InsB9-23 Gene Transfer to Hepatocyte-Based Combined Therapy Abrogates Recurrence of Type 1 Diabetes After Islet Transplantation

Erratum. Abrogation of MMP-9 Gene Protects Against the Development of Retinopathy in Diabetic Mice by Preventing Mitochondrial Damage. Diabetes 2011;60:3023–3033

AbstractThe induction of antigen (Ag)-specific tolerance represents a therapeutic option for autoimmune diabetes. We demonstrated that administration of a lentiviral vector enabling expression of insulin B chain 9-23 (InsB9-23) (LV.InsB) in hepatocytes arrests β-cell destruction in prediabetic NOD mice by generating InsB9-23–specific FoxP3+ T regulatory cells (Tregs). LV.InsB in combination with a suboptimal dose of anti-CD3 monoclonal antibody (combined therapy [CT], 1 × 5 μg [CT5]) reverts diabetes and prevents recurrence of autoimmunity after islet transplantation in ∼50% of NOD mice. We investigated whether CT optimization could lead to abrogation of recurrence of autoimmunity. Therefore, alloislets were transplanted after optimized CT tolerogenic conditioning (1 × 25 μg [CT25]). Diabetic NOD mice conditioned with CT25 when glycemia was <500 mg/dL remained normoglycemic for 100 days after alloislet transplantation and displayed reduced insulitis, but independently from the graft. Accordingly, cured mice showed T-cell unresponsiveness to InsB9-23 stimulation and increased Treg frequency in islet infiltration and pancreatic lymph nodes. Additional studies revealed a complex mechanism of Ag-specific immune regulation driven by CT25, in which both Tregs and PDL1 costimulation cooperate to control diabetogenic cells, while transplanted islets play a crucial role, although transient, recruiting diabetogenic cells. Therefore, CT25 before alloislet transplantation represents an Ag-specific immunotherapy to resolve autoimmune diabetes in the presence of residual endogenous β-cell mass.IntroductionType 1 diabetes (T1D) results from destruction of insulin-producing pancreatic β-cells mediated by autoreactive T cells (1). CD4+ and CD8+ T cells have both been associated to the autoimmune destruction of pancreatic β-cells, although an increasing body of evidences suggests a primary role of CD8+ T cells infiltrating pancreatic islets for β-cell mass erosion (2,3). Insulin is the primitive β-cell–related antigen (Ag) targeted by T cells, determining the spreading of the response to a broader family of β-cell Ags both in human and nonobese diabetic (NOD) mice (4), which spontaneously develop the disease resembling human T1D. Autoreactive T cells are detectable at birth with a naive phenotype and progressively differentiate into memory T cells in autoantibody-positive subjects and patients at onset (5). After the active phase of autoimmunity, a pool of autoreactive T cells with a memory phenotype persists in a quiescent state for decades. This pool of autoreactive T cells can be reactivated by Ag exposure in patients undergoing islet allotransplantation (6). Islet transplantation represents an immunological challenge, since allogenic graft rejection and autoimmunity recurrence coexist, with the potential for reactivation of autoreactive memory T cells posing an additional set of therapeutic obstacles. Autoimmunity recurrence is indeed refractory to standard immune-suppressive regimens and is an important cause of graft loss (7,8). Therefore, there is a strong need to develop novel immunotherapy approaches that specifically target β-cell Ags and their immunodominant epitopes (9) to keep autoimmunity recurrence under control in patients undergoing islet transplantation.Insulin B chain 9-23 (InsB9-23) is the immunodominant insulin region in both humans and NOD mice, targeted by CD4+ (InsB9-23/MHC-II) and CD8+ T cells (InsB15-23/MHC-I) (9). We previously demonstrated that imposing InsB9-23 expression in hepatocytes by lentiviral vector (LV)-mediated in vivo gene delivery triggers an immunoregulatory program, including abortive activation of InsB-specific CD8+ T cells and induction of InsB-specific FoxP3+ T regulatory cells (Tregs), which are responsible for the maintenance of a long-lasting state of InsB-specific active tolerance (10–12). InsB9-23 LV (LV.InsB) gene transfer in late presymptomatic NOD mice halted disease development and reverted T1D in early-onset NOD mice when combined with a suboptimal dose (1 × 5 μg) of anti-CD3 monoclonal antibody (mAb) (13–15). The induction of InsB-specific Tregs was required to keep effector T cells under control (16).Here, we evaluated the combination therapy (CT: LV.InsB + anti-CD3 mAb) to regulate diabetogenic responses at the late stage of T1D progression, when residual endogenous β-cell mass needs to be expanded by islet transplantation to restore glucose homeostasis. CT was optimized in the formulation and administered 1 week before allogeneic islet transplantation. Recurrence of autoimmunity was abrogated, and stable normoglycemia was achieved in NOD mice, where the optimized CT pretreatment impeded T1D progression.Research Design and MethodsStudy DesignTo investigate the capacity of the combination LV-mediated hepatocyte-targeted InsB9-23 expression and anti-CD3 mAb to cure T1D when a syngeneic or allogeneic islet transplantation is performed, NOD mice received intravenously the CT (LV.InsB and anti-CD3 mAb) and islet transplantation under the kidney capsule according to different doses and schedules. Mice were treated at advanced stage of T1D development to cure the disease. Experimental groups were dimensioned to allow statistical analysis. Mice were randomly assigned to each group, but the experimenter was not blinded to group identity.LV VectorsInsulin B9-23 (SHLVEALYLVCGERG) coding sequence, including a Kozak sequence upstream of the transcription start site and a stop codon at the terminus, was generated and cloned into linearized vector for hepatocyte-targeted transgene expression pCCLsin.ET.wpre142-3pT (LV.ET.142T), which contains a hepatocyte-specific promoter (ET) and four repeated, miRNA142 target (142T) sequences (17). Third-generation LVs were produced by Ca3PO4 transfection into 293T cells as previously described (18).For LV.ET.142T titration, 105 293T cells were transduced with serial vector dilutions. Genomic DNA (gDNA) was extracted 14 days after transduction by using Maxwell 16 Cell DNA Purification Kit (Promega), according to the manufacturer’s instructions. Vector copies per diploid genome (vector copy number [VCN]) were quantified by quantitative PCR (qPCR) starting from 100 ng of template gDNA using primers (HIV sense: 5′-TACTGACGCTCTCGCACC-3′; HIV antisense: 5′-TCTCGACGCAGGACTCG-3′) and a probe (FAM 5′-ATCTCTCTCCTTCTAGCCTC-3′) against the primer binding site region of LVs. Endogenous DNA amount was quantified by a primers/probe set against the human telomerase gene (Telo sense: 5′-GGCACACGTGGCTTTTCG-3′; Telo antisense: 5′-GGTGAACCTCGTAAGTTTATGCAA-3′; Telo probe: VIC 5′-TCAGGACGTCGAGTGGACACGGTG-3′ TAMRA). Copies per genome were calculated by the formula = [ng LV/ng endogenous DNA] × [n of LV integrations in the standard curve]. The standard curve was generated by using a CEM cell line stably carrying four vector integrants, which were previously determined by Southern blot and fluorescence in situ hybridization analysis. All reactions were performed in duplicate or triplicate in an ABI Prism 7900HT Real-Time PCR thermal cycler (Applied Biosystems). Each qPCR run carries an internal control generated by using a CEM cell line stably carrying one vector integrant, which was previously determined by Southern blot and fluorescence in situ hybridization analysis. Titer is expressed as 293T transducing units (TU)/mL and calculated using the formula TU/mL = [VCN × 105 × dilution factor−1]. Vector particles were measured by HIV-1 Gag p24 antigen immunocapture assay (PerkinElmer) according to the manufacturer’s instructions. Vector infectivity is calculated as the ratio between 293T TU and total particles as assessed by p24 immunoassay.Islet Isolation, TransplantationUnder anesthesia with Avertin (Sigma-Aldrich), pancreata were swelled with 2 mL of cold RPMI medium (Sigma-Aldrich) containing 1.5 mg/dL of collagenase (type P; Boehringer Mannheim), excised, and incubated in a stationary bath at 37°C.The islets were separated by density gradient (Histopaque-1077) and hand-picked under a stereomicroscope. An islet mass of 600 equivalent islets was aspirated into a 200-μL tip connected with a 1-mL Hamilton syringe (Hamilton) and settled by gravity. After sedimentation, islets were transplanted under the left kidney capsule of diabetic recipient mice. Depending on the experimental design, mice received syngeneic (NSG) or allogeneic (Balb-c) islet transplantation or exogenous insulin therapy (LinBit pellet; LinShin).Mice and TreatmentsFemale nonobese diabetic (NOD/LtJ), NSG (NOD scid γ, NOD.Cg-Prkdcscid Il2rgtm1WjI/SzJ), and Balb-c (BALB/cAnNCrl) mice were purchased (Charles River, Calco, Italy) and housed in specific pathogen-free conditions. Diabetes was determined by two consecutive measures of glycemia ≥250 mg/dL. Blood glucose measurements were determined by a Bayer Breeze blood glucose monitoring system (Bayer). All procedures were reviewed and approved by the San Raffaele Institute (Milan) Institutional Animal Care and Use Committee (IACUC #831).All treated mice were administered LV by tail vein injection (i.v.) using 1 × 109 TU/mouse. Anti-CD3 (2-C11 F[ab′]2) (BioExpress) was administrated i.v. at the indicated doses. Depending on the experimental design, mice received syngeneic (NSG) or allogeneic (Balb-c) islet transplantation under the left kidney capsule or a subcutaneous insulin implant (LinShin), which provided sustained release of insulin for ∼2–3 weeks. Nephrectomy was performed under anesthesia with Avertin, and purified anti-mPDL1 mAb (500 µg/dose clone 10F.9G2) or control rat IgG2a (clone 2A3; BioXcell) was intraperitoneally injected at the indicated times.Lymphocyte Isolation and FACS StainingSpleens, livers, and pancreatic lymph nodes (PLN) were homogenized using frosted glass slides and filtered after red blood cell lyses (where needed) to obtain a single-cell suspension. Single-cell suspensions were stained with the following antibodies: CD4 (RM4-5), CD8 (53-6.7), CD25 (PC61), CD3 (17A2), CD45 (30F11) (BD Pharmingen), CD127 (A7R34), and Foxp3 (FjK-16s) (eBioscience). Labeled cells were analyzed with a BD Symphony cytometer (BD Biosciences) and analyzed using FlowJo 10 software. For intracellular staining, cells were stained for surface proteins, fixed, permeabilized, and stained for intracellular proteins, according to the manufacturer’s instructions (eBioscience). Class II MHC tetramer IAg7 Ins B9-23 class II MHC (HLVERLYLVCGEEG) and class I H2Kd Ins B15-23 (LYLVCGERG) were provided by the National Institutes of Health Tetramer Core (Emory University, Atlanta, GA).In Vitro T-Lymphocyte Stimulation and Cytokine Release DeterminationAll in vitro T-cell stimulations were performed in RPMI medium with l-glutamine, 10% FBS, 2-mercaptoethanol, nonessential amino acids, sodium pyruvate, and penicillin/streptomycin. Splenocytes were in vitro stimulated with InsB9-23 peptide, and supernatants were collected at 96 h and stored at −80°C. Cytokines released in the supernatants were quantified by mouse interferon-γ (IFN-γ), interleukin (IL) 17A Bio-Plex assay kit (Bio-Rad). Responsiveness to allo-Ag was determined by mixed leukocytes reaction coculturing in triplicate splenocytes from experimental and control NOD mice with irradiated (30 Gy) splenocytes isolated from syngeneic NOD mice or allogeneic from Balb-c and C57Bl6 naive mice. T-cell proliferation was determined by 3H-thymidine incorporation following the indicated time of stimulation.Immunohistochemistry and Islet ScoringPancreata were harvested, fixed in 10% buffered formalin (Bio Optica), and embedded in paraffin. Serial sections (4 μm) were stained with hematoxylin and eosin for morphological analysis and insulitis scoring. Pancreatic islet percentages of infiltration were assigned, as reported in Supplementary Fig. 2, by performing an unbiased quantification analysis based on the Aperio Scan Software System.To detect insulin, paraffin-embedded sections were stained with polyclonal guinea pig anti-insulin antibody dilution, 1:100 (Dako Cytomation), followed by a peroxidase-labeled anti–guinea pig IgG secondary antibody, and counterstained with hematoxylin. To asses immunolocalization of FoxP3, rabbit anti-FoxP3 (FjK-16s) antibody was used after antigen retrieval with Tris-EDTA (pH 9) in a warm bath and revealed by rat-on-rodent horseradish peroxidase-polymer and rabbit-on-rodent horseradish peroxidase-polymer (Biocare Medical), using 3,3 diaminobenzidine as chromogen (BioGenex). Slides were counterstained with hematoxylin.Statistical AnalysesStatistical analyses were performed using GraphPad Prism software. Incidence of diabetes was compared by Kaplan-Meier log-rank test. The Fisher exact test (χ2) was used to compare levels of infiltration between experimental groups. ANOVA test was used to determine statistical differences between multiple experimental groups, after verification of normal distribution of measurements in each group by the Shapiro-Wilk test. The Mann-Whitney test was used to verify means between two independent groups. Findings were considered significant with values for P ≤ 0.05.Data and Resource AvailabilityAll data generated and analyzed during this study are included in this article and its Supplementary Material. The pCCLsin.ET.InsB9-23.wpre142-3 pT generated for the current study is available from the corresponding author upon reasonable request.ResultsLV.InsB/Anti-CD3 CT After Islet Transplantation in NOD Diabetic Mice Controls Recurrence of AutoimmunityTo determine whether a single systemic administration of LV.InsB in combination with anti-CD3 mAb (1 × 5 μg) (CT5) is effective to control recurrence of autoimmunity after pancreatic islet transplantation, NOD mice (n = 19) in advanced stage of disease progression (blood glucose level [bgl] 394 ± 17 mg/dL) (Fig. 1A) received syngeneic islets under the kidney capsule, which normalized glycemic levels. Five of nine NOD mice receiving CT5 the day after transplantation stably maintained normoglycemia (bgl <250 mg/dL) up to 250 days. Conversely, all control mice untreated after islet transplantation or treated with anti-CD3 mAb (1 × 5 μg) alone returned hyperglycemic in 2 weeks (Fig. 1B). These results indicate that CT5 after syngeneic islet transplantation allows effective maintenance of glucose homeostasis in ∼50% of NOD treated mice. To determine the relative contribution of endogenous and exogenous β-cell mass in glucose homeostasis, histological analysis was performed on pancreata and kidneys of CT5-treated NOD mice that achieved insulin independence. These studies revealed loss of transplanted islets, while pancreatic insulin-producing cells persisted (Fig. 1C), suggesting that syngeneic islet transplant is required to achieve normoglycemia but is not per se necessary to revert diabetes. We then tested the efficacy of CT5 treatment in the context of allogeneic islet transplantation, which mirrors the clinical setting. A cohort of NOD mice (n = 10) in advanced stage of disease progression (bgl 358 ± 25 mg/dL) (Fig. 2A) was transplanted with allogeneic (Balb-c) islets. Three of seven NOD mice (∼40%) that received CT5 stably maintained normoglycemia up to 250 days. Conversely, all untreated mice returned hyperglycemic within 10 days posttransplant (Fig. 2B). Overall, these data indicate CT5 treatment 1 day after syngeneic or allogeneic islet transplantation leads to maintenance of a persistent state of insulin independence supported by residual endogenous β-cell mass in 50% (8 of 16) of treated mice.Figure 1 LV.InsB and anti-CD3 mAb CT (CT5) after syngeneic islet transplantation (Tx) protects from recurrence of T1D. Glycemia was measured in NOD female mice (n = 19) to monitor T1D progression. Six hundred equivalent islets isolated from syngeneic (Syn) NSG mice were transplanted under the kidney capsule of diabetic NOD mice (average bgl ± SE 394 ± 17 mg/dL). One day after transplantation, all mice displayed a normal bgl and were treated with CT5 (LV.InsB + anti-CD3 mAb 1 × 5 μg) (n = 9), anti-CD3 alone (n = 3), or left untreated (n = 7). A: Baseline bgl of each mouse is reported, as well as average bgl ± SE. In the CT5 group, filled squares represent mice returned to hyperglycemia. B: Glycemia was monitored over time posttreatment, and percentages of mice maintaining normoglycemia (bgl <250 mg/dL) are reported (log-rank Mantel-Cox test P = 2 × 10−4, P <1 × 10−4). C: Pancreas and kidney hematoxylin and eosin (HE) and insulin immunostaining were performed, and one representative field is reported.Figure 2 LV.InsB and anti-CD3 mAb CT (CT5) after allogeneic islet transplantation protects from recurrence of T1D. Glycemia was measured in NOD female mice (n = 13) to monitor T1D progression. Six hundred equivalent islets isolated from allogeneic Balb-c mice were transplanted under the kidney capsule of recipient diabetic NOD mice (average ± SE bgl 358 ± 26 mg/dL, n = 10) or treated with anti-CD3 mAb alone (1 × 5 µg, no return to euglycemia, n = 3). One day after transplantation, mice displaying a normal bgl were treated with CT5 (LV.InsB9-23 + anti-CD3 mAb 1 × 5 µg, n = 7) or left untreated as control (n = 3). A: Baseline bgl of each mouse is reported as well as average bgl ± SE. In the CT5 group, filled squares represent mice returned to hyperglycemia. B: Glycemia was monitored over time posttreatment and percentages of mice maintaining normoglycemia (bgl <250 mg/dL) are reported (log-rank Mantel-Cox test P = 3.2 × 10−3).Optimized CT Before Allogeneic Islet Transplantation Allows Stable Insulin Independence in NOD Mice With Residual β-Cell MassWe previously observed that induction of Ag-specific Foxp3+ Tregs mediated by hepatocyte-restricted Ag expression requires several days (10). Therefore, to improve the efficacy of the CT, we administered CT before allogeneic islet transplant as a tolerogenic conditioning of the host immune system. We first tested whether CT could stabilize T1D in NOD mice, avoiding further loss of endogenous β-cells and consequent increase of glycemia. With CT5 (LV.InsB + 1 × 5 μg anti-CD3 mAb; n = 3, baseline bgl 398 ± 48 mg/dL), all NOD mice reached bgl >500 mg/dL, whereas treatment with CT25 (CT25: LV.InsB + 1 × 25 μg anti-CD3 mAb; n = 14, baseline bgl 331 ± 13 mg/dL) resulted in disease stabilization with only 5 of 14 NOD mice reaching bgl >500 mg/dL 7 days after treatment (Supplementary Fig. 1). Thus, CT25 was administered to diabetic NOD mice (n = 7, baseline bgl 379 ± 21 mg/dL) (Fig. 3A) 7–10 days before transplantation of allogeneic islets (bgl at the time of allotransplantation was 386 ± 55 mg/dL) (Fig. 3A). CT25 stabilized glycemia in five of seven NOD mice with glycemic levels lower than or comparable to baseline. Allogeneic islet transplantation normalized bgl (<250 mg/dL), which remained below the symptomatic threshold for 100 days (Fig. 3B). Nevertheless, CT25 failed in arresting T1D progression in two of seven NOD mice (Fig. 3A). In those mice, allogeneic islet transplant resulted in transient normoglycemia, similar to control untreated transplanted mice, or null as in mice treated with anti-CD3 alone (Fig. 3C). Histological analysis of pancreata and kidneys isolated from stable insulin-independent CT25-treated and transplanted (Tx) NOD mice (in Fig. 3A and B, T3-T7 hereafter will be named as normo-CT25+Allo-Tx) showed loss of transplanted islets and presence of endogenous insulin-producing cells in the pancreas (Fig. 3D). Furthermore, we determined the severity of pancreatic islet infiltration (insulitis) by scoring the grade of islet infiltration in normo-CT25+Allo-Tx (n = 4, 22 sections, 137 islets, bgl <196 ± 28 mg/dL) NOD mice. As a control, we analyzed diabetic (n = 6, 24 sections, 132 islets, bgl 416 ± 45 mg/dL), 10-week-old normoglycemic (n = 4, 13 sections, 368 islets, bgl <124 ± 17 mg/dL), and 5-week-old normoglycemic (n = 3, 12 sections, 321 islets, bgl ∼100 mg/dL) NOD mice (Fig. 3E). Although infiltrated, insulitis was comparable between normo-CT25+Allo-Tx and normoglycemic 10-week-old NOD mice and significantly less severe compared with diabetic NOD mice, mirroring the baseline at CT25 administration. Moreover, a comparable number of islets was found in normo-CT25+Allo-Tx mice and diabetic controls, indicating that the treatment blocked progression of insulitis rather than inducing β-cell mass regeneration. Overall, these data indicate that CT25 conditioning in mice having residual β-cells mass (bgl <500 mg/dL) when islet transplantation is performed induces a permanent recovery to a state of normal glucose homeostasis supported by resident pancreatic β-cells.Figure 3 Optimized LV.InsB and anti-CD3 mAb CT (CT25) provided before allogeneic islet transplantation protects mice from T1D recurrence. A: Glycemia was measured in NOD female mice (n = 10) to monitor T1D progression. Diabetic NOD mice (n = 7, baseline bgl 379 ± 21 mg/dL) were treated with CT25 (LV.InsB + anti-CD3 mAb 1 × 25 μg), and 7 days later, 600 equivalent islets isolated from allogeneic Balb-c mice were transplanted under the kidney capsule of CT25-treated NOD mice (n = 7) or untreated NOD mice as control (n = 3). B and C: After transplantation, glycemia was monitored over time and reported for each single mouse. C: The impact on T1D progression of anti-CD3 alone (1 × 25 μg) was also evaluated. D: Pancreas and kidney hematoxylin and eosin (HE) and insulin immunostaining were performed, and representative images are reported. E: Pancreatic islet infiltration (insulitis) was evaluated by Aperio Scan Software System in normo-CT25+Allo-Tx (n = 4, 22 sections, 137 islets, bgl <196 ± 28 mg/dL), diabetic (n = 6, 24 sections, 132 islets, bgl 416 ± 46), 10-week-old normoglycemic (n = 4, 13 sections, 368 islets, bgl <124 ± 17 mg/dL), and 5-week-old normoglycemic (n = 3, 12 sections, 321 islets, bgl ∼100 mg/dL) NOD mice (P = 1.85 × 10−6, P = 2.02 × 10−7 Fisher exact test [χ2]).Foxp3+ Tregs Homing to the Pancreas and PDL1 Costimulation Are Required for Active Control of Recurrence of Autoimmunity in CT25-Treated Diabetic NOD Mice Receiving Allogenic IsletsThe frequency of Foxp3+ Tregs was evaluated in islets and PLN as relevant sites for T1D autoimmune responses. Foxp3 expression was quantified in pancreatic islets of normo-CT25+Allo-Tx and in control diabetic NOD mice by immunohistochemistry (Supplementary Fig. 2). Results showed that Foxp3-expressing cells were significantly enriched within the infiltrated regions of islets in normo-CT25+Allo-Tx compared with infiltrated islets of diabetic NOD mice (Fig. 4A–C). Similarly, the ratio between the percentages of Foxp3-expressing Tregs and total T cells was significantly higher in PLN, but not in the spleen, of normo-CT25+Allo-Tx compared with normoglycemic mice and control mice receiving only allogeneic islet transplantation (Fig. 4D and E). In addition, the presence and expansion of InsB9-23–specific Tregs was investigated by IAg7.InsB9-23 tetramer immunostaining of PLN-lymphocytes from normo-CT25+Allo-Tx, and normoglycemic or diabetic untreated NOD mice as the control. Although the frequency of CD4+ IAg7.InsB9-23+ T cells was not different among the groups, the proportion of CD4+ IAg7.InsB9-23 Tregs expressing CD25 and low-negative for CD127 was significantly increased in normo-CT25+Allo-Tx (Supplementary Fig. 3).Figure 4 The proportion Foxp3+ Tregs within the T-cell compartment islet infiltration and in PLN is expanded in normo-CT25+Allo-Tx NOD mice. Foxp3 immunostaining was performed in pancreas sections from normo-CT25+Allo-Tx NOD mice (n = 4, 53 islets) and from diabetic controls (n = 5, 78 islets), and areas covered by Foxp3 staining were calculated by Aperio Scan Software System. The ratios between Foxp3 area and the area of the infiltration of each single mouse (P = 1.59 × 10−2 Mann-Whitney test) (A) and of each single islet are plotted (P = 9.2 × 10−7 Mann-Whitney test) (B), and representative images are reported (C). At the end of the observation time (>100 days of normoglycemia after islet transplantation) T-cell immunophenotyping of PLN (D) cells and splenocytes (E) was performed in normo-CT25+Allo-Tx (n = 5), normoglycemic (5-week-old n = 4, 10- to 14-week-old n = 12), diabetic (n = 7), CT25-treated (n = 3), or alloislet transplanted (n = 3) NOD mice. The ratio between the percentage of Foxp3+ CD4 T cells and percentage of total T cells within CD45+ cells is reported for each single mouse (P = 1.03 × 10−2 ANOVA test).To further support the induction of Ag-specific immunoregulation, splenic T cells isolated from normo-CT25+Allo-Tx and diabetic NOD mice were restimulated in vitro to quantify responsiveness to InsB9-23 and to Allo-Ags (Balb-c splenocytes). These studies showed that T cells isolated from normo-CT25+Allo-Tx mice retained a minimal capacity to release IFN-γ and IL-17 in response to InsB9-23 stimulation, while displaying a strong allo-Ag–specific reactivity (Supplementary Figs. 4 and 5).To determine the role of Foxp3+ Tregs in promoting stable normoglycemia, splenocytes (5 × 106 cells/mouse) from normo-CT25+Allo-Tx at 127 days after treatment were adoptively transferred into NSG mice. In parallel, another group of NSG mice received splenocytes from normo-CT25+Allo-Tx NOD mice depleted of CD4+CD25+CD127− Tregs. As control, splenocytes isolated from 10-week-old normoglycemic and diabetic (bgl 340 mg/dL) NOD mice were transferred into NSG mice (Fig. 5A). As expected, splenocytes from 10-week-old normoglycemic NOD mice did not transfer T1D, while mice receiving cells from diabetic NOD mice developed T1D within 30 days. Conversely, a significant reduction in T1D incidence, with 60% of mice disease-free for >50 days after transfer, was observed in NSG mice injected with splenocytes from normo-CT25+Allo-Tx NOD mice (Fig. 5B). In line with the described T-cell hyporesponsiveness to InsB9-23, a significant delay in T1D onset was observed in NSG mice transferred with Treg-depleted splenocytes from a CT25+Allo-Tx mouse compared with mice receiving diabetogenic control cells (Supplementary Fig. 4).Figure 5 A: Foxp3+ Tregs in normo-CT25+Allo-Tx are necessary to abrogate T1D recurrence. Gating strategy to identify and deplete Foxp3+ Tregs from total splenocytes without intranuclear Foxp3 staining (upper panel). Of note, Foxp3+ Tregs fall in the CD4+CD127−CD25+ region (red contour plot) (lower panel, control Foxp3 staining), while CD4+CD127+CD25− T conv (blue contour plot) and splenocytes CD4− (green contour plot) are low/negative for FoxP3 and CD25 expression. B: Total splenocytes (red dashed line) or splenocytes from Treg-depleted (red solid line) normoglycemic CT25-treated and transplanted mice (normoglycemic for >100 days after bgl = 146 mg/dL) and splenocytes from normoglycemic 10-week-old NOD mice (pink line) or diabetic NOD mice (blue line, bgl = 340 mg/dL) were transferred (5 × 106 cells) into female NSG recipient mice, and T1D development was monitored measuring bgl over time.Studies conducted in vitro demonstrated that hepatocytes upregulate both PDL1 mRNA and protein expression in response to type II interferon (IFN-γ) (Supplementary Fig. 6B and C). Therefore, we tested in vivo the relevance of the PD1-PDL1 pathway in the mechanism of tolerance induction providing anti-PDL1 blocking mAb after LV gene transfer to hepatocytes. Data indicate that blockade of the PDL1 costimulatory pathway results in the loss of Ag-expressing hepatocytes (Supplementary Fig. 7A–C), due to expansion of Ag-specific CD8+ T cells (Supplementary Fig. 7D and E), thus demonstrating that the PD1-PDL1 pathway plays a key role in the establishment of Ag-specific tolerance after hepatocyte-directed gene transfer. These data provided a strong rationale to investigate the role of the PDL1 costimulatory pathway in the tolerogenic mechanism acting in normo-CT25+Allo-Tx NOD mice. Therefore, we treated diabetic mice with the optimized CT25, followed by allogeneic islet transplantation (as done in Fig. 3A and B), and monitored the efficacy of the treatment for at least 30 days after transplantation. Thereafter, normo-CT25+Allo-Tx NOD mice underwent nephrectomy to remove the transplantation site (Fig. 6). Blood glucose level of normo-CT25+Allo-Tx NOD mice remained unaltered for 1–4 days after nephrectomy, thus demonstrating that the endogenous β-cell mass supported glucose homeostasis and confirming our conclusions driven by the histological analysis of pancreas and kidney (Fig. 3D and E). Subsequently, we treated normo-CT25+Allo-Tx NOD mice with anti-PDL1 mAb, and the immunoregulation imposed after CT25+Allo-Tx was abrogated (Fig. 6).Figure 6 Glucose homeostasis in normo-CT25+Allo-Tx NOD mice is supported by endogenous β-cell mass, and the PD1-PDL1 costimulatory pathway plays a role controlling recurrence of autoimmunity. Glycemia was measured in NOD female mice (n = 4) to monitor T1D progression. Diabetic NOD mice (n = 4, baseline bgl 315 ± 8 mg/dL) were treated with CT25 (LV.InsB + anti-CD3 mAb 1 × 25 μg). Seven to ten days later, CT25-treated NOD mice were transplanted (bgl before the transplant 315 ± 49 mg/dL) with allogeneic pancreatic islets isolated from Balb-c mice under the kidney capsule. After transplantation, glycemia was monitored over time and reported for each single mouse. After at least 30 days of stable glycemia <250 mg/dL, the transplantation site was removed by nephrectomy to determine the role of exogenous transplanted islets. Glycemia was monitored after nephrectomy for 1–4 days (gray areas), and then mice were treated intraperitoneally with anti-PDL1 blocking mAb to assess the role of the PD1-PDL1 costimulatory pathway in impeding the recurrence of autoimmunity in normo-CT25+Allo-Tx NOD mice.Immunophenotyping of mononuclear cells isolated from the kidney after surgery revealed a significant enrichment of CD8+ T cells compared with normoglycemic or diabetic untreated NOD controls (Supplementary Fig. 8A). Moreover, we investigated the presence of InsB9-23–specific CD4+ T cells and InsB15-23–specific CD8+ T cells using IAg7.InsB9-23 and H2Kd.InsB15-23 tetramers, respectively. Results indicate that insulin-specific autoreactive T cells are present at the site of transplantation (Supplementary Fig. 8B–D) and, to a lesser extent, also in the liver (harvested at the end of the experiment) (Supplementary Fig. 8E–H).Overall, these results revealed a complex mechanism of immune regulation driven by CT25 in which both Tregs and the PDL1 costimulatory pathway cooperate to keep under control diabetogenic cells, while the lack/blockade of one of these two components leads to recurrence of autoimmunity. The Ag-specific nature of the therapy is highlighted by loss of T-cell reactivity to InsB9-23 but not to allo-Ags. Conversely, transplanted islets play a crucial role, although transient, as site mediating recruitment of diabetogenic cells.Allogeneic Islet Transplantation Is Necessary for Stable Glucose HomeostasisThe absence of a persistent exogenous β-cell mass transplanted under the kidney capsule (Fig. 1C and Fig. 3D) upon CT treatment prompted us to test whether islet transplantation might be replaced by temporary insulin replacement therapy (IRT). Slow-release insulin pellets were implanted subcutaneously 7 days after CT25 conditioning of diabetic NOD mice (bgl 376 ± 42 mg/dL) (Fig. 7A). Despite IRT allowing an efficient normalization of bgl in four of five mice for 14–21 days, all CT25-treated mice became hyperglycemic 28 days after the insulin pellet implantation (Fig. 7B). In agreement with the lack of bgl normalization after CT25/allotransplantation treatment (see Fig. 3B mice T1 and T2), NOD mice displaying bgl >500 mg/dL at the time of IRT experienced a suboptimal glycemic normalization, which was lost in <3 weeks after insulin pellets implantation. These data indicate that the transplant procedure is required to achieve active suppression of autoimmunity leading to definitive diabetes reversal, which occurs regardless of insulin provision or the induction of an efficient alloresponse.Figure 7 IRT in CT25-conditioned diabetic NOD mice does not control autoimmunity. Glycemia was measured in NOD female mice (n = 5) to monitor T1D progression. Diabetic NOD mice (average ± SE bgl 376 ± 42 mg/dL) were treated with CT25 (LV.InsB9–23 + anti-CD3 mAb 1 × 25 μg) (n = 5) 7 days before the intradermal implantation of insulin pellets. A: Bgl of each mouse is reported, as well as average bgl ± SE of each group of treatment before CT25 (baseline) and before pellet implant. B: After pellet implant, glycemia was monitored over time and is reported for each single mouse. sc, subcutaneous.DiscussionSuccessful transplantation requires the prevention of allograft rejection and, in the case of transplantation to treat autoimmune disease, the suppression of autoimmune responses (8). Allograft rejection is successfully controlled by standard immune-suppressive regimens that, in the case of islet transplantation, are usually several alternative Edmonton-like protocols (19). With respect to autoimmunity recurrence, an Ag-specific treatment to promote self-tolerance would greatly improve the outcome of islet transplantation. In this scenario, our LV gene transfer–based Ag-specific therapy was applied to control autoimmunity by targeting to hepatocytes the expression of a single epitope of insulin coupled with a minimal dose of T cell–depleting/inactivating mAb (CT). Our study shows that NOD mice at a very advanced stage of disease progression (up to bgl 500 mg/dL) still have residual β-cell mass, potentially sufficient to recover glucose homeostasis, if autoimmunity is suppressed. Indeed, we observed that below that bgl threshold, CT25 administered 1 week before islet transplantation promoted an increase of the ratio between Tregs and T cells in islet infiltrates and in pancreas-draining LNs, minimizing T-cell effector functions evocated by InsB9-23 stimulation and resulting in milder insulitis and stable normoglycemia, even beyond the persistence of transplanted allogeneic islets.The concomitant induction of allospecific responses and active suppression of autoimmunity in cured NOD mice by CT25 treatment and allogeneic islet transplantation explains how transplanted islets are eradicated and in turn supports Ag-specificity of immune regulation mediated by LV.InsB gene transfer to hepatocytes (16).T cells responsive to β-cell–related Ags preferentially home to sites where the cognate Ag is presented (20–22). We hypothesize that transplantation of exogenous β-cells and InsB9-23 expression in hepatocytes in NOD mice treated with CT25 and transplanted may favor spreading of diabetogenic T cells, thus favoring the reduction of immune pressure toward endogenous islets via a dilution of diabetogenic effector T cells toward multiple target organs and lymphoid draining stations (i.e., pancreas-PLN, liver-hepatic LN, and transplant under kidney capsule and renal LN). In a second phase when transplanted islets are deleted by allo- and islet-specific T cells, the ratio between Tregs and effector T cells may rise in PLN and pancreas. This mechanism might explain why responses toward insulin and alternative autoAgs did not prevail, leading to long-term disease-free survival.Gagliani et al. (23) previously proposed the combination of allogeneic islet transplantation and a full dose of anti-thymoglobulin (mALS), a T cell–depleting agent, as resolutive treatment to revert T1D in NOD mice at disease onset (bgl >250 mg/dL). Although this treatment was effective when administered at the beginning of the symptomatic phase, it led to an almost complete T-cell depletion, severely impacting the host immune functions. On the contrary, the CT25 formulation includes a single dose of anti-CD3 mAb with a proven minimal impact on the host immune system (15,24), synergizing with LV gene transfer to hepatocytes for Ag-specific Treg induction and effector T-cell compartment contraction, rather than massive T-cell depletion.Assuming that NOD mice with 350–500 mg/dL of blood glucose level resemble patients with overt T1D with a minimal but residual β-cell mass, the strategy proposed here might be applied to patients with T1D in advanced stage of the disease with a residual insulin production capacity. Moreover, the application of the CT alone or CT in conjunction with subtherapeutic islet transplantation can be envisaged as a therapeutic option, respectively, for subjects at risk or early-onset patients, respectively. Herold et al. (25) recently showed clinical findings (phase II randomized clinical trial) of the 14-day teplizumab (non–Fc-binding anti-CD3 mAb) treatment in genetically predisposed high-risk subjects (autoantibodies and metabolic dysfunction). Treatment with teplizumab alone led to a significant delay of T1D onset but did not protect patients from disease progression (25). Instead, the synergy of anti-CD3 mAb and InsB9-23 gene transfer to hepatocytes relies on several common mechanisms of action, including active Ag-specific tolerance via Ag-specific Tregs (10,14), depletion of T effector (10,26), transforming growth factor-β upregulation (27) (Supplementary Fig. 6A), and PDL1 inhibitory costimulation (28–30) (Supplementary Fig. 7). Based on this, we believe that it could potentially be exploited to definitively arrest disease development (or revert T1D at early onset) in genetically predisposed high-risk subjects who still have a fraction of endogenous pancreatic β-cells.Potential limitations of our strategy may derive from in vivo administration of LV and from the choice of the autoAg. LVs are integrating vectors with intrinsic genotoxic risk. However, hematopoietic stem cell ex vivo gene therapy clinical trials conducted to correct genetic defects in primary inherited immunodeficiency (31) and in monogenic hematologic and lysosomal storage disorders (32–34) were proven effective and safe after several years of follow-up. Specifically, these studies did not show signs of genotoxicity in highly replicating cells, such as hematopoietic stem cells, with a pattern of integration in active genes, typical of the LV (35). Moreover, we recently reported safety and efficacy of the LV platform for in vivo gene transfer to hepatocytes. We showed that in vivo administration of LV is well tolerated in nonhuman primates and results in stable expression of the transgene with a safe pattern of integration without signs of clonal expansion, thus paving the way for clinical trials of in vivo LV gene therapy (36).The hierarchy of the autoAgs is not as well defined in humans as it is for NOD mice, in which T1D development is mostly driven by insulin-specific T-cell responses. However, we foresee a clinical translation of the CT approach with the possibility of personalizing the protocol according to immunodominant autoAgs identified by prescreening of patient’s T-cell reactivity.Overall, our studies indicate that targeting the expression of an autoAg immunodominant epitope to the liver parenchyma can be the base for a “one shot” therapy to control autoimmunity, which can be combined with minimal amounts of a T cell–depleting/inactivating mAb to prevent disease progression in subjects at risk, to reverse diabetes in patients with early-onset T1D, and to improve the efficacy of allogeneic islet transplantation for patients in advanced stage of disease development.Article InformationAcknowledgments. The authors thank R. Norata and V. Mauro for technical support in histological analyses (San Raffaele Telethon Institute for Gene Therapy-GLP team), A. Nonis, F. Cugnata, and C. Di Serio for statistical consulting (Centro Universitario di Statistica per le Scienze Biomediche), L. Passerini and F.R. Santoni De Sio and all other members of the Mechanisms of Peripheral Tolerance Unit at San Raffaele Telethon Institute for Gene Therapy for helpful discussions, and T. Jofra and M. Battaglia at San Raffaele Diabetes Research Institute for supporting us in the acquisition of murine islet transplantation methodology. The authors thank L. Passeri and R. Curto for technical support; L. Naldini, A. Cantore, and M. Milani at San Raffaele Telethon Institute for Gene Therapy in vivo gene therapy unit for scientific discussion; and the National Institutes of Health (NIH) Tetramer Core Facility (TCF) at Emory University for provision of tetramers.Funding. This work was supported by the Italian Telethon Foundation (SR-Tiget Core Grant G2 2015-2020).Duality of Interest. A.A. and M.G.R. are inventors on a filed patent (WO2010055413 A1) on induction of Ag-specific tolerance by LV gene transfer to hepatocytes. No other potential conflicts of interest relevant to this article were reported.Author Contributions. F.R., A.C., G.S., and A.A. designed and performed experiments and analyzed data. F.S. coordinated histological analyses. P.M., S.G., and M.G.R. interpreted data and edited the manuscript. A.A. coordinated the project, interpreted data, and wrote the manuscript. A.A. is the guarantor of this work, and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.Received December 18, 2019.Accepted October 22, 2020.© 2020 by the American Diabetes Association

Via Source link