Novel engineered B lymphocytes targeting islet-specific T cells inhibit the development of type 1 diabetes in non-obese diabetic Scid mice

Abstract

Introduction: In this study, we report a novel therapeutic approach using B lymphocytes to attract islet-specific T cells in the non-obese diabetic (NOD) mouse model and prevent the development of autoimmune diabetes. Rather than using the antibody receptor of B cells, this approach utilizes their properties as antigen-presenting cells to T cells.

Methods: Purified splenic B cells were treated with lipopolysaccharide, which increases regulatory B (Breg) cell function, then electroporated with mRNA encoding either chimeric MHC-I or MHC-II molecules covalently linked to antigenic peptides. Immunoregulatory functions of these engineered B cells (e-B cells) were tested by in vitro assays and in vivo co-transfer experiments with beta-cell-antigen-specific CD8⁺ or CD4⁺ T cells in NOD.Scid mice, respectively.

Results: The e-B cells expressing chimeric MHC-I-peptide inhibited antigen-specific CD8⁺ T-cell cytotoxicity in vitro. The e-B cells expressing chimeric MHC-II-peptide induced antigen-specific CD4⁺ T cells to express the regulatory markers, PD-1, ICOS, CTLA-4, Lag3, and Nrp1. Furthermore, e-B cells encoding the chimeric MHC-I and MHC-II peptide constructs protected NOD.Scid mice from autoimmune diabetes induced by transfer of antigen-specific CD8⁺ and CD4⁺ T cells.

Discussion: MHC-peptide chimeric e-B cells interacted with pathogenic T cells, and protected the host from autoimmune diabetes, in a mouse model. Thus, we have successfully expressed MHC-peptide constructs in B cells that selectively targeted antigen-specific cells, raising the possibility that this strategy could be used to endow different protective cell types to specifically regulate/remove pathogenic cells.

Keywords: CD4+ T cells; CD8+ T cells; NOD mice; regulatory B cells; type 1 diabetes.

Figure 1

The expression of INS_B15-23/hβ₂m construct. (A) Scheme of the cell-surface expression of INS_B15-23/hβ₂m construct on e-B cells and the interaction with G9Cα^−/− CD8⁺ T cells. (B) The expression of INS_B15-23 construct was tracked by extracellular staining of hβ₂m. B cells from NOD mice were activated with 5 μg/ml LPS for 16–18 h Activated B cells were washed and electroporated without (Mock) or with mRNA of the INS_B15-23/hβ₂m construct, and rested for a minimum of 3 h at room temperature. After resting, 5×10⁶ cells/mouse of B cells were washed in saline and injected intravenously into NOD.Scid female mice. Spleen tissues of the recipient mice were collected 24 and 48 h post-injection. The plots are representative of three independent experiments, with six mice in total. (C) The expression of CD69, as an activation marker, on G9Cα^−/− CD8⁺ T cells co-cultured with NOD B cells with no construct (Mock) or non-relevant-antigen construct–influenza Nucleoprotein antigen (NP_50-57) or expressing the INS_B15-23/hβ2m construct. The B cells were electroporated without (Mock) or with the non-relevant NP_50-57 construct or the INS_B15-23/hβ2m construct and co-cultured with G9Cα^−/− CD8⁺ T cells at different ratios for 24 h. The numbers in these plots show the frequency of CD69⁺G9CD8⁺ T cells. The plots are representative of three independent experiments. (D) Line graphs showing the percentage of CD69⁺CD8⁺ T cells after 24 h of co-incubation with B cells expressing the INS_B15-23/hβ₂m construct or mock-transfected, which were pre-activated using LPS or anti-CD40 antibody. The two mock-transfected conditions were very close to each other and appear superimposed as there was no/little activation for either of these. The plots show mean ± SEM of three independent experiments, with seven mice in total. (E) Summary graph showing significant difference in IL-10 secretion between LPS-stimulated B cells and anti-CD40 stimulated B cells. B cells were stimulated and electroporated without construct. After 24 h culture, supernatants were collected and IL-10 concentration in the supernatant was detected by ELISA. Statistical analysis was performed using Mann–Whitney test. *p < 0.05, ****p < 0.0001.

Figure 2

Engineered B cells (e-B cells) expressing 2.5HIP/I-A^g7 construct activate BDC2.5 CD4⁺ T cells. (A) The scheme of the 2.5HIP/I-A^g7 construct expressed on the cell surface of e-B cells and the interaction with BDC2.5 CD4⁺ T cells. (B) The expression of CD69 and CD44, as activation markers, on BDC2.5 CD4⁺ T cells after 24 h co-culture with e-B cells transfected without (mock) or with non-relevant or 2.5HIP/I-A^g7 construct. The B cells were electroporated without (mock) or with the 2.5HIP/I-A^g7 construct, or a non-relevant antigen (NP50-57) construct. e-B cells were co-cultured with BDC2.5 CD4⁺ T cells at different ratios for 24 h. The plots are representative of three independent experiments and the line graph is the mean ± SEM of the three experiments. Statistical analysis was performed using two-way ANOVA and multiple comparisons. Significance shown indicates the result between mock-transfected and 2.5HIP/I-A^g7 construct group. ***p < 0.001, ****p < 0.0001. (C) Proliferation of BDC2.5 CD4⁺ T cells co-cultured with e-B cells transfected without (control) or with 2.5HIP/I-A^g7 construct for 24 h. BDC2.5 CD4⁺ T cells were labeled with CellTrace Violet and co-cultured with e-B cells at B cell:T cell 1:1 and 5:1 ratios. The frequency of gated proliferated cells (in duplicate), with loss of CellTrace violet indicating proliferation, is plotted in the line-graph, shown as the mean ± SEM of three independent experiments. Statistical analysis was done by Student’s t test. *p < 0.05. **p <0.01.

Figure 3

Enriched IL-10-competent CD9⁺ B cells. (A) CD9⁺ B cell enrichment. B cells were isolated, then stimulated with LPS for 18 h. Activated B cells were harvested and stained using FITC-conjugated anti-CD9 antibodies. CD9⁺ B cells were positively selected using anti-FITC magnetic beads. CD9⁺ and CD9⁻ populations were then restimulated using PMA, ionomycin, and monensin for 3.5 h to perform intracellular staining for cytokines, followed by staining with the appropriate antibodies for IL-10 and extracellular LAP. (B) A scatter plot shows the differential expression of intracellular IL-10 and TGF-β for the CD9⁺ and CD9⁻ B cell populations; the individual points on the graph show results from three independent experiments. (C) Characterization of CD9⁺ and CD9⁻ B cell populations, showing distribution within MZ, T1, T2MZ, and FO zones (36), from three independent experiments. Statistical analysis was performed by multiple comparisons. *p <0.05, **p < 0.01. ns not statistically significant.

Figure 4

CD9⁺ B cells inhibit the cytotoxicity of CD8⁺ T cells in vitro. (A) Representative flow cytometry plots demonstrating cytotoxicity assay using naïve G9Cα^−/− CD8⁺ T cells co-cultured with PKH-labeled P815 target cells coated with INS_B15-23 peptide (1 or 5 μg/ml) at an effector:target ratio of 10:1. CD9⁺ or CD9^– e-B cells (mock-transfected or expressing INS_B15-23/hβ₂m construct) were added to the cytotoxicity assay using an equal number of e-B cells and CD8⁺ T cells. The controls did not have added e-B cells; the negative control was target cells alone and the positive control was G9Cα^−/− CD8⁺ T cells together with target cells. The cells were co-cultured at 37°C, 5% CO₂, for 16 h. After incubation, cells were collected and cytotoxicity was determined by the percentage of dead PKH-labeled P815 target cells. (B) Summary of cytotoxicity assays with added CD9⁺ or CD9⁻ e-B cell subsets to demonstrate effect on cytotoxicity by G9Cα^−/− CD8 T cells towards INS_B15-23 peptide-coated target cells. The data are a summary of three independent experiments. Statistical analysis was performed by multiple comparisons. *p < 0.05. ns, not statistically significant.

Figure 5

Expression of cell surface markers and intracellular cytokines in BDC2.5 CD4⁺ T cells co-cultured with e-B cells transfected with or without 2.5HIP/I-A^g7 construct. The percentage of cell surface markers (CD62L, PD-1, Lag3, Nrp1, CTLA-4, ICOS, FoxP3, and LAP) and intracellular cytokine (IL-10) in the BDC2.5 CD4⁺ T cells co-cultured with CD9⁺, CD9⁻, or total e-B cells transfected with 2.5HIP/I-A^g7 construct or without (mock-transfected). Naïve BDC2.5 CD4⁺ T cells were co-cultured with e-B cells with or without the 2.5HIP/I-A^g7 construct at a 1:1 ratio for 3 days. Cells were treated using PMA, ionomycin, and monensin for 3.5 h before harvest for intracellular staining of cytokines. The graphs show the percentage expression of each of the indicated markers in the CD4⁺ T cells. These data show mean values ± SEM of five independent experiments. Statistical analysis was performed using two-way ANOVA, with multiple comparisons for the difference between time points, and multiple t-tests were used to compare the difference between e-B cell types. ns non-significant, +p < 0.05, ++p < 0.01, +++p < 0.001, ++++p < 0.0001, comparisons between 2.5HIP/I-A^g7 construct vs. mock transfection within each B-cell group and are shown in black on the figure. Where there was no statistical significance (ns) between the 2.5HIP/I-A^g7 construct vs. mock transfection within each B-cell group, at each time point, as for PD-1, ICOS, CTLA-4, Foxp3, IL-10, and LAP, these were performed but are not shown on the graph. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, comparisons between different B-cell groups with 2.5HIP construct and are shown in color on the figure. All comparisons that showed any statistical significance are shown on the graph. ns, not statistically significant.

Figure 6

Cytokine production after co-culture of e-B cells and antigen-specific BDC2.5 CD4⁺ T cells. Cytokine levels in supernatant from e-B cells that have been either Mock or 2.5HIP/I-A^g7 transfected and cultured with BDC2.5 CD4⁺ T cells. Naïve BDC2.5 CD4⁺ T cells were co-cultured with e-B cells with or without the 2.5HIP/I-A^g7 construct at a 1:1 ratio for 3 days. Supernatants were collected on Day 3 and analyzed using MSD. The dotted line shows levels of cytokines from CD4 T cells alone. (A) IFN-γ. (B) IL-10. (C) IL-4. (D) IL-6. (E) TNF-α. Statistical analysis was performed using two-way ANOVA with multiple comparisons. Significance was determined by the p values using multiple comparisons between Mock and HIP/I-A^g7 groups with the same cell types. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 7

e-B cells delay the development of T1D induced by pathogenic CD8⁺ T cells in NOD.Scid mice. (A) Schematic diagram illustrating the in vivo adoptive transfer. (B) Diabetes incidence in mice co-adoptively transferred at a 1:1 e-B cell:T cell ratio, with the CD9⁺ or CD9⁻ e-B cells either expressing the INS_B15-23 construct or Mock-transfected (5 × 10⁶ cells) together with insulin-reactive CD8⁺ T cells (5 × 10⁶ cells). Insulin-reactive CD8⁺ T cells (5 × 10⁶ cells) were transferred alone as a control. (C) Diabetes incidence in mice co-adoptively transferred at a 5:1 e-B cell:T cell ratio with e-B cells (25 × 10⁶ cells) expressing the INS_B15-23 construct and insulin-reactive CD8⁺ T cells (5 × 10⁶ cells) per recipient mouse. Insulin-reactive CD8⁺ T cells (5 × 10⁶ cells) were transferred alone as a control. Log rank test **p < 0.01.

Figure 8

e-B cells prevent diabetes development induced by pathogenic CD4⁺ T cells in NOD.Scid mice. (A) Diabetes incidence in mice co-adoptively transferred at a 1:1 e-B cell:T cell ratio, with CD9⁺ or CD9⁻ or total e-B cells either expressing the 2.5HIP/I-A^g7 or mock-transfected (5 × 10⁶ cells) together with BDC2.5 CD4⁺ T cells (5 × 10⁶ cells). BDC2.5 CD4⁺ T cells (5 × 10⁶ cells) were transferred alone as a control. (B) The results from different B-cell groups shown in A were pooled to show overall protection. * p<0.05, ** p<0.01, **** p<0.0001.

Figure 9

Insulitis scoring of the pancreas tissue from adoptive transfer of mock-transfected or 2.5HIP/I-A^g7 e-B cells and BDC2.5 CD4⁺ T cells at the end point (diabetes) or 28 days (non-diabetes). All the samples from the mice transferred with mock-transfected e-B cells were diabetic; all the samples from the mice transferred with e-B cells transfected with 2.5HIP/I-A^g7 were non-diabetic. (A) Representative examples of the insulitis within the pancreatic islets from each group, with separate staining (shown as red) for insulin, B cells (B220), and CD4 T cells (CD4). (B) Insulitis scores of the mice groups that received BDC2.5 CD4⁺ T cells together with CD9^+/− e-B cells with or without the 2.5HIP/I-A^g7 construct. We examined 73–119 islets from two to three mice in each group: CD9⁺ e-B cell Mock (n = 2, 85 islets), CD9⁻ e-B cell Mock (n = 2, 116 islets), CD9⁺ B cell 2.5HIP/I-A^g7 (n = 3, 73 islets), and CD9⁻ B cell 2.5HIP/I-A^g7 (n = 3, 119 islets). Insulitis was scored separately for the CD4⁺ T cells and the B cells. Insulitis scores were as follows: 0 = no insulitis, 1 = peri-insulitis, 2 = less than 50% infiltration, 3 = greater than 50% infiltration. Data were analyzed by the Chi-square test.