Localized Immunomodulation with PD-L1 Results in Sustained Survival and Function of Allogeneic Islets without Chronic Immunosuppression¹

Abstract

Allogeneic islet transplantation is limited by adverse effects of chronic immunosuppression used to control rejection. The PD-1 pathway as an important immune checkpoint has potential to obviate the need for chronic immunosuppression. We generated an oligomeric form of PD-1 ligand chimeric with streptavidin, SA-PDL1, that inhibited the T effector cell response to alloantigens and converted T conventional cells into CD4⁺FoxP3⁺ T regulatory (Terg) cells. The SA-PDL1 protein was effectively displayed on the surface of biotinylated mouse islets without a negative impact islet viability and insulin secretion. Transplantation of SA-PDL1-engineered islet grafts with a short course of rapamycin regimen resulted in sustained graft survival and function in > 90% of allogeneic recipients over a 100-day observation period. Long-term survival was associated with increased levels of intragraft transcripts for innate and adaptive immune regulatory factors, including indoleamine dioxygenase-1, arginase-1, FoxP3, TGF-β, IL-10, and decreased levels of proinflammatory T-bet, IL-1β, TNF-α, and IFN-γ as assessed on day 3 post-transplantation. T cells of long-term graft recipients generated a proliferative response to donor antigens at a similar magnitude to T cells of naïve animals, suggestive of the localized nature of tolerance. Immunohistochemical analyses showed intense peri-islet infiltration of Treg cells in long-term grafts and systemic depletion of this cell population resulted in prompt rejection. The transient display of SA-PDL1 protein on the surface of islets serves as a practical means of localized immunomodulation that accomplishes sustained graft survival in the absence of chronic immunosuppression with potential clinical implications.

Introduction

The potential of pancreatic islet grafts as a treatment option for type 1 diabetes (T1D) has been demonstrated in a recent multicenter phase 3 clinical trial (1). Two significant limitations of clinical islet transplantation include the paucity of cadaveric pancreata and the adverse effects of chronic immunosuppression to control rejection. Significant effort has recently been devoted to the generation of replenishable supply of insulin-producing cells, such as porcine pancreatic islets (2) or beta cells derived from stem cells (3), as an alternative to cadaveric human islets for transplantation. Regardless of the beta-cell source, the widespread use of insulin-producing cells as an effective treatment for T1D will require immunomodulatory approaches that obviate or mitigate the need for chronic immunosuppression.

Programmed cell death-1 (PD-1, CD279) is a member of the CD28/B7 superfamily of costimulatory molecules that shows low levels of expression on resting T cells and high levels of expression on activated CD4⁺ and CD8⁺ T cells, B cells, NK T cells, and monocytes (4). PD-1 ligand (PD-L1), one of the two physiological ligands, is constitutively expressed on a variety of hematopoietic and non-hematopoietic cells (4, 5). PD-1 is an important immune checkpoint pathway that modulates innate, adaptive, and regulatory immune responses, and as such, plays a critical role in immune homeostasis and tolerance to self-antigens (6). Mice deficient for PD-1 exhibit a breakdown of peripheral tolerance and manifest multiple autoimmune features, such as lupus and cardiomyopathy (7, 8). The role of this immune inhibitory pathway in tolerance to self-antigens was further substantiated by the demonstration that PD-L1 knockout NOD mice develop a rapid onset of diabetes as compared with wild type NOD (9, 10). PD-1 signaling is also extensively exploited by chronic infections and tumors for immune evasion, providing additional evidence for the importance of this pathway in negative regulation of immune responses (11, 12). The PD-1 pathway as an important therapeutic target in immuno-oncology has been verified by the reported remarkable clinical efficacy of blocking antibodies for various tumor types (13).

The PD-1 pathway also regulates alloreactive immune responses. PD-L1 blockade was shown to result in enhanced alloreactive T cell proliferation, Th1 cell differentiation, and accelerated MHC class II-mismatched skin graft rejection in mice (14). A dimeric form of PD-L1 and Ig fusion protein (PD-L1.Ig) blocked T cell proliferation in vitro and prevented cardiac allograft rejection in combination with anti-CD154 blockade, providing direct evidence for the potential of this pathway to induce allograft tolerance (15). The combination of PD-L1.Ig and anti-CD154 Ab was shown to have robust efficacy in an allogeneic islet transplantation setting where tolerance, rather than prolonged survival, was achieved (16). Treatment with another form of PD-L1.Ig fusion protein alone was shown to significantly prolong the survival of orthotopic corneal allografts in mice (17). Thus, the PD-1 pathway has significant potential for modulating alloreactive responses to overcome graft rejection.

We have previously reported an approach that allows the transient display of immunological ligands on biologic and nonbiologic surfaces for localized immunomodulation (18-20). This concept relies on the generation of recombinant proteins that constitute functional domains of immunological ligands fused to a modified form of core-streptavidin and the positional display of these molecules on biotinylated surfaces by taking advantage of the high affinity interaction between biotin and streptavidin (21). The transient display of a modified form of ligand for the Fas death receptor, SA-FasL, on pancreatic islets resulted in indefinite graft survival in allogeneic hosts in the absence of chronic immunosuppression (20). The importance of the PD-1 checkpoint pathway in self-tolerance, taken together with the clinical success of blocking antibodies to PD-L1 in cancer immunotherapy, served as the impetus to generate a recombinant form of PD-L1 chimeric with core streptavidin, SA-PDL1, and test its immunomodulatory function in a preclinical model of allogeneic islet transplantation. SA-PDL1 blocked the proliferation of T cells in response to alloantigens and facilitated the conversion of naïve T conventional (Tconv) cells into CD4⁺FoxP3⁺ Treg cells. The transient display of SA-PDL1 on allogeneic islet grafts resulted in sustained survival and function in the absence of chronic immunosuppression. Immune protection was localized to the graft and associated with intragraft expression of various innate and adaptive immunoregulatory factors and accumulation of CD4⁺FoxP3⁺ Treg cells. Depletion of Treg cells resulted in prompt rejection of long-term islet grafts. Thus, the transient display of immunomodulatory molecules on the surface of islets serves as an effective approach to prevent rejection in the absence of chronic immunosuppression with significant translational potential.

Materials and Methods

Animals

C57Bl/6, BALB/c and C3H animals were purchased from the Jackson Laboratory. C57BL/6.FoxP3^hCD2 (22) and B6.SJL-4C.TCR-tg (23) animals were generaously provided by Drs. H. Waldmann of Oxford University and TV Brennan of Duke University. All animals were maintained in our specific pathogen free vivarium at the University of Louisville. All experiments were performed in accordance to approved protocols by Institutional Animal Care and Use Committee, University of Louisville.

Construction, Expression, and Characterization of SA-PDL1 Protein

A synthetic gene was constructed to include the extracellular domain of mouse PD-L1 (68-728 bp, GI: AF233517.1) N-terminus to a modified form of core streptavidin (SA) and a 6xHis tag for purification. The synthetic gene was then subcloned into the pMT/BiP/V5-His A CuSO₄-inducible expression vector (Invitrogen, San Diego, CA) for stable expression in Drosophila S2 cells following published protocols (20, 21). SA-PDL1 protein was purified using metal affinity chromatography (GE, Amersham) and assessed for structure and purity using SDS-PAGE and Western blots. Purified protein was tested for concentration and endotoxin using the bicinchoninic acid and limulus amebocyte lysate tests respectively, and aliquoted, and frozen in −80°C until use.

SA-PDL1-Mediated T Conventional Cell Conversion into T Regulatory Cells

Splenocytes harvested from C57BL/6.hCD2 mice transgenic for human CD2 expressed under the control of FoxP3 (22), stained with anti-mouse CD4-APC and anti-human CD2-PE Abs, and suspended in cell sorting media (Hank’s balanced salt solution with 2% FBS). T conventional (Tconv; CD4⁺hCD2⁻) cells were sorted using FACSAria (> 99% purity) and cultured in 96-well U bottom plates (0.2 x 10⁶ cells/well) coated with anti-CD3 Ab (5 μg/ml). Cultures were supplemented with anti-CD28 Ab (1 μg/ml), various concentrations of SA-PDL1 protein, human-TGF-β1 (1 ng/ml, R&D Systems), and 20 U/ml recombinant human IL-2 (PeproTech) in complete mixed lymphocyte reaction (MLR) medium (20). Cultures were incubated for 72 hrs at 37°C in a 5% CO₂ incubator. Cells were harvested, stained with anti-mouse CD4-Alexa 700 and anti-human CD2-PE Abs and run on a BD LSR II flow cytometry. Data were analyzed using FlowJo (Tree Star Inc., San Carlos, CA) software.

SA-PDL1-Mediated Inhibition of Alloreactive T Cell Proliferation

Splenocytes from 4C mice (C57BL/6 transgenic for a TCR recognizing the BALB/c MHC I-A^d) (23) were used as responders against irradiated BALB/c splenocytes as stimulators in a standard MLR assay (19). Briefly, 4C splenocytes were cultured in a Petri dish for 45 minutes at 37°C to enrich for T cells. Non-adherent cells were collected, washed, and incubated with 2000 cGy irradiated BALB/c splenocytes in 96-well U-bottom plates (10⁵ cells/well/each) in complete MLR medium (19). After 48 hrs of incubation, cultures were supplemented with varying concentrations of SA-PDL1 protein or equimolar concentrations of control SA protein. Cultures were incubated for a total of 72 hrs, with the last 16 hrs pulsed with [³H] thymidine (1 μCi/well). Cultures were then harvested with Tomtec cell harvester to assess DNA-associated radioactivity [counts per minute (cpm)] as the measure of cell proliferation using a Beta plate counter. The percentage inhibition of T cell proliferation was calculated using the formula of 1- [cpm in test proliferation / cpm in control proliferation] x 100.

Mixed Lymphocyte Reaction (MLR) Assay

Splenocytes from various C57BL/6 graft recipients were labeled with carboxyfluorescein succinimidyl ester (CFSE) dye and used as responders against irradiated (2000 cGy) allogeneic BALB/c donor or third-party C3H splenocytes in a standard MLR assay (20). The cultures were harvested 4 days later and run on a BD LSR II flow cytometer to assess the proliferation of CD4⁺ and CD8⁺ T cells by gating on live cells. Data were analyzed using FACS Diva software and graphed using GraphPad Prism.

Modification of Cell Membrane with Biotin and Engineering with the SA-PDL1 Protein

Splenocytes were prepared from the BALB/c mice and incubated in 5 μM EZ-Link™ Sulfo-NHS-LC-Biotin solution (Thermo Scientific) in PBS for 30 minutes at room temperature in the dark per a published protocol (21). After washing to remove free biotin, cells were incubated in PBS supplemented with various amounts of SA-PDL1 protein (10-1280 ng/10⁶ cells) for 30 min at 4°C. Cells were then washed, stained with anti-SA-FITC Ab or SA-APC protein, run on a BD LSR II flow cytometry, and data were analyzed using FlowJo.

Islet Isolation, Engineering with the SA-PDL1 Protein, and Transplantation

Pancreatic islets were harvested from 8 to 12-week-old BALB/c mice and engineered with SA-PDL1 protein per published protocols (20). Briefly, the pancreas was perfused with cold Liberase TL (0.18 mg/ml, Roche Diagnostics) and digested for 17-18 minutes at 37°C. Islets were isolated on a discontinuous Ficoll (Sigma-Aldrich) gradient and maintained overnight in RPMI-1640 medium supplemented with 10% fetal bovine serum and penicillin/streptomycin (100 U/ml and 100 μg/ml) in a 5% CO₂ incubator at 37°C. Islets were collected in a 15-ml polypropylene round bottom tube, washed with PBS, and incubated in 5 μM EZ-Link™ Sulfo-NHS-LC-Biotin solution at room temperature for 30 min. Biotinylated islets were then engineered with SA-PDL1 protein (~400 ng/500-550 islets) for 30 min at room temperature. Engineered islets were washed and then transplanted under the kidney capsule of streptozotocin-diabetic C57BL/6 recipients per published protocols (19, 20). Selected groups of graft recipients were intraperitoneally injected with rapamycin (0.2 mg/kg, LC Company) starting on the day of transplantation for 15 days. Animals were monitored for blood glucose levels, and those with two consecutive daily measurements of ≥ 250 mg/dl were considered diabetic and confirmation of graft rejection.

To assess the role of Treg cells in the maintenance of long-term grafts, SA-PDL1-engineered islets were transplanted into streptozotocin-diabetic C57BL/6.FoxP3^EGFP/DTR mice (Jackson Lab) transgenic for human diphtheria toxin receptor expressed under the control of FoxP3 promoter. Graft recipients were treated i.p. with diphtheria toxin (50 ng/gm) for two consecutive days at the indicated days post-transplantation. Treg depletion was confirmed in peripheral blood using flow cytometry.

Flow Cytometry

Abs against mouse CD4 (RM4-5), CD3 (500A2), CD8 (53-6.7), CD62L (MEL-14), CD44 (IM7), CD25 (PC61.5), PD1 (RMP1-30), FoxP3 (FJK-16s), IFN-γ (XMG1.2), TNF-α (MP6-XT22), IL2 (JES6-5H4) were purchased from BD or eBioscience. Optimal concentrations for all fluorochrome-conjugated Abs were determined by titration. For surface staining, cells were incubated with respective Abs for 30 min at 4° C. For intra-nuclear FoxP3 staining, Fixation-Permeabilization buffer (eBioscience) was used as per manufacturer’s instruction followed by incubation with anti-FoxP3 Ab for 30 mins at 4° C. For intracellular cytokine analysis, spleen cells were incubated for 6 hours at 37° C in complete MLR media with PMA (50 ng/ml), Ionomycin (1 μg /ml) and Golgi plug (BD biosciences, for last 4 hours). After surface staining, Cytofix (BD biosciences) was used according to manufacturer’s instruction to fix and home-made perm buffer (0.1% saponin in PBS) for permeabilization followed by incubation with Abs against cytokines and analysis using flow cytometry as reported (19).

Glucose-Stimulated Insulin Secretion (GSIS) Assay

GSIS was performed for naïve and SA-PDL1 engineered BALB/c islets using a static incubation protocol (24). Briefly, low (3 mM) and high (11 mM) glucose solutions were prepared in Krebs ringer bicarbonate buffer. Islets were handpicked and equilibrated at low glucose (100 islets/ml) in a millicell cell culture insert (Merck) for 1 hr at 37 °C in a 5% CO₂ incubator. Islets were then challenged with high glucose solution for stimulation by incubating the inserts holding islets in the well containing high glucose solution for 1 hr at 37 °C in a 5% CO₂ incubator. The supernatant was collected from each well at different time points and stored at −80°C until insulin content analysis by ELISA (Mercodia). DNA quantification (Qiagen) was done from each pelleted islet mass for normalization against individual insulin content. The stimulation index was calculated as the ratio of insulin secreted in high glucose stimulation to low glucose stimulation.

Intraperitoneal Glucose Tolerance Test (IPGTT)

IPGTT was performed on mice fasted for 6 hrs. Blood samples were obtained at 0 and 10, 20, 30, 45, 60, 90, and 120 min post-intraportal glucose injection (2 g/kg body weight). Blood glucose levels were measured using an Accu-Check Nano glucometer and Smart View test strips (Roche). IPGTT was calculated by estimating the total area under the curve using GraphPad trapezoid rule.

Immunohistochemical Analysis

Kidneys harboring islet grafts frozen in Tissue-Tek O.C.T. compound (Sakura FineTek) were cut into 5-8μm sections and fixed with ice-cold acetone or 4% paraformaldehyde on frosted slides. Slides were incubated in blocking solution (0.5% Triton X-100, 1% bovine serum albumin (BSA), 5% goat serum and rat anti-mouse CD16/CD32 FcγII/III receptor block) for 30 minutes at room temperature. Sections were then incubated for 1 hr at room temperature in a cocktail of rat anti-mouse CD4 Ab (1:40; BD Pharmingen) and guinea pig anti-insulin polyclonal Ab (1:100; Dako) in PBS supplemented with 1% BSA. After washing twice with PBS, tissue sections were incubated in a secondary Ab cocktail of Alexa Fluor 647-conjugated (CD4) goat anti-rabbit polyclonal Ab (1:100; Life Technologies) and Alexa Fluor 555-conjuagted (insulin) anti-guinea pig Ab (1:300; Invitrogen) for 1 hr at room temperature. Slides were then washed with PBS and incubated for 1 hr at room temperature with rat anti-FoxP3 Ab conjugated with FITC (1:20; eBioscience) to visualize the Treg cells. Fluorescent images were obtained using a Leica TCS SP5 confocal microscope under 20X magnification.

RNA Isolation and Quantitative RT-qPCR

Islet grafts under the kidney capsule were surgically removed on day 3 post-transplantation and preserved in the RNAlater solution for stabilization and immediate RNase protection. RNA was extracted with Trizol reagent (Thermo Scientific), and cDNA was synthesized from 4 μg of the total RNA using SuperScript™ IV VILO™ Master Mix (Thermo Scientific). Quantitative real-time RT-PCR was performed for different genes using the TaqMan probe assay on QuantStudio 3 Real-Time PCR system (Applied Biosystems). Each PCR reaction consisted of 5 μl of TaqMan Fast Advanced master mix, 1 μl of Taqman primer and FAM/VIC labeled MGB probes, 1 μl of cDNA sample (from 1:10 dilution), and 3 μl of nuclease-free water in a final volume of 10 μl. The thermal cycling conditions were as follows: 1 cycle of 95°C for 20 s, followed by 40 cycles of 95°C for 1 s and 60°C for 20 s. The relative quantitation was analyzed using Comparative Ct (threshold cycle) protocol based on the 2^−ΔΔC_T method with DataAssist Software (Thermo Scientific). Briefly, differences in the C_T for the target gene and the C_T for the GAPDH gene control were calculated as ΔC_T, to normalize for differences in the amount of total nucleic acid added to each reaction and the efficiency of the RT step. Finally, the ΔC_T values were normalized to unmodified islet only group to be expressed as 2^−ΔΔC_T. Thus, all the experimental samples were expressed as n-fold change relative to the control (unmodified islet only group) as published (25).

Statistical Analysis

Student’s t-test (two tailed) was used to assess differences between two groups, whereas one-way ANOVA with Bonferroni’s multiple comparisons or Tukey’s multiple comparisons were used to determine differences among three or more groups. Graft survival was assessed using the log-rank (Mantel-cox) test. Data were expressed as mean ± SEM where indicated, p values of < 0.05 were considered statistically significant. Statistical analyses were performed using GraphPad Prism 9 software.

Results

Construction and Production of the SA-PDL1 Recombinant Protein.

A synthetic gene containing the coding sequences for the extracellular domain of mouse PD-L1 and a modified form of core streptavidin was subcloned in frame with the BiP secretion signal in the pMT/BiP/V5-HisA expression vector (Fig. 1A). Drosophila S2 cells stably transfected with this construct secreted the SA-PDL1 protein following stimulation with CuSO₄. The protein was successfully purified from the culture supernatants using metal affinity chromatography taking advantage of a 6xHis tag engineered into the construct. SA-PDL1 ran as oligomers in a reducing SDS-PAGE gel without heat (Fig. 1B). Heating the protein at 100°C resulted in the dissociation of oligomers into monomers, a structural feature dictated by the native streptavidin molecule that forms oligomers in solution (21, 26). Western blot analysis using polyclonal anti-streptavidin antibodies confirmed the identity and integrity of the protein (Fig. 1C).

FIGURE 1.

Construction, expression, and characterization of the SA-PDL1 protein. (A) Schematic presentation of pMT-Bip-SA-PDL1 construct. A synthetic gene encoding the extracellular domain of mouse PD-L1 linked to a modified form of streptavidin and 6His tag was cloned under the CuSO₄-inducible metallothionein (MT) promoter and the Bip secretion signal in the Drosophila S2 pMT/Bip/V5-HisA expression vector. (B) SDS-PAGE analysis of the purified SA-PDL1 protein. SA-PDL1 was purified from culture supernatants of S2 stable transfectants using immobilized metal affinity chromatography. Protein samples in reducing and denaturing loading buffer were either left unheated or heated in boiling water for 5 minutes and run on a 12.5% SDS-PAGE gel. The SA-PDL1 protein migrates as a single monomeric band (~52 kDa) with heat and as an oligomeric band (> 250 kDa) without heat. (C) Western blot analysis of the SA-PDL1 protein probed with anti-streptavidin Ab.

SA-PDL1 Protein Enhances the Conversion of Tconv Into Treg Cells and Blocks Proliferation of Alloreactive T Cells.

The interaction of PD-L1 with PD-1 receptor on CD4⁺ Tconv cells was shown to augment TGF-β-mediated in vitro conversion of these cells into CD4⁺CD25⁺FoxP3⁺ induced Treg cells (27). To assess if SA-PDL1 protein has a similar function, Tconv cells from a transgenic C57BL/6 mouse expressing the human CD2 gene (hCD2) under the mouse FoxP3 promoter (22) were sorted by flow cytometry gating on CD3⁺CD4⁺hCD2⁻ cells. The sorted Tconv cells were then cultured in the presence of agonistic Abs to CD3, CD28, and a cocktail of human TGF-β1 and IL-2 cytokines. There was a significant (p = 0.004) increase in the number of induced Treg cells in cultures supplemented with the highest dose of SA-PDL1 as compared with equimolar SA used as a control protein (Fig. 2A). As compared with cultures without proteins, cultures supplemented with the SA had lower percentage of Treg conversion, although not significant, plausibly due to nonspecific inhibition caused by SA protein or contaminants derived from insect cell proteins.

FIGURE 2.

SA-PDL1 protein enhances TGF-β-mediated conversion of Tconv into Treg cells and blocks alloantigen-mediated proliferation of T effector cells. (A) SA-PDL1 mediated augmentation of Tconv conversion into CD4⁺FoxP3⁺ Treg cells. Tconv cells (CD4⁺hCD2⁻) were flow sorted from C57BL/6 mice transgenic for the human CD2 molecule under the control of mouse FoxP3 promotor and cultured in medium supplemented with agonistic Ab to CD3 (5 μg/ml) and CD28 (1 μg/ml) and TGF-β (1 ng/ml), IL-2 (20 U/ml), and the indicated amounts of SA-PDL1 protein for 3 days. Streptavidin (SA) was used at equimolar levels as a control protein. The frequency of Treg cells (hCD2⁺CD4⁺FoxP3⁺) was assessed using flow cytometry. (B) SA-PDL1 suppresses proliferation of alloreactive T cells ex vivo. Splenocytes from 4C mice (C57BL/6 transgenic for a TCR specific for I-A^b) were co-cultured with irradiated BALB/c splenocytes as stimulators in a mixed lymphocyte reaction assay for 72 hrs. At 48 hrs of incubation, cultures were supplemented with the indicated amounts of SA-PDL1 or equimolar concentrations of SA as control and incubated for additional 24 hrs. Cultures were pulsed with [³H]-thymidine for the last 16 hrs of incubation and harvested to assess the DNA associated radioactivity. Data was graphed as percent inhibition. Both sets of these studies were performed in triplicate and repeated at least three times. Each data point is indicative of Mean ± SEM. Statistical analysis was performed using one-way ANOVA with Bonferroni’s multiple comparison test. *p < 0.05. **p < 0.01, ***p < 0.001.

PD-L1, as an immune checkpoint negative regulator, blocks T cell proliferative responses (28, 29). Thus, we tested the SA-PDL1 protein for blocking alloreactive T cell responses in vitro. Splenocytes from 4C mice transgenic for a TCR on the C57BL/6 background recognizing BALB/c H-2 I-A^d molecule (23) were used as responders to irradiated BALB/c splenocytes in the presence of various doses of SA-PDL1 protein in a 3-day proliferation assay. There was robust inhibition of 4C proliferation in cultures supplemented with higher doses of SA-PDL1 (10 and 20 μg/ml) as compared with cultures supplemented with equimolar amounts of SA control protein (Fig. 2B; p < 0.001). Taken together, these results demonstrate the immunomodulatory function of SA-PDL1 both on Treg and T effector (Teff) cells, consistent with published literature (27-29).

Pancreatic Islets Are Effectively Engineered to Display the SA-PDL1 Protein on Their Surface Without a Major Impact on Their Viability and Function.

To assess if SA-PDL1 protein can be transiently displayed on the surface of pancreatic islets, splenocytes were used as a rapid and quantitative platform to establish engineering conditions. Our previous studies have shown that surface modification of cells and islets with 5-15 μM biotin was optimum for the transient display of various immunomodulatory molecules at desired densities (21, 30-32). Thus, splenocytes modified with 5 μM EZ-Link sulfo-NHS-LC-Biotin were incubated with various concentrations of SA-PDL1 protein and cells were analyzed in flow cytometry. We observed a dose-dependent binding of SA-PDL1 to biotinylated splenocytes that plateaued at 640 ng protein/1x10⁶ cells (~ 98.5% of targeted cells with MFI values of > 13,000; Fig. 3A, B).

FIGURE 3.

Islets are effectively engineered with SA-PDL1 protein without a significant impact on their function. (A) Assessing cell surface engineering conditions with the SA-PDL1 protein. Mouse splenocytes were surface modified with 5 μM EZ-Link sulfo-NHS-LC-Biotin and engineered with the indicated amounts of SA-PDL1 protein (ng/10⁶ cells). The level of SA-PDL1 on the cell surface was assessed using an anti-SA Ab in flow cytometry. Biotinylated splenocytes without SA-PDL1 engineering served as control. (B) Mean fluorescence intensity (MFI) plotted against varying concentrations of SA-PDL1 protein. Data were tabulated from five independent experiments. (C) Engineering mouse pancreatic islets with the SA-PDL1 protein. Mouse islets were biotinylated (5 μM) followed by engineering with the SA-PDL1 protein (400 ng/500 islets). Biotinylation and the presence of SA-PDL1 on the islet surface were assessed using the streptavidin protein conjugated with APC (SA-APC, red) and anti-SA Ab (Anti-SA-FITC, green), respectively, in confocal microscopy. Islets positive for both molecules appear as yellow. Original magnification X 20. Staining patterns were consistent for samples across independent runs. (D) Engineering islets with SA-PDL1 protein does not impact insulin secretion. SA-PDL1-engineered and unmodified islets as control were stimulated with low (3 mM) and high (11 mM) glucose concentrations in a glucose stimulated insulin secretion assay. (E) Stimulation indices of studies conducted in D. Stimulation index was calculated by dividing the mean DNA normalized insulin value measured from high glucose samples by the low glucose samples. No significant difference (p = 0.73) was observed between the two groups using unpaired student t test.

Our previous studies with FasL as another immunomodulator used for islet engineering (20) and the broad range of SA-PDL1 protein displayed on the surface of splenocytes led us to use 5 μM biotin and 800 ng of SA-PDL1 protein per 1000 islets for engineering. Confocal microscopy analysis demonstrated the intense presence of both biotin and SA-PDL1 protein on the surface of islets (Fig. 3C) with moderate levels of SA-PDL1 in the inner core as assessed by Z-stack analysis (Fig. S1). To assess the impact of engineering on function, islets were tested in a glucose-stimulated insulin secretion assay. Although there was a nonsignificant reduction, plausibly due to the engineering process, in insulin secretion from the SA-PDL1-engineered islets as compared with unmodified islets (Fig. 3D), stimulation indices showed comparable functions (Fig. 3E).

SA-PDL1-Engineered Islets Survive Indefinitely in Allogeneic Hosts Under a Short Course of Rapamycin Regimen.

Pancreatic islets from mice genetically modified to lack PD-L1 were shown to undergo accelerated rejection in allogeneic recipients, emphasizing the importance of this immune checkpoint ligand in modulating alloreactive responses (33). Thus, we assessed the impact of SA-PDL1 protein displayed on the surface of islet grafts on survival in allogeneic hosts (Fig. 4A). Chemically diabetic C57BL/6 mice receiving unmodified BALB/c islets under the kidney capsule acutely rejected all grafts with median survival time (MST) of 14 days (Fig. 4B). SA-PDL1-engineered islet grafts showed significantly prolonged survival, but all grafts eventually rejected (MST = 28 days). We have previously shown that treatment with a short course of rapamycin in a similar model using FasL as an immunomodulator resulted in tolerance to allogeneic islet grafts (20). Treatment of SA-PDL1-engineered islet graft recipients with a 15-day course of rapamycin regimen starting the day of transplantation resulted in sustained survival of > 90% of grafts over an observation period of 100 days. In marked contrast, naïve unmanipulated islet grafts under the same rapamycin regimen were rejected in an acute fashion (MST = 19 days).

FIGURE 4.

SA-PDL1-engineered pancreatic islets show sustained long-term survival and function in allogeneic graft recipients. (A) Experimental scheme showing islet engineering, transplantation, and the transient use of immunosuppression. (B) Survival of SA-PDL1-engineered islet grafts in allogeneic recipients. SA-PDL1-engineered or unmodified islets were transplanted under the kidney capsule of chemically diabetic recipients without or with a short course of rapamycin (0.2 mg kg/daily for 15 doses). Animals were monitored for blood glucose levels, and those with two consecutive daily readings of ≥ 250 mg/dl were considered diabetic. Graft survival was assessed using the log-rank (Mantel-cox) test, ***p < 0.001. (C) Intraperitoneal glucose tolerance test (IPGTT) showing sufficient mass and function of transplanted islets. Recipients with long-term (≥ 150 days) graft survival were subjected to IPGTT with naïve mice serving as controls. (D) Area-under-the-curve for each animal glucose clearance was computed using the trapezoid rule in GraphPad Prism and compared using the student two-tailed t test. (E) Long-term euglycemia is maintained by the transplanted SA-PDL1-enginered islet grafts. Blood glucose levels of mice transplanted with SA-PDL1-engineered islets with or without rapamycin. Surgical removal of the kidney harboring the long-term SA-PDL1-engineered islets results in prompt hyperglycemia, confirming graft function.

Intraperitoneal glucose tolerance test (IPGTT) performed on long-term graft recipients revealed a similar pattern of blood glucose clearance as compared with naïve mice (Fig. 4C). The area-under-the-curve (AUC) of the blood glucose concentration during the IPGTT showed no differences between naïve and long-term graft recipients (Fig. 4D, p = 0.23), demonstrating sufficient functional mass of transplanted islets. Importantly, surgical removal of the kidney harboring long-term SA-PDL1-engineered grafts resulted in prompt hyperglycemia, confirming the role of transplanted islets in normalizing the blood glucose levels (Fig. 4E).

Recipients of Long-Term Allogeneic Grafts Generate a Systemic Response to Donor Antigens.

To assess if localized immunomodulation with the transient display of SA-PDL1 protein on allogeneic islets results in systemic unresponsiveness, splenocytes from long-term graft recipients were tested in vitro against donor antigens in a standard in vitro proliferation assay (18, 19). Both CD4⁺ and CD8⁺ T cells from long-term graft recipients responded to BALB/c donor and C3H third-party antigens at levels comparable to T cells from naïve mice or from rejecting recipients (SA-PDL1-engineered islet graft without rapamycin; Fig. 5A). T cells from long-term graft recipients also expressed similar levels of IL-2 and proinflammatory TNF-α and IFN-γ cytokines as compared with T cells isolated from naïve or recipients with graft rejection (Fig. 5B). Taken together, these data demonstrate the lack of systemic tolerance in long-term recipients of surviving islet allografts, thereby revealing the graft-localized nature of immune protection.

FIGURE 5.

Immunomodulation with SA-PDL1 results in islet graft-localized tolerance. (A) T cells of long-term islet graft survivors generates a normal response to donor antigens. Splenocytes from naïve (Naïve), rejecting (SA-PDL1-islet), and long-term survivors (SA-PDL1-islet + rapa) were used as responders against irradiated donor-matched BALB/c or C3H third party splenocytes in a standard CFSE-based in vitro mixed lymphocyte reaction assay. Dilution of CFSE in CD4⁺ and CD8⁺ T cells was assessed using Abs to CD4 and CD8 molecules in flow cytometry and plotted as the percent division for each cell population. (B) Intracellular cytokine response. Splenocytes from groups in (A) were stimulated with PMA and Ionomycin for 6 hrs, stained with fluorescence-labelled Abs to CD4, CD8, CD44 (to define activated cells) along with IFN-γ, TNF-α, and IL-2 and analyzed using flow cytometry. Data was tabulated as percentage of the indicated cell population gated on total CD4⁺ T or CD8⁺ T cells. The experiments were performed in duplicates and each data point is indicative of Mean ± SEM. Statistical analysis was performed using one-way ANOVA with Tukey’s multiple comparison test.

Long-Term Graft Survival Is Associated with Increased Intragraft Levels of Transcripts for Immunoregulatory Factors.

To provide further insight into the mechanistic basis of the observed localized immune protection, quantitative RT-PCR was performed on total RNA isolated from various groups on day 3 post-transplantation to assess the intragraft expression of transcripts for pro- and anti-inflammatory genes. We observed a significant increase in intragraft expression of the regulatory transcriptional factors FoxP3 (Treg; p < 0.05) and GATA-3 (Th2; p < 0.001) and the regulatory cytokines TGF-β (p < 0.001), IL-10 (p < 0.001), and IL-4 (p < 0.001) in SA-PDL1-engineered islets plus rapamycin group as compared with SA-PDL1-islets or unmodified islets plus rapamycin groups (Fig. 6). The levels of transcripts for the pro-inflammatory transcriptional factor T-bet (p < 0.01) and cytokines IFN-γ (p < 0.01), TNF-α (p < 0.001) and IL-1β (p < 0.0001) were significantly lower in the SA-PDL1-islet and SA-PDL1-islets plus rapamycin versus the islet plus rapamycin group. Importantly, we also observed increased expression of regulatory indolamine dioxygenase-1 (IDO-1, p < 0.01) and arginase-1 (ARG-1, p < 0.0001) as markers for M2 macrophages in the SA-PDL1-islets plus rapamycin group as compared with the SA-PDL1-islets or unmodified islets plus rapamycin groups (Fig. 6). There was not a significant difference in the expression of pro-inflammatory cytokines IL-6 among the groups, whereas nitric oxide synthase-2 (NOS-2) as a marker for M1 type macrophages and RORγT transcriptional factor for Th17 cells showed 1.5 to 2-fold increased expression in both rapamycin and SA-PDL1-islets plus rapamycin groups over the SA-PDL1-islet group. These results demonstrate that SA-PDL1 works in synergy with rapamycin immediate post-transplantation to change the balance of regulatory and effector responses in favor of sustained islet graft survival.

FIGURE 6.

Long-term survival of SA-PDL1-engineered islet grafts is associated with upregulated expression of immunoregulatory factors. Total RNA from unmodified islet grafts, unmodified islet grafts plus rapamycin (Rapa), SA-PDL1-engineered islet grafts (SA-PDL1), and SA-PDL1-engineered islet grafts plus rapamycin (SA-PDL1 + rapa) was extracted 3 days post-transplantation. The RNA was subjected to a TaqMan probe (FAM/VIC) based quantitative real-time RT-PCR using TaqMan primers to the indicated cytokines and transcriptional factors. Data were analyzed using DataAssist Software and plotted as fold change (2^−ΔΔCT) relative to GAPDH and unmodified islet graft only group. The experiments were performed in triplicate and repeated two times. Each data point is indicative of Mean ± SEM. *p < 0.05, **p <0.01, ***p <0.001 as assessed by one-way ANOVA with Tukey’s multiple comparison test.

CD4⁺Foxp3⁺ T Regulatory Cells Play an Important Role in Sustained Graft Survival.

We next assessed whether SA-PDL1-engineered allogeneic islets impact T cell frequencies in lymphoid organs early post-transplantation. Flow analysis of various T cell populations isolated from islet graft-draining lymph nodes on days 3 and 7 post-transplantation did not reveal significant differences in the absolute cell numbers (Fig. 7A) or percentages (Fig. S2), except that the ratio of CD4⁺CD25⁺FoxP3⁺ Treg to CD4⁺PD1⁺CD44^hiCD62L⁻ activated Teff cells was higher in the SA-PDL1-islets plus rapamycin group compared to the unmodified islets plus rapamycin group (p < 0.01) (Fig. 7B). A similar profile was also observed for splenic T cell subtypes, except that the increase in the ratio of CD4⁺CD25⁺FoxP3⁺ Treg to CD4⁺PD1⁺CD44^hiCD62L⁻ activated Teff cells was not significant on either day 3 or 7 post-transplantation (Fig. S3). Immunohistochemical analysis of long-surviving (> 100 days) grafts revealed dense peri-islet presence of CD4⁺FoxP3⁺ Treg cells occuring in patches (Fig. 7C).

FIGURE 7.

CD4⁺FoxP3⁺ Treg cells contribute to the maintenance of localized tolerance. T cells isolated from graft-draining lymph nodes of the indicated groups were analyzed on day 3 (A) or 7 (B) post-transplantation using flow cytometry with Abs to various cell surface markers. Absolute cell numbers for the indicated T cell populations are graphed. The ratio of CD4⁺CD25⁺FoxP3⁺ Treg to activated CD4⁺PD1⁺CD44^hi CD62L⁻ Teff cells on day 7 post-transplantation for SA-PDL1-engineered islets is significant when compared with the rapamycin control group, **p < 0.01 using unpaired t-test with Welch’s correction. Data (mean ± SEM) is representative of two independent experiments. (C) Long-term surviving SA-PDL1-engineered allogeneic islets show increased numbers of CD4⁺FoxP3⁺ Treg cells localized at the periphery of the graft. Tissue sections of kidney harboring long-term (> 100 days) islet grafts were stained with Abs against CD4 (red), Foxp3 (green), and insulin (blue) and analyzed using confocal microscopy (lower panel). Tissues stained with the same Abs, except an isotype Ab to Foxp3, served as control for Treg staining (upper panel). Circles indicate Treg cells in patches at the periphery of islet grafts. Upper right two panels are higher magnification of the different areas shown by circles. (D) Depletion of Treg cells results in rejection of long-term islet allografts. Streptozotocin diabetic C57BL/6.FoxP3^EGFP/DTR mice (n = 6) were transplanted with SA-PDL1-engineered islet grafts under the cover of a 15-day rapamycin regimen. Two out of 6 mice rejected within 33 days of transplantation (red solid lines). Four mice with sustained euglycemia were treated with diphtheria toxin (DT; 50 ng/gm) for two consecutive days on day 60 (n = 2, upward green arrows) and 80 (n = 2, upward blue arrows) post-transplantation. One animal in each treatment rejected (shown in respective broken lines) following DT treatment. One of the two euglycemic mice was treated again with DT (25 ng/gm two consecutive days; blue broken line, injection shown as downward grey arrows) 32 days post first treatment that resulted in graft rejection. The remaining one recipient (green unbroken line) maintained euglycemia for 140 days experimental end-point.

To assess the contribution of Treg cells to sustained graft survival, BALB/c allogeneic islets engineered with the SA-PDL1 protein were transplanted under the kidney capsule of chemically diabetic C57BL/6.FoxP3^EGFP/DTR recipients (n = 6) under the same rapamycin regimen. Two out of 6 recipients rejected their grafts on day 29 and 33, whereas the other 4 showed sustained euglycemia and were treated with two consecutive injections of diphtheria toxin on day 60 or 80 post-transplantation to deplete Treg cells (Fig. 7D). Two of 4 recipients rejected islet grafts following the treatment with diphtheria toxin, whereas the other 2 remained euglycemic despite significant depletion of Treg cells in the blood (Fig. S4). Treatment of one of these mice with diphtheria toxin for a second time resulted in prompt graft rejection, indicating insufficient depletion of Treg cells with one treatment in the target tissues, plausibly the islet graft. Taken together, these data signify the contribution of Treg cells to the observed long-term islet graft survival.

Discussion

The PD-1 pathway is an important immune checkpoint that regulates self-reactive responses and is extensively exploited by chronic infections and tumors for immune evasion (11, 34). These traits of PD-1 led to intensive efforts to modulate this pathway for therapeutic purposes. Indeed, PD-1 has been proven as an important therapeutic target in oncology with impressive clinical outcomes for various malignancies. However, the efficacy of this immune checkpoint for modulating alloreactive responses in transplantation has not been extensively explored, which is the subject of this manuscript. As a practical alternative to the DNA-based ectopic expression of immunomodulatory proteins, we previously reported the concept of generating recombinant chimeric immune ligands with a modified form of streptavidin and their transient display on biologic surfaces, including islets, for regulating immune responses with demonstrated efficacy in various transplantation and autoimmune settings (20, 21, 35-37). Using this scheme, we generated a functional form of mouse PD-L1 protein chimeric with streptavidin and assessed its efficacy when displayed on the surface of allogeneic islets in establishing sustained euglycemia in the absence of chronic immunosuppression in chemically diabetic mice.

SA-PDL1 exists as an oligomeric molecule with immunoregulatory functions. SA-PDL1 augmented the TGF-β-mediated conversion of CD4⁺ Tconv into induced CD4⁺Foxp3⁺ Treg cells and blocked the proliferative response of Teff cells to alloantigens, consistent with the reported regulatory functions of the native ligand (27-29). SA-PDL1 was efficiently displayed on the surface of pancreatic islets modified with biotin without a significant negative impact on the viability and insulin secretion function of islets. This observation is also consistent with our previous studies demonstrating that the transient display of a chimeric form of FasL, SA-FasL, on the surface of various cells, including hematopoietic stem cells, pancreatic islets, and cardiac graft vasculature did not negatively impact their viability or function (21, 31, 35, 36, 38, 39).

Transplantation of SA-PDL1-engineered islets into chemically diabetic allogeneic recipients resulted in sustained survival and function in the absence of chronic immunosuppression. This observation is consistent with several reports targeting the PD-1 pathway for modulation in the transplantation setting (14-17, 40). Pancreatic islet or solid organ grafts genetically modified to lack PD-L1 expression showed accelerated rejection following transplantation into allogeneic recipients (33, 40). Hematopoeitc stem cells genetically or pharmacologically modulated to overexpress PD-L1, but not wild cells with low levels of expression, were effective in reversing type 1 diabetes in NOD mouse model (41). Blockade of PD-L1 using an Ab resulted in the rejection of allogeneic fetus, implicating this pathway in fetomaternal tolerance (42). Our findings are also consistent with published studies reporting the efficacy of a dimeric form of PD-L1 and Ig fusion protein, PD-L1.Ig, when used in combination with CD40 blockade in prolonging cardiac allografts survival (15) and supporting sustained survival of islet allografts (16). Furthermore, infusions of ethylene carbodiimide-treated donor splenic antigen-presenting cells resulted in tolerance to allogeneic islets and the induction of tolereance required a synergistic action of the PD-1 pathway and Treg cells (43).

SA-PDL1-mediated long-term graft survival required a short course (15 days) of rapamycin as SA-PDL1-engineered grafts without rapamycin were rejected, albeit in a delayed fashion as compared with untreated controls. This observation is consistent with our reported findings using SA-FasL as an immunomodulator in different islet transplantation settings (18-20), and with those of others using a short course of rapamycin in conjunction with various immunomodulatory protocols in transplantation and autoimmunity settings (44-46). The mechanisms underlying the synergy between rapamycin and SA-PDL1 are presently unknown. Rapamycin has pleiotropic immune regulatory functions on cells of innate, adaptive, and regulatory immunity, and as such may contribute to the SA-PDL1-induced graft survival by various means (47-57) . Particularly, rapamycin has opposite effects on Teff and Treg cells; it blocks Teff proliferation by inhibiting p34cdc2 kinase (58), while facilitating the generation, expansion, and sustained survival of CD4⁺CD25⁺FoxP3⁺ Treg cells (53-57). PD-L1 converts Teff into Treg cells as shown in the present study and reported by others (59). The depletion of Treg cells resulted in acute rejection of SA-PDL1-engineered long-term islet grafts, thus implicating the generation of Treg cells as the potential basis of synergy between SA-PDL1 and rapamycin.

T cells from long-term graft recipients generated a proliferative response to donor antigens that was comparable to the response generated by naïve animals, demonstrating the lack of systemic tolerance. This finding is not surprising given the localized nature of immunomodulation by presenting SA-PDL1 on the surface of transplanted islets. Furthermore, the localized nature of graft protection is consistent with our previous studies with SA-FasL displayed on the surface of islets (20), poly(ethylene glycol) microgels co-transplanted with unmodified islets (19), or poly(lactic-co-glycolic acid) scaffolds housing unmodified islets (18). Real-time quantitative RT-PCR performed on day 3 post-transplantation showed significantly (p < 0.001) higher levels of intragraft transcripts for FoxP3, TGF-β, and IL-10 in the SA-PDL1-engineered islet grafts plus rapamycin group as compared with SA-PDL1-engineered islet grafts or unmodified islet grafts plus rapamycin. The TGF-β/Treg axis has been shown to play a dominant role in systemic tolerance to autoantigens. Treatment with an anti-CD3 Ab was shown to result in the prevention of experimental autoimmune encephalomyelitis through a TGF-β/Treg axis (60). Phagocytes engulfing apoptotic bodies produced by anti-CD3 Ab-mediated apoptosis of T cells were shown to produce TGF-β that in turn facilitates the generation of Treg cells. Similarly, the treatment of NOD mice with the anti-CD3 Ab was shown to prevent and reverse diabetes through a TGF-β/Treg axis (61). The direct display of SA-FasL on pancreatic islets or PEG microgels co-transplanted with unmodified islets were also shown to results in localized tolerance maintained by Treg cells (19, 20). Immunohistochemical analysis of long-term grafts revealed the peri-islet presence of CD4⁺FoxP3⁺ Treg cells occurring in patches. Depletion of Treg cells in long-term graft recipients resulted in the rejection of islet grafts confirming the dominant role of Treg cells in sustaining allograft survival.

The SA-PDL1 group with or without rapamycin treatment on day 3 post-transplantation also showed decreased levels of intragraft transcripts for IFN-γ and its transcriptional regulator T-bet as compared with the rapamycin group. Importantly, SA-PDL1 had a synergistic effect with rapamycin in significantly increasing the levels of intragraft transcripts for IDO and ARG-1, two canonical markers of M2 regulatory macrophages (62). Both SA-PDL1 and SA-PDL1 plus rapamycin groups had significant lower levels of transcripts for proinflammatory TNFα and IL-1β, two predominantly M1 cytokines (63). The predominant M2 response is also consistent with increased levels of transcripts for IL-4 and its transcriptional regulator GATA-3 in the SA-PDL1 plus rapamycin as compared with the other two groups. IL-4 induces the expression of ARG-1 and plays a critical role in polarizing macrophages towards an M2 phenotype (64). Also, co-transplantation of allogeneic islets with recipient autologous mesenchymal stem cells under the kidney capsule resulted in prolonged graft survival with immune responses skewed towards a Th2 profile (65). In addition to playing a critical role in the suppression of Teff cells and generation, homeostasis, and regulatory function of Treg cells (27, 29, 34, 66), PD-L1 has also been shown to polarize macrophages towards M2 regulatory phenotype both in mice and humans (67, 68) and block the generation and immune stimulatory function of M1 macrophage (69, 70). M1 macrophages have been shown to play a critical role in graft rejection by direct damage or through priming the adaptive immunity for a Th1 response (50-52). In marked contrast, M2 macrophages expressing Arg-1 and IDO prevent graft rejection by generating anti-inflammatory and regulatory factors and biasing adaptive immune responses towards a graft-protective regulatory response (49, 50).

The increased ~ 2 fold-expressionof NOS-2 in the combination and rapamycin group over the SA-PDL1 group is interesting as the role of NOS-2 in immune responses is contextual. Under inflammatory conditions, IFN-γ and LPS, macrophages upregulates NOS-2 and release high amount of nitric oxide (NO) that eventually leads to tissue damage, thus defining NOS-2 as M1 phenotype (71). However, NOS-2 gene product, inducible nitric oxide synthase (iNOS), expressed by myeloid derived suppressor cell (MDSC) is critical to their suppression function (72) . iNOS mediates T cell inhibition and downregulation of MHC class II molecules on APCs through phosphorylation of STAT5 and Janus Kinase 3 (73, 74). iNOS also blocks IFN-γ signaling pathway by nitration of STAT1 (72). Importantly, tolerance to kidney allografts was shown to associate with accumulation of MDSC expressing iNOS in the graft (75). Indeed, co-transplantation of iNOS^+/+, but not iNOS^−/−, MDSC with allogenic islets under the kidney capsule results in sustained graft survival (76).

Rapamycin directly binds to the mTORC1 component of mTOR and inhibits its activity, without a significant impact on the mTORC2 complex (77). Importantly, mTORC2 plays a major role in the polarization of macrophages to M2 and their function (48), whereas mTORC1 was shown to polarizes macrophages to M1 under select conditions (47). Unmodified islet grafts under the rapamycin treatment had significantly higher levels of intragraft transcripts for proinflammatory IL-1β, TNF-α, IFN-γ cytokines and T-bet transcriptional factor as compared with SA-PDL1 islet group with or without rapamycin. IFN-γ promote polarization of macrophages to M1 that express IL-1β and TNF-α in various infection settings (78). The observed expression of proinflammatory cytokines in the rapamycin group is consistent with study demonstrating that drug induces the expression of proinflammatory cytokines in innate immune cells (79-81). PD-L1 has also been shown to induce plasmacytoid DCs expressing IDO, which directly activates Treg cells and improves their suppressive function (82). Treg cells activated by IDO increases the expression of PD-L1 and PDL-2 on target DCs, further contributing to an immunosuppressive environment (82).

In conclusion, we demonstrate that the transient display of a novel form of PD-L1 protein, SA-PDL1, on the surface of islet grafts is effective in preventing rejection in allogeneic recipients when used in combination with a brief course of rapamycin. Long-term protection is localized to the graft and induced by a complex immunoregulatory loop involving polarization of macrophages to M2 and maintained by CD4⁺FoxP3⁺ Treg cells. The transient display of SA-PDL1 on the surface of islet grafts presents an attractive means of modulating alloreactive immune responses locally for sustained graft survival with significant translational potential.

Key points.

1. Islets are engineered with SA-PDL1 protein without impacting viability/function

2. SA-PDL1-enginered islets show indefinite survival in allogeneic hosts

3. Survival is associated with elevated intragraft Th2, Treg, and M2 transcripts

Abbreviations:

T1D

Type 1 diabetes

PD1

Programmed cell death-1

hCD2

human CD2

IPGTT

Intraperitoneal glucose tolerance test

SA

Streptavidin

SA-FasL

Fas ligand chimeric with streptavidin

SA-PDL1

PDL1 chimeric with streptavidin

MST

Median survival time

mTOR

Mechanistic target of rapamycin

GSIS

Glucose stimulated insulin secretion

Treg

Regulatory T cells

RT-qPCR

Real time-quantitative polymerase chain reaction