Ethyl Pyruvate Stimulates Regulatory T Cells and Ameliorates Type 1 Diabetes Development in Mice

Abstract

Type 1 diabetes (T1D) is an autoimmune disease in which a strong inflammatory response causes the death of insulin-producing pancreatic β-cells, while inefficient regulatory mechanisms allow that response to become chronic. Ethyl pyruvate (EP), a stable pyruvate derivate and certified inhibitor of an alarmin-high mobility group box 1 (HMGB1), exerts anti-oxidant and anti-inflammatory properties in animal models of rheumatoid arthritis and encephalomyelitis. To test its therapeutic potential in T1D, EP was administered intraperitoneally to C57BL/6 mice with multiple low-dose streptozotocin (MLDS)-induced T1D. EP treatment decreased T1D incidence, reduced the infiltration of cells into the pancreatic islets and preserved β-cell function. Apart from reducing HMGB1 expression, EP treatment successfully interfered with the inflammatory response within the local pancreatic lymph nodes and in the pancreas. Its effect was restricted to boosting the regulatory arm of the immune response through up-regulation of tolerogenic dendritic cells (CD11c⁺CD11b^-CD103⁺) within the pancreatic infiltrates and through the enhancement of regulatory T cell (Treg) levels (CD4⁺CD25^highFoxP3⁺). These EP-stimulated Treg displayed enhanced suppressive capacity reflected in increased levels of CTLA-4, secreted TGF-β, and IL-10 and in the more efficient inhibition of effector T cell proliferation compared to Treg from diabetic animals. Higher levels of Treg were a result of increased differentiation and proliferation (Ki67⁺ cells), but also of the heightened potency for migration due to increased expression of adhesion molecules (CD11a and CD62L) and CXCR3 chemokine receptor. Treg isolated from EP-treated mice had the activated phenotype and T-bet expression more frequently, suggesting that they readily suppressed IFN-γ-producing cells. The effect of EP on Treg was also reproduced in vitro. Overall, our results show that EP treatment reduced T1D incidence in C57BL/6 mice predominantly by enhancing Treg differentiation, proliferation, their suppressive capacity, and recruitment into the pancreas.

Keywords: ethyl pyruvate; immunoregulation; inflammation; regulatory T cells (Treg); type 1 diabetes (T1D).

Figure 1

The effect of EP on T1D development. (A) Diagram of diabetes induction, EP treatment and diabetes monitoring by weekly measurements of blood glucose levels. (B) T1D incidence presented as a proportion of C57BL/6 mice with glycemia higher than 11 mmol/L. (C) The proportion of pancreatic islets without immune cell infiltrates, with infiltrates surrounding the islets (peri-insulitis) and with infiltrates within the islet (insulitis). Representative images of pancreatic islets from MLDS-treated mice (D) or EP-treated mice (E), stained with hematoxylin. (F) Insulin expression was determined by analyzing fluorescence intensity with Fiji software. Representative images of pancreatic islets from a control healthy animal (G), MLDS-treated mice (H) or EP-treated mice (I) stained for insulin visualization (red) and with Hoechst 33342 (nucleus – blue). (J) HMGB1 expression within pancreatic β-cells was determined using the Color Picker Threshold Plugin within Icy software. Representative images of pancreatic islets from a control healthy animal (K), MLDS-treated mice (L) or EP-treated mice (M) stained for HMGB1. (N) The proportion of mice with ketones in urine. All groups consisted of 7-10 animals. *p < 0.05 represents a statistically significant difference between MLDS+EP-treated compared to MLDS-treated mice.

Figure 2

The influence of EP on innate antigen-presenting cells. The proportion of all cells, isolated from the spleen, pancreatic lymph nodes (PLN) or pancreatic infiltrates, was measured by flow cytometry. (A) The proportion of CD11c⁺ dendritic cells, along with representative dot plots. (B) The proportion of tolerogenic DC (CD11c⁺CD11b⁻CD103⁺), along with representative dot plots (first gated on live CD103⁺ cells, followed by the gate on CD11c⁺CD11b⁻). Statistical analysis for CD11c⁺ and tolerogenic DC was performed by Mann-Whitney non-parametric test. (C) Proportion of IL-10⁺ cells within MHC II⁺ population, along with representative dot plots. Secretion of IL-1β (D) and TNF (E) from LPS-treated splenocytes cultured ex vivo for 48 h and measured by ELISA. All measurements were performed on samples from at least 7 animals per group. *p < 0.05 represents a statistically significant difference between cells of MLDS+EP compared to those of MLDS-treated mice.

Figure 3

Phenotypic analysis of adaptive immune cells after EP treatment. All cell proportions were measured by flow cytometry. (A) The proportion of Th (CD4⁺), cytotoxic lymphocytes (CD8⁺) or B lymphocytes (B220⁺ or CD19⁺) in spleen, PLN or pancreatic infiltrates. (B) The proportion of activated cytotoxic lymphocytes (CD8⁺CD44⁺) in the pancreatic infiltrates. (C) The proportion of regulatory B cells (CD19⁺CD5⁺IL-10⁺) within PLN and pancreatic infiltrates. The proportion of Th subsets: Th1 (CD4⁺IFN-γ⁺), Th2 (CD4⁺IL-4⁺), Th17 (CD4⁺IL-17⁺) and Treg (CD4⁺CD25^high) within the spleen (D), PLN (E), and pancreatic infiltrates (F) of MLDS or MLDS+EP-treated mice. Representative dot plots for CD4⁺CD25^high on the right hand side. All measurements were performed on samples from 7 animals per group. *p < 0.05 represents a statistically significant difference between cells of MLDS+EP compared to those of MLDS-treated mice.

Figure 4

Phenotypical characterization of Treg after EP treatment. (A) Proportion of Treg (CD4⁺CD25^highFoxP3⁺) within PLN and pancreatic infiltrates, along with representative dot plots (first gated on live FoxP3⁺ cells, followed by the gate on CD4⁺CD25^high). (B) The proportion of Treg within PLN that express GITR and are negative for CD127, along with the representative dot plots (first gated on live CD127⁻GITR⁺ cells, followed by the gate on CD4⁺CD25^high). All measurements were performed on samples from at least 7 animals per group. Statistical analysis was performed by Mann-Whitney non-parametric test. *p < 0.05 represents a statistically significant difference between cells of MLDS+EP compared to those of MLDS-treated mice.

Figure 5

Functional characterization of Treg after EP treatment. (A) CTLA-4 protein expression in CD25⁺ cells purified from the pool of lymphoid tissues, normalized to the expression of β-actin, along with the representative blot. (B) The proportion of TGF-β-expressing cells within Treg, along with representative dot plots. (C) IL-10 protein expression in CD25⁺ cells purified from a pool of lymphoid tissues, normalized to the expression of β-actin, along with the representative blot. (D) The proportion of IL-10-expressing Treg within pancreatic infiltrates, along with representative dot plots. (E) The level of inhibition of effector T cell (Teff–CD4⁺CD25⁻) proliferation after co-culture with Treg (CD4⁺CD25⁺) cells purified from a pool of lymphoid tissues of MLDS or MLDS+EP-treated mice. Proliferation was measured after 72 h of incubation by CFSE dilution (in Teff). (F) Proportion of divided Teff, or Teff cultured in the presence of Treg (1:1 ratio). Representative dot plots are given below. All measurements were performed on samples from 7 animals per group. *p < 0.05 represents a statistically significant difference between cells of MLDS+EP compared to those of MLDS-treated mice.

Figure 6

The effect of EP on Treg differentiation and proliferation in vivo. (A) IL-2 protein expression in PLN, along with the representative blot. (B) The proportion of non-Treg producers of TGF-β in the pancreatic infiltrates, along with representative dot plots (first gated on live non-CD25^high cells, followed by the gate on TGF-β⁺). (C) The proportion of proliferating Treg (CD25^highKi67⁺) in PLN and pancreatic infiltrates. Representative dot plots (first gated on live CD25^high cells, followed by the gate on Ki67⁺) for pancreatic infiltrates are shown. All measurements were performed on samples from 7 animals per group. *p < 0.05 represents statistically significant difference between cells of MLDS+EP compared to those of MLDS-treated mice.

Figure 7

The effect of EP on Treg migratory abilities. (A) CD11a and CD62L expression on Treg within PLN, measured by mean fluorescence intensity (MFI). Representative histograms are shown. (B) The proportion of CXCR3⁺ cells within activated Th cells (CD4⁺CD25^med) or within Treg (CD4⁺CD25^high) from PLN. Representative dot plots show the first gate on either live CD4⁺CD25^med or live CD4⁺CD25^high cells, followed by the gate on CXCR3⁺. (C) Migration of CD25⁺ cells purified from a pool of lymphoid tissues of MLDS or MLDS+EP-treated mice in a chemotaxis assay toward pancreatic islets or CXCL12 (10 ng/ml). (D) CD25^highCD103⁺ proportion within PLN and pancreatic infiltrates. Representative dot plots for pancreatic infiltrates are shown. All measurements were performed on samples from 7 animals per group. *p < 0.05 represents a statistically significant difference between cells of MLDS+EP compared to those of MLDS-treated mice.

Figure 8

Activation and phenotype of Treg and effector T cells. The proportion of activated (CD44⁺), T-bet⁺, and RORγT⁺ cells within Treg (CD4⁺CD25^high) (A,C,E) or within effector T population (CD4⁺CD25^med) (B,D,F) was determined in PLN or infiltrates by flow cytometry. Representative dot plots are shown on the right hand side. Cells were first gated on live CD4⁺, followed by the gate on T-bet⁺ in CD4⁺CD25^med or CD4⁺CD25^high population. All measurements were performed on samples from 7 animals per group. *p < 0.05 represents a statistically significant difference between cells of MLDS+EP compared to those of MLDS-treated mice.

Figure 9

In vitro effect of EP on Th cell differentiation. (A) The proportion of Th1 (CD4⁺IFN-γ⁺), Th17 (CD4⁺IL-17⁺), and Treg (CD4⁺CD25^highFoxP3⁺) cells after 96 h of incubation of CD4⁺CD25⁻ cells with adequate stimulation (described in Material and methods) and in the presence of EP (125 μM), added 24 h after the start of cell cultivation. Representative dot plots are shown. Cells were first gated on live CD4⁺, followed by the gate on IFN-γ⁺, IL-17⁺ or CD25^highFoxpP3⁺. (B) The proportion of Treg that expressed GITR, CTLA-4, and PD-1, determined by flow cytometry. Representative dot plots show the first gate on live CTLA-4⁺, GITR⁺ or PD-1⁺ population, followed by the gate on CD4⁺CD25^high. (C) The proportion of Treg after incomplete stimulation (anti-CD3+anti-CD28+IL-2) after 5 days of incubation of CD4⁺ cells in the presence or absence of EP. Representative dot plots show the first gate on live CD4⁺ cells, followed by the gate on CD25^highFoxpP3⁺. (D) IL-10 concentration in supernatants of EP-treated CD4⁺ cells in the presence of complete Treg stimulation (anti-CD3/anti-CD28+IL-2+TGF-β), after 5 days of incubation, measured by ELISA. *p < 0.05 represents a statistically significant difference between cells of MLDS+EP compared to those of MLDS-treated mice.

Figure 10

Proposed mechanism of EP's beneficial effect in T1D. EP enhances the regulatory arm of the immune response during T1D pathogenesis. From pancreatic lymph nodes, tolDC (CD11c⁺CD11b⁻CD103⁺) migrate into the pancreatic islets and their proportion becomes significantly higher. Also, Treg (CD4⁺CD25^high) in the pancreatic lymph nodes increase their suppressive properties (CTLA-4, IL-10, and TGF-β) and migrate into the inflamed pancreatic islet (increased CD11a and CD62L expression and proportion of CXCR3⁺). In the islet, since Treg are already activated (CD44⁺), they proliferate (Ki67⁺), and are retained at the site (CD103⁺). Also, some of Treg express T-bet (increased proportion after EP treatment) and thereby inhibit T-bet⁺ effector cells (CD25^med) within the PLN and pancreatic infiltrates. Finally, EP may stimulate in vitro and in vivo differentiation of Treg through enhancement of TGF-β production. All these events result in the preservation of β-cell function (insulin production) and reduction of HMGB1 expression.