Immunological Shielding by Induced Recruitment of Regulatory T-Lymphocytes Delays Rejection of Islets Transplanted in Muscle

Abstract

The only clinically available curative treatment of type 1 diabetes mellitus is replacement of the pancreatic islets by allogeneic transplantation, which requires immunosuppressive therapies. Regimens used today are associated with serious adverse effects and impaired islet engraftment and function. The aim of the current study was to induce local immune privilege by accumulating immune-suppressive regulatory T-lymphocytes (Tregs) at the site of intramuscular islet transplantation to reduce the need of immunosuppressive therapy during engraftment. Islets were cotransplanted with a plasmid encoding the chemokine CCL22 into the muscle of MHC-mismatched mice, after which pCCL22 expression and leukocyte recruitment were studied in parallel with graft functionality. Myocyte pCCL22 expression and secretion resulted in local accumulation of Tregs. When islets were cotransplanted with pCCL22, significantly fewer effector T-lymphocytes were observed in close proximity to the islets, leading to delayed graft rejection. As a result, diabetic recipients cotransplanted with islets and pCCL22 intramuscularly became normoglycemic for 10 consecutive days, while grafts cotransplanted with control plasmid were rejected immediately, leaving recipients severely hyperglycemic. Here we propose a simple method to initially shield MHC-mismatched islets by the recruitment of endogenous Tregs during engraftment in order to improve early islet survival. Using this approach, the very high doses of systemic immunosuppression used initially following transplantation can thereby be avoided.

Introduction

The apparent drawback with all allogeneic transplantations is the required lifelong immunosuppressive therapy with associated systemic adverse effects. When treating type 1 diabetes mellitus patients with unstable glycemic control by transplanting insulin-producing pancreatic islets, the immunosuppressive therapy also negatively influences the graft directly by hampering islet graft function (30). Introduction of the Edmonton protocol, which excluded glucocorticoids, improved the success rate for islet transplantation. However, the remaining components of the currently used immunosuppressive therapy have antiangiogenic and prodiabetogenic effects that impede islet engraftment and long-term function (26,31).

Approaches to circumvent immunosuppressive therapy following islet transplantation include islet encapsulation (17) and induction of peripheral graft tolerance (19,23), which both have yet to be proven clinically feasible. The current study investigates a novel nonviral strategy to avoid the immunosuppressive therapy-coupled graft toxicity triggered by the high initial doses required early posttransplantation. This strategy includes local induction of endogenous immune privilege in muscle during islet engraftment and is based on the finding that tumors escape host antitumor immunity by producing high levels of the chemokine C-C motif ligand 22 (CCL22) to specifically recruit cluster of differentiation 4-positive (CD4⁺), CD25⁺, Forkhead box P3⁺ (CD4⁺CD25⁺FoxP3⁺) regulatory T-lymphocytes (Tregs) (9,20). Tregs play a key role in immune homeostasis due to their ability to suppress both activation and function of effector T-lymphocytes (14,20). This immune-modulating cell population is therefore crucial to control autoimmune reactions, prevent exaggerated responses to pathogens, and maintain a balance with the commensal microflora. Tregs express the C-C motif receptor type 4 (CCR4) and chemotax toward chemokine C-C motif 22 (CCL22) produced by macrophages or dendritic cells, and also by certain tumors (14,20). Induced overexpression of CCL22 by intraductal injection of viral vector carriers has been shown to delay diabetes onset in the NOD (nonobese diabetic) mouse model by increasing Treg mobilization and thereby reduce autoimmune islet destruction by suppressing effector T-lymphocytes (24). In the current study, DNA plasmids encoding CCL22 were chosen to avoid plasmid immunogenicity. The high translational threshold associated with viral vectors was reduced when using DNA plasmids (35).

Islets are traditionally transplanted intraportally, where they become scattered throughout the liver (12). Even though bioengineering of the liver is possible, a local effect is difficult to attain, and noninvasive high-resolution bioimaging of the scattered islet grafts is precluded. In contrast, striated muscle allows for noninvasive bioimaging (27) and tissue engineering before, during, and after islet implantation (13), in addition to reported high plasmid uptake (35). The intramuscular site was therefore preferred for the current investigation. Striated muscle was, in contrast to the liver, recently shown to promote prompt restoration of the intraislet capillary network and improved islet functionality in experimental models (8). Therefore, the clinical relevance for muscle as a site for islet transplantation needs to be further explored.

In this study, DNA plasmid transfer was used for the first time to curb rejection of transplanted pancreatic islets to muscle by inducing myocyte expression of the Tregrecruiting chemokine CCL22. This strategy created an immune-privileged site for transplanted islets that became shielded from adaptive immune recognition during engraftment in an allogeneic mouse model.

Materials and Methods

Animals

All experiments were approved by the Uppsala Laboratory Animal Ethical Committee. Mice, C57Bl/6 (H2^b) and Balb/c (H2^d) (Taconic, Ejby, Denmark), were monitored daily and had free access to pelleted food (Type R36; Lantmännen, Stockholm, Sweden) and tap water throughout the experiments. Animals were rendered diabetic by intravenous administration of alloxan, 75 mg/kg (Sigma-Aldrich, St. Louis, MO, USA). Plasma glucose levels were measured using a blood glucose monitor (Freestyle; Abbott, Alameda, CA, USA). Levels <11.1 and >16.7 mmol/L were considered normal and hyperglycemic, respectively.

Plasmid Construction and Expression Analysis

Plasmids were constructed on the pVAX1 backbone with cytomegalovirus promotor (Invitrogen, Waltham, MA, USA), and either insert -CCL22-porcine teschovirus-1 (P2A)-copGFP-thosea asigna virus 2A (T2A)-luciferase (Luc2)-, referred to as pCCL22, or -copGFP-T2A-Luc2-, referred to as pCTR, was introduced. Thus, pCTR plasmids encode GFP (green fluorescent protein) and luciferase, but not CCL22, while pCCL22 plasmids encode GFP, luciferase, and CCL22. Plasmids, 40 or 100 μg, were injected in the abdominal external oblique muscle in a volume of 200 μl saline. Transgene expression was measured over time based on luciferase activity following intraperitoneal injection of luciferin (150 mg/kg; PerkinElmer, Waltham, MA, USA) 10 min prior to anesthesia and image acquisition using a bioimaging device (IVIS Spectrum; PerkinElmer). Data was quantified using Living Image 3.1 software (PerkinElmer), and imaging parameters were maintained for comparative analysis (4). Settings were also maintained when selecting settings and the contralateral reference area was subtracted.

pCCL22 Protein Secretion

Abdominal external oblique muscles expressing pCTR or pCCL22 (40 μg) were excised 3 days postplasmid delivery. Tissues were manually homogenized in radioimmunoprecipitation assay buffer [Tris-HCl 50 mM, pH 7.4, Tritox X-100 1%, NaCl 150 mM, ethylenediaminetetraacetic acid 1 mM (20 μl buffer/mg muscle); Triton X-100 from Kebo, Stockholm, Sweden; the rest are from Sigma-Aldrich] with cOmplete mini protease inhibitor cocktail tablets (1 tablet/10 ml buffer; Roche Diagnostics, Basel, Switzerland). Total protein concentration was determined using DC Protein Assay (Bio-Rad, Hercules, CA, USA). Serum (400 μl) was collected from mice expressing pCTR or pCCL22 (40 and 100 μg) on day 3 or day 14. CCL22 protein levels were measured using the Quantikine mouse CCL22 immunoassay (R&D Systems, Minneapolis, MN, USA).

FoxP3 and GITR mRNA

Levels of the classic Treg marker FoxP3 (15) and Treg-associated marker glucocorticoid-induced tumor necrosis factor receptor (TNFR) family-related gene (GITR) (21,32) were measured in excised transfected muscle tissue 3 days postplasmid delivery. Tissues were snap frozen and homogenized in TRIzol (Invitrogen) according to the manufacturer's protocol. Total RNA concentrations and purity were determined using a Nanodrop 2000c spectrophotometer (Thermo-Scientific, Waltham, MA, USA). cDNA was synthesized using a Taqman High-Capacity cDNA kit (Applied Biosystems) and analyzed (MyIQ2 Real-Time PCR, Bio-Rad) using EvaGreen (primers; Bio-Rad) or Universal (probes; Applied Biosystems) master mix: FoxP3, Mm00475162_m1; GITR; forward: GCTGCATGCATTT TCTTCCTA, reverse: GAATGGCTGGGTCTCTCG; transcription initiation factor IIB (TFIIB); Mm01323562_g1, forward: TGGAGATTTGTCCACCATGA, reverse: GAA TTGCCAAACTCATCAAAACT. mRNA levels were normalized to TFIIB mRNA (ΔCt) (29).

Immunohistochemistry

Transfected muscles were excised and fixated in ice-cold methanol before permeabilization (0.05% Triton X). Optical clearing, after incubation with 10–15 μg antibodies per sample [anti-luciferase (rabbit anti-mouse polyclonal; Abcam, Cambridge, UK), anti-CD4 (clone GK1.5; R&D Systems), and anti-CD31 (clone 390; eBioscience, San Diego, CA, USA)], was performed by gradual liquid replacement to first ethanol and then to methyl salicylate (Sigma-Aldrich). For quantitative analysis, transfected muscle tissues were snap frozen in liquid nitrogen and sectioned (7 μm) before incubation with antibodies (0.5–1 μg) targeting the following antigens: CD4 (clone GK1.5), CD8 (clone 53–6.7), macrophage mannose receptor (MMR) (clone 310301) (all R&D Systems); CD25 (clone PC61.5), F4/80 (clone BM8), CD115 (clone AT598), CD31 (clone 390) (all from eBioscience); lymphocyte antigen 6G (Ly6G, clone 1A8; BD Bioscience, Franklin Lakes, NJ, USA); and FoxP3 (rabbit anti-mouse polyclonal; Thermo Scientific). In muscle tissues with islet grafts, every fifth section was stained for insulin (rabbit anti-mouse polyclonal; Abcam) and CD4 or CD8. Nuclei were stained with Hoechst 33345 (Invitrogen). Images were acquired using a laser scanning confocal microscope Nikon C1 on a TE2000-U base with Plan APO VC 20X/0.75 (Nikon, Tokyo, Japan) and analyzed using EZ-C1 software (Nikon) and ImageJ (NIH, Bethesda, MD, USA).

Single-Cell Suspension

Transfected muscle tissue was excised 3 days post-plasmid delivery. Single-cell suspension was prepared as described previously (37). Briefly, muscle tissues were minced and placed in Collagenase II solution, 500 U/ml (Sigma-Aldrich) in Hank's balanced salt solution (HBSS; Sigma-Aldrich), 0.250 mM CaCl₂ for 30 min in 37°C, 5% CO₂. Suspension was retrieved, washed in phosphate-buffered saline (EMD Millipore, Billeriva, MA, USA), and remaining tissue was placed in Collagenase D, 1.5 U/ml (Roche Diagnostics), Dispase II, 2.4 U/ml (Roche Diagnostics) in HBSS with 0.250 mM CaCl₂ for 60 min in 37°C, 5% CO₂. The homogenate was transferred through a 40-μm cell strainer filter (BD Bioscience). Finally, cells were fixated, permeabilized (Cytofix/CytoPermkit; BD Bioscience), and stained for lymphocyte markers CD4 (clone GK1.5), CD25 (clone PC61.5, both eBioscience), and FoxP3 (rabbit anti-mouse polyclonal, Thermo Scientific). Samples were analyzed by flow cytometry using FACSCalibur, four-color configuration, and CellQuest Pro software (BD Bioscience).

Pancreatic Islet Isolation, Culture, and Transplantation

Mouse islets were isolated using collagenase digestion and density gradient purification as described previously (5). Briefly, mice were killed by cervical dislocation, and ice-cold Collagenase A solution, 0.18 U/mg (Roche Diagnostics) in HBSS (Sigma-Aldrich), was injected into the pancreas via the common bile duct, after which the pancreas was removed and placed in 37°C for 18 min. Islets and exocrine tissue were separated by density gradient separation [Histopaque-1077 and Roswell Park Memorial Institute (RPMI)-1640; Sigma-Aldrich]. Single islets were kept in culture medium [RPMI-1640 (Gibco, Waltham, MA, USA) supplemented with D-glucose (11.1 mmol/L; Fresenius Kabi, Bad Homburg, Germany), l-glutamine (2 mmol/L), streptomycin (0.1 mg/ml), 10% (v/v) fetal calf serum (all from Sigma-Aldrich), and benzyl penicillin (100 U/ml; Roche Diagnostics)] overnight or for 4 days in 37°C, 5% CO₂. Islets (400 or 500) were then transplanted like pearls on a string between the fibers in the abdominal external oblique muscle of major histocompatibility complex (MHC)-mismatched diabetic recipients.

Statistics

Data are presented as mean ± SEM. Two groups were analyzed with Student's two tailed t test, and more than two groups were analyzed with one-way ANOVA and Bonferroni's multiple comparison test (GraphPad Prism 5, La Jolla, CA, USA); a value of p < 0.05 was considered statistically significant.

Results

Myocytes Express Injected Plasmid Genes

Following injection of plasmids into the abdominal external oblique muscle, the expression of the genes in the construct was visualized by longitudinal noninvasive in vivo bioimaging (Fig. 1A, D). The signal from the firefly luciferase (Luc2) was demonstrated to originate only from the muscle tissue as seen in the acutely excised muscle (Fig. 1B). Examination of the site of injection by immunohistochemical analysis confirmed the presence of the luciferase enzyme within the myocytes (Fig. 1C). Expression onset occurred within 2 h postplasmid delivery, confirming strong promoter sequence in the plasmid construct. Expression declined over 50 days, during which expression levels were significantly higher and peaked later (at day 10 vs. day 3, respectively) when using 100 μg plasmid compared to 40 μg plasmid (Fig. 1D). No signal was detected from mice injected with plasmids not containing the Luc2 insert (data not shown).

Figure 1.

Intramuscular plasmid expression and pCCL22 production and secretion. Noninvasive detection of transgene expression by bioimaging (IVIS Spectrum) (A) with the luminescent signal originating from the intramuscular site of plasmid delivery shown in excised abdominal wall muscle (B). The white line indicates the position of linea alba. (C) Plasmid-encoded luciferase is seen in blue (Alexa Fluor 405) and showed distribution within the myocytes of transfected striated muscle tissue, where muscle vasculature (CD31) is seen in green (Alexa Fluor 488). Image acquired by laser scanning confocal microscopy; scale bar: 100 μm, z = 60 μm. (D) Plasmid expression over time using 40 μg (n = 28) and 100 μg (n = 15) plasmid detected by noninvasive luminescence bioimaging. (E) Levels of CCL22 in homogenate of control plasmid (pCTR) (n = 5) or pCCL22 (n = 5) transfected muscle 3 days postplasmid delivery (d post m. transf.) using 40 μg plasmid. (F) Circulatory levels of CCL22 in untreated mice (n = 5) or mice transfected with 40 μg pCTR (n = 5) or pCCL22 3 days (n = 5) and 14 days (n = 5) postplasmid delivery and circulatory levels of CCL22 in mice transfected with 100 μg pCCL22 3 days (n = 5) and 14 days (n = 5) postplasmid delivery. #Statistically significant difference from the group receiving 40 μg plasmid (D). *Statistically significant difference from the pCTR group (E). #Statistically significant difference from the pCCL22, 100 μg, 14 days group (F). p < 0.05, Student's t test (D, E); one-way ANOVA with Bonferroni's multiple comparison test (F).

Plasmid-Encoded CCL22 Protein Is Secreted by Transfected Muscle

To confirm that plasmid-encoded CCL22 protein was produced at the site of plasmid administration and transfection, levels of CCL22 in homogenates of muscle receiving pCTR (plasmid-encoding reporter but no chemokine) or pCCL22 were measured 3 days postplasmid delivery (40 μg). While very low levels of CCL22 were detected in muscles treated with pCTR, significantly increased levels of the chemokine were measured in muscles expressing pCCL22 (Fig. 1E). Circulating levels of CCL22 were measured in serum from untreated, pCTR-, and pCCL22-treated mice, where levels were slightly elevated only after 14 days using 100 μg pCCL22 (Fig. 1F). These results indicate that plasmid-delivered CCL22 is produced and exerts predominantly local effect using doses of both 40 and 100 μg plasmid.

Myocyte Expression of CCL22 Induces Local Treg Recruitment to Muscle

Recruitment of Tregs to the pCCL22-expressing muscle was investigated by measuring mRNA levels of the Treg-specific markers FoxP3 and GITR. Indeed, significantly increased mRNA expressions of both markers were detected in muscle 3 days post-pCCL22 delivery (40 μg) (Fig. 2A, B), compared to pCTR-treated muscles, confirming recruitment of Tregs by CCL22. Immunohistochemistry of transfected muscles showed accumulation of CD4⁺ cells in pCCL22- but never in pCTR-transfected muscles (Fig. 2C). To determine the identity of recruited cells, quantitative analyses of single-cell suspensions from transfected muscles were performed 3 days postplasmid delivery (40 μg). Using this approach, a resident muscle population of CD4⁺CD25⁺ T-lymphocytes was identified (Fig. 2D), of which some were also FoxP3⁺ (Fig. 2E), demonstrating the Treg population. Quantifications revealed a significant increase in CD4⁺CD25⁺ (Fig. 2F) and CD4⁺CD25⁺FoxP3⁺ (Fig. 2G) T-lymphocytes in pCCL22-expressing muscle as well as a dramatic alteration of the distribution of the populations where 39% were FoxP3⁺ in the pCCL22-expressing muscles compared to only 10% in the pCTR-expressing muscles (Fig. 2E).

Figure 2.

Myocyte expression of CCL22 induces local Treg recruitment to muscle 3 days postintramuscular pCCL22 plasmid delivery. (A, B) mRNA levels of the regulatory T-cell (Treg)-associated genes forkhead box P3 (FoxP3) and glucocorticoid-induced TNFR family-related gene (GITR), respectively, in pCTR (n = 11) or pCCL22 (40 μg, n = 9) transfected muscle. (C) Recruitment of CD4⁺ T-lymphocytes to pCCL22, but not to pCTR-expressing muscles, visualized by confocal microscopy 3 days postplasmid delivery. Scale bars: 100 μm, z = 100 μm. (D, E) Flow cytometry analysis of CD4⁺CD25⁺ T-lymphocytes (D) and their FoxP3⁺ expression (E) in single-cell suspensions of pCTR (n = 5) or pCCL22 (n = 4)-expressing muscles (40 μg, 3 days postplasmid delivery). FSC, forward scatter. (F, G) Quantifications of CD4⁺CD25⁺ (F) and CD4⁺CD25⁺FoxP3⁺ (G) T-lymphocytes per milligram muscle tissue. *Statistically significant difference from the pCTR group (p < 0.05, Student's t test).

A wider analysis of leukocyte recruitment to pCCL22 was then performed using immunohistochemistry 3 days postplasmid delivery (40 μg). Again, accumulation of CD4⁺ cells was observed in pCCL22-expressing muscles (Fig. 3A, B), while the inflammatory (F4/80⁺MMR^-) or alternatively activated (F4/80⁺MMR⁺) macrophage populations were not altered by pCCL22 expression in muscle (Fig. 3C, E). CD8⁺ T-lymphocytes, neutrophils (Ly6G⁺), or monocytes (CD115⁺) were only detected in very few numbers in muscle irrespective of treatment (Fig. 3F–H).

Figure 3.

Leukocytes present in transfected muscle 3 days postintramuscular delivery of 40 μg plasmid. Analysis of intramuscular leukocyte populations at the site of pCTR (n = 5) or pCCL22 (n = 5) injection 3 days postplasmid delivery using immunohistochemistry in tissue sections (7 μm) where leukocytes expressing CD4⁺ (A, B), F4/80⁺MMR^- (inflammatory macrophages), F4/80⁺MMR⁺ (alternatively activated macrophages) (C, D, E), CD8⁺ (F), lymphocyte antigen 6G positive (Ly6G⁺) (G), and CD115⁺ (H) were analyzed. Representative images were acquired by laser scanning confocal microscopy. Scale bar: 100 μm. *Statistically significant difference from the pCTR group (p < 0.05, Student's t test).

pCCL22 Delays Pancreatic Islet Graft Rejection

The effect of pCCL22 on suppressing effector T-lymphocytes and concomitant islet graft rejection was investigated in alloxan-induced diabetic recipients using MHC-mismatched islet transplantations into the abdominal external oblique muscle (Figs. 4, 5). Islets isolated from C57Bl/6 donors were transplanted to Balb/c recipients. The reverse transplantation setting (Balb/c islets to C57Bl/6 recipients) elicited substantial initial recruitment of phagocytes to the islets using no plasmids (data not shown), as described previously (16), which obstructs studies of lymphocyte trafficking in the system. Allogeneic intramuscular transplantation of acutely isolated islets or islets precultured for 4 days, using no or control (pCTR) plasmid, completely failed to reduce hyperglycemia in diabetic recipients (Figs. 4A, 5A). In contrast, transplantation of 500 precultured islets together with pCCL22-expressing plasmids resulted in normoglycemia for 10 consecutive days before the mice returned to hyperglycemia (Fig. 4A). Leukocyte infiltration within and surrounding islets was substantially decreased 7 days postislet transplantation when cotransplanted with pCCL22, but not pCTR (Fig. 4B). Significantly larger insulin-positive areas were detected 7 days, but not 14 days, following cotransplantation with pCCL22 compared to pCTR (Fig. 4C). Further, the proportion of transplanted islets with central necrosis was strongly reduced when cotransplanted with pCCL22 compared to with pCTR (7 days, Fig. 4D). Improved engraftment 7 days posttransplantation was confirmed for pCCL22 cotransplanted islets (Fig. 4E), since intraislet vascular densities had returned to levels previously reported in syngeneic intramuscular islet transplantation (8). No differences in total number of CD4⁺ T-cells present at the graft were detected at the different time points following islet transplantation between pCTR-transfected or pCCL22-transfected muscles (Fig. 4F). However, the CD8⁺ T-cell population present at the site of engraftment (Fig. 4G) or in direct contact with the islets (Fig. 4H) was strongly reduced in the CCL22-expressing muscle at 7 days, but not at 14 days, posttransplantation. The ratio of CD8/CD4 T-cells was four times higher for the pCTR-treated compared to pCCL22-treated muscles 7 days following islet transplantation (0.61 vs. 0.16, respectively), demonstrating a shift in the graft microenvironment toward CD4⁺ T-cells by pCCL22 expression, suggesting a less rejection-prone microenvironment.

Figure 4.

pCCL22 delays pancreatic islet graft rejection. Islets (C57Bl/6) were transplanted to the muscle of diabetic MHC-mismatched recipients (Balb/c) together with 100 μg pCTR or pCCL22 plasmids. (A) Blood glucose levels following intramuscular transplantation of 500 precultured islets together with either 100 μg pCTR (n = 8) or pCCL22 (n = 11). (B) Representative images from sections (7 μm) of islet grafts [Insulin-A, Alexa Fluor 488 (green)] in muscle 7 days posttransplantation showing cell infiltration [Hoechst (blue)], CD8⁺ T-lymphocytes [Northern Light 557 (red)] (B, top) and CD31⁺ vasculature [Alexa Fluor 555 (red)] (B, bottom). Scale bar: 100 μm. Quantification of insulin-positive area in grafts of precultured islets cotransplanted with 100 μg plasmid 7 days and 14 days posttransplantation (C) and examination of islet central necrosis 7 days posttransplantation (D). Engraftment assessed by intraislet vascular density 7 days posttransplantation (E). Quantification of T-lymphocyte density, CD4⁺ (F) and CD8⁺ (G) present at the graft and the number of CD8⁺ T-lymphocytes in direct contact with the transplanted islet (H) 7 days and 14 days posttransplantation. *Statistically significant difference from the pCTR group (p < 0.05, one-way ANOVA with Bonferroni's multiple comparison test).

When 400 acutely isolated islets were transplanted into pCCL22-expressing muscle (pretreated 3 days earlier with 40 μg plasmid) of diabetic recipients, blood glucose levels decreased to about 15 mmol/L for 6 days before returning to severe hyperglycemia (Fig. 5A). Abundant leukocyte infiltration at the graft was detected at 7 days posttransplantation (Fig. 5B), and there were no differences in insulin-positive area at 7 or 14 days posttransplantation in muscles pretreated with pCTR or pCCL22 (Fig. 5C). These observations indicate insufficient graft shielding using the lower plasmid dose, resulting in graft destruction during the first week. However, despite the inadequate immune privilege, suppressive effects by the pCCL22 treatment were uncovered 7 days posttransplantation, since significantly fewer CD8⁺ T-lymphocytes were found around (Fig. 5D) and in direct contact with the graft (Fig. 5E) in pCCL22-treated grafts, while no differences in the CD4⁺ T-lymphocyte populations present at the grafts were observed (Fig. 5F).

Figure 5.

Insufficient shielding of islet grafts transplanted to striated muscle of diabetic MHC-mismatched recipients pretreated with plasmids. (A) Blood glucose levels following transplantation of 400 acutely isolated islets to abdominal wall muscle expressing pCTR (n = 12) or pCCL22 (40 μg, n = 11). (B) Representative images from sections (7 μm) of islet grafts [Insulin-A, Alexa Fluor 488 (green)] in muscle 7 days posttransplantation to pCTR-expressing muscle where F4/80⁺, Ly6G⁺, CD4⁺, and CD8⁺ [all Northern Light 557 (red)] leukocytes were present at the graft at this time point. Scale bar: 100 μm. Quantification of insulin-positive area in grafts of acutely isolated islets transplanted to pretreated with pCTR and pCCL22 7 days and 14 days posttransplantation (C). Assessment of T-lymphocyte density, CD8⁺ (D) and CD4⁺ (F) present at the graft and the number of CD8⁺ T-lymphocytes in direct contact with the transplanted islet (E) 7 days posttransplantation to either pCTR (n = 5) or pCCL22 (n = 5) expressing muscle. *Statistically significant difference from the pCTR group (p < 0.05, Student's t test).

Acutely Isolated Islets Trigger an Earlier Graft Rejection

For safety and logistical reasons, isolated human islets are cultured prior to clinical transplantation for an average of 4 days, which has been suggested to decrease islet immunogenicity (18). To challenge this in the current experimental setup, acutely isolated islets were cotransplanted with pCCL22 (100 μg) to diabetic recipients. These islets were rejected already by 6 days posttransplantation (Fig. 6A). Also, only 50% of the mice receiving acutely isolated islets together with 100 μg pCCL22 transiently reversed the diabetes following transplantation compared to all the mice where precultured islets were cotransplanted with 100 μg pCCL22 (Fig. 6B). These results indicate that the potential immunogenicity of noncultured islets overrides the immunosuppressing activities of recruited Tregs.

Figure 6.

Acutely isolated islets trigger earlier graft rejection when cotransplanted intramuscularly with pCCL22 plasmid. Islets (C57Bl/6) were transplanted to muscle of diabetic MHC-mismatched recipients (Balb/c). (A) Blood glucose levels following intramuscular transplantation of 500 acutely isolated islets cotransplanted together with pCCL22 (100 μg, n = 4). (B) The percentage of mice remaining diabetic post-islet transplantation where 500 precultured islets were cotransplanted with pCTR (n = 8) or pCCL22 (100 μg, n = 5) or where 500 acutely isolated islets were cotransplanted with pCCL22 (100 μg, n = 4). Blood glucose level <11.1 mmol/L for 3 consecutive days were considered as diabetes reversal.

Antigen Exposure Impairs the pCCL22-Induced Immune Privilege and Enhances Islet Rejection

To investigate if the immunological shielding induced by overexpression of CCL22 would be impaired by exposure of islets from the same donors to the adaptive immunity, 100 precultured islets were transplanted to untreated contralateral abdominal muscle simultaneously as 500 precultured islets were cotransplanted with pCCL22 (Fig. 7). This additional graft disrupted the normoglycemia at 6 days, compared to the 10 consecutive normoglycemic days observed in mice cotreated with pCCL22 and 500 precultured islets (Fig. 4A), indicating that antigen exposure accelerates graft recognition and destruction and interrupts the immune privilege created by pCCL22 over-expression. Thus, cotransplantation of islets and DNA plasmids, as opposed to only pretreating the engraftment site is preferred not only from the logistic point of view but also to ensure the full coverage of transplanted islets by the bioengineered endogenous immunosuppression during engraftment.

Figure 7.

Antigen exposure impairs the pCCL22-induced immune privilege and enhances islet rejection. Blood glucose levels following intramuscular allogeneic transplantation of 500 precultured islets cotransplanted with pCCL22 (100 μg) in addition to a contralateral graft of 100 islets transplanted with no plasmids (n = 5). ¤Statistically significant difference from the group transplanted with 500 precultured islets cotransplanted with pCCL22 in Figure 4A (p < 0.05, one-way ANOVA with Bonferroni's multiple comparison test).

Discussion

This study demonstrates means to circumvent immunosuppressive therapy at the time of islet engraftment following allogeneic islet transplantation to muscle. A local immune privilege was constructed by inducing myocyte expression and secretion of the Treg-recruiting chemokine CCL22 by intramuscular injection of pCCL22-encoding DNA plasmids. Thus, when isolated pancreatic islets were allogeneically transplanted together with pCCL22 plasmids to diabetic recipients, the grafts were protected from early CD8⁺ T-cell-driven rejection by recruited Tregs, and the recipients became normoglycemic for 10 consecutive days.

To improve patient safety and islet survival, several ongoing investigations explore the possibilities to elude or decrease immunosuppressive therapies following allogeneic transplantations (31). Means to avoid general immune suppression and to specifically prevent graft recognition are being addressed using different strategies including immune-modulating (19) or cellular therapies (11) and islet encapsulation (25). One strategy involves cotransplantation of islets with immune-regulating cells (myeloid-derived suppressor cells) and this was demonstrated to protect against graft rejection (7). Leukocytes from both myeloid and lymphoid origin have been reported to exert tolerogenic immune responses and thereby promote graft survival (36). The most well-studied immuneregulating immune cells are CD4⁺CD25⁺FoxP3⁺ regulatory T-lymphocytes. Cellular therapies using in vitro expanded Tregs were recently shown to prevent islet graft rejection in xenotransplanted mice (38). However, Treg cell transfer therapy is still associated with important concerns and logistic problems, that is, few Tregs are found in circulation, lack of stable Treg surface markers impede discrimination from other T-lymphocytes, and possible implications of in vitro expansion on in vivo suppressor effects (33,34). These issues highlight the limitations for lymphocyte characterization through detection of protein markers as well as the importance of functional lymphocyte studies. In addition, systemic infusion of a large number, as opposed to local accumulation of Tregs, causes general immune suppression, as do current immunosuppressive regimens, even though human studies of systemic Treg administration have so far not revealed increased susceptibility to infections (10).

To avoid systemic effects, induction of local immunosuppression is an attractive approach. In the current study this was achieved by induced local expression by DNA plasmid gene transfer of the chemokine CCL22, previously reported to be chemotactic for Tregs (9,20). Myocytes at the site of injection were demonstrated to express and secrete plasmids shortly after delivery, and the induced chemokine expression resulted in local accumulation of Tregs. By delivering a highenough dose of plasmids, a functional immune privilege was transiently constructed at the site of islet transplantation, without additional means to amplify uptake, such as electroporation. When insulin-producing allogeneic islets were transplanted together with the pCCL22 plasmids, acute graft rejection was prevented, and recruitment of CD8⁺ T-cells was strongly counteracted. This resulted in reduced immune-derived graft destruction early after transplantation with preserved insulin-positive islets. The ability of these islets to engraft, revascularize, and regulate glucose homeostasis was confirmed with respect to increased intraislet vessel densities and normoglycemia for 10 consecutive days. DNA plasmid gene transfer was employed in the current study based on the clinically attractive arguments of low immunogenic effects, low costs, and transient expression since the plasmids are not incorporated into the somatic DNA (35). Whether repeated plasmid delivery or increased dosage could expand the period of suppressed graft recognition remains to be investigated. However, the goal of the current study, to allow for islet engraftment prior to introduction of immunosuppressing therapies, was achieved.

The muscle is a promising site for islet transplantation. Preclinical data demonstrated improved islet revascularization and function following syngeneic transplantation to muscle, results that were confirmed in the same study in three patients (8), and also in a case report of a girl receiving islet transplantation to the forearm muscle (28). The suitability of the muscle as a site for allogeneic islet transplantation is currently being evaluated in an ongoing clinical study. In the current study, diabetic mice receiving intramuscularly transplanted islets together with pCTR remained hyperglycemic. This is in contrast to what is reported for experimental allogeneic islet transplantation to the kidney subcapsular space of diabetic recipients, which resulted in normoglycemia for a brief period shortly after the surgical procedure (3). This discrepancy can be the result of the different degree of insult between the techniques, and indeed mice receiving intramuscular islet transplantation were observed to be fully mobile and to start eating directly following anesthesia. An alternative explanation would be that a faster rejection process is initiated for muscle-transplanted allogeneic grafts compared to allogeneic islets transplanted intrahepatically and to the kidney subcapsular space (22). However, the muscle provides another advantage compared to the liver in regard to being a site for islet transplantation, since it enables easy modification by tissue engineering, which can be visualized longitudinally and noninvasively due to its accessibility (13,27). Intraportal injection of plasmids leads to systemic dissemination in addition to transfected liver cells scattered throughout the organ (35), which precludes high-resolution monitoring of expression levels (12). When designing the present study, considerations were made to allow transfer of results to the clinic in regard to both patient safety and the actual logistics during clinical islet isolation and transplantation. Gene transfer by DNA plasmids were chosen before viral vectors, which may induce sustained gene expression, but are not easily transferred into the clinic (6). However, viral vectors are widely used in experimental settings. Induced expression of CCL22 by adeno-associated viral vectors injected intraductally into young NOD mice was recently shown to postpone development of onset of autoimmune diabetes for up to 160 days. Using the NOD mouse model of diabetes, prolonged resistance to autoimmune destruction was demonstrated for virally transfected CCL22-expressing islets transplanted to the kidney subcapsular space, demonstrating a role of CCL22 also in the onset of disease (24). Viral vectors have also been used to curb islet rejection and prolong allograft survival following transduction of islets to induce expression of tryptophan-degrading enzymes that prevent T-cell proliferation (1,2,17). Further, codelivery of plasmids with islets was compared to pretreatment of the site for islet transplantation, the former being preferred due to less intervention for the recipient. In addition, cotransplantation of islets and plasmids lowers the risk of transplanting islets to areas outside the induced immune privilege, which was confirmed in this study to accelerate graft recognition and destruction of the pCCL22 cotransplanted islets. Owing to the strong promoter used in the current study and the rapid onset of CCL22 expression and secretion during the first hours following administration, cotransplantation of plasmids and islets prevented graft rejection and rendered the diabetic recipient normoglycemic for a period of 10 consecutive days.

In conclusion, this study applied simple bioengineering of the intramuscular transplantation site to induce an immune-privileged microenvironment that promoted islet engraftment and enabled graft function while curbing early rejection without the need for systemic immunosuppression. Owing to direct islet toxicity and antiangiogenic effects associated with immunosuppressive drugs, a period free from these pharmaceuticals during islet engraftment would most certainly improve islet survival and long-term function as well as increase patient safety following transplantation, all of which is necessary to establish islet transplantation as a functional treatment available for more people with diabetes.