ASSESSMENT OF INSULIN SECRETION IN VITRO FROM MICROENCAPSULATED FETAL PORCINE ISLET-LIKE CELL CLUSTERS AND RAT, MOUSE, AND HUMAN PANCREATIC ISLETS¹

Abstract

Background. 

The possibility of transplanting microencapsulated pancreatic islets into patients with insulin-dependent diabetes mellitus, either as allografts or xenografts, has attracted great interest. A critical evaluation of the results obtained reveals that the success has been very limited. The aim of the present study was to compare the in vitro function of microencapsulated islets obtained from adult humans, adult mice, adult rats, and fetal pigs.

Methods. 

Human pancreatic islets were isolated at β-Cell Transplant in Brussels, Belgium, and sent to the Department of Medical Cell Biology, Uppsala University in Uppsala, Sweden. Rat and mouse pancreatic islets and fetal porcine islet-like cell clusters (ICC) were prepared in Uppsala. All groups of islets were subsequently sent to the Department of Biotechnology, Norwegian Institute of Biotechnology, University of Trondheim, Trondheim, Norway. After 1 day in tissue culture, the islets were microencapsulated in alginate then cultured and sent back to Uppsala the next day. After either overnight culture (day 1) or 6 days of culture (day 6), the microencapsulated islets were examined for their insulin content and insulin release. Nonencapsulated islets from the same isolations were used as controls.

Results. 

The insulin content of rodent and human islets was not affected by microencapsulation, whereas porcine ICC showed a diminished insulin content. Microencapsulated porcine ICC also had a marked reduction in their insulin secretion in response to stimulation with glucose or glucose + theophylline both on days 1 and 6 in tissue culture. Mouse islets showed a reduced insulin response at both time points. Rat islets exhibited an inhibition of insulin secretion on day 1, but this had been restored by day 6. Human islets had well-preserved insulin secretion after both days 1 and 6. Microencapsulated human islets showed a normal morphology 3-4 weeks after intraperitoneal transplantation to nude mice.

Conclusions. 

Pancreatic islets isolated from human, rat, and mouse donors show a glucose-stimulated insulin release in vitro after microencapsulation and repeated transports between laboratories. The insulin secretory capacity of microencapsulated human and rat islets was preserved best, whereas mouse islets and particularly fetal porcine ICC were impaired by microencapsulation.

The possibility of patients with insulin-dependent diabetes mellitus receiving microencapsulated pancreatic islet transplants, either as allografts or xenografts, has attracted great interest for more than 10 years ^(1-7). A critical evaluation of the literature, however, reveals that success has so far been very limited. This failure is presumably related to several issues, e.g., insufficient immunoprotection, a microenvironment inside the microcapsules leading to islet dysfunction, and foreign body reactions outside the microcapsules causing diffusion problems and eventually impaired islet function. Recently it has become apparent that major species differences exist in the sensitivity of β cells against various noxious treatments ^{(8, 9)}. By the same token, there may be species differences in how islets obtained from different species function after microencapsulation.

The aim of the present investigation was to test the function of microencapsulated islets obtained from humans, rats, and mice and islet-like cell clusters (ICC^*) prepared from fetal pigs. Most of the study was performed in vitro in order to obviate the reactions taking place in vivo upon transplantation of microencapsulated islets. Since the microencapsulated human islets showed excellent function in vitro, the morphology of such islets was also assessed after transplantation to nude mice.

MATERIALS AND METHODS

Islet isolation and culture. Rodent islets: Non-inbred, adult, male Sprague-Dawley rats (Biomedical Center, Uppsala, Sweden) and male inbred C57BL/6 mice originally obtained from Jackson Laboratories (Bar Harbor, ME) and subsequently bred at the Biomedical Center were used. Pancreatic islets were isolated by collagenase digestion as described elsewhere ⁽¹⁰⁾. After handpicking, groups of 150-200 islets were cultured free-floating in RPMI 1640 medium (11.1 mM glucose) supplemented with 10% fetal calf serum (Sigma Chemical Co., St. Louis, MO). The culture medium was changed every second day.

Fetal porcine ICC: Pregnant sows, obtained from a local stock outside Uppsala, were killed with a slaughtering mask. The length of gestation was 65-75 days. The fetuses, 7-15 in a litter, were immediately removed from the uterus and placed on ice for transport to the laboratory. After aseptic removal, the pancreatic glands were placed in cold Hanks' solution, minced into 1- to 2-mm³ fragments, and incubated with collagenase from Clostridium histolyticum (≅10 mg/ml; Boehringer Mannheim, Mannheim, Germany) during vigorous shaking, as described elsewhere ⁽¹¹⁾. The digested tissue was washed and explanted into tissue culture dishes to allow cellular attachment (Nunclon 50 mm Ø; Nunc, Kamstrup, Denmark). The culture medium was 9 ml of RPMI 1640 medium containing 11.1 mM glucose and 10 mM nicotinamide plus 1 ml of pooled human serum (The Blood Center, Huddinge Hospital, Huddinge, Sweden) ⁽¹²⁾. The medium was exchanged every second day of culture. During the subsequent days of culture, spherical cell aggregates, with a diameter of 200-400 μm, were formed. In the stereomicroscope, the cell aggregates resembled isolated pancreatic islets, but since only a minority of the cells at this developmental stage stain positively for the different islet hormones, they have been designated as ICC ⁽¹¹⁾.

Human islets: Nine human pancreata were excised from adult heartbeating organ donors, and islets were isolated in the β-Cell Transplant Central Unit (Vrije Universiteit Brussels, Brussels, Belgium) according to a previously described method ^{(13, 14)}. The mean age of the donors (± SEM) was 35.7±6.7 years (range, 8-63 years). Aliquots of the purified suspensions containing human islets from the different preparations were examined by electron microscopy, which showed 8±1% dead cells and <0.2% exocrine cells. The prevalence of insulin-positive cells and glucagon-positive cells, as evaluated by light microscopic examination of immunocytochemically stained islets, was 62±4% and 15±3%, respectively, and the islet insulin content was 1.50±0.22 ng insulin/ng DNA. The human islets were cultured in Ham's F-10 medium containing 7.5 mM glucose and supplemented with 1.0% bovine serum albumin. The culture period varied between 2 and 19 days (mean, 8.2±1.8 days) before islets were sent by air to the Department of Medical Cell Biology, Uppsala University, Uppsala, Sweden. A detailed description of this procedure, and of functional tests of human islets maintained under the different culture conditions, has been published ^{(13, 15)}. After arrival in Uppsala, groups of 150-200 islets were cultured free-floating for 3 to 4 days in RPMI 1640 medium (5.6 mM glucose) supplemented with 10% (v/v) fetal calf serum at 37°C in an atmosphere of humidified air plus 5% CO₂. Medium was changed every second day.

Shipment of ICC and islets between Uppsala-Trondheim-Uppsala. After 4 to 5 days of culture in Uppsala, rodent and human islets and ICC were harvested and transferred to sterile plastic tubes containing culture medium (same as above) and sent by air to the Norwegian Institute of Biotechnology, University of Trondheim, Trondheim, Norway. The islets were subsequently re-explanted into tissue culture (see below). On the next day, all islets were subjected to microencapsulation (see below). From each preparation, a fraction of islets or ICC was retained in culture in Uppsala and served as cultured nonencapsulated controls.

After encapsulation, the islets were cultured and on the next day harvested and sent back to Uppsala. Upon arrival in the laboratory in Uppsala, the encapsulated preparations were again maintained in tissue culture before being functionally examined either on the following day (day 1) or after 6 days of culture (day 6) together with corresponding nonencapsulated control groups, which were cultured for the same amount of time.

Encapsulation of islets and ICC. An alginate with a high content of guluronic acid (G), prepared from stipes of Laminaria hyperborea (Pronova UP-LVG), was obtained from Pronova A/S (Drammen, Norway). This high-G alginate has a G content of 63%, an average length of blocks consisting of more than two continuous G units (N_G>1=12.6), an average molecular mass of 243 kDa, and a molecular weight distribution (M_w/M_n) of 1.80. Poly-L-lysine hydrochloride with a molecular weight of 21 kDa (low-angle laser light scatter) and polydispersity index of M_w/M_n 1.25 (low-angle laser light scatter/size exclusion chromatography) was purchased from Sigma.

After arrival in Trondheim (about 6-hr transport), the islets and ICC were distributed into culture dishes (Nunclon 90 mm; Nunc). Each dish contained 9 ml of RPMI 1640 medium (Flow Laboratories, Irvine, UK) supplemented with 1 ml of heat-inactivated human serum. The culture dishes were kept at 37°C in a gas phase consisting of air + 5% CO₂ until encapsulation. The islets were suspended in a sterile filtered 1.8% (w/v) alginate solution. Alginate beads with a solid inhomogeneous gel core ⁽¹⁶⁾ were formed by pressing the viscous solution containing islets through a 22-gauge syringe. The droplets fell into an aqueous solution containing Ca²⁺ ions (0.8% CaCl₂·2H₂O) in the presence of mannitol as osmolythe. The size of the alginate beads was controlled by a coaxial airstream. The beads had an average diameter of 0.7 mm and the standard deviation of the average was 0.05 mm. The alginate beads were collected from the gelling solution and washed three times in saline. Polylysine coating of the beads was achieved by suspending the gel beads in a solution of polylysine 0.1% (w/v) in saline for 10 min. An outer coating of alginate was applied by suspension of the beads for 10 min in 0.1% alginate solution. The capsules were finally washed three times in saline, resuspended in culture medium, and sent back to Uppsala.

Insulin release and insulin content. Groups of 10 islets or ICC, both nonencapsulated and encapsulated, were transferred in triplicate to glass vials containing 0.25 ml of Krebs-Ringer bicarbonate buffer supplemented with 10 mM HEPES (Sigma) and 2 mg/ml bovine serum albumin (fraction V; Miles, Slough, UK), hereafter referred to as KRBH buffer, and incubated for 1 hr at 37°C (O₂:CO₂, 95:5). The KRBH medium contained either 1.7 mM glucose, 16.7 mM glucose, or 16.7 mM glucose + 5 mM theophylline. After the incubations, the islets were harvested, pooled in groups of 30, and homogenized in 0.2 ml of redistilled water. The aqueous solution was then mixed with 0.5 ml of acid ethanol, and insulin was extracted overnight at +4°C for determination of the islet insulin content. The insulin concentrations in the incubation media were determined by radioimmunoassay ⁽¹⁷⁾.

Transplantation into nude mice. Normoglycemic, male, inbred, athymic nude C57BL/6 mice (Bomholdtgard, Ry, Denmark) were used as recipients of microencapsulated human islets. After induction of anesthesia with tribromoethanol (Avertin, Sigma; ¹⁸), a small incision was made through the abdominal skin and peritoneum. Approximately 500 microencapsulated human islets that had been cultured for 6 days were deposited intraperitoneally by means of a sterile pipette; finally, the peritoneal membrane and skin were closed with sutures. In two mice, microspheres containing human islets were also transplanted beneath the kidney capsule, according to a previously described procedure for islet transplantation ^{(11, 12)}. The mice were subsequently housed separately and allowed free access to tap water and pelleted food (R34, Analycen, Lidköping, Sweden). After 3 to 4 weeks, the mice were killed by cervical neck distension, their abdomens were opened, and groups of microspheres, partly forming larger clusters, were collected from the pelvic part of the abdominal cavity and fixed in 10% formalin, as were the graft-bearing kidneys. After fixation, the microspheres and the kidneys were embedded, cut, and stained as described in detail elsewhere ⁽¹⁴⁾. For immunocytochemistry, a primary guinea pig anti-bovine insulin antiserum (Biomakor, Rehovot, Israel) was used, and insulin-positive cells were subsequently demonstrated by the peroxidase antiperoxidase technique ⁽¹⁹⁾.

Statistical analysis. Each individual observation represents different islet donors or different pig litters. In each experiment, the insulin secretion was calculated as a mean of the values obtained from the three incubation vials, and was considered as one separate observation. Means ± SEM were subsequently calculated, and groups of data were compared using Student's unpaired or paired t test.

RESULTS

Insulin content. When measured on day 1 and day 6 of culture following the microencapsulation procedure, the insulin content of rat pancreatic islets was similar in control islets and encapsulated islets (Fig. 1A). The insulin content of mouse pancreatic islets also remained unchanged, irrespective of whether the islets were microencapsulated (Fig. 1B). However, after microencapsulation, the porcine ICC showed a lowering of their insulin content that was statistically significant after 6 days (Fig. 1C). The microencapsulated human pancreatic islets showed an elevated insulin content compared with the nonencapsulated controls on day 1, but on day 6 this difference did not attain statistical significance (Fig. 1D). During this investigation, we had also aimed to assess on a quantitative basis the amount of cells inside the microcapsules, i.e., to measure DNA content after homogenization. However, it turned out that the microcapsules were so resistant that even repeated freeze-thawing cycles and sonication did not destroy them. Nevertheless, the presently performed insulin content measurements may give an estimate of the β-cell survival, although it cannot be excluded that some insulin might be present in nonviable cells.

Insulin release. When examined on day 1, the insulin release of rat pancreatic islets at 1.7 mM glucose was equal in control and microencapsulated islets (Fig. 2A). Both microencapsulated and control rat islets increased their insulin release in response to glucose stimulation (16.7 mM glucose), and this response was further potentiated by addition of the phosphodiesterase inhibitor theophylline (Fig. 2A). However, the stimulated secretory rates of the microencapsulated islets were reduced by 43% (16.7 mM glucose) and 46% (16.7 mM glucose + 5 mM theophylline). On day 6, the nonencapsulated rat islets maintained their secretory rates in response to the various stimuli, and at this stage the microencapsulated islets secreted as much insulin as the controls (Fig. 2B). The nonencapsulated mouse pancreatic islets showed a marked insulin release in response to high glucose and high glucose + theophylline, both on day 1 (Fig. 3A) and on day 6 (Fig. 3B). The microencapsulated mouse islets, however, exhibited a clear-cut reduction in the insulin response compared with the control group, and this difference remained during the 6-day culture.

The control porcine ICC showed a low insulin response to 16.7 mM glucose but a marked elevation when 16.7 mM glucose + 5 mM theophylline was present (Fig. 4, A and B). However, the microencapsulated porcine ICC secreted small amounts of insulin both on day 1 and day 6. Microencapsulated human pancreatic islets secreted somewhat more insulin at low glucose on day 1 than the corresponding nonencapsulated islets, but this difference was not observed on day 6 of culture (Fig. 5, A and B). When exposed to 16.7 mM glucose on culture day 6, both groups of human islets showed an increased insulin release, which was further potentiated by the addition of theophylline. For both encapsulated and nonencapsulated human islets, the insulin output seemed to be higher on day 6 than on day 1. There were no differences in the amounts of insulin secreted during high-glucose stimulation when nonencapsulated and encapsulated human islets were compared.

Transplantation of microencapsulated human islets. Four nude mice each received intraperitoneal grafts from four different human islet preparations. When histologically examined after 3 to 4 weeks, well-preserved islets (Fig. 6A) containing insulin-positive cells were found in all animals. A mild fibrotic reaction was frequently observed on the outside of the capsules. The microencapsulated islets implanted under the kidney capsule showed a normal structure (Fig. 6 B), with several insulin-positive cells. The fibrotic reaction on the outside of the alginate beads appeared to be less prominent than for the intraperitoneally implanted microcapsules.

DISCUSSION

The present study shows that microencapsulated islets maintain appropriate function for at least 6-8 hr after transportation in culture medium. The lowest rate of insulin release from the microcapsules was observed with the ICC explants. It should be noted in this context that the fetal porcine ICC represents a fairly undifferentiated tissue in terms of β-cell function, as evidenced by the lack of insulin response to glucose stimulation in vitro ^{(11, 12)}. The microenvironment inside the microcapsules may be insufficient to further promote differentiation of fetal cells, possibly because of lack of different growth factors and a suitable extracellular matrix ⁽²⁰⁾. On the other hand, during prolonged primary culture, nonencapsulated ICC tend to become overgrown and disappear due to fibroblasts proliferating on the bottom of the culture dishes.

Microcapsules containing mouse islets showed a reduced insulin release on days 1 and 6, whereas the microencapsulated rat islets showed a lower insulin release only on day 1. However, these effects do not seem to be related to a reduction in the β-cell stores of insulin, as evidenced by the insulin concentrations of the acid ethanol extracts. The functional improvement of the rat islets may indicate a restoration after minor damage induced by the microencapsulation process and the transportation. Indeed, we have observed a number of situations in which an impaired β-cell function can be restored following an initial injury ⁽²¹⁾.

The human islets appeared to have endured the microencapsulation process without any notable impairment. This is in line with previous observations that human β-cells seem to be less susceptible to a number of noxious treatments as compared with rodent β-cells ⁽⁸⁾. It has been reported that human islets contain, in comparison with islets of other species, increased levels of the stress protein heat-shock protein 70 ⁽⁹⁾, and it is possible that this may have contributed to the preserved function of the human islets. In the present study, microencapsulated human islets were exposed to a number of preparatory steps and transports between four different laboratories. Therefore, the microcapsules with human islets may have contained a subpopulation of surviving cells particularly resistant to the various treatments. On the other hand, this may not be a disadvantage if microencapsulated human islets will be used as allografts in clinical insulin-dependent diabetes mellitus. The morphology of the encapsulated human islets transplanted into nude mice both beneath the kidney capsule and intraperitoneally suggested that the human islets had survived inside the microcapsules. These limited observations indicate that the subcapsular renal space can be used successfully as an implantation site for microencapsulated human islets. This will facilitate a future experimental characterization of the function of transplanted microencapsulated islets.

In conclusion, the present study suggests that islets isolated from rat, mouse, and human donors retain their functional competence with respect to insulin content and release after microencapsulation and transportation. As evaluated by their insulin secretory capacity, the human and rat islets showed the best preservation, whereas mouse islets and fetal porcine ICC in particular were impaired by the procedure.

Acknowledgments. The authors thank M. Engkvist, E. Forsbeck, A. Nordin, and E. Törnelius for excellent technical assistance and A. King for linguistic revision of the manuscript. The human islets were prepared by the Central Unit of β-Cell Transplant.

Figure 1: Insulin content of cultured nonencapsulated ([dotted square]) and cultured microencapsulated ([square with lower left to upper right fill]) (A) rat islets, (B) mouse islets, (C) fetal porcine ICC, and (D) human islets, either after 1 day or after 6 days of culture. Bars are means ± SEM for the number of observations given within parentheses. *P<0.05 vs. corresponding nonencapsulated group, using Student's unpaired t test.

Figure 2: Insulin release of cultured nonencapsulated ([dotted square]) and cultured microencapsulated ([square with lower left to upper right fill]) rat islets on (A) day 1 and (B) day 6 of culture. The islets were exposed for 1 hr to either 1.7 mM glucose, 16.7 mM glucose, or 16.7 mM glucose + 5 mM theophylline, as indicated. Bars are means ± SEM for the number of observations given within parentheses in Figure 1A. **P<0.01 vs. corresponding nonencapsulated group, using Student's paired t test.

Figure 3: Insulin release of cultured nonencapsulated ([dotted square]) and microencapsulated and cultured ([square with lower left to upper right fill]) mouse islets on (A) day 1 and (B) day 6 of culture. The islets were exposed to either 1.7 mM glucose, 16.7 mM glucose, or 16.7 mM glucose + 5 mM theophylline, as indicated. Bars are means ± SEM for the number of observations given within parentheses in Figure 1B. *P<0.01 vs. corresponding nonencapsulated group, using Student's unpaired t test.

Figure 4: Insulin release of cultured nonencapsulated ([dotted square]) and microencapsulated and cultured ([square with lower left to upper right fill]) fetal porcine ICC on (A) day 1 and (B) day 6 of culture. The ICC were exposed to either 1.7 mM glucose, 16.7 mM glucose, or 16.7 mM glucose + 5 mM theophylline, as indicated. Bars are means ± SEM for the number of observations given within parentheses in Figure 1C. *P<0.05 and **P<0.01 vs. corresponding nonencapsulated group, using Student's unpaired t test.

Figure 5: Insulin release of cultured nonencapsulated ([dotted square]) and microencapsulated and cultured ([square with lower left to upper right fill]) human islets on (A) day 1 and (B) day 6 of culture. The islets were exposed to either 1.7 mM glucose, 16.7 mM glucose, or 16.7 mM glucose + 5 mM theophylline, as indicated. Bars are means ± SEM for the number of observations given within parentheses in Figure 1D. *P<0.05 vs. corresponding nonencapsulated group, using Student's paired t test.

Figure 6: Light micrograph of a microencapsulated human islet transplanted (A) intraperitoneally or (B) under the kidney capsule of a normal nude mouse, 3-4 weeks after implantation. Some fibrotic tissue can be seen on the outside of the alginate capsule in the intraperitoneally implanted islet. During the histologic processing, the alginate microcapsules become deformed. Hematoxylin and eosin, magnification ×300.