Mixed chimerism established by hematopoietic stem cell transplantation is maintained by host and donor T regulatory cells

Abstract

Transplantation is an effective treatment of many clinical disorders, but the mechanisms that regulate immunological tolerance are uncertain and remain central to improving patient outcome. Hemopoietic stem cell transplantation (SCT) often establishes "mixed chimerism" in which immune cells from both the donor and patient coexist in vivo in a setting of immunological tolerance. We studied immune function in 69 patients within 2 months following SCT; 37 were fully donor and 32 displayed mixed chimerism. The proportion of T regulatory (Treg) cells was increased during mixed chimerism and comprised equal numbers of donor and host-derived regulatory cells. This was associated with a tolerogenic PD-L1⁺ profile on dendritic cells. Importantly, effector T cells from patients with mixed chimerism exhibited reduced cytotoxicity against host target cells in vitro, but this was restored following depletion of CD4⁺ Treg cells. These data show that Treg cells play a major role in sustaining immunological tolerance during mixed chimerism. These insights should help to guide novel interventions to improve clinical transplantation.

Graphical abstract

Figure 1.

PBMC and T-cell donor chimerism following HSCT. (A) Bubble chart demonstrating PBMC and T-cell donor chimerism levels at day 50 posttransplant. The area of each bubble represents the number of events. The first group (n = 37, red), displayed FC, in which both PBMC and T-cell donor chimerism was ≥99%. The second group (MC, n = 32, blue) demonstrated MC in both PBMC and T-cell fractions (PBMC and T-cell chimerism ≤98%). The pattern of PBMC donor chimerism up to day 365 according to donor chimerism at day 50 posttransplant. PBMC chimerism remained stable at complete (or near complete) donor chimerism in patients who demonstrated FC by day 50 posttransplant. In contrast, PBMC donor chimerism remained significantly lower in those found to have MC at day 50, even beyond the initiation of donor lymphocyte infusion (DLI) at day 180 (****P < .0001; 2-way ANOVA, with post hoc Tukey test, mean, and standard error shown for all data points; FC, n = 37; MC, n = 32). (B) The pattern of T-cell donor chimerism up to day 365 according to donor chimerism at day 50 posttransplant. (C) T-cell chimerism remained stable at complete (or near complete) donor chimerism for the first 100 days in patients who demonstrated FC by day 50 posttransplant, but displayed an average drop of 10% by day 180. In contrast, T-cell donor chimerism remained significantly lower in those found to have MC at day 50 with a trend toward partial correction following instigation of DLI at day 180 (****P < .0001, 2-way ANOVA, with post hoc Tukey test, mean, and standard error shown for all data points; FC, n = 37; MC, n = 32).

Figure 2.

DCs are less activated and express a tolerogenic PD-L1⁺phenotype in patients with MC. (A) The contribution and phenotype of cDC and pDC within the peripheral mononuclear cell compartment at 4 to 8 weeks after allo-HSCT was assessed by flow cytometry. The ratio of cDC:pDC was similar in patients with FC (2.5:1; n = 15) or MC (2.6:1; n = 15). (B) Expression of CD86 on DC in relation to chimerism status. A total of 81% of DC in FC expressed CD86 (n = 15) compared with only 58% of DC in MC (n = 15; **P = .0068). (C) Expression of PD-L1 and PD-L2 on DC in relation to chimerism status. A total of 41% of DC in MC expressed PD-L1 compared with 18% of DC in FC (**P = .0052; FC, n = 15; MC, n = 15). No difference was observed in the level of PD-L2 expression (FC, 4.3%; MC, 7% [not significant]). DCs therefore appear to have a tolerogenic phenotype in the setting of MC. (D) Expression of PD-L1 on DC subsets. Increased PD-L1 expression was most pronounced on the cDC subset in MC (47%; n = 15) compared with cDC in FC (20%; n = 15; **P = .0053), whereas a trend was observed toward increased PD-L1 expression on pDC in MC. (E) Expression of PD-L1 on host and donor DC in MC. This was determined by combining cell surface phenotyping with RNA probes specific for the KDM5D (SMCY) gene, derived from the Y chromosome. Host and donor cells were thereby distinguishable in the context of patients undergoing gender-mismatched allo-HSCT. PD-L1 expression was observed on a mean of 67% donor DC compared with only 29% of host DC (n = 10; ***P = .0002). For graphs with unpaired data, the mean and standard error of the mean are shown; analyses were undertaken with 2-tailed Mann-Whitney U tests. Paired analysis used a Wilcoxon matched pairs signed rank test.

Figure 3.

Treg cells comprise a larger proportion of the CD4⁺repertoire in patients with MC. (A) The percentage of CD4⁺CD25⁺CD127^lowFoxP3⁺ Treg cells within the CD4⁺ T-cell pool is shown in relation to chimerism status. Patients with MC demonstrate a higher frequency of Treg cells (8.7%, n = 29) compared with patients with FC (4.7%, n = 35; *P = .03). (B) The absolute number of Treg cells was similar between FC and MC cohorts at 2.6 × 10⁶/mL (n = 35) and 2.7 × 10⁶/mL (n = 29), respectively. (C) The number of Treg cells is inversely correlated to the degree of donor T-cell chimerism. Pearson’s coefficient r = −0.25, *P = .04 (n = 64). (D) The number of host and donor-derived Treg cells within patients with MC calculated using flowRNA probes specific to the KDM5D gene on the Y chromosome. No differences were seen between the 2 groups (n = 11). (E) Interleukin-10 and transforming growth factor-β production by Treg cells from patients with FC and MC. (F) The number of CD4⁺ T cells within peripheral blood in the first year after transplant in relation to chimerism status. No differences were seen between the 2 groups (FC, n = 35; MC, n = 29). (G) The number of CD8⁺ T cells within peripheral blood in the first year after transplant in relation to chimerism status. No differences were seen between the 2 groups (FC, n = 35; MC, n = 29). All graphs show the mean and standard error. Analyses were undertaken with 2-tailed Mann-Whitney U tests (A-B) and Wilcoxon matched pairs signed rank test (C).

Figure 4.

Treg cells limit cytotoxic alloreactive immune responses in the setting of MC. (A) Irradiated host HLA-DR⁺ APCs were cocultured with responder T cells isolated posttransplant. Cells were cocultured at a ratio of 1:1 for 5 days and then the number of residual live cells was determined by propidium iodide and Accu-Chek counting beads. An index of in vitro cytotoxicity was then defined as the percentage of remaining stimulator host HLA-DR⁺⁺ cells divided by the percentage of remaining responder T cells. (B) Ratio of host HLA-DR⁺ cells: post-HSCT T cells, 1.3: 1 in the context of FC compared with 5:1 in the context of MC (*P = .011), indicating that T cells from patients with MC are less able to eliminate HLA-DR⁺⁺ APCs. The mean and standard error are shown, and analysis was undertaken with a 2-tailed Mann-Whitney U test (FC, n = 11; MC, n = 12). (C) Coculture assays were reestablished from patients with MC following depletion of CD25⁺ cells (Treg cells) before in vitro culture. The ratio of host HLA-DR⁺⁺ cells to T cells was significantly decreased, indicating that Treg cells act directly to suppress donor alloreactive cytotoxicity in the context of MC (*P = .0312). The mean and standard error are shown; analysis was undertaken with a Wilcoxon matched pairs signed rank test (n = 7).

Figure 5.

A model for the maintenance of MC following allo-HSCT. (A) In an immunocompetent host, donor cells are processed by host DC and presented to host T cells, which then eliminate residual donor cells. This scenario can occur following entry of fetal cells into the maternal circulation during pregnancy or during rejection of solid organ grafts. (B) In the setting of profound host immune suppression following allo-HSCT, all host T cells are eliminated by conditioning so that a host alloreactive response to donor cells is abrogated. Instead, donor T cells recognize host peptide/major histocompatibility complex complexes on DC; this alloreactive response leads to both GVHD and the graft-versus-leukemia effect. (C) MC may be established in the setting of moderate host immune suppression following allo-HSCT. Here, residual host Treg cells may limit the activation of DCs, which then fail to induce a cytotoxic allogeneic immune response by donor T cells. Propagation of the allogeneic immune response is inhibited further by the presence of donor Treg and donor DCs that are rendered tolerogenic.