Local Expression of Indoleamine 2,3 Dioxygenase in Syngeneic Fibroblasts Significantly Prolongs Survival of an Engineered Three-Dimensional Islet Allograft

Islets were isolated from 6- to 8-week-old male BALB/c mice (The Jackson Laboratories, Bar Harbor, ME), and mouse dermal fibroblasts were explanted from B6 mice skin and transduced with a recombinant adenoviral vector carrying human IDO cDNA as previously described (26). Control fibroblasts were infected with a mock vector or left untreated. Fibroblast-populated collagen gel (FPCG) matrices were prepared as described by Sarkhosh et al. (23) using IDO-expressing or control fibroblasts. Mouse islets were added to FPCG before solidification in 24-well plates. Care and maintenance of all animals were in accordance with the principals of laboratory animal care and the guidelines of the institutional Animal Policy and Welfare Committee.

Recipient C57BL/6 (B6) mice were rendered diabetic by a single intraperitoneal injection of 200 mg/kg streptozotocin (Sigma), and diabetes was defined as a minimum of two consecutive blood glucose measurements ≥20 mmol/l. Islet plus IDO-expressing or control fibroblast composite grafts (∼500 islets) were transplanted under the left kidney capsule of isoflurane-anesthetized diabetic mice. After transplantation, blood glucose levels were measured using an Accu-Chek Compact Plus blood glucose monitoring system, and grafts were deemed functioning when blood glucose levels decreased to <10 mmol/l. Intraperitoneal glucose tolerance tests (IPGTT) were performed as previously described (27) 2 and 4 weeks after transplantation. All animals were cared for according to the guidelines of the institutional Animal Policy and Welfare Committee.

Graft-bearing kidneys or draining lymph nodes were harvested at indicated time points, fixed in 10% buffered formalin solution, and embedded in paraffin. Sections 5-μm thick from graft area were stained with hematoxylin-eosin (H-E). Immunofluorescence staining was performed for insulin, CD3, FOXP3, and IDO. Sections were rehydrated, and nonspecific binding was blocked. Sections were incubated overnight at 4°C with appropriate primary antibodies and then washed and incubated with relevant secondary antibodies for 45 min. Primary antibodies include guinea pig anti-insulin antibody (1:500 dilution; Dako Laboratories, Mississauga, Ontario, Canada), rabbit anti-CD3 antibody (1:100 dilution; abcam, Cambridge, MA), rat anti-mouse FOXP3 (1:200 dilution; eBioscience, San Diego, CA), and rabbit anti-IDO antibody (1:1,000 dilution, raised in rabbit; Washington Biotechnology, Baltimore, MD). Secondary antibodies (1:2000 dilution) used were fluoroisothiocyanate (FITC) goat anti-guinea pig IgG (abcam), rhodamine goat anti-rabbit IgG (Chemicon International, Temecula, CA), and rhodamine goat anti-rat IgG (Jackson ImmunoResearch, West Grove, PA).

Composite grafts were harvested at indicated time points, and collagen matrices were digested using type I collagenase (1 mg/ml; Sigma) at 37°C for 10 min. Cell suspensions were washed with PBS and passed through a 40-μm cell strainer. Cells were then incubated for 30 min at 4°C with fluorescent-conjugated antibodies (1:100 dilutions; eBioscience) specific for particular lymphocyte and antigen-presenting cell (APC) markers. Fluorescence dot plots were created using a BD FACS Calibur flow cytometry machine (BD Biosciences Pharmingen, Mississauga, Ontario, Canada) and were used to determine the percentage of positive cells labeled with the corresponding antibodies.

Total RNA was isolated from harvested grafts at indicated time points using an RNeasy kit (Qiagen). cDNA was synthesized using the SuperScript first-strand synthesis system for RT-PCR (Invitrogen, Carlsbad, CA) according to the manufacturer's protocol. Quantitative RT-PCR was performed in a reaction mixture of 25 μl using the Platinum SYBR Green qPCR SuperMix-UDG with ROX (Invitrogen) PCR kit according to the manufacturer's instructions. The reactions were run on a 7900 HT Fast Real time PCR system (Applied biosystems, Foster City, CA). Human IDO-specific and mouse cytokine and chemokine mRNAs were quantified relative to mouse glyceraldehyde-3-phosphate dehydrogenase (GAPDH) using primers listed in supplementary Table 1 (available in an online appendix [http://diabetes.diabetesjournals.org/cgi/content/full/db09-1560/DC1]). The analyses were performed with the Sequence Detection system 7900HT software (version 2.3; Applied biosystems). All measurements were performed in duplicate. Results are expressed as the percentage or fold increase in expression of mRNA of interest at the indicated time point compared with day one posttransplantation.

Blood serum samples were collected from composite graft recipients B6 mice (H-2b) at indicated time points posttransplantation. Donor BALB/c mice (H-2d) thymocytes (1 × 10⁶) were incubated with recipient B6 mouse serum at 1:128 dilution for 1 h at 37°C. Thymocytes were then washed with PBS and incubated with FITC-conjugated goat anti-mouse IgG1 or IgG2c (1:200 dilution; AbD Serotec, Raleigh, NC) for 1 h at 4°C. Binding of cells to antibody was detected from single-parameter fluorescence histograms using a BD FACSCanto flow cytometer (BD Biosciences) after gating on viable cells. Percentage of cells shifting into an appositive region on fluorescence histogram was calculated.

Stimulator cells were splenocytes from islet donor (BALB/c; H-2d) or third-party (C3H; H-2K) mouse strains, and responder cells were lymphocytes from graft-draining (renal) lymph nodes of graft-recipient mice (B6; H-2b). Responder cells (2 × 10⁵) were cultured with irradiated (2000 rad) stimulator cells in 96-well round-bottom microtiter plates (Costar, Corning, NY) for 96 h. Individual wells were pulsed with 1 μCi/well of [³H]-thymidine for the last 18 h. Culture plates were then harvested, and [³H]-thymidine incorporation was measured using a Wallac Jet 1450 Microbeta scintillation counter (Perkin Elmer, Waltham, MA). Mean counts per minute were determined from cultures set up in triplicate.

Data are means ± SEM of three or more independent set of experiments. Survival of islet grafts in IDO-expressing versus control groups was compared using the Kaplan-Meier log-rank test. The statistical differences of mean values among treated and control groups were tested with one-way ANOVA. Post hoc comparisons were done using Student's t test with Bonferroni correction for multiple comparisons. P values <0.05 were considered statistically significant.

To investigate the local immunosuppressive effect of IDO, three-dimensional grafts were engineered by embedding 500 BALB/c mouse islets within collagen matrix populated with adenoviral-transduced IDO-expressing or control (mock vector infected or untreated) B6 mouse fibroblasts. IDO overexpression was validated in composite grafts (supplementary Fig. 1). These composite grafts where then transplanted to renal subcapsular space of streptozotocin-induced diabetic immune-competent B6 mice. Another control group of mice received only islets. Islet graft function was checked by measuring blood glucose in graft-recipient mice. Composite IDO-expressing grafts showed a significant prolongation of graft survival (41.2 ± 1.64 days; P < 0.001; n = 10 (Table 1 and Fig. 1,A). In contrast, as shown in Table 1 and Fig. 1,A, control grafts were rejected within 2 weeks after transplantation. Mean duration of graft survival in islet-alone group and grafts with untreated and mock virus–treated fibroblast were 12.9 ± 0.73, 13.5 ± 0.79, and 13.2 ± 0.61 days, respectively (Table 1). Mean survival duration for mock vector–infected composite grafts was not significantly different from those of grafts with untreated fibroblasts and islet-alone grafts, which confirmed that adenoviral infection by itself had no effect on graft survival.

IPGTT was performed on mice that received IDO-expressing composite grafts after 2 and 4 weeks posttransplantation. As shown in Fig. 1 B and C, glucose clearance rates in mice received IDO-expressing grafts were comparable with age-matched naïve animals after 2 weeks (area under the curve 1,671.0 ± 289.1 vs. 1,700.1 ± 300.3 mmol · l⁻¹ · min, respectively; P > 0.05) and 4 weeks (1,870.7 ± 327.8 vs. 1,824.2 ± 325.4 mmol · l⁻¹ · min; P > 0.05) posttransplantation. Normal response of islet grafts to IPGTT demonstrated that islets in IDO-expressing composite grafts are able to function normally in response to glucose load, suggesting preservation of islet mass. These data collectively confirm that local IDO expression significantly prolongs islet allograft survival.

A set of graft recipient mice were killed at the end of each week posttransplantation to examine histopathological changes in islet allografts. Composite grafts were then recovered and stained with H-E or subjected to immunofluorescence staining for insulin, CD3, or FOXP3. Histological studies demonstrated that islets architecture was well preserved in the IDO-expressing composite grafts for up to 6 weeks (Fig. 2,G–L), whereas islets in the untreated (Fig. 2,A–C) or mock vector–infected grafts (Fig. 2,D–F) were almost completely destroyed before the third week posttransplantation. Control grafts were diffusely infiltrated with mononuclear cells as early as the second week posttransplantation (Fig. 2,B and E), while IDO-expressing grafts remained intact for up to 6 weeks (Fig. 2 G–K). To investigate whether a component of immune response generated against adenovirus-infected syngeneic fibroblasts and/or human transgenic IDO, islet-free IDO-expressing or mock vector fibroblast–populated collagen matrices were transplanted to B6 mice. T-cell infiltration was not detected in IDO-expressing grafts after 40 days posttransplant, whereas IDO transgene was still expressed (supplementary Fig. 2).

Double staining of grafts for CD3 and insulin showed very few insulin-producing cells and dense T-cell infiltration in both untreated and mock vector–infected controls at the end of the second week posttransplantation (Fig. 3,A). In contrast, islets in IDO-expressing grafts were strongly stained for insulin and maintained their normal architecture with minimal infiltration of T-cells at the same time point (two weeks posttransplant; Fig. 3,A). A remarkable finding was that T-cells densely accumulated at the interface between IDO-expressing graft and kidney tissue but did not penetrate into the composite graft (IDO 2 weeks posttransplant [Figs. 2,I and 3,A]). However, T-cells started to infiltrate IDO vector–infected grafts by the end of the fifth week posttransplant (Fig. 3A). While FOXP3 immunostaining revealed very few intragraft FOXP3⁺ cells in all experimental groups (Fig. 3,B), more FOXP3⁺ cells were present in graft-draining lymph nodes of IDO-expressing recipients at week two but not week five posttransplant (Fig. 3 C).

Flow cytometry was used to determine the specific T-cell and APC composition of the leukocyte infiltrates to the grafts. As shown in Fig. 4,A, CD4⁺ and CD8⁺ T-cells were much more abundant in untreated (25.4 and 12.3%, respectively) and mock vector grafts (24.7 and 10.4%) compared with IDO-expressing grafts (3.3 and 3.4%) at 2 weeks posttransplant. Nonetheless, IDO grafts were accumulated with CD4⁺ and CD8⁺ T-cells by week five posttransplant (24.1 and 11.2%, respectively). Figure 4,B summarizes flow cytometry data on graft-infiltrating cells composition. Because IDO has been shown to regulate APC function, the frequency of dendritic cells present in grafts was investigated. Results showed that although the frequency of CD11c⁺ cells was less in IDO overexpressing grafts after 2 weeks (Fig. 4,A and B), the ratio of CD11c⁺ to CD3⁺ cells was significantly higher in this group (Fig. 4 C), suggesting that dendritic cells are more resistant than T-cells to IDO effects.

Taken together, these data show that local expression of IDO significantly prevents infiltration of T-cells into islet allografts and suggest local depletion of effctor CD4⁺ and CD8⁺ T-cells as a mechanism for an antirejection function of IDO.

To further investigate the mechanism of immunosuppression induced by IDO, we examined the local cytokine/chemokine expression profile and production of donor-specific alloantibodies in composite graft recipients. Quantitative PCR showed significantly higher proinflammatory cytokines γ-interferon, interleukin (IL)-2, and IL-17 as well as T-cell–attracting chemokines CXCL9 and CXCL10 in untreated and mock vector grafts versus IDO-expressing grafts two weeks posttransplantation (Fig. 5,A). Conversely, anti-inflammatory cytokines IL-4 and IL-10 were significantly overexpressed in IDO grafts at the same time point. However, cytokine/chemokine profile switched to a proinflammatory pattern after 5 weeks in the IDO group (Fig. 5 A). Thus, low expression of T-cell–attracting chemokines and high levels of anti-inflammatory cytokines can be considered one of the reasons for strict containment of leukocyte infiltration into IDO-expressing grafts.

To investigate alloantibody production, flow cytometry was used to test sera from graft recipient animals for presence of antibodies against donor antigens. In control groups (i.e., mice receiving untreated or mock vector composite grafts), high levels of alloantibodies were detectable three weeks after transplantation (Fig. 5,B and C). Higher levels of IGg2c versus IgG1 isotype was indicative of a Th1 dominant humoral immune response in these groups. In contrast, sera from IDO-expressing graft recipients showed low levels of donor-specific antibodies until the fifth week posttransplant, and it took two more weeks to reach to a level equal to that of the control groups at week three posttransplant (Fig. 5 D). Moreover, early alloantibodies in IDO group were mainly from IgG1 subclass, suggesting a Th2 immune response shift in the presence of local IDO expression. These findings show that local IDO expression promotes an anti-inflammatory environment and shifts the Thl/Th2 balance toward Th2 response.

As shown in Fig. 3,C, FOXP3⁺ cells in graft-draining lymph nodes of IDO-expressing animals were more abundant than those in control groups at 2 weeks posttransplant. Counting the number of FOXP3⁺ cells in immunostained sections showed significantly a higher frequency of FOXP3⁺ cells in the IDO group (9.8 ± 0.9%; P < 0.001) compared with untreated (5.6 ± 1.1%) and mock vector (6.2 ± 1.2%) groups (Fig. 6,A). FOXP3⁺ cell frequency, however, decreased in the IDO group after 5 weeks (7.0 ± 1.1%; Fig. 7,A). To investigate the impact of an increased number of FOXP3⁺ cells and other possible effects of local IDO overexpression on T-cell priming capacity, mixed lymphocyte reactions were performed. The results showed a significant hyporesponsiveness of lymph node cells of IDO-expressing group to stimulation with BALB/c (islet donor) mouse splenocytes but not to a third-party mouse strain (Fig. 6 B and C). This anergy was reverted after 5 weeks posttransplant. These findings suggest that, possibly because of an increased number of regulatory T-cells, antigen-specific incompetence of T-cell priming occurs in draining lymph nodes of IDO-expressing graft–recipient mice.

In view of the limited length of an antirejection effect of IDO in this study, we hypothesized that a decline in the level of IDO transgene expression can be considered a possible cause of cessation of immunosuppression and, eventually, late graft rejection in IDO group. We therefore tested the length of intragraft IDO transgene expression using IDO immunofluorescence staining and quantitative PCR. Composite grafts were harvested at indicated time points and subjected to IDO immunofluorescence staining or quantitative PCR (qPCR). As shown in Fig. 7,A, IDO protein is expressed uniformly and widely in fibroblasts throughout the graft when tested 2 weeks after transplantation. However, after 5 weeks, only a limited number of cells in the graft were still positive for IDO protein (Fig. 7 B). The anti-IDO antibody used in this experiment was able to detect IDO protein from both human and mouse origins.

qPCR results showed that the IDO transgene was expressed at high levels in composite grafts containing IDO vector–infected fibroblasts for up to 4 weeks posttransplantation; then, the expression level started to decrease. As presented in Fig. 7 C, significant decline in the intragraft IDO transgene expression were observed from 100% on day one to 47.6 ± 13.4, 37.2 ± 12.3, and 30.6 ± 10.7% on days 35, 42, and 49 posttransplantation, respectively (P < 0.001; n = 3). This finding was consistent with an in vitro lasting effect of IDO expression in fibroblasts following adenoviral IDO transduction (data not shown). IDO transgene mRNA was not found in the control grafts (data not shown). Because the IDO transgene was of human origin, by using human specific primers we expected that intrinsic IDO expression (i.e., mouse IDO that expressed in grafts as a result of inflammation) did not interfere with IDO transgene qPCR.

These data collectively show a transient pattern for IDO transgene expression in composite grafts following adenoviral gene transfer. The time course of intragraft IDO expression closely corresponds to the duration of graft survival and suppression of cellular and humoral alloimmune responses. We therefore suggest that late graft rejection in this study was probably due to time-dependent loss of transient IDO expression.

In the present study, we showed that local IDO expression suppresses cellular and humoral alloimmune responses against islets and significantly prolongs islet allograft survival without systemic anti-rejection treatments in diabetic immune-competent mice. IDO plays a crucial role in suppression of immune responses. Several previous studies implicated an immunomodulatory effect of IDO in different settings including islet transplantation (28,29,37). However, this is the first time that IDO-expressing bystander syngeneic fibroblast–populated collagen scaffold has been used to protect islet allograft. The engineered composite graft developed and transplanted in this study has several unique features. First, we avoided direct transfer of IDO transgene to islets to reduce the risk of cytotoxicity and loss of islet function. There have been reports showing that adenoviral gene transfer to islets interferes with β-cell function and increases apoptosis rate (30,–32,38–39). Moreover, gene expression after direct transduction has been observed only in the periphery of islets and not in the islet core (40,41). Therefore, by using bystander fibroblasts and not islets as the target for IDO gene transfer, we were able to induce high levels of IDO expression as a local immunosuppressive factor while avoiding any deleterious consequence of gene transfer on islet survival and function.

Embedding islets within a three-dimensional extracellular matrix (ECM) is another advantage of this composite graft. ECM is one of the most important constituents of the islet microenvironment. Several studies have demonstrated that entrapment of islets within a collagen matrix results in satisfactory morphology, enhanced viability, and improved insulin secretory capability in islets (42,–44). As such, embedding islets within ECM, by itself, can improve islet graft function. Moreover, as shown in supplementary Fig. 1, collagen matrix provides a supportive scaffold and helps IDO-expressing fibroblasts surround and protect islets very efficiently in a three-dimensional structure.

Cotransplantation of fibroblasts is also a beneficial aspect of this composite graft. Fibroblasts are cells of choice to be used as bystander IDO-expressing cells because 1) they can easily be induced to express IDO, 2) they show low levels of DNA synthesis and become quiescent when embedded in a collagen matrix (45), and 3) they are resistant to apoptosis in an IDO-induced low-tryptophan environment (25,27). Fibroblasts can also improve islet cell viability. The essential role of fibroblasts in islet physiological competence has previously been reported, and it was shown that some fibroblast-produced factors can promote islet survival in culture (35,36). In addition, availability of syngeneic or autologous fibroblasts eliminates the risk of immune response against these cells. Collectively, application of this novel composite islet graft can address a number of common obstacles in islet allotransplantation by preventing immunological rejection, reestablishing islet cell-ECM interaction, and improving islet survival and function.

Findings of the present study suggest that there are several mechanisms through which local IDO overexpression mounts its antirejection effects. These include suppression of effector T-cells at the graft site, induction of a Th2 immune response shift, generation of an anti-inflammatory cytokine profile, and an increased number of regulatory T-cells in draining lymph nodes, which results in antigen-specific impairment of T-cell priming. As shown in Figs. 2 and 3, T-cells accumulated at the margins of the IDO-expressing grafts but did not invade or infiltrate islets. This accords with the well-known function of IDO: generation of a low-tryptophan and high-kynurenine microenvironment within which activated T-cells are not able to proliferate and survive. Data presented in the Fig. 1 show that such a microenvironment protects islet grafts rejection but does not negatively affect islet function.

The transient increase in the number of FOXP3⁺ cells in our model is very similar to the pattern seen in pregnancy-induced tolerance. It has been shown that due to local expression of IDO in maternal-fetal interface, the number of regulatory T-cells increases during early pregnancy, peaking during the second trimester and followed by a decline postpartum (46). In our model, the number of FOXP3⁺ cells increases as long as local IDO level is high, and upon cessation of IDO overexpression in the grafts the number of FOXP3⁺ cells decreases. This is similar to Treg downregulation after IDO levels decrease in maternal-fetal interface postpartum.

This study for the first time shows that local IDO expression can inhibit production of donor specific alloantibodies. The mechanism(s) underlying this phenomenon needs further elucidation. It is well documented that humoral response against donor-specific antigens is clearly dependent on help provided by CD4⁺ T-cells (47). Only CD4⁺ T-cells recognizing alloantigen through the indirect pathway are able to provide cognate help to allospecific B-cells for the development of alloantibody (48). As such, inhibition of alloantibody production in our study might be due to a T-cell suppressive effect of IDO that significantly reduces the number of donor-specific T-cells and consequently results in insufficient cognate help to B-cells. Moreover, the Th1/Th2 shift seen in humoral immune response in this study may contribute to the IDO-induced deviation of cytokine expression profile given that IL-4 promotes IgG1 isotype switching (49), whereas IFN-γ favors the IgG2a and IgG2c switch (50).

The finite survival of islet grafts in this study may appear discouraging at first glance. However, data presented here clearly show that as long as high levels of intragraft IDO expression was maintained, islet grafts survived and functioned normally and cellular and humoral immune responses against islets were suppressed very efficiently. This finding suggests that long-term protection of islet grafts is feasible if stable IDO expression in bystander fibroblasts is achieved.

In conclusion, the present study proves the feasibility of the development of a viable and functional IDO-expressing composite islet graft and confirms that IDO efficiently and significantly improves islet allograft survival. This promising finding proves that there is a local islet-directed immunosuppressive effect of IDO and sets the stage for development of a nonrejecatable islet allograft using stable IDO induction in bystander fibroblasts. We are currently investigating this approach.

This study was supported by the Canadian Institutes of Health Research. R.B.J. was supported by University of British Columbia Graduate Fellowship (UGF) and Transplantation Scholarship Training awards.

No potential conflicts of interest relevant to this article were reported.

R.B.J. researched data, contributed to discussion, wrote the manuscript, and reviewed and edited the manuscript. F.F. researched data, contributed to discussion, and reviewed and edited the manuscript. A.M.R. researched data and contributed to discussion. R.H. researched data and contributed to discussion. A.M. researched data. B.L. contributed to discussion and reviewed and edited the manuscript. G.L.W. contributed to discussion and reviewed and edited manuscript. A.G. contributed to discussion, wrote the manuscript, and reviewed and edited the manuscript.

The authors are grateful to Dr. Y. Li (University of British Columbia, Vancouver, British Columbia, Canada) for constructing the IDO adenoviral vector. They thank Dr. C. B. Verchere (University of British Columbia) for useful discussion and advice and also Dr. M. K. Levings and A. N. McMurchy (University of British Columbia) for their help with flow cytometry experiments.