Survival and Maturation of Microencapsulated Porcine Neonatal Pancreatic Cell Clusters Transplanted into Immunocompetent Diabetic Mice

NPCCs were prepared using a modification (11) of a technique described by Korbutt et al. (10). Male Yorkshire pigs (1–3 days old; Parsons Farm, Hadley, MA) were anesthetized with 0.15 ml of Telazol (tiletamine HCL; Fort Dodge Laboratories, Ft. Dodge, IA) and 0.30 ml of Xylaject (xylazine HCL; Phoenix Pharmaceutical St. Joseph, MO) given by intramuscular injection. After partial exsanguination, the pancreas was removed through a midline incision. Pancreata were then minced into 1- to 2-mm³ pieces in a 50-ml plastic Falcon centrifuge tube containing 25 ml of M199 medium. After centrifugation for 1 min at 600 rpm, the supernatant was removed and 5 mg/ml collagenase (collagenase P; Roche, Indianapolis, IN) was added to the pellet. The pancreata were digested in a water bath at 37°C. After filtration, NPCCs were washed two times and were cultured in Ham’s F10 media (Life Technologies, Grand Island, NY) supplemented with 10 mmol/l glucose, 10 μmol/l 3-isobutyl-l-methylxanthine (Sigma, St. Louis, MO), 2 mmol/l l-glutamine (Life Technologies), 10 mmol/l nicotinamide (Sigma), 25 mg/l CaCl₂:H₂O (Sigma), 0.5% BSA (RIA grade Fraction V, Sigma), and antibiotics (100 units/ml penicillin, 100 μg/ml streptomycin; Life Technologies) in 150 × 50 mm bacteriological plates (Becton Dickinson, Franklin Lakes, NJ) at 37°C for 8 days. Media changes were performed on days 2, 4, 5, and 7.

Encapsulation was performed with a minor modification of a previously described technique (15), using a highly purified alginate containing 61% mannuronic acid and 39% guluronic acid (High M alginate 3.3%, the gift of John Holahan, Pharmacia, Peapack, NJ). At day 8 after preparation, NPCCs were washed with M199 medium and calcium-free Krebs Ringer buffer. The number (given as islet equivalent [IE]) and the viability (determined with trypan blue staining) of the NPCCs were determined. NPCCs were suspended in an alginate solution (7,000 cell cluster/ml alginate) and then transferred into tubing connected to an infusion pump and an air-driven droplet generator. Each microcapsule generally contained two to four NPCCs; however, as many as seven NPCCs were observed in some capsules. The droplets fell into a BaCl₂ solution (20 mmol/l), which cross-links alginate molecules. Microencapsulated NPCCs were then cultured overnight in serum-free medium (Ultraculture; BioWhittaker, Walkerville, MD) supplemented with penicillin-streptomycin and glutamine, as described above, at 37°C in humidified atmosphere containing 5% CO₂. NPCCs from nine preparations were used for transplantations.

Male B6AF1 mice (6–7 weeks old; Jackson Laboratories, Bar Harbor, ME) were used as recipients of nonencapsulated (controls) or microencapsulated NPCCs. One week before transplantation, animals were made diabetic by intraperitoneal injections (180 mg/kg body wt) of STZ (Sigma). Only animals with blood glucose levels >350 mg/dl on two consecutive measurements were used as recipients.

Under methoxyflurane (Metofane; Shering-Plough Animal Health, Union, NJ) anesthesia, which was carried out in a 350-ml jar containing evaporated methoxyflurane (produced by gauze pads being soaked in the 3–5 ml of the liquid anesthetic in the jar), NPCCs were transplanted under the capsule of the left kidney by using PE-50 tubing (Becton Dickinson) connected to a 1-ml Hamilton syringe as previously described (11). Each mouse received 10,000 IE of NPCCs.

After washing with Ultraculture medium, microcapsules were aliquotted and the volume of the capsules in each aliquot was determined in a 3-ml syringe (Becton Dickinson). Microencapsulated NPCCs were collected in a sterile transfer pipette and transplanted into the peritoneal cavity of Metofane-anesthetized mice (n = 32) through a small abdominal incision. Each mouse received 10,000 IE of NPCCs contained in 2.25 ± 0.06 ml of microcapsules.

The body weight and the blood glucose levels of the mice were monitored twice weekly between 9:00 and 11:00 a.m. Glucose levels were measured with a portable glucometer (Precision QID; Medisense, Bedford, MA) on blood samples obtained from snipped tails. The mice were considered normoglycemic when blood glucose levels were <200 mg/dl. Twenty weeks after transplantation, an intravenous glucose tolerance test (IVGTT) (1.5 g/kg body wt) was performed in normoglycemic-transplanted mice and in body weight matched (having similar body weight but being at younger age than mice that received a transplant) and age-matched nondiabetic animals. Blood glucose samples were taken before and 2, 5, 10, 15, 30, 60, 90, 120, and 180 min after a glucose solution (30%) was injected through the orbital sinus of nonanesthetized mice. Blood samples for plasma C-peptide assessments were taken at the same time points except 2, 5, and 15 min after glucose injection.

At 2, 6, and 20 weeks after transplantation, microcapsules were removed under Metofane anesthesia, by peritoneal lavage using warmed Krebs Ringer HEPES buffer. The abdomen was sutured, and blood glucose levels were monitored during the next 2 weeks. The removed capsules were assessed for volume (in a 3-ml syringe), degree of fibrotic overgrowth (under a phase contrast microscope by systematic examination of 500 consecutive capsules for each mouse), structural integrity (by counting the number of the broken capsules as percentage of total capsules), and dithizone (dimethylthiocarbazone; Sigma) staining. Images were captured by a Magnafire digital camera (Optronics, Goleta, CA) connected to a microscope (CK2, Olympus, Tokyo, Japan) and are stored as .tiff files.

Microencapsulated NPCCs were fixed in 10% buffered formalin (Fisher Scientific, Fairlawn, NJ) for 3 h and stored in 0.1 mmol/l Sorensen’s buffer. NPCCs were separated from alginate gel by incubation in 1 mol/l sodium citrate and filtration through 250-μm polyester mesh filter (Small Parts, Miami Lakes, FL). After washing two times in PBS NPCCs were suspended in 0.5 ml of an agar solution (2%; Difco, Grayson, GA) and pelleted by centrifugation. The pellet-containing agar was stored in Sorensen’s buffer and embedded in paraffin.

Double staining for Ki-67, a marker of cell proliferation (16–18) that is expressed during the cell cycle between the middle of the G1 phase and mitosis, and insulin was performed to quantify the percentage of proliferating β-cells as previously described (11,19). Briefly, 5-μm sections were incubated overnight at 4°C with anti- Ki-67 polyclonal antibody (1:200, rabbit antihuman; Novocastra Laboratory, Newcastle Upon Tyne, U.K.). Indirect immunofluorescent staining was performed by incubation with biotinylated secondary antibody (Vector, Burlingame, CA) for 1 h at room temperature (RT), followed by incubation for 1 h at RT with fluorescein isothiocyanate (FITC)-streptavidin (1:200; Jackson ImmunoResearch Laboratories, West Grove, PA). Insulin immunostaining was performed after that of Ki-67 with anti-insulin antibody (1:200, Guinea pig anti human; Linco Research) for 1-h incubation at RT followed by incubation with Texas Red conjugated anti- Guinea pig IgG for 1 h at RT (1:200; Jackson ImmunoResearch Laboratories). The proliferation of β-cells was expressed as a proliferation index that was determined as the number of Ki-67 positive β-cells over the total number of β-cells; at least 300 β-cells were counted per sample, with the exception of one sample each of pretransplantation NPCC and 20-week graft.

To assess β-cell area, we performed double staining for insulin and pan-cytokeratin (pan-CK), a marker for epithelial cells. In preliminary experiments, an effort was made to stain CK7 in transplanted NPCCs using a mouse monoclonal antibody against CK7. However, there was an unacceptable amount of noncellular staining as a result of binding of the FITC-labeled anti-mouse IgG (donkey) secondary antibody (Jackson ImmunoResearch Laboratories) with IgG deposited in the alginate and NPCCs as a result of a xenogeneic immune reaction. For that reason, a polyclonal rabbit anti-pan-CK (1:200; DAKO, Glostrup, Denmark) was used with an overnight incubation at 4°C. After washing, sections were incubated with biotinylated anti-rabbit secondary antibody (Vector) for 1 h at RT, followed by FITC-streptavidin (1:200, 1 h at rt; Jackson ImmunoResearch). Whereas the monoclonal antibody against CK7 is relatively specific for duct cells (17), the anti-pan-CK antibody reacts with all pancreatic epithelial cells. Immunostaining for insulin staining was performed in addition to that of pan-CK as described above. From each section, double-stained fluorescent images of nonoverlapping fields were captured systematically. Immunofluorescent images were acquired on a Zeiss LSM 410 microscope. Images were analyzed using IP Lab Spectrum software (Scanalystis, Fairfax, VA), a computer-assisted image analysis program. Maturation of NPCCs was defined with an index, calculated as the percentage of the cellular area that was stained for insulin divided by the total cellular area stained by pan-CK, with samples analyzed before and 2, 6, and 20 weeks after transplantation.

To assess the insulin secretory capacity of the encapsulated NPCCs, we performed a 12-h static incubation before transplantation and after retrieval. For each animal, triplicate samples of 20 nonovergrown capsules were randomly handpicked, and the IE number was determined. Capsules were then washed three times and incubated with RPMI 1640 media without glucose (Life Technologies) supplemented with 10% neonatal calf serum for 3 h. Because of the anticipated low insulin output from NPCCs as a result of their immaturity in the pretransplantation period, microcapsules were incubated sequentially for 12 h as opposed to more conventional 1-h time periods. Microcapsules were first incubated for 12 h in media containing 2.8 mmol/l glucose, then for 12 h in media containing 16.7 mmol/l glucose, and finally in media containing 16.7 mmol/l glucose + 10 mmol/l theophylline (Sigma). At the end of each incubation period, supernatant medium was collected and kept at −20°C until measurement of insulin with a radioimmunoassay kit (Linco Research, St. Charles, MO). In addition, insulin content of microencapsulated NPCCs was measured for samples from each recipient. After the NPCCs were counted, capsules were added to 0.5 ml of acid-ethanol solution, sonicated, and kept overnight at 4°C. After centrifugation, supernatants were stored at −20°C until measurement. Results are given as nanograms of insulin per/IE.

Results are expressed as means ± SE. For the static incubations, results are the mean (nanograms of insulin per IE per 12 h) of triplicate samples measured for each animal at pretransplantation and 2, 6, and 20 weeks after transplantation. The insulin response to glucose and to glucose plus theophylline was expressed as a stimulation index defined as the ratio of the insulin secretion at 16.7 mmol/l glucose and at 16.7 mmol/l glucose + 10 mmol/l theophylline over the insulin secretion at 2.8 mmol/l glucose. Student’s t tests for paired and unpaired data were used to compare data for IVGTT. Statistical analysis for graft insulin content and insulin response was performed with one-way ANOVA.

Mice that received a transplant of 10,000 nonencapsulated (naked) NPCCs under the kidney capsule did not gain weight after transplantation (18.9 ± 2.4 g before transplantation vs. 19.3 ± 3.2 g at 4 weeks after transplantation; NS; Fig. 1A) and remained hyperglycemic after transplantation (Fig. 1B). After transplantation of microencapsulated NPCCs, all animals (n = 32) gained weight: 20.7 ± 2.3 g before transplantation vs. 30.9 ± 2.9 g 20 weeks after transplantation (Fig. 1A). In the mice with normalized blood glucose levels (n = 26), the weight gain was 11 ± 1% at 2 weeks, 20 ± 1% at 6 weeks, and 51 ± 2% at 20 weeks compared with pretransplantation levels. Whereas 26 of 32 transplanted mice (81%) normalized blood glucose within 15 ± 12 days (range from 3 to 27 days; Fig. 1B), 6 mice from two preparations (of nine total preparations) did not reverse hyperglycemia, even at 20 weeks. It should be noted that none of the unsuccessful transplants with naked NPCCs used NPCCs from the apparently substandard preparations that never reversed hyperglycemia in the encapsulated group. Normoglycemia was maintained over the 20-week period in all other mice. After removal of capsules at 2, 6, and 20 weeks, animals (7, 7, and 12 mice, respectively) returned to hyperglycemic levels, showing the efficacy of the microencapsulated NPCCs.

An IVGTT was performed in all mice that received a transplant of microencapsulated NPCCs (20 weeks after transplantation) and body weight- and age-matched nondiabetic controls (Fig. 2). The fasting blood glucose level of mice that received a transplant of encapsulated NPCCs was 100 ± 21 mg/dl, which was similar to body weight-matched controls (102 ± 13 mg/dl) and lower than age-matched controls (138 ± 7 mg/dl). In all groups, blood glucose levels peaked at 2 min and started decreasing at 15 min. From this time point to the end of the IVGTT, significantly lower glucose levels were observed in the mice that received a transplant of microencapsulated NPCCs compared with the control groups (P < 0.05 at 15 min and P < 0.001 thereafter; Fig. 2).

Capsules were removed 2, 6, and 20 weeks after transplantation and were incubated in the presence of different concentrations of glucose to assess the function of the removed capsules. Increased insulin secretory responses to 16.7 mmol/l glucose (0.09 ± 0.03, 0.26 ± 0.03, 0.32 ± 0.05, and 0.69 ± 0.1 ng · IE⁻¹ · 12 h⁻¹ at pretransplantation and 2, 6, and 20 weeks, respectively) and 16.7 mmol/l glucose plus theophylline (0.26 ± 0.05, 0.4 ± 0.04, 1.3 ± 0.3, and 2 ± 0.6 ng · IE⁻¹ · 12 h⁻¹ at the same time points) compared with insulin responses to 2.8 mmol/l glucose (0.06 ± 0.03, 0.15 ± 0.03, 0.18 ± 0.09, and 0.30 ± 0.04 ng · IE⁻¹ · 12 h⁻¹ at comparable time points) were observed. The stimulation index of encapsulated NPCCs in response to glucose alone increased compared with pretransplantation levels (1.1 ± 0.1 before transplantation; 2.3 ± 0.06 [P < 0.001], 2.9 ± 1.2 [P < 0.05], and 2.6 ± 0.2 [P < 0.001] at 2, 6, and 20 weeks, respectively; Fig. 3). A more pronounced increase in the stimulation index was observed with stimulation of glucose and theophylline (2.7 ± 0.5 before transplantation; 3.7 ± 1.1 at 2 weeks [P < 0.05], 10 ± 2.2 at 6 weeks [P < 0.001], and 9.4 ± 1.8 [P < 0.001] at 20 weeks).

The recovery percentage of capsules, corresponding to the ratio of the total capsular volume recovered over the volume before transplantation, was 91 ± 6, 85 ± 6, and 93 ± 2%, respectively at 2, 6, and 20 weeks, showing the stability of the Ba-alginate microcapsules. The majority of capsules were free of overgrowth at 2 and 20 weeks, with the percentage of clean capsules being 87 ± 11 and 80 ± 5%, respectively. However, at 6 weeks, 43 ± 9% of the capsules were moderately to severely overgrown (Fig. 4).

During the assessment of overgrowth, the presence of insulin was indicated by dithizone staining (Fig. 5A). From pretransplantation to 20 weeks, there was an obvious increase in area of insulin-positive cells (Fig. 5B and C). The amount of β-cells was quantified as maturation index on immunostained sections. The maturation index (percentage of insulin-positive area divided by pan-CK- positive area) of encapsulated NPCCs, shown in Fig. 5B, was significantly increased with time (18 ± 5% before transplantation, 48 ± 13% at 6 weeks [P < 0.01], and 58 ± 2% at 20 weeks [P < 0.001]). Furthermore, the graft insulin content of NPCCs increased (Fig. 5D). Compared with pretransplantation levels (1 ± 0.2 ng/IE), insulin content of NPCCs removed from normalized animals was increased 4-fold at 2 weeks (4 ± 0.1 ng/IE), 12-fold at 6 weeks (11 ± 2 ng/IE), and 10-fold at 20 weeks (9 ± 2 ng/IE), which is in agreement with the increased number of β-cells in the NPCCs. In contrast, encapsulated NPCCs removed from noncured animals (n = 6 mice from two preparations) had a significantly lower graft insulin content than cured animals at two time points: before transplantation and at the end of 20 weeks (0.13 ng/IE, mean of results from two samples, and 2.6 ng ± 1.0 ng/IE, respectively).

Deposition of IgG antibody was shown whenever FITC-labeled anti-mouse IgG antibody was used as a secondary antibody for immunostaining. Even incubation with no primary antibody and only FITC anti-mouse IgG antibody gave a similar pattern of fluorescence. The staining of the surrounding alginate was particularly intense, but staining throughout the extracellular portions of the NPCC can also be appreciated (Fig. 6).

β-Cell proliferation was assessed by double staining for Ki-67 and insulin (Fig. 7A). The proliferation index of the β-cells was 7.3 ± 1.1% before transplantation, which was essentially the same as that found at 6 weeks after transplantation (6.8 ± 0.01%, NS), but a marked decrease to 1.7 ± 0.8% was found at 20 weeks (Fig. 7B).

Islet microencapsulation, first described in 1980 (20), could represent a valuable therapeutic approach for the treatment of diabetes. This technique allows transplantation of islets enclosed in a capsule without immunosuppression, thus creating the potential to use xenogeneic islets. Previous studies have demonstrated protection of encapsulated pig islets from immune destruction and long-term maintenance of normoglycemia without immunosuppression. Lanza et al. (21) showed that transplantation of microencapsulated adult pig islets, enclosed in poly-L-lysine (PLL)-coated alginate capsules, could normalize the blood glucose levels of STZ-induced diabetic mice for 10 weeks. Furthermore, Murakami et al. (22) showed that encapsulation with agarose prolongs the survival of NPCCs for 6–8 weeks after transplantation in BALB/c mice. Sun et al. (4) reported long-term survival of microencapsulated adult porcine islets transplanted into STZ-diabetic Cynomolgus monkeys, although these results have yet to be reproduced.

In this study, we assessed the protective role of microcapsules made of highly purified high-M alginate cross-linked with BaCl₂ without an additional permselective PLL coating. We cannot fully explain why these simple capsules work so well for islet xenografts. Previous studies have shown that barium cross-linkage provides stronger alginate gels than those found with calcium (23). In addition, the avoidance of PLL is probably helpful because the macrophage activation so often seen with PLL might act synergistically with a xenograft reaction. Thus, uncoated alginate capsules may provide consistently better biocompatibility and stability (14). In addition, fabrication of microcapsules with highly purified and very low endotoxin-containing alginate probably contributed to the good results of our study. Finally, although not well documented, the mice used in these experiments may have a weaker reaction to xenografts than other species. Our results show that microencapsulated NPCCs sustained normoglycemia for >20 weeks in the majority of the mice that received a transplant. All of the mice that initially reversed hyperglycemia maintained normoglycemic limits until end of the experiments at 2, 6, and 20 weeks. Reversal of hyperglycemia after explantation of the capsules demonstrated the efficacy of encapsulated NPCCs. Six mice that did not become normoglycemic after transplantation all received NPCCs from the two preparations. These preparations of NPCCs had lower insulin content than the other preparations, which were successful. The factors responsible for batch-to-batch variation with NPCCs are unknown, but any number of the multiple variables could be involved.

The period of normalization of the glycemia after transplantation (15 ± 12 days; range 3–27 days) was shorter than has been observed in different studies in which nonencapsulated (10,11, 19) or encapsulated (13) NPCCs were transplanted into diabetic nude mice. However, our results are consistent with the study of Murakami et al. (22), in which glucose levels were normalized 3–4 weeks after transplantation of 8,000 encapsulated NPCCs in the peritoneal cavity of diabetic mice. It is noteworthy that the study of Rayat et al. (13) found that only 2,000 encapsulated NPCCs could reverse diabetes in nude mice, with this result being better than that obtained with similar numbers of unencapsulated NPCCs. However, in other studies, the encapsulation approach seems to require larger numbers of either islets or NPCCs to normalize glucose levels (11,14, 19,22). The reasons that so many NPCCs are required are that the peritoneal cavity is not a particularly efficient site for transplantation and NPCCs contain precursor cells and immature islets, which limit insulin production. In our preliminary studies, 5,000 IE of encapsulated NPCCs were unable to produce normoglycemia (A.O., unpublished data). The capsules used in this study were relatively large at 0.8–1.0 mm in diameter, but this diffusion distance did not seem to prevent adequate provision of nutrients, glucose, and oxygen. It is noteworthy that glucose clearance after intravenous glucose was much faster than in control mice. This may be because porcine islets have a lower set point for glucose-induced insulin secretion than mouse islets, which would make the mice relatively hyperinsulinized. We were surprised, however, by the slow rate of clearance in the control mice compared with experiments in which younger mice of a different strain were used (14).

In assessing the immune reaction on the microcapsules after removal, it was found that at 2 and 20 weeks after transplantation, the majority of the microcapsules (87 ± 7 and 80 ± 5%, respectively) were free of any fibrotic or cellular reaction, indicating the biocompatibility of the uncoated capsules. Unexpectedly, 6 weeks after transplantation in cured animals, the majority of the capsules had a mild to moderate reaction on their surface. This result raises interesting questions about the amount of shed antigens and the time of shedding required to develop an immune reaction on the microcapsules, as well as the possibility of a later accommodation of the immune system to the presence of the encapsulated xenogeneic islets. It has been shown that a strong immune reaction to the nonencapsulated fetal porcine islet cell clusters occurs during the first week of transplantation (24). In addition to the cellular infiltrate, we observed IgG deposition on the NPCCs contained in the microcapsules at 6 and 20 weeks after transplantation (Fig. 6, 6-week example not shown). These results support the idea that IgG can diffuse through the capsule and yet not exert any adverse effect on the encapsulated NPCCs. Similar observations have been reported by Friedman et al. (25) who found that rejection of fetal islet cell clusters seemed to be mediated only by CD4⁺ cells without a contribution from IgG.

The second aim of this study was to assess the proliferation and differentiation of encapsulated NPCCs after transplantation. NPCCs contain more mature islet cells than fetal porcine islet clusters, yet they still possess considerable capacity for further differentiation, thus providing a potentially valuable source of β-cells for transplantation (10). The present results demonstrate maturation of NPCCs contained in the capsules as evidenced by graft insulin content and immunostaining results. We observed a 12-fold increase in the graft insulin content at 6 weeks, with a similar increase at 20 weeks compared with pretransplantation insulin content. These results are similar to those observed with nonencapsulated NPCCs transplanted in the subrenal space of nude mice (10,11,19), in which it seems that the increase in β-cell mass after transplantation was due to both proliferation of β-cells and differentiation of duct cells into β-cells. Although we were not able to quantify duct cells with specific staining for CK7 at 6 and 20 weeks because of the interference from IgG deposition, we showed a significant increase in the β-cell area, which we assume results from a developmental pattern similar to that which occurs with nonencapsulated NPCCs.

To our knowledge, this is the first study to show differentiation of encapsulated duct cells and expansion of β-cells in the NPCCs after transplantation into immunocompetent mice. As was the case with nonencapsulated NPCCs (19), we found a high β-cell proliferation index before transplantation and 6 weeks afterward, with a decline at 20 weeks. However, those results are not in agreement with another study that showed increased proliferation index and apoptosis of encapsulated rat islets followed by a fall in β-cell mass (26). Discrepancies in results may be due to different encapsulation models (PLL versus non-PLL), differences in the maturity of islets, and species differences. We do not know how long encapsulated NPCCs can continue to function, but at 20 weeks after transplantation, they seem to be stable and healthy.

In conclusion, these results indicate that an alginate barrier, without a traditional permselective membrane, can provide protection against xenorejection in mice. Furthermore, the encapsulated NPCCs can differentiate into β-cells. These results provide support for the idea that neonatal porcine pancreas could provide large numbers of insulin-producing cells for transplantation (2).

This research was supported by Animal, Media, Advanced Microscopy and Specialized Assay Cores of Joslin’s Diabetes and Endocrinology Research Center (DERC grant DK 36836) and grants PO1 DK 53087 and RO1 DK 50657 from the National Institutes of Health. Support was also provided by the Juvenile Diabetes Research Foundation Center for Islet Transplantation at Harvard Medical School, the Diabetes Research and Wellness Foundation, and an important group of private donors. A.O. is a recipient of a mentor-based fellowship of the American Diabetes Association.

We appreciate the expert technical assistance of Jennifer Lock.