CD99-targeted immunomagnetic negative selection: A novel strategy for high-purity pancreatic islet isolation in murine models
Jiao Liu, Dong Li, Jiayong Huang, Yangxin Hu, Lei Zhang
TL;DR
A new method using CD99-targeted immunomagnetic selection improves the purity and quality of isolated pancreatic islets for diabetes research in mice.
Contribution
A novel immunomagnetic negative selection protocol using CD99 for high-purity islet isolation in murine models.
Findings
Islet purity increased from 10.4 ± 3.9% to 93.0 ± 1.4% using CD99-targeted selection.
Purified islets showed structural integrity and glucose-stimulated insulin secretion comparable to conventional methods.
CD99-purified islets effectively reversed diabetes in a syngeneic transplantation model.
Abstract
Islet transplantation represents a promising therapeutic approach for type 1 diabetes through restoration of endogenous insulin production. However, efficient purification of islets from surrounding exocrine tissue remains a critical challenge, as current methodologies often compromise purity, yield, or islet viability. We exploited the differential expression of CD99 between murine pancreatic exocrine tissue (high expression) and islets (negligible expression) to develop a novel immunomagnetic negative selection protocol. Expression patterns were validated using immunofluorescence, immunohistochemistry, and western blot. Subsequently, streptavidin-conjugated magnetic beads coupled with biotinylated anti-CD99 antibodies were employed to selectively deplete CD99-positive exocrine cells from pancreatic digests, thereby enriching viable islets. This approach achieved a remarkable…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Fig 1
Fig 2
Fig 3
Fig 4- —Jiangsu Provincial Health Commission Scientific Research Project
- —http://dx.doi.org/10.13039/501100004608Natural Science Foundation of Jiangsu Province
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsPancreatic function and diabetes · Diabetes and associated disorders · Diabetes Management and Research
Introduction
Type 1 diabetes (T1D) constitutes a significant global health burden, affecting 8.42 million individuals worldwide in 2021, with projections indicating an increase to 13.5–17.4 million by 2040 [1]. This autoimmune disorder, characterized by progressive β-cell destruction, necessitates lifelong exogenous insulin therapy [2,3]. While pancreatic transplantation can restore euglycemia, the associated surgical morbidity and complications limit its widespread application [4–6]. Islet transplantation has emerged as a minimally invasive alternative, requiring transplantation of only a few milliliters of cellular suspension, making it particularly suitable for patients with brittle diabetes [7–9].
Despite initial improvements in glycemic control, maintaining long-term graft function remains challenging [10–12]. A critical determinant of transplant success is the purity of islet preparations. Residual exocrine tissue within grafts triggers innate immune activation and subsequent adaptive rejection, significantly compromising long-term outcomes [13,14]. Current purification methods face inherent limitations regarding purity, yield, efficiency, and cost-effectiveness.
Density gradient centrifugation, the clinical gold standard for islet isolation, enables large-scale processing [15]. However, its efficacy critically depends on unstable and subtle density gradients, resulting in inconsistent yield and purity. Moreover, the requisite centrifugal forces and media (e.g., Ficoll) can compromise islet viability [16,17]. In contrast, manual hand-picking achieves near-perfect purity without chemical insult, but its painstakingly low throughput makes it prohibitively inefficient for clinical use [18]. Although emerging microfluidic platforms provide high precision and gentle manipulation, their low technical maturity and currently limited throughput preclude their clinical translation [19]. Consequently, developing rapid, efficient, and economical islet purification strategies represents a crucial bottleneck limiting clinical translation.
Recent advances in immunomagnetic separation technology offer novel opportunities for high-precision cell purification [20–22]. This approach enables targeted enrichment while minimizing mechanical and chemical damage, thereby enhancing operational efficiency [23]. When suitable targets are available, negative selection achieves a level of specificity comparable to positive selection. Furthermore, by avoiding direct antibody binding to target cells, this gentler approach minimizes activation-related stress and better preserves cellular function and integrity [24,25].
CD99, a 32-kDa transmembrane glycoprotein [26,27], exhibits striking differential expression between murine pancreatic compartments. Our preliminary investigations revealed robust CD99 expression in exocrine tissue with virtually absent expression in islets. This distinctive expression profile positions CD99 as an ideal target for negative selection-based islet purification. Here, we describe the development and validation of a CD99-targeted immunomagnetic negative selection strategy that significantly enhances islet purity while preserving functionality, providing a novel, highly specific, and efficient alternative to current islet isolation technology.
Materials and methods
Reagents and materials
Liberase TL was obtained from Roche. Biotinylated and unconjugated anti-CD99 antibodies were purchased from Beijing Solarbio Science & Technology Co., Ltd. Anti-insulin antibodies, enhanced chemiluminescence (ECL) reagents, and fluorophore-conjugated secondary antibodies were acquired from Proteintech. Streptozotocin (STZ) and Ficoll 400 were purchased from Sigma-Aldrich. Mouse insulin enzyme-linked immunosorbent assay (ELISA) kits were obtained from Crystal Chem. Streptavidin-conjugated magnetic beads were purchased from BioSharp Life Sciences Limited. Magnetic separation stands were acquired from Miltenyi Biotec.
Animals
Male BALB/c and C57BL/6 mice (12 weeks old) were obtained from SPF Biotechnology (Beijing,China). All experimental procedures were approved by the Animal Ethics Committee of Kangda College, Nanjing Medical University (Approval No: IACUC-23XS006), and all personnel involved in the experiments had received standard laboratory animal training and passed the required assessments. The mice were housed under standard specific pathogen-free conditions with a 12-h light/dark cycle, ambient temperature maintained at 20–22°C, relative humidity of 40–60%, and provided with standard diet and water ad libitum. The experimental period lasted for 60 days. Health status was assessed at least once daily. Euthanasia was performed by cervical dislocation under deep anesthesia induced by pentobarbital sodium if any of the following humane endpoints were observed: (1) persistent anorexia or refusal to fluids for more than 24 hours; (2) severe incapacitation, such as inability to stand or ambulate; or (3) signs of severe distress including lethargy, hunched posture, and hypothermia (<37°C). A total of 60 BALB/c mice and 10 C57BL/6 mice were used in this study, and all experimental animals were euthanized upon completion of the study. Following islet transplantation, mice were maintained on a heating pad until they regained consciousness and exhibited normal postural reflexes. All efforts were made to minimize animal suffering and to adhere to the principles of replacement, reduction, and refinement (the 3Rs).
Tissue processing
For tissue analysis, mice were first deeply anesthetized with pentobarbital sodium (50 mg/kg, i.p.) and then euthanized by cervical dislocation, and pancreata were rapidly excised. Tissues designated for histological analysis were fixed in 4% paraformaldehyde for 24 hours prior to paraffin embedding. For islet isolation, the common bile duct was cannulated and perfused with 3 mL of Liberase TL solution (0.18 mg/mL in Hanks’ Balanced Salt Solution (HBSS)). Following 15-minute digestion at 37°C, enzymatic activity was terminated with 10% fetal bovine serum (FBS). The digest was filtered through a 500-μm mesh, yielding a heterogeneous suspension containing islets and exocrine fragments.
CD99 expression analysis
Immunofluorescence microscopy.
Deparaffinized sections underwent antigen retrieval (citrate buffer, pH 6.0, 95°C, 20 minutes) and blocking (5% bovine serum albumin (BSA), 1 hour). Primary antibodies (anti-insulin 1:200, anti-CD99 1:100) were applied overnight at 4°C. Following extensive washing, fluorophore-conjugated secondary antibodies were applied (1 hour, room temperature). Nuclei were counterstained with DAPI (4’,6-diamidino-2-phenylindole). Images were acquired using a Nikon N-STORM confocal microscope.
Immunohistochemistry.
Following standard deparaffinization and antigen retrieval, sections were incubated with anti-CD99 antibody (1:100) overnight at 4°C. Detection employed horseradish peroxidase-conjugated secondary antibodies with 3,3’-diaminobenzidine (DAB) chromogenic substrate. Hematoxylin counterstaining was performed prior to imaging.
Western blot analysis.
Protein lysates from purified islets and exocrine tissue were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to nitrocellulose membranes. Following blocking (5% BSA), membranes were probed with anti-CD99 (1:1000) and anti-β-actin (1:10000) antibodies. Protein bands were visualized using enhanced chemiluminescence (ECL) detection.
Islet purification protocols
CD99-targeted immunomagnetic negative selection.
Streptavidin magnetic beads (200 μg) were conjugated with biotinylated anti-CD99 antibody (1 μg) for 30 minutes on ice. The islet-exocrine mixture was blocked with 5% BSA (30 minutes), then incubated with antibody-bead complexes (20 minutes, 4°C). Following magnetic separation (5 minutes), the unbound fraction containing purified islets was collected. This process was repeated twice to maximize recovery.
Ficoll density gradient centrifugation.
Pancreatic digests were layered onto discontinuous Ficoll gradients (25%, 23%, 20.5%, 11%) and centrifuged (750 × g, 20 minutes, 4°C). Islets were harvested from the 11%/20.5% and 20.5%/23% interfaces, followed by extensive washing in HBSS.
Functional assessment
Purity determination.
Dithizone (DTZ) staining was performed to identify insulin-containing cells. Purity was quantified using ImageJ software (NIH, Bethesda, MD) by calculating the ratio of DTZ-positive area to total cellular area.
Glucose-stimulated insulin secretion (GSIS).
Groups of 50 islets were pre-incubated in Krebs-Ringer buffer containing 3.3 mM glucose (1 hour), then sequentially exposed to low (3.3 mM) and high (16.7 mM) glucose concentrations (1 hour each). Insulin release was quantified by ELISA. Stimulation index (SI) was calculated as the ratio of insulin secretion at high versus low glucose concentrations.
In Vivo transplantation.
C57BL/6 mice were used as recipients for diabetes induction and transplantation studies, with diabetes induced via a single tail vein injection of streptozotocin (STZ; 200 mg/kg); a fresh 20 mg/mL STZ solution in 0.9% saline was prepared immediately before use. Non-fasting blood glucose levels were monitored daily starting on day 3 post-injection, and mice with persistent hyperglycemia (>300 mg/dL) for two consecutive days were defined as diabetic and selected as islet transplantation recipients, while BALB/c mice served as islet donors. For the in vivo assessment, islets were isolated and pooled from 30 BALB/c donor mice and then purified by either the novel CD99-targeted immunomagnetic negative selection or the conventional Ficoll density gradient centrifugation method. Subsequently, 10 STZ-induced diabetic C57BL/6 mice were randomly divided into two groups (n = 5 per group), and each recipient was transplanted with 400 islet equivalents (IEQ) of the corresponding purified islets under the renal capsule, which allowed for a direct comparison of the therapeutic efficacy between the two purification methods. Blood glucose and body weight were monitored for 40 days post-transplantation. Blood glucose and body weight were monitored for 40 days post-transplantation.
Statistical analysis
Data are presented as mean ± standard error of the mean (SEM). Statistical analyses were performed using GraphPad Prism 8.0. One-way analysis of variance (ANOVA) with post-hoc Student-Newman-Keuls test (for homogeneous variances) or Dunnett’s T3 test (for heterogeneous variances) was employed for multiple comparisons. Two-group comparisons utilized unpaired Student’s t-test. Statistical significance was defined as P < 0.05.
Results
Differential CD99 expression in pancreatic compartments
Immunofluorescence analysis revealed distinct compartmentalization of CD99 expression within murine pancreatic tissue. While insulin immunoreactivity was confined to islet structures, CD99 expression was exclusively localized to exocrine tissue, with no detectable signal within islets (Fig 1A). This differential expression pattern was corroborated by immunohistochemical staining, which demonstrated intense CD99 immunoreactivity throughout exocrine regions with complete absence in islet clusters (Fig 1B). Western blot analysis provided quantitative confirmation, revealing robust CD99 protein expression in exocrine lysates while islet lysates showed no detectable signal (Fig 1C). These convergent findings establish CD99 as an exocrine-specific marker in murine pancreas.
Differential expression of CD99 in mouse pancreatic exocrine tissue and islets.(A) Immunofluorescence co-localization analysis. Paraffin-embedded mouse pancreatic sections were immunostained with antibodies against insulin (green) and CD99 (red), followed by DAPI counterstaining (blue) for nuclear visualization. CD99 immunoreactivity (red) was exclusively localized to exocrine tissue and did not co-localize with insulin-positive β-cells (green), magnification, 20 × ; scale bar, 100 μm. (B) Immunohistochemical analysis. CD99 expression (brown) was evaluated in paraffin-embedded sections of intact mouse pancreas, purified islets, and isolated exocrine fragments. Robust CD99 immunoreactivity was observed in exocrine regions of intact pancreatic tissue, while islets remained consistently negative. Purified islets exhibited no detectable CD99 expression, magnification, 40 × ; scale bars, 50 μm (pancreas sections), 10 μm (purified islets and exocrine fragments). (C) western blot analysis. CD99 protein expression was quantified in purified islets and isolated exocrine fragments. CD99 was selectively detected in exocrine samples while remaining undetectable in islet preparations. β-Actin served as the loading control.
Enhanced islet purification via CD99-targeted negative selection
Microscopic examination confirmed well-preserved islet architecture. The islets presented as intact, well-demarcated clusters with characteristic rounded morphology and smooth surfaces. Initial pancreatic digests contained substantial exocrine contamination, resulting in a baseline islet purity of 10.4 ± 3.9% as determined by DTZ staining and morphometric analysis (Fig 2A and 2E). CD99-targeted immunomagnetic depletion dramatically improved purity to 93.0 ± 1.4% (P < 0.0001; Fig 2B and 2E), while maintaining a high islet recovery rate of 81.7 ± 2.2% (Fig 2F). For comparison, Ficoll density gradient purification was also assessed. The initial purity (12 ± 1.7%) was comparable to that of digests for immunomagnetic purification (P > 0.05) (Fig 2C and 2E). After Ficoll purification, islet purity increased to 69.2 ± 1.1% (Fig 2D and 2E), which was significantly lower than the purity achieved with the immunomagnetic method (P < 0.0001). Moreover, the islet recovery rate using the Ficoll method was only 31 ± 1.5%, also significantly lower than that of the immunomagnetic approach (P < 0.0001) (Fig 2F).
*Purification of pancreatic islets via CD99-targeted immunomagnetic depletion markedly improves purity and recovery compared to Ficoll gradient centrifugation.(A, B) Representative images of dithizone-stained pancreatic digest samples. (A) Unpurified sample demonstrating mixed cellular composition (islets stained red, exocrine tissue grayish-white). (B) Sample after CD99-targeted immunomagnetic purification, showing substantial enrichment of islets.(C, D) Representative images of dithizone-stained samples processed by Ficoll density gradient centrifugation. (C) Initial unpurified digest. (D) Sample after Ficoll purification.(E) Quantitative analysis of islet purity. Initial purity of digests was comparable between groups. CD99-targeted purification achieved a final purity of 93.0 ± 1.4%, significantly higher than the 69.2 ± 1.1% achieved by Ficoll purification. Data represent mean ± SEM; n = 3 per group. ****p < 0.0001.(F) Quantitative analysis of islet recovery rate. The recovery rate after CD99-targeted purification (81.7 ± 2.2%) was significantly higher than that after Ficoll purification (31 ± 1.5%). Data represent mean ± SEM; n = 3 per group. ***p < 0.0001.
Functional competence of purified islets
GSIS assays confirmed that β-cell function was preserved in CD99-purified islets. Both basal (3.3 mM glucose) and high-glucose (16.7 mM) stimulated insulin secretion were comparable to levels in Ficoll-purified islets (P > 0.05; Fig 3A). The robust response to glucose stimulation (P < 0.0001) in both groups resulted in equivalent stimulation indices (P > 0.05; Fig 3B), indicating intact glucose-sensing-secretory coupling after CD99-based purification.
*CD99-targeted immunomagnetic purification preserves glucose-responsive insulin secretion function comparable to Ficoll gradient separation.Pancreatic islet functionality was assessed through glucose-stimulated insulin secretion (GSIS) assays. (A) Insulin secretion response. High glucose stimulation (16.7 mM) significantly enhanced insulin secretion compared to low glucose conditions (3.3 mM) in islets purified by both methodologies, demonstrating preserved β-cell responsiveness. No significant differences in insulin secretion were observed between purification methods under either glucose concentration. n = 6; ***p < 0.0001 vs. low glucose; ns, not significant between methods, p > 0.05. (B) Stimulation index. The stimulation index, calculated as the ratio of insulin secretion under high glucose to low glucose conditions, showed no significant difference between purification methods, confirming equivalent functional preservation. ns, not significant, p > 0.05.
In vivo efficacy was demonstrated through syngeneic transplantation studies. CD99-purified islets (400 IEQ) successfully reversed STZ-induced diabetes, achieving normoglycemia within 3 days post-transplantation. Glycemic control and weight recovery profiles were indistinguishable from Ficoll-purified controls throughout the 40-day observation period (Fig 4A and 4B).
CD99-targeted immunomagnetic purification achieves therapeutic efficacy equivalent to Ficoll separation in diabetic reversal.Pancreatic islets (400 IEQ) purified via CD99-based immunomagnetic separation or Ficoll density gradient centrifugation were transplanted beneath the renal capsule of streptozotocin (STZ)-induced diabetic mice. Glycemic control and body weight were monitored over 40 days post-transplantation. (A) Glycemic response. All recipients exhibited hyperglycemia prior to transplantation (Day 0). Both treatment groups demonstrated rapid glucose normalization, achieving euglycemia (<200 mg/dL) by day 3 post-transplantation, which was sustained throughout the observation period. Area under the curve (AUC) analysis revealed no significant difference in overall glycemic control between purification methods, confirming equivalent therapeutic efficacy. n = 5; ns, not significant, p > 0.05. (B) Body weight dynamics. Both groups exhibited initial transient weight loss followed by progressive recovery, consistent with metabolic restoration. No significant inter-group differences were observed throughout the study period. n = 5; ns, not significant, p > 0.05.
Cost-effectiveness analysis
The two purification methods were compared based on the direct cost of core reagents. According to current market prices, for processing each batch of islets using Ficoll density gradient centrifugation (Sigma F9378-500G, 30,748 CNY per 500 g), approximately 5 mL of a 25% stock solution is required, resulting in a core reagent cost of about 77 CNY. In contrast, the CD99-targeted immunomagnetic negative selection method is significantly more economical: the biotinylated anti-CD99 antibody (Solarbio, catalog number K113581P‑Biotin, priced at 960 CNY per 50 μL) costs approximately 9.6 CNY per single use of 0.5 μL, while the streptavidin‑conjugated magnetic beads (Biosharp, BL1262A, 659 CNY per mL) cost about 13.18 CNY per single use of 20 μL. Together, these amount to a total cost of approximately 23 CNY per batch. The above calculation includes only differential consumables; shared materials such as collagenase, buffers, serum, and common laboratory consumables were not taken into account.
Discussion
The present study proposes a CD99-targeted immunomagnetic bead based negative selection strategy for the efficient purification of pancreatic islets. This approach originates from a key experimental observation: in mouse models, CD99 shows an expression pattern opposite to that in human tissues—it is highly enriched in exocrine cells but nearly absent in islet endocrine cells [28]. Based on this finding, we have innovatively repurposed the CD99 antibody, originally designed for positive selection in human tissues, into an effective negative selection tool for mouse models. This method enables the preparation of high purity islet preparations while effectively preserving islet viability and function.
The presence of exocrine contamination in islet preparations represents a major impediment to successful transplantation outcomes. Residual exocrine tissue increases graft volume, exacerbating local hypoxia [29,30], impedes revascularization [19], releases proteolytic enzymes that damage islets [31], and triggers robust inflammatory responses that accelerate graft rejection [19,32]. Therefore, developing efficient purification strategies is essential for optimizing transplantation outcomes.
Despite its widespread use, density gradient centrifugation often yields islet preparations with inconsistent and highly variable purity levels [29,33]. Moreover, the associated mechanical stress and osmotic fluctuations can impair islet function [15,17].
Our CD99-targeted immunomagnetic negative selection strategy demonstrates several distinct advantages. First, we comprehensively validated the differential expression of CD99 at the molecular level using multiple techniques, confirming its exclusive and abundant presence in exocrine tissue with negligible expression in islets. This high specificity minimizes the risk of off-target binding and ensures precise enrichment of islets, in contrast to other exocrine markers that may exhibit overlapping expression patterns. Secondly, negative selection effectively prevents cell activation, damage, or functional interference by avoiding direct antibody labeling of the target cells, thereby maximally preserving their viability and functional integrity [22,34]. Third, this method significantly enhanced islet purity, increasing it from approximately 10% to over 90%, while maintaining excellent structural integrity and functional performance of the purified islets.
In summary, the core rationale for selecting CD99 as a negative selection target resides in its demonstrated specificity and high expression in exocrine tissues in mouse models. We then employ the well-established immunomagnetic bead technology to efficiently and specifically eliminate contaminants, rather than directly manipulating the fragile islet cells and this is precisely the core innovation and improvement of the present method in comparison with the positive selection strategy that directly targets pancreatic islets.
Importantly, functional integrity was preserved throughout the purification process. The maintained glucose responsiveness, as evidenced by equivalent GSIS profiles and stimulation indices, indicates preservation of stimulus-secretion coupling machinery. Moreover, the successful reversal of diabetes in transplant recipients demonstrates that the purification process does not compromise in vivo engraftment or long-term function.
Beyond its role as a purification target, CD99 functions as an immunomodulatory molecule, regulating leukocyte adhesion, migration, and activation [35–40], while promoting inflammatory cytokine production (such as IL-6, TNF-α) [27,41,42]. Thus, depletion of CD99-positive cells may confer dual benefits: physical removal of contaminating tissue and reduction of the pro-inflammatory milieu within grafts. This potential “purification-immunomodulation synergy” warrants further investigation.
A notable limitation concerns species-specific expression patterns. Human pancreatic tissue exhibits reversed CD99 distribution, with high islet expression and low exocrine expression [28]. This precludes direct clinical translation of our murine protocol. Future investigations should focus on identifying conserved markers exhibiting similar differential expression across species or developing human-specific targeting strategies.
Conclusions
We have established a CD99-targeted immunomagnetic negative selection protocol that achieves exceptional islet purity while preserving functional competence. Although species-specific expression patterns limit immediate clinical translation, this proof-of-concept study establishes the feasibility of molecular-guided negative selection for islet purification and highlights a promising direction for developing efficient and potentially immunomodulatory islet purification technologies.
Supporting information
S1 TableRow data.Raw data for all bar graphs and statistical analyses.(XLSX)
S1 FileRaw image.Uncropped western blot images corresponding to Fig 1C.(DOCX)
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Gregory GA, Robinson TIG, Linklater SE, Wang F, Colagiuri S, de Beaufort C, et al. Global incidence, prevalence, and mortality of type 1 diabetes in 2021 with projection to 2040: a modelling study. Lancet Diabetes Endocrinol. 2022;10(10):741–60. doi: 10.1016/S 2213-8587(22)00218-2 36113507 · doi ↗ · pubmed ↗
- 2Katsarou A, Gudbjörnsdottir S, Rawshani A, Dabelea D, Bonifacio E, Anderson BJ, et al. Type 1 diabetes mellitus. Nat Rev Dis Primers. 2017;3:17016. doi: 10.1038/nrdp.2017.16 28358037 · doi ↗ · pubmed ↗
- 3Quattrin T, Mastrandrea LD, Walker LSK. Type 1 diabetes. Lancet. 2023;401(10394):2149–62. doi: 10.1016/S 0140-6736(23)00223-4 37030316 · doi ↗ · pubmed ↗
- 4D’Alessandro C, Todisco M, Di Bella C, CrimìF, Furian L, Quaia E, et al. Surgical complications after pancreatic transplantation: a computed tomography imaging pictorial review. World J Gastroenterol. 2023;29(46):6049–59. doi: 10.3748/wjg.v 29.i 46.6049 38130739 PMC 10731157 · doi ↗ · pubmed ↗
- 5Gruessner AC, Gruessner RWG. Long-term outcome after pancreas transplantation: a registry analysis. Curr Opin Organ Transplant. 2016;21(4):377–85. doi: 10.1097/MOT.0000000000000331 27258580 · doi ↗ · pubmed ↗
- 6Redfield RR, Rickels MR, Naji A, Odorico JS. Pancreas transplantation in the modern era. Gastroenterol Clin North Am. 2016;45(1):145–66. doi: 10.1016/j.gtc.2015.10.008 26895686 · doi ↗ · pubmed ↗
- 7Shapiro AMJ, Pokrywczynska M, Ricordi C. Clinical pancreatic islet transplantation. Nat Rev Endocrinol. 2017;13(5):268–77. doi: 10.1038/nrendo.2016.178 27834384 · doi ↗ · pubmed ↗
- 8Langlois A, Pinget M, Kessler L, Bouzakri K. Islet transplantation: current limitations and challenges for successful outcomes. Cells. 2024;13(21):1783. doi: 10.3390/cells 13211783 39513890 PMC 11544954 · doi ↗ · pubmed ↗