Central memory CD8+ T lymphocytes mediate lung allograft acceptance

Abstract

Memory T lymphocytes are commonly viewed as a major barrier for long-term survival of organ allografts and are thought to accelerate rejection responses due to their rapid infiltration into allografts, low threshold for activation, and ability to produce inflammatory mediators. Because memory T cells are usually associated with rejection, preclinical protocols have been developed to target this population in transplant recipients. Here, using a murine model, we found that costimulatory blockade-mediated lung allograft acceptance depended on the rapid infiltration of the graft by central memory CD8+ T cells (CD44(hi)CD62L(hi)CCR7+). Chemokine receptor signaling and alloantigen recognition were required for trafficking of these memory T cells to lung allografts. Intravital 2-photon imaging revealed that CCR7 expression on CD8+ T cells was critical for formation of stable synapses with antigen-presenting cells, resulting in IFN-γ production, which induced NO and downregulated alloimmune responses. Thus, we describe a critical role for CD8+ central memory T cells in lung allograft acceptance and highlight the need for tailored approaches for tolerance induction in the lung.

Figure 1. Cellular mechanisms of lung allograft rejection in the absence of immunosuppression.

(A) BALB/c→nude lung grafts remain ventilated and free of inflammation, as demonstrated by gross appearance, histology, and ISHLT A grade. (B) BALB/c→B6 Cd8^–/– lung grafts are rejected acutely within a week of transplantation, as evidenced by graft collapse due to loss of ventilation and by severe perivascular infiltration with inflammatory cells. (C) Perivascular infiltration in BALB/c→B6 Cd8^–/– lung grafts was composed of CD4⁺ but not CD8⁺ cells. (D) Reconstitution of nude mice with B6 CD4⁺ T cells results in rejection of transplanted BALB/c grafts (P = 0.0039 compared with A by Mantel-Haenszel χ² test). All gross and histological appearances as well as rejection grades represent grafts at 7 days after transplantation (original magnification, ×200 [histology, H&E staining and immunohistochemical staining]). TXP denotes graft, and arrows point to perivascular infiltrates.

Figure 2. Cellular mechanisms of CSB-mediated lung acceptance.

BALB/c lungs transplanted into (A) B6 or (B) CD4-depleted B6 recipients remain ventilated with minimal inflammation, with gross appearance, histology, and numerical ISHLT A rejection grade shown. Rejection grades were not significantly different (Mantel-Haenszel χ² test, P = 0.500). BALB/c lungs transplanted into (C) CD8-depleted or (D) Cd8^–/– B6 recipients are rejected (P < 0.002 compared to A by Mantel-Haenszel χ² test). (E) Acceptance is restored in B6 Cd8^–/– mice after CD8⁺ T cell injection (P = 0.613 vs. A by Mantel-Haenszel χ² test). All gross and histological appearances as well as rejection grades represent grafts at 7 days after transplantation (original magnification, ×200 [histology, H&E staining]). TXP denotes graft, and arrows point to perivascular infiltrates.

Figure 3. CD4⁺ T lymphocyte responses in CSB-treated lung transplant recipients.

(A) The proportion of lung-resident CD4⁺Foxp3⁺ T cells is not different in resting B6 and B6 Cd8^–/– lungs. These numbers do increase, however, after allograft transplantation, and a higher abundance of graft-infiltrating CD4⁺Foxp3⁺ T cells is detectable in B6 compared with B6 Cd8^–/– lung graft recipients (comparison between resting and transplanted lungs by ANOVA and comparison between B6 and B6 Cd8^–/– groups by unpaired t test). (B) Proliferation of B6 CD4⁺CD45.1⁺ T cells was greater after injection into B6 Cd8^–/– (51.3% ± 5%) than B6 wild-type (20.6% ± 4%) recipients (P = 0.0017 by unpaired t test). Proliferating CD4⁺CD45.1⁺ T cells in B6 Cd8^–/– recipients upregulated CD27, ICOS, and OX40 but not CD28 or CD154 compared with wild-type mice. Numbers in contour plots represent the percentages of adoptively transferred CD4⁺CD45.1⁺ T cells that have undergone proliferation (shaded gray, isotype controls; black lines, B6 wild-type; red lines, B6 Cd8^–/– recipients). (C) Inhibiting CD27/CD70, ICOS/ICOS ligand, and OX40/OX40 ligand in addition to blocking CD40/CD154 and CD28/B7 does not prevent rejection in the absence of CD8⁺ T cells (P = 0.00074 vs. Figure 2A by Mantel-Haenszel χ² test). All gross and histological appearances as well as rejection grades represent grafts at 7 days after transplantation (original magnification, ×200 [histology, H&E staining]). TXP denotes graft, and the arrow points to perivascular infiltrates.

Figure 4. Graft-infiltrating CD8⁺ T cells play a critical role in downregulating alloimmune responses.

(A) In vitro MLRs were established by isolating CD8⁺ T cells from CSB-treated BALB/c→B6 lung transplants and adding them as “regulators” to cocultures of BALB/c splenocytes (stimulators) and CFSE-labeled B6 CD45.1⁺ T cells (responders). (B) After 5 days of coculture the majority of (B) B6 CD4⁺CD45.1⁺ T cells or (C) B6 CD8⁺CD45.1⁺ T cells proliferate and blast, as evidenced by size (forward scatter) (top row). Proliferation and blasting is inhibited if CD8⁺ T cells isolated from accepting lung allografts are added to the MLRs (second row). No inhibition is evident if CD8⁺ T cells are isolated from the spleens of accepting mice (third row). Numbers in histograms represent percentages of (B) CD4⁺CD45.1⁺ or (C) CD8⁺CD45.1⁺ T cells that have undergone proliferation. Proliferation and size (forward scatter) in the respective groups are summarized in the bottom panels of B and C, with pair-wise comparison between groups performed by t test.

Figure 5. IFN-γ–producing central memory CD8⁺CD44^hiCD62L^hiCCR7⁺ T cells infiltrate accepting lung allografts.

(A) Flow cytometry of CD8⁺ T lymphocytes in lung allografts of acceptors demonstrated few Foxp3⁺– or IL-10–producing cells. A large proportion of lung-resident CD8⁺ T cells had the capacity to produce IFN-γ and expressed a central memory phenotype (CD44^hiCD62L^hiCCR7⁺). (B) Fewer cells in spleens of lung graft recipients had the capacity to produce IFN-γ, and only few cells had a central memory T cell phenotype. Numbers in density plots in A and B represent percentages of CD8⁺ T cells expressing indicated markers. Phenotype of CD8⁺ T cells is representative of at least 4 separate experiments.

Figure 6. Central memory CD8⁺ T cells are abundant in the lung and can suppress alloimmune responses both in vitro and in vivo.

(A) Compared with other solid organs, such as heart, kidney, and pancreas, the lung contains a relative abundance of CD8⁺ T lymphocytes, including central memory cells. Central memory (CM) cells are defined as CD44^hi62L^hi, effector memory (EM) cells are defined as CD44^hi62L^lo, and naive cells are defined as CD44^lo62L^hi. Data is representative of 4 separate animals. (B) Freshly isolated central memory CD8⁺ T cells from resting B6 mice suppress proliferation of B6 CD4⁺CD45.1⁺ T cells stimulated with BALB/c splenocytes using methodology similar to that described in Figure 4A. Pair-wise comparison between proliferation profiles of responder CD4⁺CD45.1⁺ T cells in wells containing no CD8⁺ T cells, effector memory CD8⁺ T cells, and central memory CD8⁺ T cells was performed by unpaired t test. (C) Adoptive transfer of in vitro–generated B6 anti-BALB/c central memory cells into B6 Cd8^–/– recipients prevents rejection of BALB/c lung allografts after costimulatory blockade (P = 0.751 compared to Figure 2E by Mantel-Haenszel χ² test). (D) BALB/c lungs are rejected by B6 Cd8^–/– recipient mice reconstituted with in vitro–generated anti-BALB/c CD8⁺ effector memory T lymphocytes despite costimulatory blockade (P = 0.00105 compared to Figure 2E by Mantel-Haenszel χ² test). TXP denotes graft, and the arrow points to perivascular infiltrates All gross and histological appearances as well as rejection grades represent grafts at 7 days after transplantation (original magnification, ×200 [histology, H&E staining]).

Figure 7. CD8⁺ T cell–mediated lung allograft acceptance is dependent on IFN-γ production.

(A) Blocking IFN-γ prevents acceptance of BALB/c lung grafts by B6 recipients despite CSB. Gross appearance, histology, and ISHLT A rejection grade are shown (P = 0.000258 vs. Figure 2A by Mantel-Haenszel χ² test). (B) In vitro proliferation of CD4⁺CD45.1⁺ T cells (CFSE) stimulated by BALB/c splenocytes in the presence of accepting allograft-derived CD8⁺ T cells after addition of IFN-γ–blocking (red) or control antibody (black) (n = 3 separate experiments). (C) Ifng levels in allografts are significantly higher 4 days after transplantation into wild-type vs. Cd8^–/– B6 recipients (n = 4 each; unpaired t test). (D) Injection of Ifng^–/– CD8⁺ T cells does not restore lung allograft acceptance (P = 0.0066 vs. Figure 2E using Mantel-Haenszel χ² test). All gross and histological appearances as well as rejection grades represent grafts at 7 days after transplantation (original magnification, ×200 [histology, H&E staining]). TXP denotes graft, and the arrow points to perivascular infiltrates.

Figure 8. CD8⁺ T cells suppress through IFN-γ–mediated production of NO.

(A) After 5 days, the majority of CD4⁺CD45.1⁺ T cell “responders” are not viable if CD8⁺ T cells from accepting allografts are added. CD4⁺CD45.1⁺ T cell viability (by 7-AAD uptake) and representative plots of CFSE vs. 7-AAD are shown (groups compared by unpaired t test). (B) CD4⁺ T cell proliferation (CFSE) and viability (7-AAD) in an MLR containing Ifngr1^–/– CD4⁺ T cell responders or Ifngr1^–/– antigen-presenting cells (n ≥ 3). Numbers in density plots in A and B represent percentages of CD4⁺CD45.1⁺ T cells within the respective quadrants, assessing their proliferation (CFSE) vs. viability (7-AAD). (C) CD4⁺ T cell proliferation after stimulation with plate-bound anti-CD3 and soluble anti-CD28 in the absence or presence of accepting allograft-derived CD8⁺ T cells (P = 0.55 between the 2 groups by unpaired t test). (D) CD4⁺ T cell proliferation with inhibitors of amino acid metabolism, arginine, or Inos^–/– antigen-presenting cells (multiple group comparison performed by ANOVA). (E) NO levels in resting lungs, allografts, and right native lungs (n ≥ 3) (unpaired t test). (F) BALB/c lungs transplanted into CSB-treated Inos^–/– B6 recipients (P = 0.00059 vs. Figure 2A by Mantel-Haenszel χ² test). TXP denotes graft, and the arrow points to perivascular infiltrates All gross and histological appearances as well as rejection grades represent grafts at 7 days after transplantation (original magnification, ×200 [histology, H&E staining]).

Figure 9. Chemokine receptor expression regulates CD8⁺ T cell–mediated lung acceptance.

(A) Graft infiltration by PTX-treated or untreated anti-donor (BALB/c) central memory, anti–third party (CBA/Ca) central memory, or anti-donor (BALB/c) effector memory B6 CD8⁺CD45.1⁺ T cells (comparison within each group by unpaired t test). (B) Injection of Ccr7^–/– CD8⁺ T cells does not restore allograft acceptance in B6 Cd8^–/– recipients (P = 0.00054 vs. Figure 2E by Mantel-Haenszel χ² test). The arrow points to perivascular infiltrates. (C) Immunosuppressed BALB/c→B6 Cd8^–/– recipients reconstituted with wild-type B6 CD8⁺ T cells (n = 8) had higher graft Ifng levels than those reconstituted with B6 Ccr7^–/– CD8⁺ T cells (n = 5) (day 4) (unpaired t test). (D) Majority of recipient-derived graft-infiltrating CD11c⁺ cells in immunosuppressed BALB/c (CD45.2⁺)→B6 (CD45.1⁺) transplants express donor MHC class I (H-2K^d) (n = 3). Numbers in top density plot represent percentages of CD11c⁺ cells expressing CD45.1 (recipient) vs. CD45.2 (donor). Numbers in bottom density plot represent percentages of recipient CD45.1⁺CD45.2^–CD11c⁺ cells expressing donor (H-2K^d) vs. recipient (H-2K^b) MHC class I. Representative of 3 independent experiments. All gross and histological appearances as well as rejection grades represent grafts at 7 days after transplantation (original magnification, ×200 [histology, H&E staining]).

Figure 10. CCR7 expression on CD8⁺ T cells regulates their interactions with CD11c⁺ antigen-presenting cells within CSB-treated lung allografts.

Intravital 2-photon microscopy demonstrating wild-type B6 CD8⁺ T cells (cyan), Ccr7^–/– B6 CD8⁺ T cells (red), and CD11c⁺ cells (green) in immunosuppressed BALB/c→B6 CD11c-EYFP allografts on day 4. Collagen appears as blue. Higher-magnification views show representative T cell movement over a 1-hour interval. Cyan tracks follow the movement of wild-type CD8⁺ T cells, whereas red tracks follow Ccr7^–/– CD8⁺ T cells. Scale bars: 50 μm (top); 40 μm (bottom). Images are individual frames from a continuous time-lapse recording (Supplemental Video 1). Relative time displayed in minutes/seconds. Boxed regions are shown at high magnification in bottom panels. Wild-type T cells (blue) have higher mean retention times (mostly associated with CD11c⁺ cells) than Ccr7^–/– T cells (red) (23 vs. 16 minutes, P < 0.001, t test). Representative data are shown from two independent experiments with similar results.