Despite success of early islet allograft engraftment and survival in humans, late islet allograft loss has emerged as an important clinical problem. CD8+ T cells that are independent of CD4+ T cell help can damage allograft tissues and are resistant to conventional immunosuppressive therapies. Previous work demonstrates that islet allografts do not primarily initiate rejection by the (CD4-independent) CD8-dependent pathway. This study was performed to determine if activation of alloreactive CD4-independent, CD8+ T cells, by exogenous stimuli, can precipitate late loss of islet allografts. Recipients were induced to accept intrahepatic islet allografts (islet ‘acceptors’) by short-term immunotherapy with donor-specific transfusion (DST) and anti-CD154 mAb. Following the establishment of stable long-term islet allograft function for 60–90 days, recipients were challenged with donor-matched hepatocellular allografts, which are known to activate (CD4-independent) CD8+ T cells. Allogeneic islets engrafted long-term were vulnerable to damage when challenged locally with donor-matched hepatocytes. Islet allograft loss was due to allospecific immune damage, which was CD8- but not CD4-dependent. Selection of specific immunotherapy to suppress both CD4- and CD8-dependent immune pathways at the time of transplant protects islet allografts from both early and late immune damage.
Conventional and experimental immunosuppressive therapies generally act to promote long-term allograft survival by targeting CD4+ T-cell-dependent immune responses. However, CD8+ T-cell-mediated rejection pathways that are independent of CD4+ T cell ‘help’ have been noted for their resistance to immunoregulation by agents that control CD4-dependent rejection responses. In rodent studies costimulatory blockade-resistant, CD8+ T-cell-initiated allograft rejection has been reported in response to hepatocellular allografts (1,2), intestinal allografts (3–6), skin allografts (7,8) and cardiac allografts (9,10). Experimentally, it has been noted that islet allografts do not appear to initiate (CD4-independent) CD8-dependent immune responses in vivo (11–13). Although the importance of the (CD4-independent) CD8+ T-cell subset in the disruption of tolerance induction has been well established, it is not known to what extent this pathway may cause late immune damage of transplanted islets and jeopardizes long-term islet allograft survival.
CD8+ T cells can also adversely influence transplantation outcome through heterologous immunity arising from prior immune exposure. This can result in an increased frequency of alloreactive memory CD8+ T cells with lower costimulatory requirements and activation thresholds. These CD8+ T cells can prevent induction of allograft acceptance (14), reviewed in (15), abrogate induced tolerance (16–18) or prevent acquisition of donor cell chimerism after bone marrow transplant (14,18,19), reviewed in (15). Alloreactive subsets of CD8+ T cells can be activated directly or indirectly by bystander activation (20). A majority of these studies have focused upon the detrimental effects of alloreactive CD8+ T-cell subsets prior to or during induction immunotherapy. Few, if any, studies have been specifically designed to address the role of (CD4-independent) CD8+ T cells in late immune damage of islet allografts. Despite the recent enthusiasm associated with early success of pancreatic islet transplant in humans, progressive loss of islet allograft function over time remains an important clinical problem (21,22).
The current studies were designed to determine the susceptibility of engrafted islet allografts to damage through late activation of alloreactive (CD4-independent) CD8-dependent immune responses in vivo. The prevailing paradigm for acute rejection of islet allografts includes a central role for alloreactive CD4+ T cells (11,23–25). Furthermore, CD4+ T-cell-dependent immune responses elicited by islet allografts in rodents are highly susceptible to immunoregulation with a variety of immunosuppressive therapies. Strategies targeting CD40/CD154 costimulation using anti-CD154 mAb (either alone (11,26,27) or in combination with donor-specific transfusion [DST] of splenocytes (11,26–28); CD28/B7 costimulation using CTLA4-Ig (either alone (29) or in combination with anti-CD154 mAb (30) or LFA-1/ICAM-1 interactions using anti-LFA-1 mAb (either alone (11,31,32) or in combination with anti-CD154 mAb (33,34) induce indefinite islet allograft survival. In the current studies, DST and anti-CD154 mAb treatment was used to induce long-term islet allograft survival due to its short-term administration, established efficacy and defined mechanism of action (11,26–28,33). To elicit the activity of alloreactive CD8+ T cells (as well as alloreactive CD4+ T cells) in islet allograft recipients, a donor-matched hepatocellular cotransplantation model was used since allogeneic hepatocytes are known to readily initiate (CD4-independent) CD8-mediated as well as (CD8-independent) CD4-mediated alloimmune responses (1,11,13,35,36). Subsequent graft survival and activation of donor-specific alloimmune responses were monitored.
Mouse strains FVB/N (H-2q, Taconic), C57BL/6 (H-2b, Taconic), B10.BR (H-2k, H2-T18a/SgSnJJrep, Jackson), CD8 knockout (KO) (H-2b, C57BL/6-Cd8atm1Mak, Jackson) and CD4 KO (H-2b, C57BL/6-Cd4tm1Mak, Jackson) were used in this study at 6–9 weeks of age. All animals received humane care according to the criteria outlined in the Guide for the Care and Use of Laboratory Animals (National Academy Press, revised 1996). Transgenic C57BL/6 or FVB/N mice expressing human alpha-1 anti-trypsin (hA1AT-FVB/N, H-2q and hA1AT-C57BL/6, H-2b) were the source of ‘donor’ hepatocytes. This strain was created, bred and maintained at the Biotechnology Center and Transgenic Animal Facility (The Ohio State University) (1,37).
Chemical induction of diabetes
A 180–220 mg/kg streptozocin (STZ, Sigma) in 0.1 M citrate buffer (pH = 4.5) was injected intraperitoneal (i.p.) to induce diabetes in recipients, as previously described (11). Blood glucose was measured using a One Touch glucose analyzer. Diabetes onset was defined as two consecutive blood glucose measurements >300 mg/dL. In general, the STZ regimen induced diabetes within 2–3 days with blood glucose in the range of 300–450 mg/dL.
Donor islet isolation and purification
Mouse islets were isolated by 1% collagenase P (Boehringer Mannheim Biochemicals) pancreas digestion followed by discontinuous Dextran (Sigma) gradient purification, as previously described (11). Briefly, under inhaled isoflurane anesthetic, the pancreas was perfused via bile duct cannulation with collagenase P (1.162 mg/mL in 1× Hank's balanced salt solution (HBSS). The pancreas was excised, minced and suspended in a 15 mL tube containing 1% collagenase. Three to four pancreata per transplant were digested at 37°C in a 1% collagenase solution. Pancreata were combined to 50 mL, washed with 1×HBSS and resuspended in a discontinuous dextran gradient (stock density of 1.094). Following centrifugation (400 g), islets were collected from the interface and washed in 1×HBSS containing 5% fetal bovine serum (FBS). Residual exocrine tissue and cellular debris were excluded by handpicking islets under a stereomicroscope for transplantation.
Islet transplantation and monitoring of graft function
Six hundred purified islets were transplanted by kidney sub-capsular or intrasplenic injection, as previously described (11). Briefly, 600 purified islets were resuspended in 50–100 μL of cold 1×HBSS with 5% FBS. Recipient animals were anesthetized by inhalational delivery of isoflurane. Islets were transplanted either by injection under the left kidney capsule or by intrasplenic injection into the tip of the spleen. Islets that are transplanted by intrasplenic injection circulate to and engraft in the recipient liver. In control studies, splenectomy of intrasplenic islet isograft recipients did not disrupt glucose homeostasis, indicating that normoglycemia is a function of the islets engrafted within the liver. Islet transplants were considered functional when two consecutive blood glucose measurements <200 mg/dL occurred. The time of islet rejection was defined as the first day of two consecutive blood glucose measurements >250 mg/dL.
Immunotherapies used for induction of long-term islet allograft acceptance
All immunotherapies included the use of monoclonal antibodies. Anti-CD154 (MR1, CRL-2580, ATCC) and anti-LFA-1 (M17/18.104.22.168, ATCC) anti-mouse monoclonal antibodies (mAb) were expanded for in vivo use by ascites production in pristane-primed Nude mice (Ncr, Taconic) or were purchased in purified form (Ligocyte Pharmaceuticals, Bozeman, MO) (38,39). DST and anti-CD154 mAb immunotherapy. Hosts received 500 μg of MR1 antibody by i.p. injection on days −7, −4, 0, 4 relative to transplantation. DST consisted of 10 × 106 FVB/N splenocytes, which were isolated by mechanical dispersion, washed twice in washing buffer and injected into the recipient tail vein on day −7, as previously described (11,40). Anti-LFA-1 mAb immunotherapy. The treatment regimen with anti-LFA-1 mAb for islet allograft recipients was 150 μg i.p. on days 0–6 relative to transplantation, as previously described (11).
Hepatocyte isolation and purification
Hepatocyte isolation and purification was performed as described previously (1,37). Briefly, donor mice were anesthetized with Ketamine/Xylazine, and the liver was perfused 0.09% (EGTA)-containing calcium-free salt solution followed 0.05% collagenase (Sigma, type IV) in 1% albumin. Liver tissue was minced, filtered and washed in RPMI-1640 with 10% FBS. Hepatocytes were purified on a 50% percoll gradient (Pharmacia Biotech, Uppsala, Sweden). Hepatocyte viability and purity were consistently >99% respectively.
Hepatocyte cotransplantation and monitoring of hepatocyte graft function
Donor hepatocytes were retrieved from transgenic mice expressing human alpha-1 antitrypsin (hA1AT) under control of the liver-specific hA1AT promoter (37). Under inhalational isoflurane anesthesia recipients with long-term acceptance of pancreatic islet allografts received 2 × 106 purified allogeneic hA1AT-FVB/N (H-2q), syngeneic hA1AT-C57BL/6 (H-2b) or third-party B10.BR (H-2k) hepatocytes by injection into the tip of the spleen with circulation to the host liver, as previously described. Graft function was determined by presence of secreted transgenic reporter product hA1AT in serial recipient serum samples. Serum hA1AT levels in donor and host mice were detected by sandwich ELISA and ranged from 0.5 to 20 μg/mL (1,37). Graft survival was determined by sustained serum hA1AT levels, and graft loss was considered the time point at which host serum hA1AT was <0.5 μg/mL. Hepatocytes isolated from third-party B10.BR mice do not bear the hA1AT transgene; consequently survival of cotransplanted hepatocytes could not be functionally monitored.
Antibodies used for T-cell subset depletion in recipients with long-term survival of allogeneic islets
Recipients with long-term survival of islet allografts (after short-term treatment with DST and anti-CD154 mAb) were depleted of circulating CD4+ or CD8+ T cells using monoclonal antibodies. Anti-CD4 (GK1.5, ATCC) and anti-CD8 (2.43, ATCC) monoclonal antibodies (mAb) were obtained from purified ascites produced in pristane-primed Nude mice (Ncr, Taconic) as previously described (39). C57BL/6 recipients with long-term islet allograft survival were depleted of CD4+ T cells or CD8+ T cells by treatment with 250 μg antibody i.p. (day –4, –2, 7, 14, 21 and 28 relative to hepatocyte cotransplantation), which depletes CD4+ or CD8+ T cells from baseline 15–25% of the splenocyte population to undetectable levels within 3 days. Depletion persists for 4–6 weeks following administration of the final dose of antibody (8–10 weeks depletion total). T-cell subset depletion was monitored by flow cytometry of peripheral blood leukocytes (PBLs) (36).
Intraperitoneal glucose tolerance test (IPGTT)
Islet recipient mice were tested for metabolic function by IPGTT. Recipient mice were fasted overnight and allowed free access to water. Following overnight fasting, recipients were injected with 2 g/kg body weight glucose resuspended in 0.9% normal saline. Blood glucose levels were checked at 0, 5, 10, 15, 30, 60, 120 and 240 min relative to the initial injection using a One Touch glucose analyzer. Glucose tolerance was compared by determination of the time to normoglycemia (normal blood glucose range of 62–175 mg/dL for mice).
Liver alanine aminotransferase (ALT) serum levels
Assessment of liver inflammation was determined by the measurement of serum transaminase (ALT) mg/dL. Following hepatocyte transplant, serum (100 μL) was collected from recipient mice on days 1, 3, 5, 7, 10 and 14 posttransplant. Each serum sample was diluted 1:5 in 7% bovine serum albumin (BSA, Sigma-Aldrich, Inc., St. Louis, MO) to a final volume of 500 μL. Diluted samples were analyzed for ALT levels by the core pathology lab services at Ohio State University. Normal serum ALT in mice ranges from 17–77 mg/dL.
In vivo cytotoxicity assay
Assay for cytolytic T-cell function in vivo was modified from previously published methods (41,42), and has been previously described (43). Syngeneic or allogeneic target splenocytes were prepared by mechanical disruption followed by red blood cell lysis (0.017 M Tris and 0.75% NH4Cl, pH 7.4). Isolated splenocyte populations were washed twice, enumerated and incubated at 37°C in warm PBS containing Carboxyfluorescein Diacetate Succinimidyl Ester (CFSE, Molecular Probes, Vybrant CFDA SE Cell Tracer Kit, Eugene, OR). Syngeneic target splenocytes were isolated from C57BL/6 mice and were stained at 0.2 μM CFSE (CFSElow). Allogeneic target splenocytes were isolated from FVB/N mice and were stained at 2.0 μM CFSE (CFSEhi). Stained cells were then washed and incubated in 10 mL of pre-warmed DMEM containing 10% FBS for 30 min at 37°C. Allograft recipient mice and control naïve mice received 20 × 106 CFSE-labeled syngeneic target splenocytes and 20 × 106 CFSE-labeled allogeneic target splenocytes by tail vein injection mixed in a 1:1 ratio. Splenocyte target CFSE staining and ratios were verified by flow cytometric analysis prior to injection. Spleens from in vivo CMC recipients were harvested 18 h after CFSE-labeled target-cell injection and were analyzed by flow cytometry with gating on CFSE-positive splenocytes. Percent allospecific cytotoxicity was calculated using the following formula where #CFSEhigh represents the number of allogeneic target cells and #CFSElow represents the number of syngeneic target cells recovered from either naïve or experimental mice:
Delayed-type hypersensitivity (DTH) assay
Alloreactive DTH responses were measured as footpad swelling after injection of allogeneic splenocytes into the murine footpad. Naïve CD4 KO mice served as controls for non-antigen- specific swelling and C57BL/6 mice sensitized to FVB/N antigen by three subcutaneous injections of FVB/N splenocytes served as positive controls for the alloantigen preparation. Cellular antigen was prepared by lysis of red blood cells (0.017 M Tris and 0.75% NH4Cl, pH 7.4) in allogeneic FVB/N or control syngeneic C57BL/6 splenocyte preparations followed by two washes with medium and resuspension in HBSS at a concentration of 10 × 106 FVB/N splenocytes in 25 μL. Splenocytes were irradiated at 2000 Rad and were injected into recipient and control footpads. Subsequent footpad swelling was measured using a micrometer (0.001 inch) 24 h after antigenic challenge.
Donor-reactive alloantibody assay
Recipient serum was assayed for the presence of donor-specific IgG alloantibodies by incubating serum (diluted 1:10) with FVB/N target splenocytes followed by incubation with FITC-conjugated goat F(ab’)2 anti-mouse IgG Fc (Organon Teknika Corp, West Chester, PA). Binding of donor-reactive alloantibody was determined by flow cytometric analysis of target splenocytes, represented as the percentage of target cells labeled by secondary fluorescent antibody as described previously (2). The negative controls for the assay consisted of incubation of target cells with naïve C57BL/6 mouse serum or wash buffer during the primary incubation. Nonspecific binding of the secondary antibody to target splenocytes (approximately 1%) was used to establish the background fluorescence. Comparison of mean fluorescence intensity (MFI) results between groups mirrored the results of percentage binding to indicator cells (higher percentage binding groups have higher MFI).
Isolation and analysis of liver infiltrating cells (LICs)
Liver-infiltrating lymphocytes were isolated on days 3, 7, 10 and 14 post intrahepatic hepatocellular transplant as previously described (43). In brief, mouse liver perfusion was performed as described above for hepatocyte isolation and the resulting liver cell suspension was washed once with RPMI 1640 containing 10% FBS and digested in serum-free RPMI 1640 containing 0.05% collagenase IV (Sigma) and 0.002% DNAse I (Sigma) at 37°C for 40 min. Hepatocytes were pelleted by centrifugation at 30 g and discarded. Intrahepatic lymphocytes along with Kupffer cells and any remaining hepatocytes were then pelleted from the supernatant. Leukocyte populations were purified from the supernatant using a 36% Histodenz (Sigma) discontinuous gradient by centrifugation at 1500 g for 20 min. The interface of cells was collected, washed with media and counted by trypan blue exclusion before being stained and analyzed. Cell viability and purity was >90% on samples used for flow cytometric analysis. For flow cytometric analysis of LICs, 2 × 106 cells were resuspended in 100 μL of washing buffer (5% goat serum and 0.01% sodium azide in PBS) containing PE- or FITC-conjugated mAbs, isotype control antibodies or were left unstained (negative control). The mAbs used for cell subset analysis include anti-CD3 mAb (Hamster IgG, 145–2C11), anti-CD4 mAb (Rat IgG2b, GK1.1), anti-CD8 mAb (Rat IgG2b, 2.43), anti-γδ TCR mAb (Mouse IgG, UC7–13D5), anti-NK1.1 (Mouse IgG2a, PK136), anti-Ly6G mAb (Rat IgG2b, RB6–8C5) (all Pharmingen) and anti-F4/80 mAb (Rat IgG2b, CI:A3–1, Caltag Laboratories, Burlingame, CA). Cells were incubated for 60 min at 4°C and washed three times before resuspension in washing buffer containing 10% buffered formalin. Flow cytometry was performed using a Beckman Coulter EPICS® XL flow cytometer and the System II v. 3.0 acquisition software. Data analysis and histogram overlays were performed using Expo32 v1.2 software (Applied Cytometry Systems, Sacramento, CA). LIC subset composition was calculated based on the percentage of cells expressing cell subset markers within the total LIC population.
Graft survival between experimental groups was compared using Kaplan-Meier survival curves and log-rank statistics (SPSS version 11.5 for Windows, Chicago, IL). For IPGTT analysis, the area under the curve (AUC) was calculated using the trapezoidal method. Other statistical calculations were performed using SPSS using the Student's t test to analyze differences between experimental groups. A p value < 0.05 was considered significant.
Local allospecific immune challenge abrogates survival of established intrahepatic pancreatic islet allografts
Previous studies have demonstrated that long-term islet allograft survival (LTS) can be readily induced by treatment with DST and anti-CD154 mAb. In order to investigate the robustness of long-term acceptance of islets induced by treatment with DST and anti-CD154, we challenged LTS recipients with donor-matched hepatocyte transplant to the liver where the islets were engrafted. In this model, islets and hepatocytes are transplanted to the host liver by intrasplenic injection. Streptozocin-induced diabetic C57BL/6 (H-2b) mice (n = 6) were transplanted with 600 FVB/N islets (H-2q) by intrasplenic injection and induced to accept islets by treatment with donor-specific transfusion of splenocytes (DST, 10 × 106 FVB/N splenocytes, day –7) and anti-CD154 mAb (500 μg, day −7, −4, 0 and 4 relative to transplantation) immunotherapy, as previously reported (11,26–28,33). Islet allografts were accepted long-term in 100% of DST and anti-CD154-treated islet allograft recipients. After 90 days, islet ‘acceptor’ mice (normoglycemia for >90 days) were challenged with 2 × 106 transgenic donor hepatocytes (hA1AT-FVB/N, H-2q) by injection into the spleen, which results in circulation and engraftment of donor hepatocytes in the host liver (37,44). The function of islet transplants was monitored by serial measurement of blood glucose, and the function of hepatocellular transplants was monitored by detection of the transgenic reporter product, hA1AT, in serum by ELISA. In response to the hepatocyte immune challenge, islets were rejected in four of six intrahepatic islet acceptor mice after donor-matched hepatocellular cotransplantation (Figure 1A, B), and all hosts rejected hepatocellular allografts with MST of 14 days following cotransplantation (Figure 1A, C).
Intrahepatic C57BL/6 islet allograft recipients of donor-matched FVB/N hepatocellular cotransplant were tested for development of donor-reactive alloantibody and donor-specific DTH responses. Serum was analyzed for alloantibody 2 weeks prior to and 2 weeks following hepatocellular cotransplantation. Alloantibody was not detectable in recipient serum prior to hepatocellular challenge (day 70 post islet transplant, Figure 2A). Following hepatocellular cotransplantation, detectable levels of alloantibody were present in serum of all recipients (n = 7, p < 0.01, Figure 2A). Similarly, whereas all DST and anti-CD154 mAb-treated islet ‘acceptors’ did not demonstrate alloreactive DTH responses, all treated islet recipients developed detectable alloreactive DTH responses (equivalent to responses in control untreated islet rejector mice) after intrahepatic hepatocellular cotransplant (p = 0.04, Figure 2B).
Hepatocellular transplantation causes inflammation in the hepatic immune environment, hence it was important to determine if loss of islet allograft function occurred due to nonspecific inflammation associated with the transplant procedure. To assess this, allogeneic FVB/N (H-2q) islets were transplanted by intrasplenic injection into streptozocin-induced diabetic C57BL/6 (H-2b) hosts, and long-term acceptance of islet allografts was induced by short-term treatment with DST and anti-CD154 mAb. Islet ‘acceptor’ mice were then transplanted with C57BL/6 syngeneic hepatocytes by intrasplenic injection. All cotransplant hosts remained normoglycemic and demonstrated long-term hepatocellular allograft function for greater than 100 days (n = 5) (Figure 1D). Control studies demonstrate that transplantation of syngeneic hepatocytes induces an initial local inflammatory response, as reflected by a prominent inflammatory infiltrate (Figure 1E) and the elevation of serum ALT (Figure 1F) and an increase of liver infiltrating leukocytes (demonstrated by histologic and flow cytometric analysis). (Not shown) Collectively, these results demonstrate that local nonspecific inflammation does not perturb islet allograft survival, but immune activation induced by allogeneic hepatocyte cotransplant does disrupt survival of long-term engrafted islet allografts.
In order to determine the specificity of this hepatocyte cotransplant induced loss of islet allograft function the effect of immune challenge with third-party donor hepatocytes was next examined. Allogeneic FVB/N (H-2q) islets were transplanted by intrasplenic injection into streptozocin-induced diabetic C57BL/6 (H-2b) hosts, and long-term acceptance of islet allografts was induced by short-term treatment with DST and anti-CD154 mAb. Islet ‘acceptor’ mice were then transplanted with third-party B10.BR hepatocytes by intrasplenic injection. All cotransplant hosts remained normoglycemic for > 60 days (n = 5, Figure 1D) with normal glucose tolerance responses (n = 5, Figure 5D). The degree of liver inflammation following hepatocyte cotransplant was assessed on days 1, 3, 5, 7, 10 and 14 posttransplant. A brief transient rise in serum ALT, which peaked on day 1 following intrahepatic transplantation occurred after syngeneic (n = 3), allogeneic (n = 3) or third-party (n = 5) hepatocyte transplant into islet recipients. The degree of liver ALT increase following transplantation was not significantly different between allogeneic, syngeneic and third-party islet cotransplant recipients (Figure 1F). Overall, these data demonstrate that the loss of islet allograft function after allogeneic hepatocyte cotransplant is due to allospecific immune mechanisms.
We speculate that the relative delay in islet rejection in the cotransplant recipients reflects one or more of the following conditions: a protective microenvironment around engrafted islets induced by DST/MR1 which delays intraislet penetration or effector function of alloreactive CD8+ T cells, the time for activated alloreactive CD8+ T cells to traffick to all of the engrafted islets that are dispersed throughout the liver, an increased number of effectors necessary to destroy all of the engrafted islets, the increased vulnerability of the transplanted hepatocytes to immune damage because of their immediate proximity to the inflammatory site and/or because of the hepatocyte single cell transplant architecture (in contrast to islet cell cluster architecture).
Local activation of CD4-dependent (CD8-independent) alloimmune responses does not disrupt long-term acceptance of intrahepatic pancreatic islets
Hepatocellular allografts initiate both CD4-dependent and CD8-dependent rejection responses. It was therefore important to determine if host CD4-dependent and/or CD8-dependent alloimmune responses were responsible for abrogating islet allograft acceptance. In order to do this, CD8+ T cells were depleted from long-term intrahepatic islet allograft acceptors prior to cotransplantation of allogeneic donor-matched hepatocytes to the liver. Despite local challenge with donor-matched allogeneic hepatocytes, islet allografts continued to survive for > 63 days following hepatocellular transplant in all CD8-depleted cotransplant recipients (n = 4, Figure 3A, B). Unexpectedly, donor-matched allogeneic hepatocytes were accepted for >70 days following transplantation in all recipient mice (n = 4, Figure 3A, C). Hepatocellular allograft survival in cotransplant recipients was significantly prolonged in comparison to DST and anti-CD154 mAb-treated CD8-depleted C57BL/6 hepatocellular allograft recipients, which reject hepatocytes acutely with MST of 10 days ([n = 9, p < 0.01] (36,40) and also in comparison to untreated CD8-depleted C57BL/6 hepatocellular allograft recipients which reject hepatocytes with MST of 10 days ([n = 10, p < 0.01] (36,40) (Figure 3D). A parallel model using CD8 KO (H-2b) mice as transplant recipients was used to corroborate these data in order to avoid any unanticipated effects associated with use of the depleting monoclonal antibodies. Long-term islet allograft survival in CD8 KO mice was induced with DST and anti-CD154 mAb immunotherapy (n = 5) as in the C57BL/6 host. After establishment of islet allograft survival for >60 days, CD8 KO islet acceptor mice were challenged with donor-matched allogeneic FVB/N hepatocytes by intrasplenic injection. Islet allograft function was maintained for >60 days, with four of five CD8 KO islet acceptor mice demonstrating long-term islet allograft acceptance (>80 days) following hepatocellular cotransplantation (Figure 4A, B). Similarly, hepatocellular allograft survival was moderately prolonged (MST = 18 days with hepatocytes surviving >50 days in two of five cotransplant recipients) (Figure 4A, C). Hepatocellular allograft survival was significantly prolonged in comparison to control untreated CD8 KO hepatocellular allograft recipients (MST = 14 days, (1) (p < 0.01) and CD8 KO hepatocellular allograft recipients treated with DST and anti-CD154 mAb, (MST = 7 days, (40) (Figure 4D). These data suggest that the disruption of long-term islet allograft survival by allogeneic hepatocyte challenge in C57BL/6 recipients is not due to alloreactive CD4+ T-cell-dependent mechanisms.
Local activation of (CD4-independent) CD8-dependent alloimmune responses disrupts long-term acceptance of intrahepatic pancreatic islet allografts
To assess the influence of alloreactive CD8-dependent immune responses on long-term islet allograft survival, 600 FVB/N islets were transplanted to the liver of C57BL/6 recipients treated with short-term DST and anti-CD154 mAb. After establishment of stable long-term islet allograft function for greater than 60 days, islet allograft acceptor mice were depleted of CD4+ T cells by treatment with anti-CD4 mAb. Anti-CD4 mAb treatment was given on d-4, −2, 7, 14, 21 and 28 relative to hepatocellular allograft transplantation. This therapy resulted in complete depletion of CD4+ T cells for 8–10 weeks following hepatocellular allograft transplantation as verified by concurrent analysis of recipient PBLs. These CD4-depleted recipients were challenged by allogeneic FVB/N hepatocyte transplant to the liver where islets were previously engrafted. It has previously been demonstrated that depletion of CD4+ T cells after 50 days of successful islet allograft function under cover of short-term DST and anti-CD154 mAb therapy does not disrupt islet allograft function (45). Following cotransplantation, allogeneic hepatocytes were rejected in all islet allograft acceptor mice with MST of 35 days (Figures 5A, C). Following the rejection of hepatocytes, all islet allografts were either subsequently rejected (three of five mice with MST of 56 days, Figure 5A, B) or exhibited impaired islet allograft function, as demonstrated by glucose tolerance testing (two of five mice, Figure 5D). CD4+ T cells were not detectable by flow cytometry of PBLs in recipients at the time of islet rejection.
The influence of local activation of alloreactive CD8+ T cells on long-term islet allograft acceptance was also examined by cotransplantation of islets and hepatocytes into CD4 KO mice. CD4 KO (H-2b) recipient mice spontaneously accept intrahepatic FVB/N (H-2q) islet allografts in the absence of immunosuppression (11). In contrast, CD4 KO mice rapidly reject intrahepatic FVB/N hepatocellular allografts with MST of 10 days (1). Six Hundred FVB/N islets were transplanted to the liver of diabetic CD4 KO mice. Islet allografts were spontaneously accepted in the absence of immunosuppression for greater than 30 days following transplant. CD4 KO islet acceptor mice were then challenged by donor-matched FVB/N hepatocyte transplant to the liver. Hepatocellular allografts were rejected in all recipients with a median survival time of 14 days following cotransplantation (n = 7, Figure 6). Following hepatocellular allograft rejection, islet allografts were also rejected in six of seven hosts with MST of 28 days following hepatocellular transplantation (Figure 6). Thus, in contrast to observations regarding activation of local CD4-dependent immune responses, which did not perturb islet allograft function, local activation of (CD4-independent) CD8+ T-cell-mediated responses by allogeneic hepatocytes does interfere with the longstanding function of intrahepatic islet allografts.
Late CD8-mediated immune damage to islet allografts following hepatocyte cotransplant occurs in association with cell-mediated cytotoxic but not DTH effector mechanisms
CD4-depleted C57BL/6 recipients of sequential islet and hepatocellular cotransplants were next assayed for the development of cellular effector function. Allospecific DTH and in vivo allocytotoxicity in cotransplant recipients were measured at 1–2 weeks following islet rejection or at day 70 following challenge with allogeneic hepatocytes. CD4-depleted recipients did not demonstrate detectable allospecific DTH responses before or following hepatocellular challenge (Figure 7A). In contrast, while control DST and anti-CD154 mAb-treated islet ‘acceptor’ mice did not demonstrate in vivo allocytotoxicity (n = 2, Figure 7B), after hepatocyte cotransplant allospecific cytotoxicity was readily detectable (n = 3, Figure 7B). Neither DST and anti-CD154 mAb-treated C57BL/6 islet ‘acceptor’ mice or CD4-depleted C57BL/6 islet hepatocyte cotransplant rejector mice had alloantibody in recipient serum. Prior studies demonstrate that development of allospecific cytotoxicity after allogeneic hepatocyte transplant is CD8+ T cell mediated (46). These data indicate that local activation of CD8+ T-cell-dependent cytotoxic effector responses can precipitate islet allograft damage and/or disrupt survival of islet allografts even after long-term engraftment in the liver.
Initial immunotherapy with anti-LFA-1 mAb targeting both CD4-dependent and CD8-dependent alloimmune responses protects islet allografts from late immune damage
The preceding experiments demonstrated that although DST and anti-CD154 mAb immunotherapy effectively induces long-term acceptance of islet allografts, islets remain vulnerable to late immune damage by local activation of alloreactive CD8+ T cells. Considering that the published mechanism of action for DST and anti-CD154 mAb immunotherapy occurs by immunoregulation of CD4+ T-cell responses and early deletion of alloreactive CD8+ T cells (11,26–28,33), it is reasonable to conclude that repopulation of CD8+ alloreactive T cells late after islet engraftment allows for primary activation of these cells by hepatocellular allografts. We then hypothesized that in order to protect islet allografts from late immune damage by CD4-independent CD8+ T cells, it would be desirable to utilize immunosuppressive strategies that successfully immunoregulate both CD4+ and CD8+ T-cell alloreactivity at the time of transplant. Short-term immunotherapy targeting the adhesion molecule LFA-1 is known to induce indefinite islet allograft survival in a number of strain combinations (11,31). We have reported that this strategy suppresses both alloreactive CD4-dependent and CD8-dependent immune responses in vivo (11,38). To determine if this immunotherapy would protect intrahepatic islet allografts from late immune damage by CD8+ T cells, FVB/N (H-2q) islets were transplanted to the liver of streptozotocin-induced diabetic C57 BL/6 (H-2b) mice and long-term islet allograft acceptance was induced by short-term immunotherapy with anti-LFA-1 mAb (150 μg i.p., day 0–6). Following induction of long-term islet allograft acceptance (>60 days), islet allograft acceptor mice were challenged by donor-matched hepatocyte transplant to the liver as in previous experiments. Islet allografts continued to survive in all recipients following hepatocellular cotransplantation (n = 6, Figure 8A). Three of the six cotransplant recipients had functioning islet and hepatocyte allografts when they were death censored on days 63, 84 and 84 following hepatocyte transplantation. In addition, hepatocellular allograft survival was significantly prolonged to MST of 63 days with 50% of recipients accepting hepatocytes long-term (>80 days) even after clearance of the initial anti-LFA-1 treatment mAb and in the absence of any additional immunosuppression. In recipients with rejected hepatocellular allografts, islet allografts continued to function for >30 days. Interestingly, the metabolic reporter product (hA1AT) used to follow hepatocellular allograft survival increased more than 1000-fold in two of six recipients (Figure 8C).
The failure to preserve long-term metabolic function after pancreatic islet cell transplantation in humans despite high rates of early engraftment is an important clinical problem (21,22,47–50). CD4-independent, CD8-dependent immune responses are recognized for their resistance to suppression by conventional immunotherapies though in previous studies we have noted that pancreatic islets do not stimulate these responses primarily (11,23–25). We hypothesized that late activation of CD4-independent CD8+ T cells, perhaps by inflammatory stimuli with cross reactivity to alloantigens, could be an important pathway jeopardizing long-term islet allograft survival. Activated CD8+ T cells could directly damage islet allografts and/or indirectly cause islet allograft loss by reactivation of alloreactive CD4+ T cells, as occurs in the onset of autoimmune diabetes (51–53), reviewed in (54). In these studies, we evaluated whether CD4-dependent, CD8-dependent or both CD4- and CD8-dependent immunity contribute to late immune damage of pancreatic islet allografts.
We found that local immune activation of alloreactive CD8+ T cells but not CD4+ T cells accounts for late immune damage of islet allografts in cotransplant recipients. Immune challenge with allogeneic hepatocytes when only CD4+ T cells were present (as in CD8-depleted C57BL/6 or CD8 KO recipients, Figures 3 and 4) did not perturb islet allograft function. It is not clear whether this occurs due to local regulatory mechanisms or whether engrafted islets in DST and anti-CD154 mAb-treated recipients are less vulnerable to immune damage by alloreactive CD4+ T cells. However, we favor local immune regulation as a mechanism since under these circumstances long-term immunoregulation for intrahepatic islet allografts also appears to be protective for allogeneic hepatocytes. Median survival time for hepatocytes in cotransplant recipients was prolonged compared to untreated or DST and anti-CD154 mAb-treated hepatocyte transplant alone controls. That is, DST and anti-CD154 mAb therapy (in the absence of islet cotransplant) does not prolong primary hepatocellular allograft survival in CD8 KO or CD8-depleted C57BL/6 mice (40) whereas DST and anti-CD154-treated CD8-depleted islet recipients demonstrated prolonged survival of the primary islet transplant as well as the sequential hepatocyte transplant. In contrast, activation of CD8+ T-cell immune responses by allogeneic hepatocyte cotransplant into CD4-depleted C57 BL/6 and CD4 KO islet allograft acceptor mice did precipitate islet (and hepatocyte) allograft destruction (Figures 5 and 6). While we and others have demonstrated that islet allografts predominantly activate CD4-dependent immune responses (11,23–25), islets are known to be vulnerable to acute damage by primed (CD4-independent) CD8+ T cells (13). This highlights the concept that, while certain graft tissues do not primarily activate (CD4-independent) CD8+ T cells, they may be susceptible to damage by this immune pathway once activated by other stimuli. This study suggests that activation of alloreactive CD8-dependent cytolytic immune responses by hepatocytes cotransplanted to the liver precipitates late islet allograft damage since islet allografts were lost in almost all CD4-depleted C57BL/6 and CD4 KO cotransplant recipients. In other studies we have noted that when islet allografts are transplanted to the kidney site in diabetic C57BL/6 mice and induced with DST and anti-CD154 mAb to achieve long-term islet allograft survival (>90 days, n = 5), immune challenge by hepatocyte transplant to the liver does not precipitate islet allograft damage or loss in the majority of recipients. Instead, we noted that under these circumstances when the allogeneic stimulus is delivered peripheral to where the islets are engrafted (kidney) rather than to the same locale, four out of five recipients maintained islet allograft function > 100 days following hepatocyte cotransplant. These results could also reflect immune conditions unique to the kidney site of islet transplant. This islet acceptance occurs despite hepatocyte rejection in four of the five recipients (MST = 21 days) (not shown).
The observed scenario of CD8+ cytotoxic T-cell (CTL) immune activation and resultant late islet allograft damage in these cotransplant studies is consistent with and expands the known in vivo behaviors of effector CD8+ T cells. In primary islet allograft recipients, alloreactive CD8+ T cells are not sufficiently activated (13), and islet allografts survive long-term in CD4 KO hosts in the absence of immunosuppression (11) (Figure 6). Treatment of C57BL/6 islet allograft recipients with DST and anti-CD154 mAb is sufficient to protect islet allografts from CD4+ T-cell-dependent immune damage and induces long-term islet allograft acceptance. However, despite long-term engraftment of islet allografts, transplanted islets remain susceptible to immune damage upon local immune challenge with allogeneic hepatocytes known to activate both CD4+ and CD8+ T-cell responses (Figure 1). This locally initiated immune damage is mediated by alloreactive CD8+ (Figures 5 and 6) but not CD4+ T cells (Figures 3 and 4). Thus the DST and anti-CD154 mAb immunotherapy induces robust regulation of alloreactive CD4-dependent alloimmunity but is not sufficient to induce long-term local suppression of CD8-dependent alloimmunity in the liver. The impairment of islet allograft function in all cotransplant recipients highlights the potency of CD8+ T-cell effectors. Interestingly, our results for islet allografts using the same DST and anti-CD154 immunotherapy are in contrast to those reported by other investigators who reported that DST and anti-CD154 mAb-treated islets allograft recipients continue to accept their islet allografts long term following challenge with a second donor-matched islet allograft and donor-matched heart transplant late after transplantation (28). We suspect that the difference in these second transplant immune challenge results may reflect the propensity of islet and heart tissue to dominantly activate CD4-dependent but not CD8-dependent immunity.
In contrast to recipient treatment with DST and anti-CD154 mAb, short-term treatment with anti-LFA-1 mAb effectively induced long-term islet allograft survival, which was resistant to local immune challenge (Figure 8). Prior studies show that this treatment regimen with anti-LFA-1 mAb is associated with clearance of the antibody in recipient serum by day 60 post transplant (43). Thus short-term treatment with anti-LFA-1 mAb not only induced long-term acceptance of islet allografts, but also prevented both CD4- and/or CD8-dependent immune damage of engrafted islets in response to local immune challenge with allogeneic hepatocytes even after disappearance of the treatment antibody. Interestingly, survival of allogeneic hepatocytes sequentially cotransplanted to the liver was significantly prolonged in these islet hepatocyte cotransplant recipients (Figure 8). This is in striking contrast to the rapid CD8-dependent rejection of sequentially transplanted hepatocytes observed in all DST and anti-CD154 mAb-treated islet recipients (Figures 5 and 6). It is unclear whether this immunoprotection arises from local regulation by CD4+ and/or CD8+ T cells. Although CD8+ T cells in the absence of CD4+ T cells do not initiate islet allograft rejection, we have previously noted that transplant of allogeneic islets results in the up-regulation of CD69 (early T-cell activation marker) and CD103 (T-cell integrin) on CD8+ graft infiltrating cells, suggesting that the CD4-independent CD8+ T cells are not ignorant to the presence of islet alloantigen (13). Since LFA-1/ICAM-1 interactions can provide costimulation signals to CD8+ T cells (55–59), this early recognition of alloantigen in the presence of anti-LFA-1 mAb costimulatory blockade could program the CD8+ T cells for immunoregulation. However, this appears to occur uniquely after islet transplant. Short-term immunotherapy with anti-LFA-1 mAb after hepatocellular transplant (in the absence of prior islet transplant) in C57BL/6 mice prolongs but does not induce survival of hepatocellular allografts >60 days in any recipients (MST of 39 days) (38). In these cotransplant experiments, 50% of cotransplant recipients accepted hepatocellular allografts >60 days after clearance of the initial treatment antibody and in the absence of any additional immunosuppression. In cotransplant recipients that rejected hepatocytes, survival was prolonged to MST of 63 days. The observation that some hepatocellular and all islet allografts survived long-term while some hepatocellular allografts were eventually rejected could reflect the fact that transplanted hepatocytes, which dispersed or trafficked closest to the immune locales of protected islets had the most durable anti-LFA-1 mAb induced immunoprotection. Two of the cotransplant recipients demonstrated significantly increased production of the reporter product hA1At (Figure 8C). This increase in hepatocellular graft function after transplant has not been previously observed, and could arise due to enhanced metabolic function and/or proliferation of transplanted hepatocytes that disperse closest to the islet allografts. Pancreatic islets could secrete local growth factors that influence the function, growth and proliferation of transplanted hepatocytes, as has been observed in vitro and in vivo (60–62).
We acknowledge that the experimental model used in these studies is unconventional and has potential immunologic consequences arising from delivery of alloantigen to a peripheral lymphoid compartment (intrasplenic injection). Despite this we observe that long-term acceptance of islet allografts is readily induced by the DST and anti-CD154 mAb regimen after islet transplant by intrasplenic injection similar to the published efficacy for islets transplanted to the kidney sub-capsule. Furthermore, the pattern of immune responses to islet allografts appears to be similar regardless of whether islets are transplanted by intrasplenic injection or by injection to the kidney (e.g. we observe spontaneous acceptance of islet allografts in CD4 KO mice but rapid rejection in C57BL/6 or CD8 KO mice in the absence of immunosuppression with both intrasplenic or kidney sub-capsular islet transplantation). Finally, the differential effects of immune challenge under CD4- versus CD8-deficient conditions on long-term accepted islets engrafted in the liver (a clinically relevant site for islet transplant) is clearly demonstrable despite any unique attributes of the experimental model. Similarly, the disparate efficacy of two different immunosuppressive strategies to protect intrahepatic islet allografts from late immune damage can also be appreciated.
These studies collectively demonstrate that despite long-term engraftment of islet allografts induced by immunosuppression, islets transplanted to the liver remain vulnerable to CD8-dependent immune damage. Selection of an immunotherapy such as short-term treatment with anti-LFA-1 mAb which targets both CD4-dependent and CD8-dependent immune responses appears to protect intrahepatic islet allografts from both acute and late immune damage. This experimental model was useful in differentiating the robustness of two immunosuppressive strategies which both successfully induce primary long-term islet allograft survival in rodents. Furthermore, intrahepatic islet transplant under cover of short-term anti-LFA-1 mAb therapy appears to induce ongoing immunoregulation, which prevents rejection of a second cell transplant in the absence of additional immunosuppression. These results have direct implications for the design of immunosuppressive strategies in the setting of intrahepatic islet transplant in humans.
The authors wish to thank Brendon L. Fussnecker, Jamie K Mack, Anjuli Kolarik, Thomas Pham and Zack Zumbar for technical assistance and Heather Dziema for article preparation. This work was supported in part by grants from the American Diabetes Association, the Roche Organ Transplantation Research Foundation, the American Society of Transplant Surgeons, W.M. Keck Genetics Research Facility of the Neurobiotechnology Center, by NIH DK52920 and DK072262.