Live attenuated Salmonella msbB mutants were engineered to express of proinsulin (PI) and the immunomodulatory cytokines IL-10 and TGF-β (17–19, 25). The final vaccine formulation included Salmonella carrying PI+IL-10, PI+TGF-β, and was administered in combination with anti-CD3 monoclonal antibody (17–19, 25).

Female NOD/ShiLtJ and NSG (NOD.Cg-Prkdc^scid Il2rg^tm1Wjl/SzJ) mice aged 7–8 weeks were obtained from Jackson Laboratory (Bar Harbor, ME, USA) and maintained in a pathogen-free, AAALAC-accredited facility. All procedures were approved by the Institutional Animal Care and Use Committee of City of Hope (IACUC# 18017).

Groups of NOD mice were randomly assigned to receive:

Anti-CD3 antibody (2.5 µg i.p.) was administered for five consecutive days starting one day before vaccination (days -1 to 3) (17–20, 25).

Prevention model: Normoglycemic NOD mice (blood glucose < 200 mg/dL) were treated at 8 weeks of age and monitored for diabetes onset for 90 days.

Reversal model: Diabetic NOD mice (blood glucose > 200 mg/dL) were treated and monitored for reversal of hyperglycemia.

Blood glucose values were measured twice a week. Mice were deemed diabetic if blood glucose values were ≥ 200 mg/dL on two consecutive measurements (17–20, 25).

Pancreatic tissue was fixed, paraffin-embedded, and stained with hematoxylin and eosin. Islet infiltration was evaluated in a blinded manner across 4 non-adjacent sections per pancreas (each 5 µm thick), separated by at least 150 µm, using light microscopy at 20x magnification. Insulitis was scored as follows (0 = no insulitis; 1 = peri-insulitis; 2 = mild insulitis with <50% islet area affected; 3 = invasive insulitis with >50% islet area affected, 4 = invasive insulitis with 100% islet area affected). Islets obtained from 5 to 10 mice per group were analyzed, with 30 and 132 islets evaluated per mouse (16, 17, 45).

Immunofluorescence staining was performed using a guinea pig anti-insulin antibody (Invitrogen) and a donkey anti-guinea pig Alexa Fluor 488-conjugated secondary antibody (Jackson ImmunoResearch). Nuclei were counterstained with DAPI. Images were acquired using a ZEISS LSM700 microscope and analyzed with ZEN-lite software. The β-cell fraction was calculated as the percentage of the area of insulin-positive cells divided by the total islet area. Islets were obtained from the pancreatic tissues of 10 mice per group. For each mouse, 61 to 71 islets were evaluated across 4 slides collected from different section levels (each 5 µm thick), with at least 150 µm separating each level. Data were analyzed using QuPath software (17, 45).

Single-cell suspensions from spleens were prepared following collagenase D digestion. Cells were stained with antibodies against CD4, CD8a, CD25, and FOXP3 to identify Tregs, as well as CD49b, LAG-3 to identify type 1 regulatory T cells. Matching isotype controls (BioLegend, San Diego, CA, USA) were included, UltraComp eBeads Compensation Beads (Invitrogen) were used to set compensation controls. Dead cells were excluded using Fixable Blue Dead Cell Stain (Invitrogen). Samples were acquired on a BD LSRFortessa flow cytometer, and data were analyzed using FlowJo software (17–20, 25).

To assess antigen specificity, 5 x 10⁵ splenocytes were isolated and cultured in a 96-well plate for 72 hours with 10 µg/ml insulin B_9–23 peptide or OVA_323-339, a non-specific peptide (17). A Mouse Quantikine ELISA Kit (R&D Systems) was used to quantify IL-10, TGF-β, IFN-γ, and TNFα levels in the conditioned media of the treated splenocytes.

Splenocytes from three diabetic NOD mice were isolated, pooled, and transferred into ten immunodeficient NSG mice (18, 20). Four weeks after vaccination, CD4⁺ CD25⁺ Tregs were isolated from five NOD mice treated with vehicle, vaccine, GAST-17, or both. Ten NSG mice received either 1 x 10⁶ diabetogenic cells alone or with 5 x 10⁵ Tregs from each treatment group. GAST-17 (600 µg/kg) was given i.p. once daily to mice receiving Tregs from the GAST-17 or GAST-17 and vaccine treatment groups, starting on the day of cell transfer. Blood glucose levels were measured twice weekly in NSG mice (18, 20).

Mouse serum cytokine levels were quantified using the MILLIPLEX Mouse High Sensitivity T Cell Panel - Immunology Multiplex Assay (Millipore, Sigma) (16, 20).

Four weeks after vaccination, splenocytes and CD4⁺ CD25⁺ Tregs and CD4⁺ CD25^– cells were isolated from treated mice. RNA was extracted, converted to cDNA, and analyzed by qPCR analysis using Taq-Man assays (Applied Biosystem) to measure gene expression (20). mRNA levels of PDL-1, CTLA-4, AhR, IDO, and IL-27 were normalized to TATA box binding protein (Tbp) expression (20).

Statistical analyses were performed using GraphPad Prism 10. Kaplan-Meier plots were employed to evaluate the incidence and reversal of diabetes across treatment groups, with statistical significance assessed using the Mantel-Cox log-rank test. Data distribution and normality were assessed using the Shapiro-Wilk test. Parametric tests were applied when the data met statistical assumptions, particularly normality (p > 0.05). In contrast, non−parametric tests were used when the assumptions were not met, and the data were not normally distributed (p < 0.05). Two-way ANOVA (Bonferroni’s multiple comparison test) was used to assess immune cell infiltration in islets and cytokine secretion following peptide stimulation. Dunn’s multiple comparisons test following the Kruskal-Wallis test was employed to compare the percentage of insulin-positive β-cells per islet across treatment groups. Serum cytokine levels among treatment groups were analyzed using one-way ANOVA for normally distributed data, followed by Fisher’s LSD multiple comparisons test. For non-normally distributed data, the Kruskal-Wallis test was used, followed by Dunn’s multiple comparisons test. One-way ANOVA followed by Tukey’s multiple comparisons test was used to compare Treg and Tr1 cell populations. Differences in cytokine levels and gene expression fold changes between treatment groups were analyzed using the unpaired Welch’s t-test for parametric data and using the Mann Whitney test for non-parametric data. A p-value < 0.05 was considered statistically significant.

In the prevention model, 80% of normoglycemic NOD mice treated with the combination of oral vaccine and GAST-17 remained diabetes-free for 90 days, compared with 30% of mice receiving GAST-17 alone (Figure 1A, Supplementary Figure 1A).

In the reversal model (Figure 1B, Supplementary Figure 1A), diabetes was reversed in 80% of NOD mice treated with the combination therapy and in 63% of those receiving the vaccine alone. Both groups achieved and maintained normoglycemia for three months post-treatment (Figure 1B). In contrast, 95% of vehicle-treated and 96% of GAST-17-treated mice remained hyperglycemic (Figure 1B).

Histological analysis confirmed marked T-cell infiltration in pancreatic islets of vehicle-treated NOD mice (Figure 2, Supplementary Figure 2), both pre-diabetic (Figure 2A) and diabetic (Figure 2C), consistent with typical T1D pathology. Treatment with either the vaccine or GAST-17 alone significantly reduced islet immune infiltration compared to the vehicle group (Figure 2A-C). In the prevention model, the combined vaccine and GAST-17 therapy further decreased lymphocyte infiltration and preserved islet insulin content compared with vehicle-treated mice (Figure 2B; two-way ANOVA, p = 0.04, and p = 0.002). Vaccine alone also decreased islet inflammation compared with vehicle (Figure 2B; p = 0.0003). In the reversal model, diabetic mice receiving the combined treatment showed significantly reduced immune cell infiltration compared with mice treated with vehicle or GAST-17 alone (Figure 2C; two-way ANOVA, p < 0.0001, p = 0.03, and p < 0.0001; or p = 0.0003, respectively). Similarly, mice treated with the vaccine alone exhibited decreased islet inflammation compared with vehicle or GAST-17 alone (Figure 2C; p = 0.0002, p = 0.04, and p < 0.0001; or p = 0.003, respectively). Islets from mice treated with the vaccine or the combined vaccine and GAST-17 showed a significantly higher percentages of insulin-positive cells compared to those treated with vehicle or GAST-17 alone (Figures 2D-E; one-way ANOVA, p < 0.0001).

In the prevention model, vaccine treatment significantly increased IL-10 and CCL2 while IL-2 and IL-13 remained unchanged (Figure 3A). The vaccine also reduced inflammatory cytokines IL-1α, IL-6, and IL-12 compared to the vehicle group. Combined vaccine and GAST-17 therapy further suppressed these inflammatory cytokines as well as IFN-γ and, GM-CSF. In contrast, GAST-17 alone increased IFN-γ, GM-CSF, IL-1α, IL-6, and IL-12. No significant changes were observed in CXCL5, CXCL1, and CXCL2 under any treatment condition (Figure 3A).

In the reversal model, vaccine-treated mice showed elevated levels of IL-10, IL-2, IL-13, CCL2, and IFN-γ (Figure 3B), with no changes in GM-CSF, IL-1α, IL-6, and IL-12 or chemokines CXCL5, CXCL1, and CXCL2, regardless of treatment (Figure 3B). Combined vaccine and GAST-17 therapy increased IL-2 and IFN-γ but did not affect other cytokines. Levels of IL-4, IL-5, IL-7, IL-17, and IL-1β remained unchanged (data not shown).

Flow cytometric analysis was performed using a sequential gating strategy to identify single live lymphocytes, followed by CD4⁺ T-cells, and CD4⁺ CD25⁺ Foxp3⁺ Tregs and CD4⁺ CD49b⁺ LAG-3⁺ Tr1 cells (Supplementary Figure 3). A significantly higher percentage of CD4⁺ CD25⁺ Foxp3⁺ Tregs was observed in the spleens of mice treated with the vaccine or the combined vaccine and GAST-17 therapy, compared with those receiving vehicle (Tukey’s multiple comparisons test, p = 0.02, p = 0.04) or GAST-17 alone (p = 0.008, p = 0.02) (Figure 4A). Similarly, the frequency of CD4⁺ CD49b⁺ LAG3⁺ Tr1 cells was elevated in mice treated with the vaccine or the combination therapy relative to vehicle-treated controls (p = 0.02, p < 0.0001) or GAST-17 alone (p = 0.0007, p < 0.0001) (Figure 4B). In addition, a significant increase in Tr1 cells was detected in mice receiving the combination therapy compared with vaccine alone (p = 0.0003).

To assess the functional activity of regulatory T-cells, CD4⁺ CD25⁺ Tregs and CD4⁺ CD25^– (including Tr1) cells were isolated from the spleens of mice treated with the vaccine and GAST-17, GAST-17 alone, vaccine alone, or vehicle. After 3 days of culture, cytokine levels in the conditioned media were determined.

In mice receiving the combination therapy, the levels of IL-10 and IL-2 found in the media of CD4⁺ CD25⁺ Tregs (Welch’s t test, p =0.009, p = 0.01) and CD4⁺ CD25^– T-cells (p = 0.01, p = 0.001) were increased compared with levels found in the media of cells from GAST-17-treated mice, while IFN-γ and TNF-α levels remained unchanged (Figure 5A). IL-10 was significantly elevated in the media of cultured CD4⁺ CD25⁺ Tregs (p = 0.02) and CD4⁺ CD25^– T cells (p = 0.02), from mice given the combination treatment whereas IL-2 levels were comparable to those of vehicle controls (Figure 5B). Notably, IFN-γ and TNF-α levels were reduced in CD4⁺ CD25⁺ Tregs (p = 0.006, p = 0.03), but not in CD4⁺ CD25^– T-cells, compared with the vehicle-treated group (Figure 5B).

Analysis of cytokines in the conditioned media of splenocytes from mice revealed increased IL-10 and TGF-β levels from animals treated with the vaccine combined with GAST-17, and elevated IL-10 in splenocytes from mice treated with vaccine alone. IFN-γ, and TNF-α levels remained unchanged (Figure 5).

To evaluate antigen-specific immune responses, splenocytes from mice from each treatment group were stimulated with the insulin B_9–23 peptide or a non-specific control peptide (OVA_323-339) for 3 days. Analysis of conditioned media of insulin B_9–23 stimulated splenocytes from mice treated with the vaccine or the combination of vaccine and GAST-17 revealed higher levels of the regulatory cytokine IL-10 compared with cells stimulated with the control peptide (Figure 6A). Stimulation with insulin B_9–23 also increased TGF-β secretion in splenocyte cultured from mice receiving the combination therapy, but not in those treated with vaccine or GAST-17 alone (Figure 6A).

In contrast, IFN-γ and TNF-α levels were reduced in conditioned media from insulin B_9–23-stimulated splenocytes of mice treated with vaccine and GAST-17, compared with OVA-stimulated controls. No significant changes IFN-γ and TNF-α were observed in splenocyte culture from mice treated with vaccine or GAST-17 alone (Figure 6A).

To evaluate the in vivo suppressive function of regulatory T-cells, an adoptive transfer study was performed. Splenocytes from diabetic NOD female mice (diabetogenic cells) were injected into immunodeficient NSG mice either alone or together with CD4⁺ CD25⁺ Tregs isolated from vehicle-, vaccine-, GAST-17-, or vaccine plus GAST-17-treated mice. Recipient mice were monitored for the development of hyperglycemia.

Transfer of diabetogenic splenocytes alone induced hyperglycemia in all recipient mice within 40–50 days post-transfer (Figure 6B). Co-transfer of Tregs from vehicle-treated mice delayed or prevented hyperglycemia in approximately 50% of recipients. In contrast, Tregs isolated from mice treated with the vaccine, GAST-17, or the combination therapy completely prevented hyperglycemia in all recipient mice (Figure 6B), demonstrating potent in vivo activity.

PD-L1 and IL-27 mRNA was increased in splenocytes from mice treated with the combination of vaccine and GAST-17 (Mann-Whitney test, p = 0.006, p = 0.04) compared with those from mice treated with GAST-17 alone (Figure 7A). Furthermore, both splenic CD4⁺ CD25⁺ Tregs and CD4⁺ CD25^- T-cells from combination-treated mice showed higher expression of CTLA-4, IDO, AHR, and IL-27 mRNA relative to corresponding cells from GAST-17–treated mice (Figures 7B, C).