In contrast to current rodent models, domestic cats naturally develop a form of diabetes mellitus (DM) that shares many similarities to type 2 diabetes (T2D) in humans, including obesity-induced insulin resistance and dyslipidemia, alterations in gut microbiota, impaired β-cell function, and spontaneous pancreatic amyloid deposition (1, 2). However, the utility of cats as a research model of type 2 diabetes has been hampered by the difficulty in studying feline islets in vitro. The ability to study islets in vitro is a critical component in the advancement of research into the prevention, targeted treatment, and etiology of DM.

The incidence of DM in cats is reported as approximately 1 in 100 cats presented to veterinary teaching hospitals (3) and approximately 1 in 200 cats presented to veterinary private practice (4). Like T2D in people, obesity, physical inactivity, and increased age are risk factors for DM in cats (5–7). Feline DM does not typically involve immune-mediated destruction of beta cells, but beta cell loss and dysfunction do occur (8–10), and unfortunately most cats are insulin-dependent upon diagnosis (11). Even with excellent owner compliance, daily insulin injections can be challenging, and DM can lead to a variety of complications such as infection, polyneuropathy, nephropathy, hypoglycemia, and diabetic ketoacidosis (8, 11). Therefore, islet transplantation might be worth exploring as a treatment option in feline DM. Whole pancreas transplantation allows for precise glycemic control but necessitates lifelong immunosuppression and is accompanied by high risk of organ rejection and perioperative complications (12–14). In contrast, islet isolation from exocrine pancreatic tissue enables targeted transplantation of only functionally necessary cells while minimizing host immune response and avoiding a surgical procedure. The site of clinical islet transplantation in people is the portal vein accessed via interventional radiology (15), but infusion into the peritoneal cavity and omentum have been reported in dogs, and may be more realistic transplant sites for the veterinary patient (16, 17). Additionally, isolated islets can be encapsulated, offering protection from the host immune system, eliminating the need for immunosuppression (15, 18). In humans, islet transplantation can result in 100% insulin independence at 1 year, and while this decreases to 30% at 5 years, recipients continue to have a 90% decrease in dangerous hypoglycemic episodes (15).

Protocols for islet isolation and culture have been developed and are routinely performed in other species, but these traditional methods involve enzymatic digestion and mechanical cell separation via density gradient centrifugation (19, 20). These methods can result in low islet yield and high rates of cell death (21–23). Consequently, most diabetic patients require multiple islet transplants to achieve satisfactory results (15). Collagenases are intended to disrupt connective tissue, but can also penetrate islet cell membranes, resulting in apoptosis, necrosis, and release of pro-inflammatory cytokines and free radicals, contributing to graft failure (24, 25). Atwater et al. (26) proposed an alternative method of islet isolation to circumvent some of these negative outcomes. This method exploits the GLUT 2 transporter to selectively spare beta cells from rapid changes in osmolality mediated by glucose concentration. In a high glucose hyperosmolar solution, glucose enters beta cells freely, preventing an osmolar gradient. Meanwhile, cells that do not express this membranal transporter are subjected to rapid changes in tonicity, cell membrane damage, and subsequent cell death. This selective osmotic shock technique has been successfully applied for islet isolation from porcine, canine, and human pancreata (26–28). Moreover, it is technically simpler and yields higher quantities compared to enzymatic digestions (26, 27).

Feline islet isolation has been described in a small number of studies with limited success utilizing traditional collagenase methods (29–32). To our knowledge, there is only one publication demonstrating yield of functional feline islets as determined by glucose-stimulated insulin secretion (GSIS) testing (32), which is the gold standard for assessing islet function (33). We hypothesized that use of selective osmotic shock (SOS) for isolation of feline islets will produce glucose-responsive beta cells. Our objectives in the current study were to (1) develop a selective osmotic shock protocol for feline islet isolation (2) describe purity and morphology and (3) compare the effects of osmolality and incubation time on islet yield and glucose responsiveness.

To our knowledge, this is the first description of SOS islet isolation in cats, and our results show that the SOS isolation protocol produces glucose-responsive β-cells, supporting our hypothesis. GSIS is a well-accepted measure of islet function, and healthy islets are expected to secrete increased insulin concentrations when moved from low to high concentrations of glucose (21, 35). In other species, healthy islets typically produce stimulation index >2 (35–37), but the standard for SI in human islet transplantation is an SI >1 (21). This is because clinical outcome of human islet transplant is not correlated with SI (18). Stimulation index can also be variable due to differences in protocols between institutions (37) and potentially because of intrinsic differences between species (35–37). There is only one study that has published on GSIS in isolated feline islets, and although this study used a different incubation time (30 min) and high glucose concentration (16.7 mM), the mean SI was approximately 2 (32), which is similar to our results in treatment group C. In vivo insulin stimulation testing suggests that glucose is a less effective stimulator of insulin secretion in cats compared to humans and that arginine may be superior to glucose for assessing insulin secretory capacity (38). Our results showed that islets isolated with 600 mmol/L glucose solution had higher SI than islets isolated with 300 mmol/L glucose solution. In canine islets, no significant difference was found between these two osmolalities, however, islets isolated with 300 mmol/L glucose solutions had higher trending SI than those isolated with 600 mmol/L glucose (27). It was suggested that higher glucose concentrations in dogs may affect islet function and viability, and although no difference in islet viability was found between osmolalites, there was a higher percentage of acinar cell death in the higher osmolality treatments (26). Unlike dogs, our results suggest that a higher osmolality could be optimal for isolation by SOS in cats. However, viability was not assessed, which is one of the limitations of this study.

Feline islets may be more sensitive to collagenases than islets of other species, and one proposed reason for this is that there is less basement membrane-like material and collagen surrounding feline islets and separating them from acinar tissue (30, 31). Higher amounts of collagen directly surrounding islets has been associated with higher purity (39) after enzymatic islet isolation procedures in other species, while higher amounts of total collagen in the pancreas are also associated with more difficult islet isolation. Thus, it is possible that traditional enzymatic islet isolation methods have a causal relationship with the limited success in producing glucose-responsive feline islets in the literature. The use of collagenases inherently cause damage to islets, injuring basement membranes and extracellular matrix, which causes a localized pro-inflammatory response via free radicals and cytokines (23–25, 40). Islets have even been shown to internalize these enzymes, resulting in decreased function and cell death (24). This results in the selection for resilience to collagenase-based procedures, which is unlikely to be uniform among islet size, functionality, and micro-architecture (41–43). Conversely, SOS likely selects for islets that have a higher density of GLUT 2 transporters, which has been identified as a favorable indicator of islet function and glucose responsiveness, (26, 44, 45) desirable traits for transplantation as well as in vitro studies.

The SOS isolation method has been shown to yield higher islets per gram of tissue (13,423 islets/g) (26) compared to Liberase-treated porcine pancreata (4,210 islets/g) (46) and compared to human cadaveric pancreata (2,279 islets/g) (19). In dogs, the SOS protocol had lower yield (28×10³ islet-like equivalents) compared to traditional methods (49 to 234×10³ islet-like equivalents) (47), however, delay in quantification and warm ischemia are suggested to have affected yield (27). There is minimal information on feline islet yield, with a single publication showing a mean yield of islet-like cell clusters (±SD) to be 2,200 ± 1,400 from a mean 3.8 ± 1.0 g of pancreas (30). Other publications have not reported yield(29, 31, 32). SOS isolation in feline donors, however, seems to have similar to increased yield when compared to traditional methods. Warm ischemia during islet isolation could have affected islet yield, and performing this step under cooled conditions is recommended in the future. Our treatment groups were the same as in other studies on SOS, and we found that neither osmolality nor incubation time had an effect on islet yield. Thompson et al. had a similar finding in canine donors (27), while Atwater et al. (26) found highest yield in 600 mmol/L incubated for 20 min in porcine donors. It has been proposed that higher osmolality could increase yield while longer incubation time may increase damage to islets. However, our findings did not support this outcome in cats. Additionally, there was marked variability in yield among individual cats, despite being free of known pancreatic and circulatory disease, which could also affect our ability to perceive treatment differences. The explanation for such wide ranges in islet yields is unclear, but some potential explanations could include varying degrees of warm ischemia after harvest and subclinical pancreatic, cardiovascular, or metabolic disease that may have affected islet concentration. Unfortunately, histopathology was not performed to assess and compare pancreatic architecture and islet concentration differences among individuals, which is a limitation to this study. There was significant variability in human islet yield until Ricordi described a method for a semi-automation of the enzymatic isolation process (15, 19). It is possible that a similar system for automation of the degree of mechanical and chemical exposure of the islets will be required before the selective osmotic shock procedure results in consistent islet yield, as well.

There are no standard morphological assessment criteria for isolated islets, and morphology differs among species (48). Morphology scoring systems proposed in human islet xenotransplantation have shown to correlate with outcome, where higher scores consist of spherical islets, well-rounded borders, compact cells, and larger borders (49, 50). Peri-insular capsule integrity has also been recognized as marker of islet health, and in recent years collagenase-based methods have aimed to preserved to aim capsular integrity (25, 26, 46, 51, 52). Feline islets isolated by SOS meet the majority of high-quality morphological characteristics described in isolated islets of other species, with a notably prominent and congruent peri-insular capsule that has not been described or demonstrated in feline islets isolated by other methods (29–31).

SOS is technically simple, allows for the successful isolation of clinical grade quality islets from feline donors, and is considerably less costly than traditional methods. While this protocol will be useful in obtaining feline islets for in vitro studies of feline islet physiology, it still requires significant improvements in islet purity and yield before clinical application of islet transplantation. Following SOS, Thompson et al. (27) incorporated the use of a sieve and cell strainer prior to plating canine islets for cell culture. We found that, in feline islet isolation, this drastically compromised yield, therefore the only purification step performed was hand picking islets to separate them from acinar tissue. Visual estimation of purity is also a limitation of this study, and in the future the use of an image analysis software for quantification of purity is suggested. Clinical islet transplant materials for humans have significantly higher purity (78.5%), (50, 53) but collagenase-based techniques require an additional step in purity, such density-gradient centrifugation, which separates acinar tissue from islets using dextran or ficoll (19, 20). An additional step to improve purity and further refinement and automation of mechanical tissue disruption is likely needed for the current SOS protocol. Nevertheless, this protocol enables access to an additional islet model for human type II diabetes mellitus.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The animal study was approved by the Institutional Animal Care and Use Committee at the University of Florida. The study was conducted in accordance with the local legislation and institutional requirements.

LP: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Writing – original draft, Writing – review & editing. CA: Conceptualization, Investigation, Methodology, Project administration, Resources, Supervision, Writing – original draft, Writing – review & editing. CC: Data curation, Project administration, Writing – original draft, Writing – review & editing. JM: Data curation, Project administration, Writing – original draft, Writing – review & editing. CG: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Writing – original draft, Writing – review & editing.