At the indicated time points (0, 24 h, 1, 2, 3, and 4 weeks), peri-implant tissue was harvested to measure the concentrations of (A) IL-1β, (B) IL-6, (C) TNF-α, (D) IL-10, (E) KC, (F) IL-12p70, and (G) IFN-γ. Data are presented as mean ± SD (1-way ANOVA followed by the Tukey test). ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001 (n = 3/time point).

The duration of 3 weeks was optimal for the porosity of MST

In order to facilitate efficient metabolic exchange between the graft and host, it is necessary for the MST to possess a high level of porosity that enables nutrients and metabolites to flow in and out smoothly.^17,18 We applied scanning electron microscopy (SEM) to examine the inner surface morphology of the MST (n = 3 per time point), to investigate whether the development of FBR had any impact on the porosity of the MST (Figure 4A). The unprocessed tissue (n = 3) from the same transplantation site was used as control group. It can be seen that the FBR-induced fiber matrix exhibited a similar porous structure, but with an adjustable average pore size. Based on our analysis, the control group exhibited the highest level of porosity among all groups (Figure 4B). The MST of the 3-week group exhibited higher porosity than the other FBR groups, although there was no significant difference among the 4 FBR groups. In the 4-week group, a dense fibrous mesh was generated, which resulted in decreased porosity and pore size. This may limit efficient metabolic exchange, leading to inadequate energy supply and potentially compromising graft survival. In addition, a significant number of inflammatory cells, red blood cells, and activated platelets could be observed to distribute among the collagen fibrous network at week 1, indicating a strong inflammatory response (Figure 4C). Accordingly, the porosity of MST was greatly influenced by FBR progression. The networks generated by FBR of 2 and 3 weeks were considered as the collagen meshes having favorable porosity to facilitate sufficient dynamic metabolic diffusion.

Figure 4.

The duration of 3 weeks was optimal for the porosity of MST

(A) The surface images of the unprocessed tissue and the MST generated after 1–4 weeks of FBR. Scale bars: left 50 μm, middle 5 μm, right 2 μm.

(B and C) (B) Porosity of the MST at different time points (n = 3/time point). Data are presented as mean ± SD (1-way ANOVA followed by the Tukey test). ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 (C) Inflammatory cells, red blood cells (yellow arrowheads), and activated platelets (white arrowheads) distributed over the fiber matrix after 1 week of FBR. Scale bar: 10 μm.

Based on the above results, the matrix consisted of collagen fibers, newly formed vessels, and scattered inflammatory cells. This suggested that, by controlling the duration of FBR, MST could be created to mimic the ideal transplantation site (Table S2). The sinus tracts of 2–3 weeks’ duration were considered the physiologically optimal microenvironment for successful transplantation.

Islet grafts demonstrated glucose control function in the first month

We transplanted ∼350 islets from C57BL/6 mice into the MST of 2-week FBR (n = 7), MST of 3-week FBR (n = 9), and under the kidney capsule (KC) (n = 5) (Figure 5A). Islets were also directly transplanted into the muscles (DTM, n = 6).

Figure 5.

Islet grafts demonstrated glucose control function in the first month

(A) Diagram of the syngeneic transplantation protocols.

(B) Nonfasting blood glucose measurements of recipient mice and control mice after transplantation. Black dashed line represents a nonfasting physiological maximum (11.1 mM).

(C) The area under the blood glucose curve of each group, expressed as mmol/L/30 days (1-way ANOVA followed by the Tukey test).

(D) The proportion of recipients that achieved euglycemia. p = 0.0262 for the comparison of the 3w MST and KC groups, p = 0.0157 for the comparison of the 3w MST and DTM groups, p = 0.1608 for the comparison of the 3w MST and 2w MST groups (Log rank [Mantel-Cox] test).

(E) IPGTT blood glucose curve post dextrose bolus.

(F) Areas under the blood glucose curve of IPGTT, expressed as mmol/L/120 min (1-way ANOVA followed by the Tukey test).

(G) H&E staining of islet grafts in the 3w MST after 30 days (red arrowheads).

(H) Fluorescent staining of islet grafts in the 3w MST for insulin (red), glucagon (green), and nuclei (blue).

(I) Fluorescent staining of islet grafts in the 3w MST for insulin (red), CD31 (green), and nuclei (blue). Scale bar: 100 μm. Data are presented as mean ± SD; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗∗p < 0.0001.

The initial nonfasting blood glucose levels among the KC group (26.8 ± 3.0 mM, n = 5), 2w MST group (28.6 ± 3.3 mM, n = 7), 3w MST group (26.0 ± 4.4 mM, n = 9), and DTM group (29.0 ± 4.8 mM, n = 6) were not significantly different (p > 0.05) (Figure 5B). At the end of the first month, the blood glucose levels among the KC group (9.1 ± 1.9 mM), 2w MST group (15.0 ± 8.3 mM), and 3w MST group (11.4 ± 5.6 mM) were all lower than the DTM group (29.0 ± 8.3 mM) (p < 0.01). Figure 5C showed no significant difference in the area under the curve (AUC) of nonfasting blood glucose among the naive group (300.4 ± 20.4 mmol/L/30 days, n = 5), KC group (320.0 ± 24.4 mmol/L/30 days, n = 5), and 3w MST group (408.3 ± 127.3 mmol/L/30 days, n = 9) (p > 0.05). However, the AUC of nonfasting blood glucose was higher in the DTM group (941.0 ± 62.0 mmol/L/30 days, n = 6) compared with the 3w MST group (p < 0.0001), and in the 2w MST group (566.1 ± 162.9 mmol/L/30 days, n = 7) compared with the naive group (p < 0.01) and KC group (p < 0.01).

Syngeneic islets transplanted under the KC reversed diabetes in 100% of recipients (5/5) within 30 days (Figure 5D). Transplants in the unmodified muscle space did not reverse diabetes (0%; 0/6). Transplants in the MST space of 3w FBR reversed diabetes in 66.7% of mice (6/9), better than the unmodified muscle space (p < 0.05), but not as effective as the KC space (p < 0.05). There was no notable difference in reversal rate between the 2w (42.9%; 3/7) and 3w MST groups, but reversal of hyperglycemia in the 2w MST group (18.0 ± 2.0 days) was delayed compared with the 3w MST group (10.3 ± 2.0 days) (p < 0.001, Student t test).

We conducted intraperitoneal glucose tolerance test (IPGTT) at day 30 after transplantation (Figure 5E). KC (n = 5), naive (n = 3), and 3w MST (n = 6) groups had well-preserved glucose clearance profiles that were not significantly different. AUCs ±standard deviation (SD) for glucose clearance were similar (Figure 5F; KC: 1,452.8 ± 110.3 mmol/L/120 min, naive: 1,587.0 ± 140.0 mmol/L/120 min, 3w MST: 1,322.9 ± 291.9 mmol/L/120 min, p > 0.05, ANOVA). Glucose profile in the 2w MST group (AUC 1,874.7 ± 102.75 mmol/L/120 min) was not as good as that of the 3w MST group (p < 0.05).

The above results indicated that therapeutic efficacy was better in the 3w MST group than in the 2w MST group. Furthermore, the result of H&E staining indicated the survival of islet grafts for a duration of 30 days in the 3w MST, as depicted in Figure 5G. The presence of insulin + glucagon and insulin + CD31 immunofluorescence staining demonstrated preservation of islet morphology and vascularization (Figures 5H and 5I).

Islet grafts corrected hyperglycemia in all MST mice and demonstrated long-term function

We followed graft function for 182 days to measure long-term performance. Grafts were subsequently retrieved at day 182 (Figures 6A–6C). The results of monitoring of blood glucose were shown in Figure 6D. Both the 3w MST group (n = 9) and KC group (n = 5) mice maintained normoglycemia for more than 180 days. There was no obvious improvement in blood glucose levels in the DTM group (n = 5). Subsequently, the islet grafts in the 3w MST group were completely resected. As a result, mice returned to severe hyperglycemia. Figure 6E showed no significant difference in the AUC of nonfasting blood glucose among the naive group (1,574.5 ± 58.1 mmol/L/182 days, n = 5), KC group (1,818.2 ± 93.5 mmol/L/182 days, n = 5), and 3w MST group (1,847.9 ± 311.3 mmol/L/182 days, n = 9) (p > 0.05). The AUC of nonfasting blood glucose was higher in the DTM group (5,196.7 ± 1,080.6 mmol/L/182 days, n = 5), compared with the 3w MST group (p < 0.0001).

Figure 6.

Islet grafts corrected hyperglycemia in all MST mice and demonstrated long-term function

(A) The islet grafts were found in the MST at day 182 after transplantation (yellow arrowheads).

(B) Islet grafts in the MST space were wholly removed at day 182 after transplantation.

(C) The incision was sutured after grafts retrieval.

(D) Nonfasting blood glucose measurements of recipient mice and control mice over 182 days. Three sham mice died at day 28, day 91, and day 168, respectively. Black dashed line represents a nonfasting physiological maximum (11.1 mM).

(E) The area under the blood glucose curve, expressed as mmol/L/182 days (1-way ANOVA followed by the Tukey test).

(F) The proportion of recipients achieved euglycemia. p = 0.0006 for the comparison of the 3w MST and DTM groups, p = 0.0262 for the comparison of the 3w MST and KC groups, p = 0.0017 for the comparison of the KC and DTM groups (Log rank [Mantel-Cox] test).

(G) IPGTT blood glucose curve post dextrose bolus.

(H) Areas under the blood glucose curve of IPGTT, expressed as mmol/L/120 min (1-way ANOVA followed by the Tukey test).

(I) Body weight changes in the mice from different groups for 90 days after transplantation (1-way ANOVA followed by the Tukey test).

(J) The levels of fasting serum insulin in different groups (1-way ANOVA followed by the Tukey test). Data are presented as mean ± SD; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

All the syngeneic islet grafts in the 3w MST group reversed hyperglycemia within 105 days (100%, 9/9) (Figure 6F). However, the grafts in the KC group reversed hyperglycemia within 12 days (100%, 5/5), better than the 3w MST group (p < 0.05). Only 1 of the transplants in the DTM group reversed diabetes within 182 days (20%,1/5), not as effective as the 3w MST group (p < 0.001) and the KC group (p < 0.01).

IPGTT was conducted at day 180 (Figure 6G). As shown in Figure 6H, the AUCs of IPGTT in the 3w MST group (n = 6), the KC group (n = 5), and the naive group (n = 3) were 1,464.0 ± 60.9, 1,548.5 ± 246.6, and 1,630.8 ± 85.4 mmol/L/120 min, respectively. No significance was detected among the 3 groups (p > 0.05), showing satisfactory glucose clearance curve (Figure 6G). Conversely, the DTM group (n = 5) was intolerant to the glucose challenge (AUC, 3,574.5 ± 487.2 mmol/L/120 min vs. the 3w MST group, p < 0.0001). Data from 3 mice of the 3w MST group were excluded from the final analysis due to a technical complication during the procedure, where the injected glucose solution leaked from the peritoneal cavity, preventing a valid assessment of glucose tolerance.

In Figure 6I, the weight changes of different groups in post-transplant 90 days were compared. The weight change of mice in the 3w MST group was 29.3 ± 9.0% (n = 9). There was no significant difference between the 3w MST group and KC group (38.2 ± 10.4%, n = 5, p > 0.05), the 3w MST group and DTM group (13.2 ± 1.5%, n = 5, p > 0.05). The naive group demonstrated the optimal increasing trend (47.0 ± 6.5% vs. the 3w MST group, n = 5, p < 0.05). In the DTM group, the weight of mice increased more slowly than that of the KC group (p < 0.01) and naive group (p < 0.001).

The fasting serum insulin of mice was measured during a week before graft retrieval (Figure 6J). The serum insulin level of the 3w MST group (741.2 ± 52.88 pg/mL) was significantly higher than that observed in the DTM group (186.5 ± 118.9 pg/mL, p < 0.001), but was close to the level of serum insulin in the KC group (647.6 ± 104.2 pg/mL, p > 0.05) and the naive group (818.8 ± 208.7 pg/mL, p > 0.05).

In order to confirm that the potential of the prevascularized MST space is not unique to C57BL/6 mice, we studied BALB/c mice, the islet grafts of which also maintained normoglycemia for more than 182 days (Figure S3).

The islet grafts were functional and had a rich blood supply

Syngeneic islet grafts in the 3w MST group and the KC group were harvested at day 182 after transplantation. The Masson trichrome staining of the islet grafts in the MST was shown in Figure 7A, which demonstrated that islet grafts still survived in the MST and were surrounded by a small amount of collagen fibers. In addition, it revealed the presence of adipose tissue around the islet grafts at day 182, which was not observed before transplantation (Figure 2B). The collagen fibers surrounding the grafts were thinner when compared with that before transplantation (Figure S4). The insulin immunohistochemistry staining and insulin + glucagon double immunofluorescence staining were shown in Figures 7B and 7C, respectively, suggesting the durability of islet graft function. The insulin + CD31 double immunofluorescence staining indicated that the islet grafts and peripheral tissues had an abundant blood supply (Figure 7D). Furthermore, the H&E staining and insulin immunohistochemistry staining of the islet grafts under the kidney capsule were shown in Figures 7E and 7F, respectively.

Figure 7.

The islet grafts were functional and had a rich blood supply (182 days)

(A) Masson trichrome staining of islet grafts (red arrowhead) in the MST space. The grafts were surrounded by collagen fibers (black arrowhead). Adipose tissue (yellow arrowhead) was observed around the grafts. Scale bar: 500 μm.

(B) Immunohistochemistry staining of islet grafts in the MST space for insulin (brown). Scale bar: 200 μm.

(C) Fluorescent staining of islet grafts in the MST space for insulin (red), glucagon (green), and nuclei (blue). Scale bar: 100 μm.

(D) Fluorescent staining of islet grafts in the MST space for insulin (red), CD31 (green), and nuclei (blue). Scale bar: 200 μm.

(E) H&E staining of islet grafts (red arrowheads) under the kidney capsule. Scale bar: 200 μm.

(F) Immunohistochemistry staining of islet grafts under the kidney capsule for insulin (brown). Scale bar: 200 μm.

(G) Fluorescent staining of an islet in pancreas for insulin (red), glucagon (green), and nuclei (blue). Scale bar: 100 μm.

(H) Relative proportion of α and β cells within native and transplanted islets. There were no significant differences between native and transplanted islets (Unpaired t test) (p > 0.05).

(I) Fluorescent staining of an islet in pancreas for insulin (red), CD31 (green), and nuclei (blue). Scale bar: 100 μm.

(J) Relative area of CD31 within native and transplanted islets (Unpaired t test) (p > 0.05). Data are presented as mean ± SD.

The transplantation process and muscle site for islet grafts did not exhibit a significant impact on cell composition when compared with natural pancreatic islets (Figures 7G and 7H). The relative proportions of β cells were 89.40% ± 3.58% in normal islets and 86.80% ± 2.68% in islet grafts (p > 0.05). Additionally, an analysis was conducted on the vessel density in both normal islets and islet grafts (Figures 7I and 7J). The result showed no statistically significant variance in the relative areas of CD31. This finding was consistent with the theoretical analysis suggesting that increased neovascularization promotes the survival of grafts.

Allogeneic transplantation in MST was effective

To determine whether the MST transplant technique is efficacious across an alloimmune barrier, we transplanted ∼350 islets from BALB/c mice into the MST of C57BL/6 mice after 3 weeks FBR (Figure 8A), in the presence or absence of tacrolimus therapy (0.5 mg/kg/d for 28 days). As shown in Figure 8B, rejection occurred within 6 days after normoglycemia in the AT-MST control group (median survival: 9.0 ± 3.2 days). Tacrolimus therapy led to prolonged allograft survival (median survival: 20.67 ± 5.1 days). The allograft survival of the AT-MST control group was significantly shorter than that of the AT-MST tacrolimus group (p < 0.0001).

Figure 8.

Allogeneic transplantation in MST was effective

(A) Diagram of the allogeneic transplantation protocols.

(B) Impact of an allogeneic barrier upon diabetes reversal using the MST space, p < 0.0001 (Log rank [Mantel-Cox] test).

(C) H&E staining of islet grafts (red arrowheads) in the AT-MST tacrolimus group, 16 days post-transplant.

(D) Masson trichrome staining of islet grafts in the AT-MST tacrolimus group, 16 days post-transplant. Islet grafts (red arrowheads) surrounded and infiltrated by collagen fibers (black arrowheads).

(E) CD4 immunohistochemical staining (brown) of allografts in the AT-MST tacrolimus group, 16 days post-transplant.

(F) Insulin immunohistochemical staining (brown) of islet grafts in the AT-MST tacrolimus group, 16 days post-transplant.

(G) H&E staining of islet grafts (red arrowheads) in the AT-MST control group, 12 days post-transplant.

(H) Masson trichrome staining of islet grafts in the AT-MST control group, 12 days post-transplant. Islet grafts (red arrowheads) were infiltrated and replaced by large amount of collagen fibers (black arrowheads).

(I) Insulin immunohistochemical staining (brown) of islet grafts in the AT-MST control group, 12 days post-transplant. Scale bar: 200 μm.

The H&E staining and Masson trichrome staining of the rejected islet grafts in the AT-MST tacrolimus group were shown in Figures 8C and 8D, respectively. Islet grafts were surrounded and infiltrated by collagen fibers. In addition, immunohistochemistry revealed the presence of intra-graft CD4⁺ cell infiltration (Figure 8E), a crucial factor in the process of allograft rejection. As a result, insulin immunohistochemical staining revealed partial graft loss due to rejection (Figure 8F). In the AT-MST control group, most of the grafts were replaced by collagen fibers (Figures 8G and 8H), leading to near-total loss of the allogeneic islet (Figure 8I).

These results showed that, while the MST space was not immunologically privileged, the function of allogeneic islets could be effectively prolonged with immunosuppressive therapy.

Prevascularized MST was suitable for the transplantation of human islets

To determine whether human islets can function in the MST space, an in vivo experiment was carried out using nude mice (Figure 9A). Four cylindrical polyamide rods (diameter, 1 mm; length, 1.2 cm) were placed in chest muscles of mice. The polyamide rods were removed after 3 weeks, and ∼2,000 human islet equivalents (IEQs) were transplanted into the MST of diabetic nude mice (Figure S5). All mice in the sham group remained hyperglycemic (Figure 9B). In marked contrast, hyperglycemia of the diabetic recipient mice was improved. Blood glucose levels were compared at 30 days post-transplantation. The blood glucose level in the Nude-MST group (14.6 ± 1.0 mmol/L) was significantly lower than that in the Nude-Sham group (23.4 ± 2.2 mmol/L) (p < 0.01). However, glycemic control was superior in the Nude-KC group compared with the Nude-MST group. The AUC for nonfasting blood glucose was higher in the Nude-MST group (645.3 ± 177.3 mmol/L/30 days, n = 3) than in the Nude-KC group (294.2 ± 21.21 mmol/L/30 days, n = 3) (p < 0.01).

Figure 9.

Prevascularized MST was suitable for the transplantation of human islets

(A) Diagram of the transplantation protocols.

(B) Nonfasting blood glucose measurements of different groups.

(C and D) H&E staining of human islet grafts (red arrowheads) in the MST space after 30 days. Scale bar: 1 mm (C), 500 μm (D).

(E) Fluorescent staining of the human islet grafts for insulin (red), glucagon (green), and nuclei (blue). Scale bar: 100 μm. Data are presented as mean ± SD.

We further demonstrated that the human islet cells still survived after 30 days (Figures 9C–9E). The presence of insulin + glucagon immunofluorescence staining demonstrated preservation of islet morphology.

Efficacy of the prevascularized MST in large mammals

We investigated the FBR induced by a polyamide rod (diameter, 3 mm; length, 7 cm) implanted in porcine muscles. Our results revealed that the fibrous capsule surrounding the implant was thinnest at 3 and 4 weeks post-implantation (Figure S6). Furthermore, neovascularization was most prominent at the 3-week time point (Figure S7). Therefore, we concluded that 3 weeks represents the optimal time point for subsequent islet transplantation.

In order to evaluate the efficacy of the novel site for islet transplantation in large mammals, autologous transplantation was conducted in 2-year-old pigs (Figure 10A). A polyamide rod was surgically inserted into the muscular space following blunt dissection (Figures 10B and 10C), with each side of the chest capable of accommodating 4 polyamide rods. After a period of 3 weeks, the polyamide rods were removed (Figure 10D), and islet autotransplantation was performed (Figure 10E). Partial pancreas tissue was obtained surgically for the purpose of islet isolation (Figures S8A and S8B), resulting in approximately 10,000 IEQs with a purity exceeding 85% (Figure S8C) being transplanted into each MST space. Approximately 10,000 IEQs were transplanted into unprocessed muscular space as a control (pig-DTM group). The islet grafts were harvested 30 days later. The islet grafts within the pig-DTM group exhibited necrosis and were encased by a significant amount of fibrous tissue (Figure 10F). In contrast, pig islets were clearly visible within the MST site (Figures 10G–10I), and fluorescent staining with insulin revealed survival and function of islets (Figures 10J and 10K).

Figure 10.

Prevascularized MST effectively improved pig islet autotransplantation

(A) Diagram of the transplantation protocols.

(B) Space in muscles was created by blunt dissection.

(C) A polyamide rod was inserted into the space.

(D) Polyamide rods were pulled out after 3 weeks

(E) Pig islets were transplanted into the MST space.

(F) H&E staining of islet grafts (red arrowhead) in unprocessed muscular space. The grafts were necrotic with a lot of fibrous tissue. Scale bar: 2 mm.

(G) H&E staining of pig islet grafts (red arrowhead) in the MST space. Scale bar: 1 mm.

(H and I) Masson trichrome staining of pig islet grafts (red arrowheads) in the MST space. Scale bar: 500 μm (H), 200 μm (I).

(J and K) Fluorescent staining of the pig islet grafts (yellow arrowheads) for insulin (green) and nuclei (blue). Scale bar: 200 μm (J), 50 μm (K).

To investigate the translational potential of our approach, we evaluated the feasibility of generating a prevascularized sinus tract in non-human primates. We implanted sterile polyamide rods (4 cm in length, 3 mm in diameter) into the intermuscular space of a cynomolgus monkey (Figure S9), retrieving specimens at various time points for histological analysis. Histology revealed a stable, thin fibrous capsule surrounding the implant at weeks 3 and 4 (Figure S10). Immunofluorescence for the endothelial marker CD31 identified an initial accumulation of positive cells at week 2 (Figure S11), culminating in a peak density with extensive neovascularization at week 3. This was followed by a significant vascular regression by week 4 (p < 0.0001 vs. week 3). These results demonstrate that the FBR in primates generates a hyper-vascularized fibrous sinus tract, with the 3-week time point representing the optimal therapeutic window for subsequent islet transplantation.

Discussion

Islet transplantation is regarded as a promising therapeutic approach for individuals with type 1 diabetes due to its potential to enhance blood glucose regulation and achieve insulin independence.¹⁹ Ongoing research is focused on exploring alternative transplantation sites in order to identify more favorable and optimal locations for islet engraftment. Furthermore, the short-term reconstruction of the microenvironment and vascular network around the islets is essential for the reason that the destruction of microvessels resulting from islet isolation impairs effective nutrient delivery and metabolic waste removal.^20,21,22 The perfusion requirements of a graft tissue do not always match the slow growth rate of newly formed microvessels, which has led to the concept of prevascularization. The objective of this study was to create a vascular-rich microenvironment in muscles for islet grafts.

The prevascularized microenvironment was created in our study by temporary FBR, independent of additional administration of expensive growth factors,^23,24 other accessory cells,^25,26,27 or permanent encapsulation devices,^28,29 when compared with other prevascularization strategies. Polyamide can elicit a foreign body reaction in vivo and has been used for creating a subcutaneous prevascularized site.¹⁵ In addition, it is widely utilized in surgical materials, specifically as sutures.^30,31 The selection of 3 polyamide rods (1 mm in diameter, 1 cm in length) was based on our preliminary experimental experience. In our preliminary experiments, we observed that, when islets were injected into the sinus tract, they dispersed and became suspended in the medium. We attempted to transplant 350 islets into two sinus tracts; however, after approximately 70% of the grafts had been delivered, continued injection resulted in islets overflow. To minimize graft loss, we therefore opted to create three sinus tracts.

This configuration created sinus tracts with a volume of approximately 24 mm³, which was sufficient for 350 islets without impairing the mouse normal activities. Furthermore, a cylindrical tract is less prone to collapse than a planar pocket, thus ensuring the creation of a stable and well-defined space.

Thus, polyamide rods were selected in our study. Following implantation, blood proteins attached to the polyamide rods, leading to the formation of a porous collagen matrix as a result of FBR. The FBR process typically consists of five phases: protein adsorption, acute inflammation, chronic inflammation, foreign body giant cell formation, and fibrous capsule aggregation.^32,33 In the field of bioengineering, the overproduction of collagen fibers and persistent inflammatory reactions can detrimentally impact the functionality of materials and the success rate of grafts, which should be avoided.^33,34 There were no published reports of specific FBR process caused by polyamide materials. Thus, it is crucial to define the precise FBR duration for generating a suitable transplant site based on the systematic study of its physiological properties. Our primary goal is to establish a collagen fibers matrix that is porous and rich in vascularity, while also avoiding intense inflammatory infiltration and excessive fibrillar deposits. During the process of FBR, the newly generated tissue experienced a dynamic change. Our observations indicate that the collagen fiber thickness of MST exhibited a dynamic process characterized by an initial thickening, followed by thinning, and subsequent re-thickening. According to the existing literature,^32,35 we believe that the resolution of the acute inflammatory response, along with the progressive replacement of type III collagen by type I collagen, contributes to the observed gradual thinning of collagen fibers from the first to the third week. However, fibrous deposition persists throughout the chronic inflammatory phase.³³ Consequently, the fiber thickness observed in the fourth week exceeds that of the third week. The FBR process transformed the transplant site from a dense and avascular site with significant inflammatory cell infiltration to a well-vascularized and porous site within a period of 2–3 weeks, which is regarded as the optimal time frame for islet transplantation.

The following transplant results indicated significantly improved blood glucose levels in the 2w and 3w MST groups, compared with the DTM group. The influence of FBR duration on the transplantation efficiency was verified. In order to determine the more precise time point of transplantation, the effects of glucose control of the 2w and 3w MST groups were compared. Results indicated a delayed reversal of nonfasting hyperglycemia in the 2w MST group, as well as significantly poorer glucose clearance function when compared with the 3w MST group. Therefore, our study demonstrated that a 3-week FBR is feasible in MST. The MST space is a promising site for islet transplantation. However, our results showed a delayed reversal of hyperglycemia in the 3w MST group compared with the KC group. We speculate this is likely due to the richer microvasculature under the kidney capsule. A key future direction is therefore to investigate strategies for enhancing vascular density within the MST to improve therapeutic outcomes.

Based on the assessments of 30-day transplantation effect, the 3w MST group was chosen for observation exceeding 180 days to analyze the long-term effects of the islet grafts. Islet grafts survived in the MST and exhibited strong glycemic control capability for over 180 days, as evidenced by fasting serum insulin levels. All syngeneic islet grafts in the 3w MST group successfully reversed hyperglycemia within 105 days. The glucose clearance function of islet grafts in the 3w MST group closely resembled that of healthy mice. Furthermore, prompt reversion to a diabetic state in the 3w MST group following complete excision of the islet graft indicated that euglycemia was dependent on graft function. A similar effect on improving glycemic control was also demonstrated in the BALB/c mouse model. Our study confirmed the survival of the graft through histological analysis, revealing a significant presence of vascular structures within and surrounding the islet grafts, which may contribute to its long-term survival. Besides, adipose tissue was found surrounding the islet grafts after 182 days. We believe that there are two possible explanations. First, the formation of adipose tissue may relate to the paracrine effects of insulin released by the transplanted islets, leading to localized hyperinsulinemia and the promotion of lipogenesis.³⁶ Hepatic steatosis has also been observed in 20–40% of individuals who have undergone clinical islet transplantation.^37,38 Previous studies have found that brown adipose tissue can support islet organoid survival and delay immune rejection.^39,40 Similarly, in our study, the regenerated adipose tissue may also have promoted the survival of the islets. Second, the surplus space after transplantation facilitated the infiltration of adipose tissue.

The largest issue facing the field of transplant immunology is graft loss due to immune-mediated rejection. In recent years, researchers have employed a variety of materials for the encapsulation of islets in order to mitigate the risk of immune rejection.^28,41 Our study suggests that the MST site can support islet engraftment across a strong allogeneic barrier in mice, but only when tacrolimus immunosuppression is given to avert rejection. Therefore, the MST is not an immune-privileged site. To further investigate the clinical potential of this approach, we transplanted human islets into immunodeficient nude mice, to a certain extent, correcting hyperglycemia. We observed that the efficacy of the Nude-MST group was inferior to that of the Nude-KC group. We suggest this may be due to the following reasons. On the one hand, the subcapsular kidney space offered ample room for islet dispersion and it possesses a superior vascular network; both factors facilitate better contact and connection between islets and blood vessels. The space in the murine MSTs, however, was relatively limited, causing islet aggregation after a large-volume transplant, which led to insufficient oxygen supply to the interior of the graft. On the other hand, compared with the natural microenvironment of the kidney capsule, the sinus tracts contained a small number of aggregated inflammatory cells, which might have negatively impacted the islet graft. Our next steps will focus on creating a larger space to disperse the islets and enhancing the vascular density of the MSTs. Compared with the islet viability matrix system reported by Yu et al.,⁴²which achieved long-term euglycemia in mice with as few as 400 IEQs human islets subcutaneously, our MST site may still present a suboptimal microenvironment for early graft function. The islet viability matrix system, composed of collagen I and nutrient supplements, provides immediate extracellular matrix support and essential nutrients such as L-glutamine, which enhance islet survival and glucose-responsive insulin secretion. In contrast, while our MST site promotes vascularization, it may lack comparable early nutrient diffusion or matrix-derived survival signals, leading to initial functional impairment of islets—particularly given the known fragility of human islets and their susceptibility to isolation process and inflammatory stress. Species-specific differences in growth factors and immune reactivity may further compromise human islet function in murine hosts. For our approach, supplementing a special extracellular matrix during islet transplantation could be an effective strategy.

One significant finding of the study was the successful demonstration of long-term islet survival in a large-animal model. The porcine model shares physiological similarities in muscles with humans. In comparison with direct intramuscular islet transplantation in the porcine model, the prevascularized sinus tracts provided a beneficial microenvironment for islet survival. However, when contrasted with the mouse model, the relative insufficiency of vascular density in the porcine sinus tracts may have compromised the transplantation outcome. In subsequent experiments, we aim to improve porcine islet survival by combining polyamide rods with vascular growth factors or co-transplanting islets with microvascular fragments. In this experiment, only a part of pancreas was acquired for isolation, and the pigs sustained euglycemia, because total pancreatectomy would make the pigs very weak. We incorporated a non-human primate model to monitor the FBR process. Given the higher similarity of a non-human primate to humans in terms of muscle tissue architecture, we believe these findings provide valuable preclinical evidence for optimizing clinical islet transplantation strategies and guide the direction of our future experimental improvements.

A range of animal subjects was employed in this study. Specifically, C57BL/6 and BALB/c mice were used to demonstrate that this approach could effectively maintain long-term euglycemia. Consequently, the observation endpoint for the mice was six months, while the endpoint for other animal models was one month. The underlying rationale for the experimental design is detailed in Table 1.

Table 1.

The rationale for the animal experimental design

Logical step	Animals	Purpose
1	C57BL/6 mice BALB/c mice	Establish long-term therapeutic efficacy and allogeneic transplantation feasibility
2	Nude mice	Confirm the function of human islets
3	Pig	Validate the approach in a large mammal
4	Primate	Evaluate the FBR process in a primate for clinical translation

In addition, we reviewed the literature on islet transplantation into prevascularized subcutaneous sinus tracts, as summarized in Table 2. In most previous reports, the required waiting period for the FBR to establish adequate vascularization was 4 weeks. Notably, even with islet doses as high as approximately 500, most studies did not achieve a 100% diabetes reversal rate. In contrast, our study demonstrates that a shorter waiting period of 3 weeks is sufficient for preparing sinus tracts in the intramuscular site. Furthermore, utilizing this intramuscular approach, we achieved a 100% long-term normoglycemia rate with a moderate dose of 350 syngeneic islets. This may be due to the higher vascular density in muscles and the faster vascularization in the MST. We believe there is no significant difference in the accessibility of the subcutaneous and intramuscular sites. This is based on our practical findings that, in certain areas, such as the inframammary region, the ulnar aspect of the arm, and the inguinal area, the subcutaneous fat layer is relatively thin. After skin incision and subcutaneous tissue dissection, the underlying muscle is visible, making intramuscular implantation relatively convenient. Compared with the subcutaneous site, the intramuscular site is more invasive, carrying potential risks such as bleeding, nerve damage, and muscle damage. The application of ultrasound guidance combined with blunt dissection for space creation can effectively mitigate these risks.

Table 2.

Mouse islet transplantation in prevascularized sinus tracts formed by FBR

Site	Time for FBR	Transplant dose of syngeneic islets	Reversal of diabetes (%)	Reference
Subcutaneous	4 weeks	∼500	91.3 (100 days)	Pepper et al.15
Subcutaneous	4 weeks	150	72 (100 days)	Pepper et al.43
Subcutaneous	2 weeks	500	50 (56 days)	Okada et al.44
Subcutaneous	4-5 weeks	200	95 (100 days)	Pepper et al.45
Subcutaneous	4 weeks	500 ± 10%	100 (28 days)	Pepper et al.46
Intramuscular	3 weeks	350	100 (105 days)	This study

In summary, we report a study on generating a preprogrammed transplant site using a controlled FBR and improving the islet graft survival successfully. The optimal time point for FBR was determined to be 3 weeks. Our study illustrated that the transplantation of syngeneic islets into the MST effectively reversed the hyperglycemia in different mouse strains. The results of allogeneic transplantation suggested that the MST was not an immune-privileged site. Furthermore, our findings demonstrated the feasibility of utilizing clinical-grade human islets with the approach. Finally, porcine islet grafts successfully survived and functioned in large-animal autotransplantation. These results suggest that our innovative transplant site and the establishment of a prevascularized microenvironment may offer a promising avenue for advancing human islet therapy in the future.

Limitations of the study

There are still some limitations in our research. The blood glucose regulation achieved by the transplantation of 350 islets into the MST was comparable with that observed with islets transplanted under the kidney capsule. However, the marginal mass of islets warrants further investigation. Moreover, the 3-week duration of FBR may be a little long for clinical recipients. Therefore, optimizing the construction of the MST and reducing the prevascularization period necessitate additional research. In the study of human islet transplantation into mice, we attribute the failure to completely correct hyperglycemia to two primary factors: limited space and a high islet dosage. In subsequent research, we will aim to expand the transplantation space and enhance the vascular density of the MST to provide a more favorable environment. Additionally, it remains uncertain whether the islet grafts have the capacity to ameliorate hyperglycemia in pigs. In the subsequent phase of the research, we will investigate the potential benefits of the approach for glycemic control in large animals, with the aim of progressively transitioning to human studies.

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Sheng Yan (shengyan@zju.edu.cn).

Materials availability

This study did not generate new unique reagents.

Data and code availability

• •All data reported in this paper will be shared by the lead contact upon request.
• •This paper does not report original code.
• •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Acknowledgments

We thank Dandan Song and Guizhen Zhu in the Center of Cryo-Electron Microscopy (CCEM), Zhejiang University, for their technical assistance on Scanning Electron Microscopy. We thank Wei Yin from the Core Facilities, Zhejiang University School of Medicine, for her technical support. We thank Figdraw (www.figdraw.com) and Home for Researchers (www.home-for-researchers.com) for their technical support. This study was supported by a grant to S.Y. from the National Natural Science Foundation of China (82270684).

Author contributions

Conceptualization: M.J. and S.Y.; methodology: M.J., H.F., J.F., Z.G., and X.L.; investigation: M.J., L.L., Z.J., and X.M.; visualization: B.Z., G.L., and Y.Z.; supervision: S.Y.; writing—original draft: M.J.; writing—review & editing: M.J. and S.Y.

Declaration of interests

The authors declare no competing interests.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Insulin Antibody	Cell Signaling	Cat #4590; RRID:AB_659820
Glucagon Mouse Monoclonal Ab	Affinity	Cat #BF8043
CD31 Monoclonal Antibody	Invitrogen	Cat #MA1-26196; RRID:AB_779679
Insulin/Proinsulin Monoclonal Antibody	Invitrogen	Cat #MA1-83256; RRID:AB_2126667
Anti-F4/80 antibody	Abcam	Cat #ab300421; RRID:AB_2936298
CD4 antibody	Servicebio	Cat # GB15064
Critical commercial assays
Mouse Insulin ELISA	Mercodia	10-1247-01
Software and algorithms
mageJ	mageJ	https://imagej.nih.gov/ij/
GraphPad Prism	GraphPad Software, LLC	https://www.graphpad.com/

Experimental model and study participant details

Animals

All mice used were male, aged 8–10 weeks and specific pathogen-free (SPF) grade, with a body weight of around 25g. All mice were purchased from Shanghai SLAC Laboratory Animal, Co., Ltd., China, and housed under SPF conditions in the Zhejiang Academy of Medical Sciences, China.

Pigs in this experiment were male, 2-year-old Wuzhishan pigs, weighing 70-80 kg. Pigs were purchased from Laboratory Animal Center of Chengdu Clonorgan Biotechnology Co., Ltd., China, and housed in Zhejiang L-F Tai Biotechnology Co., Ltd., China.

In this experiment, one male cynomolgus monkey, 4 years old and with a body weight of approximately 6 kg, was used. The animal was purchased from and housed at Shanghai Jihui Laboratory Animal Care Co., Ltd.

All experiments were conducted in accordance with the National Institute of Health Guide and Use of Laboratory Animals, and the protocols were approved by the Animal Care and Use Committee of the Second Affiliated Hospital of Zhejiang University School of Medicine. (Approval No. 2023069, No. 2023129, No.2024039, No.2024395).

Human participant

This study involved the acquisition and use of pancreatic tissue from one human donor. The donor was a 72-year-old Asian male farmer. Written informed consent was obtained prior to sample collection. The isolation and use of human islets were conducted in accordance with the standard guidelines approved by the Ethics Committee of The Second Affiliated Hospital, Zhejiang University School of Medicine (Approval No. 2023#1061).

Method details

Creation of the MST

Three sterile cylindrical polyamide rods (diameter 1 mm, height 1 cm) (EOONGSNG Co., Ltd., Zhejiang, China) (Figure 1B) were implanted into the chest muscle space of C57BL/6 and BALB/c mice (Figures 1C–1E). Four sterile cylindrical polyamide rods (diameter 1 mm, height 1.2 cm) were implanted into the chest muscle space of athymic nude mice. A sterile cylindrical polyamide rod (diameter 3 mm, height 7 cm) (EOONGSNG Co., Ltd., Zhejiang, China) was implanted into the chest muscle space of pigs (Figure 10C).

In detail, intraperitoneal injection of pentobarbital sodium solution (100 mg kg⁻¹ body weight) was carried out on the mice. After anesthetization, chest skin was wiped with povidone-iodine three times. All the surgical instruments involved in this procedure were under autoclave sterilization. For the polyamide rods implantation, the left lower quadrant chest skin was pulled up with forceps, and a 7 mm transverse incision was made with scissors. The subcutaneous connective tissue was incised with sharp scissors to expose the underlying muscle. After making a 5 mm muscle incision, a space in muscles was created with blunt dissection. Then the polyamide rods were implanted into this space smoothly. The muscle and skin incisions were closed with a 5-0 absorbable suture. Once implanted, the polyamide rods became adherent with blood proteins and gradually encapsulated with densely vascularized collagen fiber tissue, induced by the FBR.

Biocompatibility

L929 cells (MeisenCTCC, Zhejiang Province, China) were cocultivated with polyamide rods. The Cell-Counting Kit-8 (Apexbio, USA) was used to test the vitality of the cells by detecting their absorbance at 450 nm. Measurements were performed using a microplate reader (Spark, Switzerlan).

Proinflammatory cytokine measurements

Three polyamide rods were placed into muscular space of C57BL/6 mice for 24h, 1 week, 2 weeks, 3 weeks and 4 weeks. Implanted polyamide rods with surrounding tissue were explanted. Similarly, tissue dissections were retrieved from the chest of nonimplanted mice, serving as background cytokine and chemokine control specimens. After the polyamide rods were removed, tissue samples were immediately placed in preweighed microcentrifuge tubes. The tissue weights were recorded, then subsequently flash frozen with liquid nitrogen and stored at −80°C before conducting the cytokine and chemokine proinflammatory analysis. Once all tissue samples from respective implantation period were collected and frozen, 900ul of PBS (Gibco, Grand Island, NY, USA) per 100 mg of tissue was added to the tissue containing microcentrifuge tube. Each tissue sample was homogenized (Servicebio, Hubei Province, China) on ice for 90s. Lysed tissue samples were centrifuged at 5000 r.p.m. for 15 min at 4°C to remove cellular debris. The resulting supernatant was collected and placed in a microcentrifuge tube. Peri-implant cytokine and chemokine (IL-1β, IL-6, TNF-α, IL-10, KC, IL-12p70 and IFN-γ) measurements were conducted using 7 Mouse Inflammatory Factor kits (Meimian Industrial Co., Ltd., Jiangsu Province, China) and analyzed on a microplate reader (Labsystems Multiskan MS, Finland).

Morphological characterization of MST

In order to investigate the FBR process in C57BL/6 mice, 5 time points (24 h, 1 week, 2 weeks, 3 weeks and 4 weeks after polyamide rods implantation) were set up to terminate the response and observe the morphological characteristics of the MST. The implanted polyamide rods together with the surrounding fiber matrix were retrieved. The polyamide rods were removed carefully by forceps, generating 3 vascularized collagen fiber capsules. In order to make a clear observation of the inner surface morphology, longitudinal incisions on the collagen fiber capsules were made. Then, the inner side of the MST was exposed and unfolded into a flat surface. All samples were sufficiently washed with phosphate-buffered saline (PBS, Gibco, Grand Island, NY, USA) for 3 times and fixed with 4°C 2.5% v/v glutaraldehyde (BIOISCO,Tianjin, China). All samples were dehydrated with gradient ethanol and coated by gold-spraying. Finally, the inner surfaces of the MST were observed by SEM (Nova Nano 450, Thermo FEI, Czech). The porosity of the MST was calculated using ImageJ software.

Mouse pancreatectomy and islet isolation

Islets were obtained from C57BL/6 and BALB/c mice, weighing 20 to 25 g, by static digestion method, followed by purification using gradient centrifugation. Specifically, after mice were euthanized by cervical dislocation, the common bile duct was cannulated with a 30-G steel needle (CONPUVON, Hebei Province, China) and the pancreas was inflated with 2ml collagenase V solution (1 mg/mL) (Sigma-Aldrich, Canada). After pancreatectomy, the pancreas was placed in a 50-ml tube and digested in a 37°C water bath for 15-20 min. Subsequently, the digestion was terminated with precooled Hank balanced salt solution (CELLTRANS, Zhejiang Province, China) containing 10% fetal bovine serum (Invigentech, USA) with vigorous agitation, and islets purified on Histopaque-11191 (Sigma-Aldrich, Canada) and Hank balanced salt solution. Finally, mouse islets were handpicked and cultured overnight at 37°C in fresh culture media (10% FBS, Invigentech, USA; 89% CMRL-1066 media, CELLTRANS, Zhejiang Province, China; 1% Penicillin-Streptomycin, Gibco, USA) (Figure S12).

Human islet isolation

Pancreas was procured from a patient who had undergone pancreatic resection. Consent for islet isolation had been granted. The pancreas was distended with collagenase NB1 blend solution (Nordmark Biochemicals, Uetersen, Germany) and digested in a Ricordi chamber. When islets were adequately dissociated from surrounding acinar tissue, the pancreatic digest was collected. Islets were purified using a continuous density gradient on a cell processor centrifuge (Cobe2991,TerumoBCT, USA). All human islet preparations were processed for clinical transplantation and were made available for research only when the islet yield fell below that of the minimal mass required for clinical transplantation. Human islets were handpicked and cultured overnight in CMRL-1066 media supplemented with human serum albumin at 22°C before transplantation. Permission for the study was granted by the Human Ethics Committee of the Second Affiliated Hospital of Zhejiang University School of Medicine. (No.20231061).

Pig pancreatectomy and islet isolation

The pigs were placed under general anesthesia through a ventilator following an intravenous induction of anesthesia. Open partial pancreatectomy was performed according to human standard operating procedures. The pancreas was distended with collagenase NB1 blend solution and digested in a Ricordi chamber. When islets were adequately dissociated from surrounding acinar tissue, the pancreatic digest was collected. The digestion was terminated with precooled Hank balanced salt solution containing 10% fetal bovine serum with vigorous agitation, and islets purified on Histopaque-11191 and Hank balanced salt solution. Finally, the islets were washed 3 times with CMRL-1066 media.

Diabetes induction and mice islet transplantation

One week before transplantation, implanted mice were rendered diabetic through administration of an intraperitoneal injection of streptozotocin (Sigma-Aldrich, Canada) at 180 mg/kg in sodium citrate buffer (pH 4.5). Mice were considered diabetic when their blood glucose levels exceeded a pre-established value of 20 mmol/L for two consecutive daily readings.

Prior to transplantation, the purity of the islets was evaluated. A subset of the islets was stained with Dithizone (DTZ) (Sigma-Aldrich, St. Louis, MO, USA) to induce a red coloration. Subsequently, digital photographs of the stained islets were captured. The purity was then analyzed using ImageJ software and quantified as the ratio of islet tissue to the total tissue.

At the time of transplantation, 350 mouse islets ±12% (≈210IEQ) with purity of >95% or ∼2000 human IEQ with purity of >80% were transferred into a 1.5-mL tube using a pipette, and centrifuged into a pellet. Then, the polyamide rods were pulled out, and the islet pellet was aspirated into a microsyringe (Gaoge, Shanghai, China) for transplantation. The islets were delivered into 3 sinus tracts. Immediately afterward, the entrance of the MST and the skin incision were closed with 5-0 suture to prevent leakage of islet.

Evaluation of islet graft function

Syngeneic islet graft function was assessed through nonfasting blood glucose measurements, using a portable glucometer (One Touch Ultra, Johnson & Johnson, USA) every 2 days for 1 month following islet transplantation and once a week for next 5 months, in all groups. Reversal of diabetes was defined as two consecutive readings <11.1 mmol/L (in accordance with the American Diabetes Association).

Allograft function was assessed through nonfasting blood glucose measurements every day until the graft was rejected. Graft rejection was defined as two consecutive readings >11.1 mmol/L. The allografts were retrieved immediately after rejection.

In addition, the IPGTT was conducted at 30 and 180 days after transplantation, in order to further assess metabolic capacity in response to a glucose bolus, mimicking postprandial stimulus. Animals were fasted overnight before receiving an intraperitoneal glucose bolus (2 g/kg) (Apexbio, USA). Blood glucose levels were monitored at 0, 15, 30, 60, 90 and 120 mins after injection, allowing for AUC of blood glucose to be calculated and analyzed between different groups.

Long-term islet graft retrieval

To confirm graft dependent euglycemia, and to eliminate residual or regenerative native pancreatic beta cell function, the 3w MST group mice with functional grafts had their islet transplants removed. Animals were placed under anesthesia, the islet grafts, which exhibited no visible profile after transplantation, were excised with a margin of surrounding chest skin and muscles containing the islet graft. The area of tissue excised from the mouse measured approximately 2cm x 3cm, a size considerably larger than the graft itself. Subsequently, surface of the underlying muscle, which could not be excised, was cauterized to ensure no graft left. The explanted graft preserved for immunohistochemistry in 4% paraformaldehyde solution (Solarbio, Beijing, China). The skin incision was closed with 5-0 suture. Finally, the nonfasting blood glucose was measured for 5 days to confirm the graft function.

Histological assessment

For histological analysis, the tissue samples with polyamide rods inside were fixed in 4% paraformaldehyde solution at 25°C for 24h, then polyamide rods were removed carefully, allowing the MST to remain its original round shape of the cross section. After that, the samples were embedded in paraffin wax and sectioned. After the sections were dewaxed in xylene and rehydrated in PBS, they were stained with Masson’s trichrome. The retrieved graft specimens were processed, embedded, sliced, and subjected to hematoxylin/eosin (HE) staining and Masson’s trichrome staining to present the morphology of the islets and the surrounding collagen fibers.

Immunohistochemistry staining was used to identify macrophage cells and islet graft for assessment of inflammation and graft function respectively. Briefly, the MST cross sections and graft sections were deparaffinized and hydrated using xylene and gradient of alcohol to water. The sections were blocked with 5% bovine serum albumin (w/v) to reduce the nonspecific binding, followed by incubation with a primary antibody of rabbit anti-mouse insulin (Cell Signaling, #4590) diluted 1:100 or rabbit anti-mouse F4/80 (Abcam, ab300421) diluted 1:500 overnight at 4°C. Then the samples were incubated with secondary antibody (Abcam, USA) at room temperature for 30 min. The sections were washed in PBS 3 times and stained with DAB (Abcam, USA). Pathological section scanner (Jiangfeng, Zhejiang Province, China) was used to obtain pictures.

Immunofluorescence was used to identify the presence of pancreatic β cells, α cells and endothelial cells using anti-insulin, anti-glucagon and anti-CD31 antibodies, respectively. Briefly, after deparaffinization and antigen heat retrieval, the graft sections were washed with phosphate-buffered saline (PBS) supplemented with 5% bovine serum albumin (w/v) to reduce the nonspecific binding, followed by incubation with primary antibodies of rabbit anti-mouse insulin (Cell Signaling, #4590) diluted 1:100, mouse anti-mouse glucagon (Affinity, BF8043) diluted 1:200 and mouse anti-mouse CD31 (Invitrogen, MA1-26196) diluted 1:100 overnight at 4°C. Antibodies of rabbit anti-human insulin (Cell Signaling, #4590) diluted 1:100, mouse anti-human glucagon (Affinity, BF8043) diluted 1:200, and mouse anti-human CD31 (Invitrogen, MA1-26196) diluted 1:100 were used for human islet grafts. Antibody of mouse anti-pig insulin (Thermo Fisher, MA1-83256) diluted 1:200 was used for pig islet grafts.

Then, samples were rinsed with PBS followed by secondary antibody treatment consisting of goat anti-mouse (Fluor 488, Affinity, USA) at 1:500 and goat anti-rabbit (Fluor 594, Affinity, USA) at 1:500 for 60 minutes at room temperature. Next, the samples were rinsed with PBS and counterstained with 4′,6-Diamidino-2-phenylindole (DAPI, Abcam, USA). Images were visualized under a confocal fluorescent microscope (Zeiss, Germany).

For histological analysis, a minimum of three samples per experimental group were evaluated. From each sample, at least three non-serial sections were stained and examined. The images presented in the figures were selected as representative of the typical morphology observed for each group. The thickness of the collagen matrix, the number of infiltrated macrophages, and the relative area of neovascularization were quantified for each tissue section using ImageJ software.

Quantification and statistical analysis

Statistical analysis

All statistical results are represented as mean ± standard deviation (sample size n = 3 or more per group). The differences between the groups were compared by parametric one-way ANOVA and Student unpaired t test with 2 tailed P values using GraphPad Prism 9 (GraphPad Software, USA). Kaplan-Meier survival function curves were compared using the log-rank statistical method. Results with p-values <0.05 were considered statistically significant. The levels of significance were labeled with ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001.

Published: November 28, 2025