Anti-TCRβ mAb in Combination With Neurogenin3 Gene Therapy Reverses Established Overt Type 1 Diabetes in Female NOD Mice

Abstract

Insulin-producing β cells in patients with type 1 diabetes (T1D) are destroyed by T lymphocytes. We investigated whether targeting the T-cell receptor (TCR) with a monoclonal antibody (mAb) abrogates T-cell response against residual and newly formed islets in overtly diabetic nonobese diabetic (NOD) mice. NOD mice with blood glucose levels of 250 to 350 mg/dL or 350 to 450 mg/dL were considered as new-onset or established overt diabetes, respectively. These diabetic NOD mice were transiently treated with an anti–TCR β chain (TCRβ) mAb, H57-597, for 5 days. Two weeks later, some NOD mice with established overt diabetes further received hepatic gene therapy using the islet-lineage determining gene Neurogenin3 (Ngn3), in combination with the islet growth factor gene betacellulin (Btc). We found that anti-TCRβ mAb (50 µg/d) reversed >80% new-onset diabetes in NOD mice for >14 weeks by reducing the number of effector T cells in the pancreas. However, anti-TCRβ mAb therapy alone reversed only ∼20% established overt diabetes in these mice. Among those overtly diabetic NOD mice whose diabetes was resistant to anti-TCRβ mAb treatment, ∼60% no longer had diabetes when they also received Ngn3-Btc hepatic gene transfer 2 weeks after initial anti-TCRβ mAb treatment. This combination of Ngn3-Btc gene therapy and anti-TCRβ mAb treatment induced the sustained formation of periportal insulin-producing cells in the liver of overtly diabetic mice. Therefore, directly targeting TCRβ with a mAb potently reverses new-onset T1D in NOD mice and protects residual and newly formed gene therapy–induced hepatic neo-islets from T-cell‒mediated destruction in mice with established overt diabetes.

Directly targeting TCRβ on T cells with a mAb reverses T1D in NOD mice by protecting residual and newly formed gene therapy–induced hepatic neo-islets from T-cell‒mediated destruction.

Insulin replacement is the standard therapy for type 1 diabetes (T1D). Unfortunately, exogenous insulin fails to recapitulate the normal pancreatic insulin dynamics and, despite the best available insulin regimens, individuals with T1D continue to display significantly increased mortality and morbidity. Patients with T1D manifest the disease when the majority of the insulin-producing pancreatic β cells are destroyed by T-cell-mediated autoimmunity. In theory, T-cell targeted therapy is the treatment of choice because it addresses the primary defect in T1D and has the potential to restore β-cell survival and function (1).

A central step to initiate a T-cell response is the engagement of T-cell receptor (TCR) on T cells with antigen-loaded major histocompatibility complex on antigen-presenting cells. In addition, costimulatory and cytokine signals also control T-cell activation (1). In the context of T1D, efforts have been made to target islet antigens (2–4), deplete T cells (5), alter TCR signaling (1, 6), block costimulation (7), and neutralize cytokines (8). Nevertheless, much less attention has been paid to the TCR itself as a target of immunotherapy.

We recently investigated the therapeutic effects of a monoclonal antibody (mAb) specific for TCR β chain (TCRβ) and showed that transient anti-TCRβ mAb treatment induced long-term cardiac allograft survival in fully major histocompatibility complex–mismatched recipient mice as well as reversed new-onset diabetes in nonobese diabetic (NOD) mice (9, 10). Moreover, anti-TCRβ mAb also significantly prolonged the survival of transplanted C57BL/6 islets in NOD mice, which exerted not only allogeneic response but also autoimmune response against the transplanted islets (11). Thus, anti-TCRβ mAb has strong therapeutic potential for treatment of T1D by abrogating the anti-islet immunity.

Patients with new-onset T1D have been shown to harbor considerable numbers of residual β cells in the pancreas. Potential immunotherapies may, in theory, partly reverse disease in these patients by protecting the residual β cells. By contrast, islet β cells in patients with late-stage T1D have mostly been destroyed. At this stage, pancreas transplantation would restore glycemic and metabolic control much more effectively. However, the limited availability of donor pancreases limits the number of pancreas transplantations that can be performed. Thus, much attention has been devoted lately to replenishing new glucose-responsive β cells, such as replication of residual β cells and generation of new β cells in vitro or in vivo (12).

We have previously shown that ectopic islet neogenesis in the periportal regions of the liver is achieved by transfer of the islet-lineage determining gene Neurogenin3 (Ngn3), and the islet growth factor gene betacellulin (Btc) (13). A helper-dependent adenoviral (HDAd) vector was used to deliver the Ngn3 and Btc genes. These de novo generated neo-islets secrete insulin in a glucose-responsive manner to restore glucose control in streptozotocin-induced diabetic B6 mice, but not in NOD mice in which islet-destructive T cells continue to exist (13, 14). The aim of this study was to determine whether anti-TCRβ mAb reverses T1D in NOD mice and protects Ngn3/Btc-induced neogenesis from T-cell‒mediated destruction. We believe effective therapies for de novo islet generation in patients with T1D must be supported by a potent immunotherapy that halts T-cell attack.

Research Design and Methods

Mice

C57BL/6, NOD, and NOD/scid mice were purchased from the Jackson Laboratory (Bar Harbor, ME). All animal procedures were approved by the Institutional Animal Care and Use Committee at Baylor College of Medicine and the Institutional Animal Care and Use Committee at Houston Methodist Research Institute.

Diabetes diagnosis and intervention

Starting from 10 weeks of age, NOD female mice were screened for hyperglycemia three times per week. NOD mice were diagnosed with new-onset diabetes when two consecutive glucose levels were between 250 and 350 mg/dL, and with established overt diabetes when two consecutive glucose levels were between 350 and 450 mg/dL. To reverse new-onset diabetes, newly diagnosed mice were injected intraperitoneally (IP) with an anti-TCRβ mAb (clone H57 597; Bio X Cell, West Lebanon, NH), an anti-TCRα mAb (clone H28 710; Santa Cruz Biotechnology, Dallas, TX), or phosphate-buffered saline (PBS) control for 5 consecutive days. To reverse overt diabetes, overtly diabetic NOD mice were treated with 50 μg/d anti-TCRβ mAb for 5 consecutive days. If overtly diabetic mice maintained blood glucose levels between 350 and 450 mg/dL at day 14 after initial anti-TCRβ mAb treatment, they were treated with 5 × 10¹¹ vector particles of HDAd-Ngn3 plus 1 × 10¹¹ HDAd-Btc to further reverse diabetes, or with empty control vectors, as previously described (13–15).

Intraperitoneal glucose tolerance test and serum insulin level

Four to 6 weeks after initial anti-TCRβ mAb treatment, NOD mice that had been fasted for 4 hours were injected IP with 1.5 g/kg d-glucose. Blood glucose levels were assessed at 0, 15, 30, 60, and 120 minutes after glucose injection. Additional blood samples were obtained at the same time points to determine the serum insulin level, using the Mercodia Ultrasensitive Mouse Insulin ELISA (enzyme-linked immunosorbent assay) kit (Winston-Salem, NC).

Insulitis score, immunohistochemistry, immunofluorescence staining, and pancreatic insulin content

Pancreata were harvested at 14 weeks after initial anti-TCRβ mAb treatment. Paraffin sections were prepared and stained with hematoxylin and eosin. An insulitis score for each islet was determined as follows: 0, no insulitis; 1, peri-insulitis; 2, intraislet insulitis affecting <50% of islet area; and 3, intraislet insulitis affecting >50% of islet area. Immunostaining of insulin was performed on paraffin-embedded pancreas and liver sections, as previously described (13, 15). Immunofluorescence staining of CD4, CD8, and insulin with 4′,6-diamidino-2-phenylindole was performed on frozen sections of pancreas. The staining method has been previously described (13, 15). For measuring pancreatic insulin content, snap-frozen pancreata were homogenized in precooled acid ethanol overnight. Insulin concentration in supernatants was measured using the aforementioned Mercodia enzyme-linked immunosorbent assay and total protein content was determined with the 2D Quant kit (GE Healthcare, Piscataway, NJ).

Isolation of pancreatic infiltrating lymphocytes and flow cytometric analysis

The pancreas of each killed NOD mouse was perfused with 2 mL of Liberase TL enzyme (1.08 Wünsch units /mL; Roche, Indianapolis, IN) through the common bile duct while the duodenum entry was clamped. The pancreas was removed and incubated in a 37°C water bath for 20 minutes, then washed and pushed through a 40-μm cell strainer to obtain a suspension. Pancreatic infiltrating lymphocytes were then isolated from the suspension using a Lympholyte-M solution (Cedarlane Laboratories, Burlington, NC).

Lymphocytes isolated from the pancreas, spleen, and pancreatic draining lymph nodes (PLNs) were surface stained with various fluorescent antibodies and intracellularly stained with Foxp3 using Mouse Regulatory T-Cell Staining Kit (eBioscience, San Diego, CA). For IFN-γ and TNF-α expression, lymphocytes isolated from the pancreas were stimulated for 4 hours with 50 ng/mL phorbol 12-myristate 13-acetate and 500 ng/mL ionomycin (both from Sigma-Aldrich, St. Louis, MO) in the presence of GolgiStop (BD Biosciences, San Diego, CA), and stained with fluorescent antibodies against cytokines using the Cytofix/Cytoperm Solution Kit (BD Biosciences). Flow cytometry analysis was performed using a FACS Canto II analyzer (BD Biosciences).

CD4 and CD8 T-cell numbers in pancreas were calculated by multiplying the total live cell counts isolated from pancreas (TC 20 Automated Cell Counter; Bio-Rad, Hercules, CA) by percentage of CD4⁺ or CD8⁺ cells within the Zombie Aqua negative living cells (FACS Canto II analyzer; BD Biosciences).

Skin transplantation

Skin transplantation was performed as described previously (9). Briefly, ear skin (1.0 cm²) from the donor mice was grafted onto the flank of recipient mice. The graft was covered with a sterile bandage, which was removed on day 6 or 7. A skin graft was considered rejected when 80% of the grafted tissue became necrotic.

Adoptive transfer of diabetes into NOD/scid mice

NOD/scid mice were injected intravenously with 1 × 10⁷ total or CD25^‒ splenocytes isolated from anti-TCRβ–treated NOD mice (at 14 weeks after initial treatment), or 1 × 10⁷ total splenocytes from nondiabetic NOD mice. After adoptive cell transfer, the incidence of diabetes in these NOD/scid mice was monitored by screening for hyperglycemia three times per week.

Statistical analysis

The results of the graft survival data were analyzed by Mann-Whitney test. Other statistical analysis was performed using an unpaired, two-tailed, Student t test to calculate P values. Calculated P < 0.05 was considered statistically significant.

Results

Anti-TCRβ mAb effectively reversed new-onset diabetes in NOD mice

NOD mice were diagnosed with new-onset diabetes when two consecutive glucose levels were between 250 and 350 mg/dL. To reverse new-onset diabetes, newly diagnosed NOD mice were injected IP daily with either an anti-TCRα or an anti-TCRβ mAb (10 or 50 μg) for 5 consecutive days. New-onset diabetic mice injected IP with PBS served as a control group. Mice were then monitored for blood glucose levels for 14 weeks after treatment (i.e., the duration of the study). The anti-TCRβ mAb treatment restored euglycemia in 83.33% (n = 6) of the mice in the high-dose group and 50% (n = 6) among the low-dose group (Fig. 1a and 1b). Anti-TCRα mAb, 10 μg/dose (Fig. 1a and 1b) or 50 μg/dose (data not shown), did not reverse new-onset diabetes, though it marginally delayed the progression of diabetes compared with the PBS group (Fig. 1a and 1b). Therefore, anti-TCRβ mAb potently induces a long-term reversal of new-onset diabetes in NOD mice.

Figure 1.

Anti-TCRβ mAb reversed new-onset diabetes in NOD mice. NOD mice diagnosed with new-onset diabetes (two consecutive blood glucose levels between 250 and 350 mg/dL) were injected IP with anti-TCRβ mAb (10 or 50 μg/d), anti-TCRα mAb (10 μg/d), or PBS control for 5 consecutive days. Blood glucose was monitored every week until 14 weeks after initial treatment. (a) Percentage of diabetes remission in indicated groups at different weeks after T1D onset. “Disease remission” is defined as glycemia <250 mg/dL after a 5-day course of treatments in diabetic NOD mice. **P < 0.01 vs PBS-treated group. (b) Blood glucose level of each NOD mouse in indicated groups at different weeks after diabetes onset.

Anti-TCRβ mAb reduced the severity of insulitis and preserved β-cell mass in NOD mice

To study the effect of anti-TCRβ mAb on preserving β-cell function, we performed an intraperitoneal glucose tolerance test (IPGTT; n = 6 per group) 4 to 6 weeks after the initial anti-TCRβ treatment. PBS-treated NOD mice with full-blown diabetes (i.e., blood glucose level >600 mg/dL) were used as controls. As shown in the left panel of Fig. 2a, blood glucose levels during IPGTT in PBS-treated NOD mice were markedly higher than that in anti-TCRβ–treated groups and nondiabetic NOD mice. Moreover, anti-TCRβ–treated mice displayed significantly higher serum insulin levels before IPGTT than PBS-treated mice, but they failed to increase insulin secretion in responsive to glucose injection (Fig. 2a, right panel). We also measured the pancreatic insulin content in anti-TCRβ–treated groups at 14 weeks after treatment or in PBS-treated mice with full-blown diabetes. We found that pancreata of anti-TCRβ–treated mice exhibited a significantly higher insulin concentration than those of PBS-treated NOD mice with full-blown diabetes (Fig. 2b; n = 6 per group).

Figure 2.

Anti-TCRβ mAb improved glucose tolerance, preserved pancreatic insulin content, and limited insulitis in NOD mice. (a) Anti-TCRβ mAb–treated NOD mice (4 to 6 weeks after initial treatment; 10 or 50 μg/dose) or PBS-treated NOD mice were fasted and injected with d-glucose for IPGTT analysis. Blood glucose levels (left panel) and serum insulin concentration (right panel) were shown at the indicated time points after glucose injection. (b–e) Anti-TCRβ mAb-treated mice were assessed at 14 weeks after initial treatment. PBS-treated mice with full-blown diabetes served as a control group. (b) Pancreatic insulin content is expressed as insulin per milligram of protein. **P < 0.01 vs PBS-treated group. (c) Representative H&E and insulin staining of pancreatic islets in each group. (d) Insulitis scores of each group after evaluating >100 islets per group. *P < 0.05 vs PBS-treated group. (e) Representative immunofluorescence staining of CD4 (red), CD8 (green), insulin (white), and 4′,6-diamidino-2-phenylindole DAPI (blue) of pancreatic islets. (f) The numbers of pancreas-infiltrating CD4 and CD8 T cells in NOD mice treated with 50 µg of anti-TCRβ mAb at 10 days after initial treatment, as well as nondiabetic NOD mice and PBS-treated NOD mice exhibiting hyperglycemia. *P < 0.05 PBS- vs anti-TCRβ mAb-treated group. H&E, hematoxylin and eosin.

We then performed histochemical and immunofluorescence analyses of pancreata from the anti-TCRβ–treated groups at 14 weeks after treatment and PBS-treated NOD mice with full-blown diabetes. Anti-TCRβ mAb treatment significantly decreased the severity of insulitis in NOD mice compared with PBS treatment (Fig. 2c–2e). Moreover, the residual islets protected by anti-TCRβ mAb treatment displayed easily detectable insulin immunoreactivity (Fig. 2c, lower panels) and a reduction of the CD4 and CD8 T-cell infiltration (Fig. 2e).

Three groups of NOD mice were used to determine the numbers of pancreas-infiltrating T cells, including NOD mice treated with the 50-µg dose of anti-TCRβ mAb at 10 days after initial treatment, nondiabetic NOD mice, and PBS-treated NOD mice exhibiting hyperglycemia. The numbers of pancreas-infiltrating CD4 and CD8 T cells were significantly reduced in anti-TCRβ mAb–treated mice than those in PBS-treated NOD mice (Fig. 2f). Taken together, transient anti-TCRβ mAb treatment during new-onset diabetes preserved functional residual islets in pancreata of NOD mice.

Anti-TCRβ mAb significantly changed the phenotype of infiltrating T cells in pancreata of NOD mice

Murine naïve T cells highly express l-selectin (also known as CD62L) but not CD44. Upon activation, T cells downregulate CD62L and upregulate CD44 expressions. Herein, to determine how anti-TCRβ mAb halts T-cell–mediated islet destruction, we measured CD62L and CD44 expression on T cells from the spleen, PLNs, and pancreas of NOD mice. Three groups of NOD mice were assessed, including anti-TCRβ mAb–treated NOD mice at 14 weeks after T1D reversal, nondiabetic NOD mice at 26 to 32 weeks of age, and PBS-treated NOD mice exhibiting hyperglycemia. We found that, in the spleen and PLN, the majority of CD62L^‒ cells were CD44⁺ in CD4 and CD8 T cells (Fig. 3a, i and ii, top and middle panels), and the frequencies of CD62L^‒CD44⁺ cells were similar between groups (Fig. 3b, i and ii, left panels). Anti-TCRβ mAb also did not dramatically alter the total T-cell numbers in spleen and PLNs of NOD mice (data not shown).

Figure 3.

Anti-TCRβ mAb changed the phenotypes of pancreatic infiltrating T cells. (a, b) Spleen, PLN, and pancreatic infiltrates were obtained from anti-TCRβ mAb–treated NOD mice (50 μg/dose) at 14 weeks after initial treatment. Nondiabetic and PBS-treated diabetic NOD mice were used as controls. (a) Flow cytometry analysis of CD44/CD62L expression on (i) CD4 and (ii) CD8 T cells, and (iii) CD25/Foxp3 expression of CD4 T cells. In (i) and (ii) bottom panels, the percentage of CD62L^‒ cells within pancreatic infiltrating CD4 and CD8 T cells are indicated. In (iii), the percentage of CD25⁺Foxp3⁺ cells among CD4 T cells are indicated. (b) In (i) and (ii), bar graphs show the percentage of CD44⁺CD62L^‒ cells within CD4 and CD8 T cells from spleen and PLN (left panels), and the percentage of CD62L^‒ cells among pancreatic infiltrating CD4 and CD8 T cells (right panels). In (iii), graphs show the percentage of CD25⁺Foxp3⁺ cells within the CD4 T-cell population. *P < 0.05 vs PBS treated group. (c–h) Pancreatic infiltrates were obtained from NOD mice treated with 50 µg of anti-TCRβ mAb at 10 days after initial treatment, as well as nondiabetic NOD mice and PBS-treated NOD mice exhibiting hyperglycemia. (c) Shown are gating strategies for analysis of pancreatic infiltrates and the percentage of CD69⁺ population within infiltrating CD4 and CD8 T cells. The black arrow in the left panel indicates the lymphocyte population. (d–h) Shown are the percentage of CD28⁺, OX40 mean fluorescence intensity, CD40L MFI, the percentage of PD-1⁺, and the percentage of IFN-γ⁺ within infiltrating CD4 and CD8 T cells. *P < 0.01; unpaired Student t test. FSC-A, forward scatter, area; FSC-H, forward scatter, height; SSC, side scatter.

Pancreatic infiltrating CD62L^‒ T cells exhibited a diverse range of CD44 expression. The frequencies of CD62L^‒ cells among CD4 and CD8 pancreatic infiltrating T cells were significantly higher in PBS-treated diabetic mice than those in the nondiabetic and anti-TCRβ mAb groups (Fig. 3a, i and ii, bottom panels, and 3b, right panels). The frequencies of CD4⁺CD25⁺Foxp3⁺ regulatory T (Treg) cells in the second lymphoid organs and pancreata were not significantly different between the three experimental groups (Fig. 3a and 3b, iii).

We further analyzed the expression of activation/dysfunction markers and cytokine production by pancreatic infiltrating T cells. Three groups of NOD mice were studied, including 50 µg anti-TCRβ mAb–treated NOD mice at 10 days after initial treatment, nondiabetic NOD mice, and PBS-treated NOD mice exhibiting hyperglycemia. The percentages of CD69⁺ and CD28⁺ populations in infiltrating CD4 T cells and the percentage of CD28⁺ population in infiltrating CD8 T cells from the anti-TCRβ mAb-treated group were significantly higher than those of the nondiabetic group, but were not significantly different from those of the PBS-treated group (Fig. 3c and 3d). OX40 expression on infiltrating CD4 T cells from the anti-TCRβ mAb–treated group increased significantly than those of the nondiabetic group and the PBS-treated group (Fig. 3e). CD40L expression was not different between any groups (Fig. 3f).

Most interestingly, pancreatic infiltrating T cells can be divided into PD-1⁺ and PD-1^‒ populations. Most infiltrating T cells from the anti-TCRβ mAb group were PD-1⁺. The percentages of PD-1⁺ population in infiltrating CD4 and CD8 T cells from the anti-TCRβ mAb group were, indeed, significantly higher than those of the other two groups (Fig. 3g), but the percentages of IFN-γ‒producing population in them were significantly lower than those of the PBS-treated group (Fig. 3h). Taken together, the findings indicate anti-TCRβ mAb significantly changes the phenotype of infiltrating T cells in pancreata of diabetic NOD mice.

Anti-TCRβ mAb–treated NOD mice were immunocompetent to reject skin allografts and contained diabetogenic cells

To determine the immunocompetence of NOD mice after T1D reversal, we transplanted C57BL/6 skin allografts onto NOD mice that had been treated with anti-TCRβ mAb 12 weeks before (n = 6). Nondiabetic NOD mice were also transplanted with C57BL/6 skin and used as a control group (n = 6). We found that the C57BL/6 skin allografts were acutely rejected in anti-TCRβ mAb–treated mice, with a mean (± standard deviation) survival time of 8 ± 1 days, which was not different from that in nondiabetic NOD mice (Fig. 4a). Hence, NOD mice treated with anti-TCRβ mAb remained immunocompetent.

Figure 4.

Anti-TCRβ mAb–treated NOD mice rejected skin allografts and contained diabetogenic cells in spleen. (a) C57BL/6 skin allografts were transplanted onto the diabetes-reversed NOD mice at 12 weeks after anti-TCRβ mAb treatment, or onto the nondiabetic NOD mice. Percentage skin allograft survival is shown. (b) We transferred 1 × 10⁷ total splenocytes or CD25^‒ splenocytes obtained from anti-TCRβ mAb-treated NOD mice (14 weeks after initial treatment) or the same number of total splenocytes from nondiabetic NOD mice into NOD/scid mice. Incidence of diabetes in each group after cell transfer is shown. (c) Splenocytes of anti-TCRβ mAb–treated NOD mice were flow cytometrically sorted into a CD25^‒ population. Shown are CD25 and Foxp3 expression by CD4 T cells in total splenocytes or in CD25^‒ splenocytes.

We then used an adoptive transfer model of diabetes to determine whether anti-TCRβ mAb selectively eliminated the diabetogenic T cells. Splenocytes were isolated from either anti-TCRβ mAb–treated NOD mice at 14 weeks after T1D reversal or nondiabetic NOD mice at 26 to 32 weeks of age, and were adoptively transferred into NOD/scid recipient mice. As shown in Fig. 4b, diabetes was induced in both groups similarly, and all NOD/scid mice became diabetic within 6 weeks after cell transfer. Moreover, we partially depleted Treg cells in the splenocytes of anti-TCRβ mAb–treated NOD mice by isolating CD25^‒ cells (Fig. 4c). Transfer of those CD25^‒ splenocytes moderately accelerated T1D onset in NOD/scid mice (Fig. 4b). Therefore, diabetogenic T cells were not eliminated in anti-TCRβ mAb–treated NOD mice, and their diabetogenic potential might be partially inhibited by Treg cells.

Ngn3-Btc islet-induction gene therapy reversed overt diabetes in NOD mice resistant to anti-TCRβ mAb treatment

To further determine the potency of targeting TCR in reversing T1D, 29 NOD mice with established overt diabetes (i.e., blood glucose level of 350 to 450 mg/dL) were treated with 50 μg/d anti-TCRβ mAb for 5 consecutive days (Fig. 5a). Two weeks after the first dose of anti-TCRβ mAb, the following 16 mice were left without further treatment. (1) Eight NOD mice responded to anti-TCRβ mAb alone with decreasing blood glucose level (<350 mg/dL). Six of those eight NOD mice that responded to the anti-TCRβ mAb–alone treatment remained diabetes free for at least an additional 12 weeks. (2) Eight NOD mice became severely hyperglycemic (i.e., blood glucose level >450 mg/dL) and developed full-blown diabetes within 2 weeks. The dynamic changes of the blood glucose levels of these 16 mice are shown in the left panel of Fig. 5b. Hence, anti-TCRβ mAb therapy reversed only ∼20% (six of 29) of established overt diabetes in NOD mice.

Figure 5.

Anti-TCRβ mAb combined with Ngn3-Btc gene transfer reversed established overt diabetes in NOD mice. (a) Schematic of experiment: NOD mice diagnosed with established overt diabetes (blood glucose levels of 350 to 400 mg/dL) were injected with 50 μg/d anti-TCRβ mAb for 5 days. Two weeks after initial anti-TCRβ mAb treatment, if the treated mice maintained blood glucose levels between 350 and 450 mg/dL, they were transferred with Ngn3-Btc genes or control vector. (b) The left panel shows the blood glucose levels in anti-TCRβ mAb–treated mice without vector injection (dashed lines) or before vector injection (short solid lines). The right-side panels display blood glucose levels in anti-TCRβ mAb–treated mice that received Ngn3-Btc genes (top panel) or control vector (bottom panel). (c) Percentage remission of overt diabetes in anti-TCRβ mAb–treated NOD mice that further received Ngn3-Btc gene transfer or control vector. *P < 0.01 vs anti-TCRβ plus control vector group.

As shown in Fig. 5b (left panel, solid lines), 13 NOD mice maintained blood glucose levels between 350 and 450 mg/dL at 2 weeks after the first dose of anti-TCRβ mAb. This relative resistance of overtly diabetic NOD mice to anti-TCRβ mAb therapy may be attributed to the severe loss of pancreatic islets. Therefore, we tested the efficacy of supplying these mice with neo-islets by Ngn3-Btc gene therapy using a method that was previously developed by our laboratory (13–15). These 13 NOD mice were selected for vector injection at 2 weeks after the first dose of anti-TCRβ mAb (Fig. 5b, left panel, short solid lines); eight of them were injected with HDAd vectors encoding Ngn3 and Btc (Fig. 5b, top right panel), whereas five other mice were injected with empty control vectors (Fig. 5b, bottom right panel). We found that all animals in the control-vector group developed full-blown diabetes in 8 weeks, whereas five of eight NOD mice in the Ngn3-Btc gene therapy group no longer were diabetic and remained diabetes free during the 14-week study period (Fig. 5b, right panels, and 5c).

We performed IPGTT (n = 5 per group) in three groups of mice at 4 to 6 weeks after initial anti-TCRβ mAb treatment: (1) overt diabetes reversed by anti-TCRβ mAb alone, (2) overt diabetes reversed by anti-TCRβ mAb plus Ngn3-Btc gene transfer, and (3) full-blown diabetes that developed after treatment with anti-TCRβ mAb plus control vectors. Blood glucose levels during IPGTT in group 3 were markedly higher than that in groups 1 and 2 (Fig. 6a, upper panel). Moreover, during IPGTT, serum insulin levels in groups 1 and 2 were also significantly higher than those in group 3, though insulin secretion did not increase in response to glucose injection in all three groups (Fig. 6a, lower panel). We also monitored the serum insulin level before diabetes onset and at different times after treatment. Both groups 2 and 3 were relatively resistant to anti-TCRβ mAb and their serum insulin levels declined dramatically at 2 weeks after initial anti-TCRβ mAb treatment. Nevertheless, at 4 weeks after initial anti-TCRβ mAb treatment (2 weeks after vector injection), the serum insulin levels in group 2 were partially restored after Ngn3-Btc gene transfer, whereas those of group 3 further declined to almost undetectable levels (Fig. 6b). Figure 6c shows that in the liver of NOD mice treated with anti-TCRβ mAb and Ngn3-Btc gene transfer, insulin-expressing cells were present surrounding the portal vein in the liver. In contrast, insulin-expressing cells were never detected in the liver of mice treated with control vectors (Fig. 6c). Taken together, anti-TCRβ mAb combined with Ngn3-Btc gene transfer reversed established overt diabetes in NOD mice.

Figure 6.

Anti-TCRβ mAb plus Ngn3-Btc gene transfer improved glucose tolerance, increased serum insulin level, and induced new insulin-producing cells in overtly diabetic NOD mice. Three groups of mice were included in this study: (1) overt diabetes reversed by anti-TCRβ mAb alone, (2) overt diabetes reversed by anti-TCRβ mAb plus Ngn3-Btc gene transfer, and (3) full-blown diabetes developed after treating with anti-TCRβ mAb plus control vectors. (a) At 4 to 6 weeks after initial anti-TCRβ mAb treatment, NOD mice were fasted and injected with d-glucose for IPGTT analysis. Blood glucose levels (upper panel) and serum insulin concentration (bottom panel) of each group were assessed after glucose injection. (b) Serum insulin levels in each group before (indicated as B) and after treatment. (c) Representative insulin staining of the liver sections from gene transferred groups. *P < 0.05 vs anti-TCRβ plus control vector group.

Discussion

In this study, we investigated the efficacy of an anti-TCRβ mAb plus Ngn3-Btc gene transfer in reversing T1D. We found that a 5-day course of 50 μg of anti-TCRβ mAb reversed >80% of new-onset T1D in NOD mice. Even in overtly diabetic NOD mice that were relatively resistant to anti-TCRβ mAb alone, we successfully reversed the diabetes in ∼60% of these mice by treating them with a combination of anti-TCRβ mAb and Ngn3-Btc gene transfer. Importantly, the enduring T1D reversal observed in this autoimmune diabetes mouse model was associated with a competent immune system as well as partial restoration of insulin production.

Human patients and NOD mice with new-onset T1D retain considerable numbers of residual pancreatic islets. The survival of these residual islets is being heavily challenged: Diabetogenic T cells are actively engaged in their destruction. An ultimate goal of T1D immunologists would be to halt islet destruction by selectively eliminating or disarming diabetogenic T cells while leaving other immune system components intact (16). However, the molecular mechanism of T-cell–mediated islet destruction remains unclear, and thus the optimal targets for T1D immunotherapy are poorly defined (1). To date, nonspecific immune suppressants, such as anti-CD3 mAb and antithymocyte globulin, are still the most potent immunotherapies to abrogate anti-islet T-cell responses (1, 4). We have recently demonstrated the potency of an anti-TCRβ mAb in inducing long-term transplant survival (9, 11, 17) and reversing new-onset T1D in mice (10). The same anti-TCRβ mAb clone has also been shown to be effective in preventing experimental autoimmune encephalomyelitis and collagen-induced arthritis in animal models (18, 19). Herein, we demonstrated that anti-TCRβ mAb preserved neurogenin3-induced neo-islets and reversed overt T1D in NOD mice.

The anti-TCRβ mAb treatment was transient but its effect on reversal of new-onset T1D was lasting. Moreover, anti-TCRβ mAb–treated NOD mice were immune competent to reject allogeneic skin grafts. Hence, tolerogenic mechanisms may be invoked by anti-TCRβ mAb to selectively abrogate anti-islet autoimmunity. Multiple mechanisms exist to induce tolerance in peripheral T cells, including functional anergy (20), peripheral deletion by apoptosis (21), clonal exhaustion (22), and suppression by Treg cells (23). In our adoptive transfer study, total splenocytes from T1D-reversed NOD mice robustly transferred diabetes in NOD/scid recipients, and partially depleting Treg cells in the splenocytes before transfer moderately accelerated T1D onset. Thus, diabetogenic T cells are not eliminated in T1D-reversed NOD mice and their function is partially suppressed by Treg cells. Interestingly, 40% to 65% of pancreatic infiltrating T cells in anti-TCRβ mAb–treated NOD mice expressed a naïve cell marker CD62L. In general, CD62L facilitates naïve T cells homing to lymphoid tissues (24) and is not required for diabetes onset and islet infiltration of T cells (25). More importantly, we previously showed that the anti-TCRβ mAb treatment increased T-cell expression of PD-1 in diabetic NOD mice (10). Herein, we further found pancreatic infiltrating T cells can be divided into PD-1⁺ and PD-1^‒ populations. The majority of pancreatic infiltrating T cells in the anti-TCRβ mAb group were PD-1⁺, which is associated with low IFN-γ production. Hence, anti-TCRβ mAb induces an exhaustion-like phenotype of T cells. In future studies, we would like to determine whether T-cell exhaustion or dysfunction can be induced in T1D by anti-TCRβ mAb, and whether induction of T-cell exhaustion is a potential approach for treating T1D.

NOD mice with overt diabetes are resistant to anti-TCRβ mAb, which may be simply attributed to the severe loss of functional islets. To test this notion, Ngn3-Btc genes were transferred into overtly diabetic NOD mice that had maintained blood glucose levels between 350 and 450 mg/dL for 2 weeks after anti-TCRβ mAb treatment. Five of eight Ngn3-Btc gene–transferred NOD mice successfully reversed disease and remained diabetes free for at least 14 weeks. In the liver of these diabetes-reversed mice, insulin-expressing cells were detected surrounding the portal vein. Therefore, anti-TCRβ mAb combined with Ngn3-Btc gene transfer potently reversed overt T1D by inducing and protecting islet neogenesis.

In summary, anti-TCRβ mAb induced an enduring reversal of new-onset or overt diabetes in NOD mice when used alone or combined with Ngn3-Btc transfer, respectively. Many efforts have been made to induce islet expansion or neogenesis (12). A potent immunotherapy must also be developed as a prerequisite for successful islet neogenesis. We suggest adding anti-TCRβ mAb to the limited list of T1D immunotherapies that effectively protect residual and newly formed islets.

Appendix.

Antibody Table

Peptide/Protein Target	Antigen Sequence	Name of Antibody	Manufacturer, Catalog No.	Species Raised in; Monoclonal or Polyclonal	Dilution Used	RRID
TCR β		H57-597	Bio X Cell	Armenian hamster	10 or 50 μg/dose	AB_2686916
TCR α		H28 710	Santa Cruz Biotechnology	Armenian hamster	10 or 50 μg/dose	AB_1130051
CD4		GK1.5	eBioscience	Rat	1/200	AB_464894
CD8a		53-6.7	eBioscience	Rat	1/200	AB_469584
CD25		PC61.5	eBioscience	Rat	1/200	AB_469366
CD44		IM7	eBioscience	Rat	1/200	AB_469715
CD62L		MEL-14	eBioscience	Rat	1/200	AB_494170
Foxp3		FJK-16s	eBioscience	Rat	1/200	AB_494217
Insulin		ab7842 (polyclonal)	Abcam	Guinea pig	1/100	AB_306130
CD69		H1.2F3	eBioscience	Hamster	1/200	AB_465733
CD28		37.51	BioLegend	Hamster	1/200	AB_312870
OX40		OX-86	BioLegend	Rat	1/200	AB_2272150
CD40L		MR1	BioLegend	Hamster	1/200	AB_313271
PD-1		29F.1A12	BioLegend	Rat	1/200	AB_2159183
Tim-3		B8.2C12	BioLegend	Rat	1/200	AB_1626177
TNF-α		MP6-XT22	eBioscience	Rat	1/200	AB_466199
IRF-γ		XMG1.2	BioLegend	Rat	1/200	AB_315404
CD4		EPR19514	Abcam	Rabbit	1/200	AB_2686917
CD8a		KT15	Bio-Rad	Rat	1/200	AB_2075236
Goat anti-rabbit		A-11010	Life Technology	Goat	1/2000	AB_10584649
Goat anti-rat		A-21247	Life Technology	Goat	1/100	AB_141778
Goat anti-guinea pig		A-21450	Life Technology	Goat	1/500	AB_141882

Abbreviation: RRID, research resource identifier.