Protection Against Type 1 Diabetes Development in Mice With 4E-BP2 Deletion

Abstract

Type 1 diabetes (T1D) is an autoimmune disease characterized by β-cell destruction promoted by autoreactive T cells. Eukaryotic translation initiation factor 4E (eIF4E)–binding protein 1 (4E-BP1) and 4E-BP2 are translational repressors and downstream targets of mammalian target of rapamycin complex 1 (mTORC1). Activation of the 4E-BP2/eIF4E pathway by 4E-BP2 deletion promotes translation initiation, inducing β-cell expansion and proliferation and regulating adaptive immunity. However, the involvement of 4E-BP2 in T1D remains unexplored. This study aimed to determine the role of 4E-BP2/eIF4E signaling in T1D prevention. We used the NOD mouse model of T1D and generated mice with global 4E-BP2 deletion in the NOD background (Eif4ebp2^−/−). We assessed T1D development, glucose homeostasis, pancreas morphometry, and immune responses in Eif4ebp2^−/− and littermate control mice. We found that Eif4ebp2^−/− male mice exhibited reduced diabetes incidence, which did not occur in female mice, as well as preserved β-cell mass, improved insulin secretion in vitro, and comparable insulitis. Characterization of T-cell compartments showed decreased splenic CD8⁺ cytotoxic T-cell proliferation and increased pancreatic regulatory T-cell infiltration in Eif4ebp2^−/− mice, potentially resulting from increased proliferation and suppressive capacity. Adoptive transfer studies demonstrated that Eif4ebp2^−/− male lymphocytes were less diabetogenic than those of controls. In conclusion, activation of 4E-BP2/eIF4E by 4E-BP2 deletion protected against T1D, supporting 4E-BP2 as a potential therapy target.

Article Highlights

• Mammalian target of rapamycin complex 1 (mTORC1) signaling is essential to β-cell mass, function, and adaptive immunity; however, its specific downstream mediators in type 1 diabetes (T1D) remain poorly defined.

• We investigated eukaryotic translation initiation factor 4E–binding protein 2 (4E-BP2), a major translational regulator downstream of mTORC1, by using global 4E-BP2–knockout mice on the NOD background.

• Loss of 4E-BP2 protected male NOD mice from T1D through preservation of β-cell mass and function, coupled with attenuation of autoimmune responses.

• These findings identify 4E-BP2 as a novel immunometabolic node, highlighting its potential as a therapeutic target for T1D prevention and treatment.

Graphical Abstract

Introduction

Type 1 diabetes (T1D) is an autoimmune disease characterized by loss of insulin production resulting from autoimmune-mediated β-cell destruction (1). Autoreactive CD4⁺ and CD8⁺ T cells target islet-associated antigens and infiltrate islets (2), adopting an inflammatory effector phenotype that promotes β-cell destruction. Impaired regulatory T cells (Tregs), critical regulators of immune tolerance, further drive disease progression (3,4). Therefore, treatments that preserve functional β-cells, stimulate β-cell regeneration, and modulate immune responses represent promising strategies for T1D treatment.

Mammalian target of rapamycin complex 1 (mTORC1) integrates growth factor and nutrient signals to regulate cell size and proliferation (5). mTORC1 phosphorylates S6K and eukaryotic translation initiation factor 4E (eIF4E)–binding proteins (4E-BPs) (6), a family of three proteins: 4E-BP1, 4E-BP2, and 4E-BP3. 4E-BPs suppress protein translation by blocking eIF4E binding to the mRNA 5′ cap structure. mTORC1 phosphorylation of 4E-BPs inhibits binding to eIF4E and allows protein synthesis. Through these interactions, mTORC1 controls cell size, proliferation, ribosomal biogenesis, protein translation, and autophagy (7). mTORC1 signaling has been shown to regulate β-cell mass, insulin secretion, and adaptation to insulin resistance (8). Downstream of mTORC1, 4E-BP2, but not 4E-BP1, promotes β-cell expansion and survival via insulin receptor substrate 2 (IRS2) modulation (9). However, the role of the mTORC1/4E-BP2 axis in T1D remains unexplored.

mTORC1 regulates multiple metabolic programs that shape adaptive immunity, supporting CD4⁺ T helper cell activation and CD8⁺ T effector functions (10,11). Previous work has shown that 4E-BP1 and 4E-BP2 are required for optimal mTORC1 activation in mouse CD8⁺ T cells (12), suggesting that 4E-BPs participate in a positive feedback loop regulating mTORC1 and eIF4E in CD8⁺ T cells (12). In contrast, the mTOR pathway negatively regulates Treg differentiation, because T cells lacking both mTORC1 and mTORC2 preferentially become Tregs (13). Notably, lymphocytes predominantly express 4E-BP2 over 4E-BP1, which is important for their growth and proliferation (14). Together, published evidence suggests a role for the mTORC1/4E-BP2 axis in both β-cell biology and adaptive immunity, making it a viable candidate for T1D studies.

Here, using newly generated global 4E-BP2–deficient mice in the NOD background (Eif4ebp2^−/−), we tested the role of 4E-BP2 in T1D pathogenesis. This study showed that Eif4ebp2 deletion preserved β-cell mass, reduced β-cell death, and prevented T1D development in 90% of male NOD mice, in association with comparable insulitis, inhibition of CD8⁺ T-cell effector functions, and increased Treg pancreas infiltration. Adoptive transfer studies demonstrated that lymphocytes derived from male Eif4ebp2^−/− mice were less diabetogenic than those from control mice. Collectively, these studies suggest that inhibition of 4E-BP2/eIF4E interaction could be a promising therapeutic strategy for T1D by preserving functional β-cells and/or reducing autoimmune injury.

Research Design and Methods

Mouse Models

Mice deficient in 4E-BP2 (Eif4ebp2^−/−) in the NOD background were generated by the University of Michigan Mouse Genetic Core using CRISPR-Cas9 technology. Eif4ebp2^−/− mice were crossed with NOD mice (The Jackson Laboratory, Bar Harbor, ME) to generate Eif4ebp2^+/− mice, which were then interbred to generate all experimental groups and control littermates. Colonies were maintained and bred at the University of Miami. Male NOD.SCID mice were obtained from The Jackson Laboratory. A SNP-based genome scanning service from The Jackson Laboratory was used to confirm 100% NOD background in the control, Eif4ebp2^−/−, and Eif4ebp2^+/− mouse lines.

Animal studies were performed in accordance with a protocol approval by the University of Miami Institutional Animal Care and Use Committee (Miami, FL).

Metabolic Studies

Blood glucose was measured using an ACCUCHECK glucose meter (Roche), with blood taken from the tail vein. Plasma insulin was measured with an insulin ELISA kit (ALPCO), and plasma proinsulin was measured with a proinsulin ELISA kit (Mercodia). Oral glucose tolerance tests were performed by measuring blood glucose before and after glucose (2 g/kg) administration via oral gavage in fasted (6 h) mice. Insulin tolerance tests were performed by measuring blood glucose before and after i.p. insulin (0.75 units/kg) injection in fasted (6 h) mice. Diabetes was determined by three consecutive blood glucose measurements >250 mg/dL.

Islet Isolation

Islet isolation was performed as previously described (15). Briefly, the pancreas was inflated with 1 mg/mL collagenase P (Roche) injected into the common duct. Islets were handpicked and maintained at 37°C overnight in RPMI containing 10% FBS, 1% penicillin/streptomycin, and 5 mmol/L glucose before performing subsequent experiments. Isolated islets were treated with interleukin-1β (50 units/mL), recombinant rat interferon-γ (1,000 units/mL), and recombinant rat tumor necrosis factor-α (1,000 units/mL) (PeproTech) in the presence of 16.7 mmol/L glucose (9). Protein isolation was performed 24 h after treatment.

Static Insulin Secretion

Islets were isolated and incubated at 37°C overnight in RPMI plus 10% FBS and 5.5 mmol/L glucose. The next day, islets were treated with high glucose (16.7 mmol/L) and cytokines (described above); 24 h after the treatment, the islets were incubated in Krebs buffer (114 mmol/L NaCl, 4.7 mmol/L KCl, 1.2 mmol/L KH₂PO₄, 1.16 mmol/L MgSO₄, 20 mmol/L HEPES, 2.5 mmol/L CaCl₂, 25.5 mmol/L NaHCO₃, and 0.2% BSA; pH 7.2) with 2 mmol/L glucose for 2 h. Groups of 10 islets in triplicate were incubated in Krebs-Ringer medium containing 16.7 mmol/L glucose for 30 min. Insulin from the media was measured with the Ultrasensitive Insulin ELISA Kit (ALPCO) and normalized to total insulin content, after lysing the cells with a 0.5% HCL/75% ethanol solution.

Insulitis Scoring

Pancreas sections were stained with hematoxylin-eosin and assessed blindly by light microscopy. Insulitis score was determined as previously described (16) using the following criteria: 0, no insulitis; 1, peri-insulitis (infiltration restricted to the periphery of islets); 2, mild insulitis (<50% of the islet area infiltrated); 3, severe insulitis (≥50% of the islet area infiltrated); and 4, massive insulitis (≥90% of the islet area infiltrated).

Immunostaining

Immunofluorescence staining was performed with antibodies listed in Supplementary Table 1. Slides were mounted with Vectashield plus DAPI (Vector Laboratories), and images were acquired using a Leica DM5500B microscope with a DFC360FX camera, OASIS-blue PCI controller, and Leica Application Suite X. β-Cell mass was calculated by quantifying insulin area/total pancreatic area using insulin- and acinar-stained areas in four sections (200 μm apart) and multiplying by pancreas weight. Proliferation was quantified by staining for Ki-67 and apoptosis using TUNEL (ApopTag Red Kit; Chemi-Con) in insulin-stained sections (from two sections 400 μm apart).

Western Blotting

Islet lysates (10–40 μg) were boiled for 5 min, run on 10–12% SDS-PAGE, and blotted with antibodies (Supplementary Table 1). Signals were detected with the Immun-Star WesternC Kit (Bio-Rad). Band densitometry was measured using National Institutes of Health ImageJ software (version 1.52a; https://rsb.info.nih.gov/ij/index.html) and normalized to tubulin. Goat anti-mouse and anti-rabbit secondary antibodies were from LI-COR.

Flow Cytometric Analysis

Analyses were performed as previously described (17). Spleens and pancreatic lymph nodes (pLNs) were disrupted into single-cell suspensions; pLNs were digested with collagenase D (0.2 mg/mL for 30 min at 37°C), washed with Hanks’ balanced salt solution, and filtered (70 μm). Pancreata were digested with collagenase P (0.5 mg/mL) and purified using Histopaque 1077 (Millipore Sigma). Peripheral blood mononuclear cells were isolated from whole blood on a Ficoll-Paque PLUS gradient (GE Life Sciences); red blood cells were lysed with ammonium–chloride–potassium buffer. Lymphocytes were stained with LIVE/DEAD Near-IR (Invitrogen/Thermo Fisher Scientific), blocked with anti-CD16/32 (clone 24G2), and labeled with fluorescence-conjugated antibodies (Supplementary Table 1). Foxp3 was stained using the Miltenyi FoxP3 buffer set. Granzyme B, Ki-67, and B-cell lymphoma 2 (BCL2) staining followed fixation and permeabilization with eBioscience Foxp3/Transcription Factor Buffer (Thermo Fisher Scientific). Data were acquired on a BD LSRII and analyzed with Kaluza (Beckman Coulter).

Adoptive Transfer Studies

Spleens from male NOD control or Eif4ebp2^−/− mice were mechanically disrupted into single cell suspensions to isolate lymphocytes, and red blood cells were lysed with ammonium–chloride–potassium. Lymphocytes were adoptively transferred by tail vein i.v. injection into male NOD.SCID mice (10 × 10⁶ per mouse) (17).

T-Cell Proliferation Studies

Spleens were mechanically disrupted and processed as described above. Cells (0.2 × 10⁶) were cultured in 96-well round-bottom microtiter plates, labeled with 5 mmol/L CellTrace Violet (Life Technologies), treated with anti-CD3 (clone 2C11), and analyzed 3 days later via flow cytometry.

Statistical Analysis

One-way ANOVA was followed by the Dunnett multiple comparisons test. Two-way ANOVA was followed by the Sidak multiple comparisons test. When data were normally distributed, the unpaired Student t test was performed, and when data were nonnormally distributed, the unpaired Mann-Whitney test was performed to compare two groups. Survival curves were analyzed with the Mantel-Cox log-rank test. P value is indicated as follows: *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Data and Resource Availability

All data generated or analyzed during this study are included in the published article and online Supplementary Material.

Results

Homozygous and Heterozygous Deletion of the 4E-BP2 Gene (Eif4ebp2) Protects Against T1D in Male NOD Mice

To assess whether 4E-BP2 depletion protects against T1D, we generated NOD mice with 4E-BP2 deletion (Eif4ebp2^−/− and Eif4ebp2^+/−) and compared them with littermate controls. 4E-BP2 protein levels were reduced by ∼50% in Eif4ebp2^+/− and by ∼90% in Eif4ebp2^−/− islets (Fig. 1A), unaffected by cytokine treatment. Diabetes incidence was tracked over 30 weeks, with diabetes determined by three consecutive blood glucose measurements >250 mg/dL. Male control mice started developing diabetes at 15 weeks of age (Fig. 1B), whereas male Eif4ebp2^−/− and Eif4ebp2^+/− mice were protected against diabetes (Fig. 1B and Supplementary Fig. 1A). Male Eif4ebp2^−/− mice had similar fed and fasting glucose levels to control mice until 19 weeks of age, after which fed glucose levels were significantly higher in control mice (Supplementary Fig. 1B and C). Notably, this protection was absent in female mice, which showed similar diabetes incidence and glucose levels over time between groups (Supplementary Fig. 2A–C), as well as similar β-cell mass and insulin tolerance (Supplementary Fig. 2D and E). At 25 weeks, Eif4ebp2^−/− male mice had elevated fed but similar fasting insulin levels compared with controls (Fig. 1C and D). Because diabetic mice were removed from the study on diagnosis and excluded from subsequent measurements, the data in Fig. 1C and D and Supplementary Fig. 1B and C represent only that specific week and omit mice previously diagnosed with diabetes. We measured fed proinsulin and the proinsulin-to-insulin ratio in mice at 19 weeks of age (Supplementary Fig. 1D and E), which showed a trend toward a decreased proinsulin-to-insulin ratio in Eif4ebp2^−/− mice (P = 0.076). Oral glucose tolerance tests performed in nondiabetic mice were similar but trended toward improvement in Eif4ebp2^−/− mice at 16 weeks (P = 0.05; 120 min after glucose administration) (Supplementary Fig. 1F–I).

Figure 1.

Male Eif4ebp2^−/− mice are protected from T1D development and have similar glucose homeostasis compared with control mice. A: Islet 4E-BP2 protein levels with and without cytokine (CTK) treatment for 24 h (n = 5 per group) and representative immunoblotting image. Bands were quantified and analyzed via two-way ANOVA, normalized to tubulin. B: Diabetes incidence tracked for 30 weeks in control, Eif4ebp2^+/−, and Eif4ebp2^−/− male mice, analyzed via log-rank (Mantel-Cox) test (n = 28–30 per group). C and D: Random fed (C) and fasted (D) glucose levels at different weeks in male mice, analyzed via two-way ANOVA. **P < 0.01, ***P < 0.001, ****P < 0.0001.

Male Eif4ebp2^−/− and Eif4ebp2^+/− NOD Mice Exhibit Preserved β-Cell Mass and Increased Resistance to In Vitro Cytokine-Induced Death

Given the diabetes protection in Eif4ebp2^+/− and Eif4ebp2^−/− mice, we focused on understanding the mechanism by which lack of 4E-BP2 protected against T1D in male Eif4ebp2^+/− and Eif4ebp2^−/− mice. β-Cell morphometry showed that β-cell mass in male mice was similar between groups at 16 weeks of age (Fig. 2A). At 20–25 weeks of age, β-cell mass was significantly lower in control male mice compared with Eif4ebp2^+/− and Eif4ebp2^−/− NOD mice (Fig. 2A), indicating a role for the mTOR/eIF4E/4E-BP2 axis in the preservation of β-cell mass in T1D. Of note, 50% of the β-cell mass among controls was from diabetic mice, as shown by the open diamonds in Fig. 2A. β-Cell proliferation, measured by Ki-67 staining, was comparable between groups at 16 and 25 weeks of age (Fig. 2B). β-Cell death assessment by TUNEL assay showed similar levels of cell death between groups at 16 weeks and a trend toward decreased cell death in Eif4ebp2^+/− and Eif4ebp2^−/− mice at 20–25 weeks (Fig. 2C). To distinguish the effect of 4E-BP2 deletion on β-cells from systemic effects, we exposed isolated islets from nondiabetic mice to a cocktail of proinflammatory cytokines (Fig. 2D). Baseline protein levels of IRS2 were increased in Eif4ebp2^+/− and Eif4ebp2^−/− islets (significant only in Eif4ebp2^+/− mice at P < 0.05), and this increase was ablated by cytokine treatment (Fig. 2E). Levels of BCL-XL, which regulates cell death by inhibiting apoptosis (18), and levels of cleaved poly (ADP-ribose) polymerase (PARP), a proapoptotic protein, remained similar among all groups (Fig. 2F and G). Apoptosis, assessed by cleaved caspase 3 levels, was significantly increased after cytokine stimulation in control and Eif4ebp2^−/−, but not in Eif4ebp2^−/−, islets, suggesting that lack of 4E-BP2 inhibits cytokine-induced cell death (Fig. 2H). These results indicate that lack of Eif4ebp2 preserves β-cell mass, potentially by attenuating β-cell death.

Figure 2.

Male Eif4ebp2-mutated mice have preserved β-cell mass and increased resistance to cytokine (CTK)-induced cell death. A: β-Cell mass in control, Eif4ebp2^+/−, and Eif4ebp2^−/− male mice at 16 and 20–25 weeks of age, analyzed via two-way ANOVA (n = 4–9 per group; open diamonds represent diabetic mice). B: Ki-67⁺ staining in insulin⁺ cells via immunostaining, analyzed via two-way ANOVA (n = 4–5 per group). C: TUNEL assay, analyzed via two-way ANOVA (n = 4–9 per group). Mice from panels A–C were nondiabetic, except the samples in the control group represented by diamonds. D: Western blot representative image of expression of IRS2, poly (ADP-ribose) polymerase (PARP), BLC-XL, cleaved caspase 3 (Cl Casp3), and tubulin from isolated islets after 24-h CTK treatment. E–H: Western blot protein expression quantification of IRS2 (E), PARP (F), BCL-XL (G), and Cl Casp3 (H), normalized to tubulin, analyzed via two-way ANOVA (n = 5 per group). Data shown as mean ± SE. *P < 0.05, ***P < 0.001.

Glucose-Stimulated Insulin Secretion Is Increased in Male Eif4ebp2^−/− NOD Islets

To assess whether 4E-BP2 affected insulin secretory responses to glucose stimulation, islets from prediabetic control, Eif4ebp2^+/−, and Eif4ebp2^−/− mice were stimulated with glucose and proinflammatory cytokines during static incubation. Under 2-mmol/L glucose conditions, insulin secretion was similar among all groups. During high (16.7 mmol/L) glucose stimulation, insulin secretion was significantly higher in islets from Eif4ebp2^−/− compared with those from control and Eif4ebp2^+/− mice (Fig. 3). The increase in insulin secretion in Eif4ebp2^−/− islets was lost after cytokine treatment, suggesting that the effects of Eif4ebp2 deficiency on the improved insulin secretory function in response to glucose challenge were sensitive to the effect of cytokines.

Figure 3.

Glucose-stimulated insulin secretion is increased in male NOD mutated islets. Glucose-stimulated insulin secretion in cytokine-treated nondiabetic control, Eif4ebp2^+/−, and Eif4ebp2^−/− islets subjected to different concentrations of glucose. Values were normalized to total insulin content, analyzed via two-way ANOVA (n = 6–8 per group). Data shown as mean ± SE. **P < 0.01, ***P < 0.001, ****P < 0.0001.

Male Eif4ebp2^−/− NOD Lymphocytes Are Less Diabetogenic Than Control Lymphocytes

Given the key role of autoimmunity in T1D, we investigated the contribution of 4E-BP2 to lymphocyte infiltration by comparing insulitis scores in male control, Eif4ebp2^+/−, and Eif4ebp2^−/− mice. Lymphocyte infiltration was similar across groups at 5, 16, and 20–25 weeks of age (Supplementary Fig. 3A). To assess whether Eif4ebp2-deficient lymphocytes contribute to T1D protection, we adoptively transferred splenic lymphocytes from diabetic or nondiabetic control or Eif4ebp2^−/− NOD male mice into NOD.SCID male mice (Fig. 4A). All mice receiving control diabetic lymphocytes developed diabetes by day 31, whereas 12.5% of mice receiving Eif4ebp2^−/− diabetic lymphocytes were still diabetes free by day 50 (Fig. 4B). Similarly, diabetes incidence was reduced in mice that received Eif4ebp2^−/− nondiabetic lymphocytes, where 87% of mice were diabetes free by day 50, compared with 33.3% among those receiving control nondiabetic lymphocytes (P = 0.07) (Fig. 4B). When comparing the nondiabetic with the diabetic group, a significant difference was seen in diabetes-free incidence between the control nondiabetic and control diabetic mice (33% vs. 0%, respectively; P = 0.04). This difference was more pronounced when comparing Eif4ebp2^−/− nondiabetic with Eif4ebp2^−/− diabetic mice (87.5% vs. 12.5%; P = 0.009). These results suggest that Eif4ebp2^−/− lymphocytes are less diabetogenic than control lymphocytes. Moreover, fed glucose levels were highest in mice receiving control diabetic and Eif4ebp2^−/− diabetic lymphocytes, followed by mice that received control nondiabetic lymphocytes and finally by mice that received Eif4ebp2^−/− nondiabetic lymphocytes (P < 0.05 compared with control diabetic group) (Fig. 4C). Higher fasting glucose levels were observed in mice with control diabetic lymphocytes, followed by mice with control nondiabetic and Eif4ebp2^−/− diabetic lymphocytes (Fig. 4D). NOD.SCID mice that received Eif4ebp2^−/− nondiabetic lymphocytes exhibited significantly lower fasting glucose levels when compared with the control diabetic lymphocyte group and a tendency toward significance when compared with the Eif4ebp2^−/− diabetic lymphocyte group (Fig. 4D). β-Cell mass was low in all groups, except in the mice that received Eif4ebp2^−/− nondiabetic lymphocytes (Fig. 4E), in which the highest diabetes free incidence was seen (Fig. 4B). Finally, insulitis was assessed in the mice with remaining islets. Mice that received Eif4ebp2^−/− nondiabetic lymphocytes had decreased insulitis, but this difference was not significant, and no difference was observed among the remaining groups (Fig. 4F). These results are consistent with a T1D protective effect produced by dampening the autoimmune response in Eif4ebp2^−/− lymphocytes.

Figure 4.

A: Adoptive transfer studies where splenic lymphocytes isolated from diabetic (D), nondiabetic (ND), control, or Eif4ebp2^−/− male mice were injected into NOD.SCID mice (control ND, n = 6; control D, n = 5; Eif4ebp2^−/− ND, n = 7; Eif4ebp2^−/− D, n = 8). B: Diabetes incidence followed for 50 days, analyzed via log-rank (Mantel-Cox) test. C: Fed glucose levels over time, analyzed via two-way ANOVA compared with the control D group. D: Fasting glucose levels, analyzed via two-way ANOVA, compared with the control D group. E: β-Cell mass, analyzed via two-way ANOVA. F: Insulitis score, analyzed via two-way ANOVA, compared with the control D group (n = 2–4 per group). Data shown as mean ± SE. ***P < 0.001.

Loss of Eif4ebp2 Alters the T-Cell Compartment, Promoting Treg Infiltration

Because lymphocytes from Eif4ebp2^−/− mice were less diabetogenic than those of control mice, we examined T-cell compartments from control and Eif4ebp2^−/− mice to identify the immune subpopulations modulated by 4E-BP2. Circulating CD8⁺, CD4⁺ FoxP3⁻, and CD4⁺ Foxp3⁺ Treg lymphocytes in the blood were comparable between Eif4ebp2^−/− and control mice (Fig. 5A). Proliferation, measured by Ki-67, was similar in these T-cell populations between groups (Fig. 5B). Survival, measured by BCL2, and suppressor function, measured by CD73 (19,20), were also similar in circulating Tregs (Fig. 5C–E). Splenic CD8⁺ and CD4⁺ T cells and Tregs were also comparable between groups (Supplementary Fig. 3B), but analyses of anti-CD3/CD28–stimulated splenic immune cell populations revealed that proliferative activated CD8⁺ T and cytotoxic CD8⁺ T cells (granzyme B⁺) were significantly decreased in Eif4ebp2^−/− compared with control mice (Fig. 5F). Effector T cells acquire expression of killing molecules, such as granzyme B, for effective cytotoxic killing activities. Although there were no significant differences in the proliferation of CD4⁺ T cells, a trend was observed toward decreased proliferation of effector CD4⁺ CD44⁺ T cells in Eif4ebp2^−/− compared with control mice (P = 0.07) (Fig. 5G). The evidence here suggests that loss of Eif4ebp2 impairs effector splenic CD8⁺ T cells, potentially dampening cytotoxic function within the immune attack.

Figure 5.

Loss of Eif4ebp2 modulates activated splenic immune cells. A: Percentage of CD8⁺, CD4⁺ Foxp3⁻, and CD4⁺ Foxp3⁺ T cells in the blood, analyzed via unpaired t test. B: Percentage of Ki-67 cells from blood CD8⁺, CD4⁺ Foxp3⁻, and CD4⁺ Foxp3⁺ T cells, analyzed via unpaired t test. C–E: Percentage of BCL2 (C), BCL2 MFI (D), and CD73 MFI in blood Tregs (E), measured via flow cytometry, analyzed via unpaired t test (n = 6 per group) F: Proliferative splenic CD8⁺ CD44⁺ T cells and CD8⁺ CD44⁺ granzyme B⁺ cytotoxic T cells, analyzed via unpaired t test, in control and Eif4ebp2^−/− mice. G: Proliferative splenic activated CD4⁺ CD44⁺ T cells and CD4⁺ CD44⁺ CD25⁺ effector T cells, analyzed via unpaired t test, in control and Eif4ebp2^−/− mice (n = 9–10 per group). Data shown as mean ± SE. *P < 0.05, **P < 0.01.

Given that insulitis scores were similar between male Eif4ebp2^−/− and control mice (Supplementary Fig. 3A), we hypothesized that the composition and/or functions of immune cell infiltrate in islets could be altered to favor regulation of effector T-cell responses. Therefore, T-cell compartments from the pancreas and pLNs were examined ex vivo from prediabetic Eif4ebp2^−/− and control male mice 16–20 weeks of age. We did not find any significant differences between control and Eif4ebp2^−/− pLN T-cell populations (Supplementary Fig. 3C), but we identified a trend toward decreased CD4⁺ FoxP3⁻ from the pLNs of Eif4ebp2^−/− compared with control mice. In the pancreas, we found a similar abundance of CD8⁺ T cells between the two groups but a significantly reduced abundance of CD4⁺ Foxp3⁻ in Eif4ebp2^−/− mice, similar to the pLNs. Eif4ebp2^−/− mice also exhibited a trend toward increased Treg populations (P = 0.05) (Fig. 6A). Significantly more proliferation was also seen in the Tregs of the pancreas in Eif4ebp2^−/− compared with control mice (P < 0.001) (Fig. 6B). BCL2 mean fluorescence intensity (MFI) (Fig. 6C), but not percentage (Fig. 6D), was higher in Tregs from Eif4ebp2^−/− compared with control mice (P = 0.0619), suggesting that Tregs from Eif4ebp2^−/− mice potentially exhibited increased survival. Moreover, CD73, an enzyme that converts AMP to adenosine, which augments immunosuppression (19,20), was significantly higher in Eif4ebp2^−/− compared with control Tregs, suggesting improved suppressive function (Fig. 6E). Assessment of Treg CD73 and CD25 levels from the spleen and pLNs showed no differences between groups (Supplementary Fig. 3D and E). Our results demonstrate a pancreas-specific immunosuppressive modulation that was not observed in other lymphoid tissues, suggesting that immunomodulatory effects in mice with global Eif4ebp2 deficiency are spatially restricted, likely as a result of T1D tissue-specific cues or local microenvironmental factors within the pancreas.

Figure 6.

Loss of 4E-BP2 alters the T-cell compartment within pancreatic islets and promotes Tregs. A: Percentage of CD8⁺, CD4⁺ Foxp3⁻, and CD4⁺ Foxp3⁺ T cells from the islets, analyzed via unpaired t test (n = 9–10 per group). B: Percentage of Ki-67⁺ cells from CD8⁺, CD4⁺ Foxp3⁻, and CD4⁺ Foxp3⁺ pancreatic T cells, analyzed via unpaired t test (n = 9–10 per group). C and D: BCL2 MFI (C) and BCL2⁺ percentage (D), analyzed via unpaired t test. E: Fold change percentage of CD73⁺ CD8⁺, CD4⁺ Foxp3⁻, and CD4⁺ Foxp3⁺ pancreatic T cells, analyzed via two-way ANOVA. Data shown as mean ± SE. *P < 0.05, ***P < 0.001.

Finally, immunostaining of pancreatic sections determined that the percentage of Tregs from CD3⁺ T cells was significantly higher in Eif4ebp2^−/− mice (Supplementary Fig. 4A and B), corroborating the flow cytometric findings from Fig. 6A. These results establish that deletion of 4E-BP2 alters the T-cell compartment within the pancreatic microenvironment, which inhibits effector CD8⁺ T cells and promotes Treg infiltration, proliferation, function, and survival.

Discussion

In this study, we demonstrate that whole-body Eif4ebp2 deletion confers protection from T1D in male, but not female, NOD mice. In males, protection was associated with preserved β-cell mass, despite comparable insulitis, accompanied by increased pancreatic Treg infiltration, reduced splenic cytotoxic CD8⁺ T cells, and fewer diabetogenic T-cell clones on adoptive transfer. Together, these findings reveal that loss of 4E-BP2 mitigates autoimmune attack while sustaining β-cell integrity, highlighting a sex-specific immune β-cell interaction. Importantly, our results position 4E-BP2 as a promising immunometabolic target for therapeutic intervention in T1D.

These findings extend our previous work in C57BL6 mice, where loss of 4E-BP2 modulated glucose homeostasis and enhanced proinsulin processing and β-cell proliferation by inducing IRS2 levels and regulating cap-dependent translation of carbopeptidase E (8,9). In a T1D model, we found that Eif4ebp2 loss confers protection by upregulation of IRS2 and reduced cleaved caspase 3 in cytokine-treated islets. Our data extend that framework by linking translational control via 4E-BP2 to enhanced survival signaling and improved secretory competence under inflammatory stress. Together, these findings suggest that targeting the 4E-BP2/IRS2 axis may reinforce β-cell resilience to autoimmune injury in T1D.

Although β-cell survival might explain part of the protection observed in this T1D model, the most striking results highlighted the immune contribution. Lymphocytes lacking Eif4ebp2^−/− displayed reduced diabetogenic potential and increased Treg pancreas infiltration. This is consistent with prior studies showing that 4E-BP2, more abundant than 4E-BP1 in lymphocytes, is critical for lymphocyte growth and proliferation (14). Indeed, the mTORC1 pathway is a central regulator of adaptive immunity, coordinating metabolic reprogramming through glycolysis and oxidative phosphorylation to enable effector T-cell differentiation. CD4⁺ T cells lacking mTOR fail to sustain glycolysis and cannot become effector cells (13). Conversely, combined Eif4ebp1/Eif4ebp2 deficiency has been shown to skew T-cell differentiation toward type 1 T helper cells with enhanced metabolic activity (21), underscoring the complex role of translational control in shaping T-cell fate. Although the precise mechanisms by which the mTORC1/4E-BP2 axis modulates T cells remain unclear, current evidence suggests that glycolysis-dependent metabolic reprogramming associated with T-cell activation is involved (22). Our observation that Eif4ebp2 deletion reduces pancreatic CD4⁺ FoxP3⁻ T cells and splenic CD4⁺ T-cell proliferation extends existing literature by pointing toward a previously underappreciated role of 4E-BP2 in restraining autoreactive effector T cells. Future studies are needed dissecting effects across specific CD4⁺ T-cell subpopulations and metabolic reprogramming, which could have been masked in our study, because we assessed a more general CD4⁺ FoxP3⁻ population.

Similar to CD4⁺ T cells, in CD8⁺ T cells mTORC1 drives glucose uptake, glycolysis, and HIF1 expression, which supports effector T-cell differentiation (23). mTORC1 activation promotes CD8⁺ T-cell terminal effector differentiation and function, whereas mTORC2 promotes memory cell generation (11). In NOD mice, splenic CD8⁺ T-cell activation significantly increases p-S6, reflecting increased mTORC1 activity, whereas mTORC1 inhibition with rapamycin reduces interferon-γ, granzyme B, and tumor necrosis factor-α production in CD8⁺ T cells (24). Although mTORC1 activation is essential for CD8⁺ T-cell effector activity, 4E-BP1 and 4E-BP2 suppressor proteins are also required for optimal protein translation, mitochondrial biogenesis, and proliferation in vitro of CD8⁺ T cells (12). This may seem paradoxical, because mTORC1 activation inhibits the translation repression of 4E-BP2. However, evidence suggests that 4E-BP1 and 4E-BP2 loss could prevent eIF4E from translocating from the nucleus to the cytoplasm (25), hindering the transport and translation of mRNA, mimicking mTOR inhibition. Aligning with these findings, activated CD8⁺ T cells increase eIF4E to support translation (26), whereas 4E-BP1 4E-BP2 loss reduces total and cytoplasmic eIF4E (12). In our study, deletion of 4E-BP2 alone was enough to reduce the proliferative and cytotoxic splenic CD8⁺ T cells in T1D, supporting the notion that 4E-BP2 may be required for CD8⁺ T-cell effector function. Additional research should assess protein translation across immune cell types lacking 4E-BP2 to further elucidate their roles and therapeutic potential in T1D.

Moreover, we observed increased pancreatic Treg infiltration in Eif4ebp2^−/− mice, associated with improved T1D outcomes, without changes in the spleen, pLN, or blood Treg count. Previous studies have shown that mTORC1 inhibition or deficiency blocks effector-like T helper cells and favors Treg differentiation (27–30), with some evidence suggesting both mTORC1 and mTORC2 need to be inhibited for optimal Treg induction (10,13). Additional findings indicate that eIF4E is translationally suppressed in Tregs and that eIF4E inhibition promotes Foxp3 induction (31). By contrast, our study showed increased Treg infiltration, proliferation, and suppressor function in the pancreas during autoimmune disease, after Eif4ebp2 deletion, which disrupts a key downstream arm of mTORC1. Together, these prior reports and our results suggest that mTORC1 is central to Treg differentiation and that loss of Eif4ebp2 enhances immunosuppression in T1D, possibly through reduced eIF4E activity resulting from 4E-BP2 deletion, as previously described (25).

Finally, sex-related differences influence diabetes and glucose homeostasis. In humans and in preclinical models, males and females differ greatly in their metabolic responses (32). Adult female C57BL6 mice display improved glucose tolerance associated with higher pS6, a downstream mTORC1 target (33), whereas female NOD mice develop T1D earlier and exhibit greater islet lymphocyte infiltration than males (34). Such differences in mTORC1/4E-BP2 signaling may account for the absence of 4E-BP2–mediated protection in females, reflecting sex-specific effects within islets or immune responses. Future studies using bone marrow transplantation in male and female NOD mice, together with conditional 4E-BP2 deletion, will be critical to define these sex-specific mechanisms and delineate the direct contributions of 4E-BP2 across target tissues.

In conclusion, this study showed that lack of Eif4ebp2 in NOD male mice protects against T1D development by dampening autoimmune responses and preserving β-cell mass, supporting 4E-BP2 as a novel target for immune modulation in T1D treatment.

This article contains supplementary material online at https://doi.org/10.2337/figshare.30417214.

Funding Statement

This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health (NIH), grant R01DK132103 to E.B.-M. and A.B. and grants R01DK133183 and R01DK138471 to E.B.-M. as well as the Medical Research Service, U.S. Department of Veterans Affairs, grant T01BX002728 to E.B.-M. Research reported in this publication was performed in part at the Cancer Modeling Shared Resource, Sylvester Comprehensive Cancer Center, University of Miami (RRID SCR_022891), which is supported by the National Cancer Institute, NIH, under award P30CA240139.