Human MAIT cells inhibit alloreactive T cell responses and protect against acute graft-versus-host disease

Abstract

Adoptive transfer of immunoregulatory cells can prevent or ameliorate graft-versus-host disease (GVHD), which remains the main cause of nonrelapse mortality after allogeneic hematopoietic stem cell transplantation. Mucosal-associated invariant T (MAIT) cells were recently associated with tissue repair capacities and with lower rates of GVHD in humans. Here, we analyzed the immunosuppressive effect of MAIT cells in an in vitro model of alloreactivity and explored their adoptive transfer in a preclinical xenogeneic GVHD model. We found that MAIT cells, whether freshly purified or short-term expanded, dose-dependently inhibited proliferation and activation of alloreactive T cells. In immunodeficient mice injected with human PBMCs, MAIT cells greatly delayed GVHD onset and decreased severity when transferred early after PBMC injection but could also control ongoing GVHD when transferred at delayed time points. This effect was associated with decreased proliferation and effector function of human T cells infiltrating tissues of diseased mice and was correlated with lower circulating IFN-γ and TNF-α levels and increased IL-10 levels. MAIT cells acted partly in a contact-dependent manner, which likely required direct interaction of their T cell receptor with MHC class I-related molecule (MR1) induced on host-reactive T cells. These results support the setup of clinical trials using MAIT cells as universal therapeutic tools to control severe GVHD or mucosal inflammatory disorders.

Keywords: Immunology; Stem cell transplantation; T cells; Transplantation.

Figure 1. Human MAIT cells dose-dependently inhibit the in vitro proliferation of alloreactive T cells.

(A) MLR was performed using total or MAIT-depleted CFSE-labeled PBMCs (“responders”) cocultured for 6 days with irradiated allogeneic PBMCs (“stimulators”). Proliferation was quantified as the percentage of CFSE^lo cells in non-MAIT (Vα7.2^–tetramer^–) responder T cells on day 6. Left panel shows a representative example, and the graph indicates the percentage of proliferating T cells in paired total or MAIT-depleted PBMCs (n = 10 different recipient/donor pairs). (B–D) FACS-sorted purified MAIT cells were added to the MLR at different MAIT/MAIT-depleted PBMC ratios. (B) Representative staining in the absence (ctrl) or presence of MAIT cells at 1:1 ratio. (C) Percentage of inhibition of responder T cell proliferation in the presence (versus absence) of MAIT cells purified from autologous (responder-derived) or allogeneic (stimulator-derived) PBMCs (n = 6 independent experiments). (D) Inhibition of proliferation by MAIT cells was analyzed separately in CD4⁺ and CD8⁺ responder T cells (n = 7 independent experiments). (E) Purified CFSE-labeled CD4⁺ T cells were cultured with allogeneic CD3^– PBMCs in the presence of purified MAIT cells or effector memory CD8 T cells (non-MAIT CD8⁺CD45RA^–CCR7^– T_EM cells, CD8_EM) at the indicated MAIT (or CD8_EM)/CD4⁺ T cell ratios. Results are shown as mean ± SEM (C–E). *P 0.05, **P 0.01, ***P 0.001, ****P 0.001, from paired t tests (A) or 2-way ANOVA followed by Holm-Šídák multiple comparisons test (D and E).

Figure 2. The inhibitory effect of MAIT cells on alloreactive T cells is partially contact dependent, requires TCR-MR1 interaction, and leads to inhibition of alloreactive T cell activation.

(A) Responder CFSE-labeled CD4⁺ T cells with stimulator allogeneic CD3^– PBMCs, and MAIT cells (or CD8_EM) at a 1:2 MAIT/CD4⁺ T cell ratio, were cultured in the upper and/or lower chamber of a Transwell plate, as indicated for each condition. Percentage (mean ± SEM) of inhibition of responder T cell proliferation (versus “condition 1” as reference) is shown for each culture setting. (B) Representative MR1 expression on responder CD4⁺ T cells at baseline (D0) and D6 of the MLR (where indicated, 5-OP-RU or Ac-6-FP was used to stabilize MR1 surface expression). (C) Human alloreactive (CFSE^–) or nonalloreactive (CFSE⁺) human CD4⁺ T cells were sorted (at day 6 of the MLR) and cocultured with murine MAIT “reporter” cells obtained by isolating splenocytes from double-transgenic mice (Vα19 and Vβ8 TCR, Cα^–/–, MR1^–/–). After 16 hours in the absence or presence of 5-OP-RU (100 nM), expression of CD69 and CD25 was measured on murine MAIT reporter cells (MR1-tetramer⁺ T cells). (D) Percentage inhibition (mean ± SEM) of responder CD4⁺ T cell proliferation by MAIT cells (1:2 ratio) in the absence or presence of the inhibitory MR1 ligand Ac-6-FP at indicated concentrations. (E) Expression of Nur77 transcription factor in responder CD4⁺ T cells at D6 of the MLR. (F) Expression level (mean %, ± SEM) of indicated markers as assessed by spectral cytometry on responder CD4⁺ T cells on day 6 of the MLR in the absence (Ctrl) or presence of MAIT cells (1:2 ratio) alone or with Ac-6-FP (1 μM). (G) Percentage inhibition (mean ± SEM) of responder CD4⁺ T cell proliferation by MAIT cells (1:2 ratio) in the absence or presence of the activating MR1 ligand 5-OP-RU at indicated concentrations. Results are representative of at least of 3 independent experiments. *P 0.05, **P 0.01, ***P 0.001, ****P 0.001, from paired t tests or 1-way ANOVA followed by Tukey’s multiple comparisons test (A) or followed by Dunnett’s (D, F, and G).

Figure 3. Adoptive transfer of human MAIT cells protects from xeno-GVHD in immunodeficient mice.

(A) Irradiated (1.3 Gy) NSG mice were injected i.v. with 5 × 10⁶ human total PBMCs (huPBMCs) or MAIT-depleted PBMCs. (B) Weight loss and GVHD scoring were assessed from day 0 to sacrifice (day 35 or when the mice reached a GVHD score of 6) (n = 11 mice with weight data, and 8 with GVHD scoring, in the total PBMC group, and n = 8 mice in the MAIT-depleted PBMC group, from 3 independent experiments). (C) Frequencies of human CD45⁺ leukocytes among living cells in blood and spleen of mice sacrificed on day 35. (D–F) Irradiated NSG mice were injected with 5 × 10⁶ huPBMCs on day 0 without (n = 7, black), or with additional transfer of 1 × 10⁶ purified MAIT cells on day 0 (n = 6, red), day 10 (n = 3, orange), or day 25 (n = 4, brown). (E) Weight loss and GVHD scoring were assessed from day 0 to death (i.e., when GVHD score reached 6). Brackets indicate that statistical comparisons were performed between PBMCs and PBMCs + MAIT pooled groups. (F) Kaplan-Meier plots showing mouse survival in the indicated groups. Results are shown as mean ± SEM (B, C, and E). *P 0.05, **P 0.01, ***P 0.001, ****P 0.0001, from 2-way ANOVA followed by Holm-Šídák multiple comparisons test (PBMCs versus MAIT-depleted PBMCs in B, PBMCs versus PBMCs + MAIT pooled group in E), t tests (C), and log-rank tests (F).

Figure 4. MAIT cells inhibit proliferation, activation, and effector function of xenoreactive T cells.

Irradiated (1.3 Gy) NSG mice were injected with 5 × 10⁶ huPBMCs on day 0 without (n = 18, black) or with additional transfer of 1 × 10⁶ purified MAIT cells on day 0 (n = 6, red), day 10 (n = 8, orange), or day 25 (n = 7, brown), or of effector memory CD8⁺ T cells on day 0 (n = 4, CD8_EM, gray). Mice were sacrificed on day 35, and cells from peripheral blood, spleen, liver, and colon were isolated. The proportion of human CD45⁺ leukocytes (huCD45⁺) among viable cells (A) and the CD4/CD8 T cell ratio (B) are shown in indicated compartments. (C) The proportion of intracellular Ki67⁺ cells (proliferating cells) in peripheral blood CD4⁺ and CD8⁺ T cells was determined by flow cytometry. (D–F) PBMCs were stimulated with phorbol 12-myristate 13-acetate (PMA)/ionomycin for 5 hours with brefeldin A/monensin added after 1 hour. Proportions of perforin⁺, IFN-γ⁺, and TNF-α⁺ cells among CD4⁺ and CD8⁺ T cells were determined by flow cytometry following intracellular staining. Results show individual values and mean ± SEM from 2 independent experiments (represented by circles or triangles). *P 0.05, **P 0.01, ***P 0.001, ****P 0.001, from 1-way ANOVA followed by Tukey’s multiple comparisons test.

Figure 5. Circulating levels of IFN-γ and IL-10 correlate with MAIT cell–mediated control of xenoreactive T cell expansion.

Irradiated (1.3 Gy) NSG mice were injected with 5 × 10⁶ huPBMCs on day 0 without (n = 14, black) or with additional transfer of 1 × 10⁶ purified MAIT cells on day 0 (n = 5, red), day 10 (n = 5, orange), or day 25 (n = 6, brown), or of effector memory CD8⁺ T cells on day 0 (n = 4, CD8_EM, gray). Mice were sacrificed on day 35 and plasma was collected. (A) Circulating IFN-γ and IL-10 levels (mean ± SEM) were quantified by cytometric bead array in different groups; the shaded area indicates kit detection limit. **P 0.01, ****P 0.001, from 1-way ANOVA followed by Tukey’s multiple comparisons test. (B) Correlation between circulating IFN-γ and IL-10 levels. (C) Correlation of circulating IFN-γ or IL-10 levels with the proportion of huCD45⁺ cells in the peripheral blood. Spearman’s rank correlation coefficients (R) and corresponding P values are indicated.

Figure 6. MAIT cells persist for at least 1 week in mice and are correlated with decreased numbers of huCD45⁺ cells.

(A) Irradiated (1.3 Gy) NSG mice were injected with 5 × 10⁶ huPBMCs on day 0 without (n = 7) or with additional transfer of 1 × 10⁶ purified MAIT cells at day 0 (n = 8), day 10 (n = 6), or day 25 (n = 7). Mice were sacrificed 1 week after MAIT injection (i.e., at day 7, 17, or 32, respectively), and cells were recovered from spleen. (B) The proportions of human CD45⁺ leukocytes (huCD45⁺) among viable cells and of TCRVα7.2⁺ tetramer⁺ (tet) MAIT cells among huCD45⁺ cells were determined by flow cytometry, as shown on representative plots. (C–E) Proportion of huCD45⁺ and MAIT cells (mean ± SEM) is shown in each group at indicated time points, and right panels show correlations between MAIT cell frequencies and number of pathogenic huCD45⁺ leukocytes infiltrating the spleen. **P 0.01, ***P 0.001, ****P 0.001, from unpaired t tests. Spearman’s rank correlation coefficients (R) and the corresponding P values are indicated (right panels). Two-way ANOVA followed by Holm-Šídák multiple comparisons test.

Figure 7. MAIT cell immunoregulatory function requires interaction of the TCR with MR1 in vivo.

(A) Human PBMCs (20 × 10⁶) were pretreated or not with Ac-6-FP (10 μM) for 18 hours, CFSE-labeled, and injected in mice with or without cotransfer of 1 × 10⁶ purified MAIT cells. Ac-6-FP (10 pmol) was injected intraperitoneally on D0, 2, 4, and 6 in mice that had or had not (as control) received MAIT transfer. Mice were sacrificed on D7 and cells were recovered from the spleen (n = 5 for each condition). (B) Percentage of proliferating (CFSE^lo) T cells (mean ± SEM) and correlation with the number of huCD45⁺ cells. Spearman’s rank correlation coefficient (R) and the corresponding P value are indicated. (C) CD4/CD8 T cell ratio in the different groups. (D) Splenic cells were stimulated with PMA/ionomycin for 5 hours, with brefeldin A/monensin added after 1 hour. Proportions of IFN-γ⁺ and TNF-α⁺ cells among T cells were determined by flow cytometry following intracellular staining. *P 0.05, **P 0.01, ***P 0.001, ****P 0.0001, from 1-way ANOVA followed by Tukey’s multiple comparisons test.