Induction of transplantation tolerance converts potential effector T cells into graft-protective regulatory T cells

Abstract

Naturally occurring FOXP3(+) CD4(+) Treg have a crucial role in self-tolerance. The ability to generate similar populations against alloantigens offers the possibility of preventing transplant rejection without indefinite global immunosuppression. Exposure of mice to donor alloantigens combined with anti-CD4 antibody induces operational tolerance to cardiac allografts, and generates Treg that prevent skin and islet allograft rejection in adoptive transfer models. If protocols that generate Treg in vivo are to be developed in the clinical setting it will be important to know the origin of the Treg population and the mechanisms responsible for their generation. In this study, we demonstrate that graft-protective Treg arise in vivo both from naturally occurring FOXP3(+) CD4(+) Treg and from non-regulatory FOXP3(-) CD4(+) cells. Importantly, tolerance induction also inhibits CD4(+) effector cell priming and T cells from tolerant mice have impaired effector function in vitro. Thus, adaptive tolerance induction shapes the immune response to alloantigen by converting potential effector cells into graft-protective Treg and by expanding alloreactive naturally occurring Treg. In relation to clinical tolerance induction, the data indicate that while the generation of alloreactive Treg may be critical for long-term allograft survival without chronic immunosuppression, successful protocols will also require strategies that target potential effector cells.

Figure 1

Tolerance induction with anti-CD4+DST is dependent on CD25⁺ Treg. (A) CBA mice (H2^k) were pre-treated with 200 μg of YTS177 (anti-CD4 mAb) i.v. on days 28 and 27 together with 250 μL whole H2^b blood (DST) on day 27. On day 14, mice in the CD25 targeting group received 1 mg of PC61 (anti-CD25 mAb) i.v. On day 0, donor-type (H2^b) vascularised heterotopic cardiac allografts were transplanted. (B) Representative plots (gated on viable lymphocytes) showing depletion of CD25⁺ T cells. Figures show percentage of CD25⁺CD4⁺ cells in the region indicated (mean±SEM, n=3 mice per time point). (C) Cardiac allograft survival for mice in (A).

Figure 2

Tolerant mice are enriched for graft-protective Treg. (A) T- and B–cell-deficient CBA.rag^−/− mice (H2^k) were reconstituted with CD25⁻CD4⁺ cells (Teff) purified from un-manipulated CBA mice and CD25⁺CD4⁺ cells (Treg) purified from either anti-CD4 + H2^b DST treated or naïve un-manipulated mice on day 1. On day 0, H2^b (donor) or H2^s (third party) skin allografts were transplanted. Data are from four independent experiments. Survival was compared using the Log-rank test. (B) Donor graft survival after reconstitution with 1×10⁵ Treg+1×10⁵ Teff. (C) Donor graft survival after reconstitution with 1×10⁵ Treg+2×10⁵ Teff. (D) Donor graft survival after reconstitution with 1×10⁵ Treg+4×10⁵ Teff. (E) Third-party graft survival after reconstitution with 1×10⁵ Treg+2×10⁵ Teff.

Figure 3

Correlation of phenotypic markers with FOXP3 expression. Splenocytes from naïve CBA mice were stained for CD4, cell surface markers and FOXP3. Histograms are gated on live CD4⁺ cells, and are representative of three independent experiments. (A) Expression of CD25 and FOXP3 in gated viable CD4⁺ cells. (B) Markers that correlated with FOXP3 expression. Figures indicate median fluorescence intensities for FOXP3⁻ and FOXP3⁺ populations. (C) Predicted yield and purity of FOXP3 cells based on the pairs of markers shown.

Figure 4

CD25⁻GITR⁻CD4⁺ cells are essentially devoid of nTreg. (A) CBA splenocytes were stained for GITR, CD25 and CD4 and gated CD4⁺ cells with the lowest expression of CD25 and GITR were separated by FACS. (B) Input CD4⁺ and sorted CD25⁻GITR⁻ cells were permeabilised and stained for intracellular FOXP3 (gated on live CD4⁺ cells). (C) Combined results from five independent FACS sorts (mean±SEM). (D) Sorted CD25⁻GITR⁻ cells were analysed for foxp3 mRNA expression by qRT-PCR. Naïve CBA FACS separated CD25⁺ cells were used as a positive control. DKK.rag TCRtg CD4⁺ cells were used as a negative control. In addition, DKK.rag CD4⁺ cells were spiked with 0.5% naïve CBA CD25⁺ cells. Results are normalised against CD3 expression and are representative of three independent experiments.

Figure 5

Graft-protective Treg develop from FOXP3⁻ precursors in vivo. (A) CBA.rag mice (H2^k) were reconstituted with 1×10⁶ sorted naïve CBA (H2^k) FOXP3 cells on day 35; 200 μg of YTS177 (anti-CD4 mAb) was administered i.v. on days 28 and 27, together with 250 μL whole H2^b blood DST on day 27; 1×10⁵ naïve CD25CD4⁺ CBA (H2^k) effector cells were adoptively transferred on day 1. A full thickness H2^b skin allograft was transplanted on day 0. Controls received anti-CD4 mAb without DST (anti-CD4 control) or adoptive transfer of cells alone (adoptive transfer control). Data are from two independent experiments. (B) Skin allograft survival. Survival was compared using the Log-rank test. (C) Representative image of skin graft at day 100 post-transplant.

Figure 6

FOXP3 analysis fails to reflect Treg generation from non-Treg precursors. (A) CBA.rag mice (H2^k) were reconstituted with 1×10⁶ sorted CBA (H2^k) FOXP3⁻ cells on day 35. The anti-CD4+DST group received 200 μg of YTS177 (anti-CD4) i.v. on days 28 and 27 together with 250 μL whole H2^b blood (DST) on day 27. Controls received adoptive transfer of cells alone. On day 0, spleens were harvested and FOXP3 expression analysed by FACS. Representative data are shown from two independent experiments. (B) Proportion of TCR-β⁺CD4⁺ cells expressing FOXP3. (C) Absolute number of FOXP3⁺ T cells per spleen (mean±SD). Statistical analysis using the t test.

Figure 7

Treg develop from FOXP3⁻ precursors in immunocompetent mice. (A) 3×10⁶ GFP⁻CD4⁺CD45.2⁺ T cells purified from FOXP3-GFP reporter mice by flow cytometry were adoptively transferred into congenic CD45.1⁺ B6 mice on day 29. Recipient mice received 200 μg anti-CD4 mAb (YTS177) on days 28 and 27 with 250 μL DBA2 (H2^d) blood transfusion (DST) on day 27. Control mice received anti-CD4 mAb alone, DST alone or adoptive transfer of GFP⁻CD4⁺CD45.2⁺ T cells only. On day 0, spleens were harvested and the number of GFP⁺CD4⁺CD45.2⁺ T cells determined by FACS. (B) Flow cytometry of sorted GFP⁻CD4⁺CD45.2⁺ T cells before injection. (C) Analysis gate for adoptively transferred CD45.2⁺ T cells on day 0. (D) Absolute numbers of FOXP3⁻GFP⁺CD4⁺CD45.2⁺ T cells (mean±SEM, n=3 mice per group).

Figure 8

Tolerance induction inhibits CD4⁺ effector cell priming. (A) CBA mice (H2^k) were pre-treated with 200 μg of YTS177 (anti-CD4) i.v. on days 28 and 27 together with 250 μL whole H2^b blood (DST) on day 27. Control mice received anti-CD4 alone, DST alone or no pre-treatment. CD4⁺ cells were purified from spleens harvested on day 0 and stimulated with H2^b+ splenocytes in an IFN-γ ELISpot. (B) Absolute number of donor-reactive IFN-γ-producing CD4⁺ cells per spleen – pooled cells from four to five mice per group, representative of two independent experiments. Representative ELISpot well images are shown. Statistical analysis using the t test.

Figure 9

Tolerance induction ameliorates in vitro cytotoxicity. (A) Total T cells from unmanipulated or anti-CD4+DST-treated mice were stimulated in vitro for 5 days with irradiated H2^b splenocytes and then harvested (effectors) and cultured for 6 h with H2^b+ splenocytes (target). The absolute number of live 7AAD⁻K^b+ target cells was determined by FACS. (B) Percentage killing (normalised to negative control). Data indicate the mean±SD of triplicate tubes and are representative of four independent experiments each using three to four mice per group.