Donor plasmacytoid dendritic cells modulate effector and regulatory T cell responses in mouse spontaneous liver transplant tolerance

Abstract

We assessed the role of donor liver non-conventional plasmacytoid dendritic cells (pDCs) in spontaneous liver transplant tolerance in a fully MHC-mismatched (C57BL/6 (H2^b) to C3H (H2^k)) mouse model. Compared with spleen pDCs, liver pDCs expressed higher levels of DNAX-activating protein of 12 kDa and its co-receptor, triggering receptor expressed by myeloid cells 2, and higher ratios of programed death ligand-1 (PD-L1):costimulatory CD80/CD86 in the steady state and after Toll-like receptor 9 ligation. Moreover, liver pDCs potently suppressed allogeneic CD4⁺ and CD8⁺ T cell proliferative responses. Survival of pDC-depleted livers was much poorer (median survival time: 25 days) than that of either untreated donor livers or pDC-depleted syngeneic donor livers that survived indefinitely. Numbers of forkhead box p3 (FoxP3)⁺ regulatory T cells in grafts and mesenteric lymph nodes of mice given pDC-depleted allogeneic livers were reduced significantly compared with those in recipients of untreated livers. Graft-infiltrating CD8⁺ T cells with an exhausted phenotype (programed cell death protein 1⁺, T cell immunoglobulin and mucin domain-containing protein 3⁺) were also reduced in recipients of pDC-depleted livers. PD1-PD-L1 pathway blockade reversed the reduction in exhausted T cells. These novel observations link immunoregulatory functions of liver interstitial pDCs, alloreactive T cell exhaustion, and spontaneous liver transplant tolerance.

1 |. INTRODUCTION

Many liver transplant patients tolerate reduced doses of immunosuppressive (IS) drugs.¹ Moreover, only in liver transplantation can about 20%–40% of carefully selected stable graft recipients (up to as high as 79%, depending on patient age, number of years posttransplant and other factors) be withdrawn from all IS therapy without rejection.^2,3 Mechanisms that underlie the induction and maintenance of clinical operational liver transplant tolerance are not well understood,⁴ and there are no validated biomarkers to reliably predict rejection or tolerance. Studies in rodents, however, in particular use of the mouse orthotopic liver transplant model,⁵ in which MHC-mismatched grafts are accepted without IS treatment,⁶ have advanced understanding of cellular and molecular pathways that regulate liver ischemia-reperfusion injury⁷ and immune-mediated rejection versus tolerance.⁸ Prominent among these studies of liver transplant tolerance in mice and rats have been observations of abortive effector T cell responses within the allograft early (from 4 to 5 days) posttransplant.^9–11

The liver’s unique constituency of parenchymal and non-parenchymal cells (NPCs), in particular conventional (c) dendritic cells (DCs) and other professional antigen (Ag) presenting cells (APCs), has been implicated in tolerance induction, as have molecular mechanisms whereby liver APCs regulate alloreactive T cell responses.¹² The latter include expression by liver APCs of programed death ligand 1 (PD-L1), anti-inflammatory IL-10, the ectoenzyme CD39 that degrades ATP to adenosine, and the immunoreceptor tyrosine-based activation motif-bearing transmembrane adaptor protein DNAX-activating protein of 12 kDa (DAP12).¹³ Each of these molecules has been implicated in the regulation of mouse spontaneous liver transplant tolerance.⁸

Plasmacytoid dendritic cells (pDCs) are bone marrow-derived, unconventional DCs found predominantly in T cell areas of peripheral lymphoid organs. Their development, differentiation and function,¹⁴ and their roles in immunity and tolerance^15,16 have been reviewed. Like cDCs, pDCs exhibit dual functionality of immunogenicity versus tolerogenicity, based on receptor ligation and their activation status. pDCs arise from the same progenitors as cDCs, yet express a genetic profile that more closely resembles lymphoid (T and B) cell development.¹⁴ Lineage commitment of pDCs is controlled by expression of the transcription factor E2–2, a member of the E protein family that plays a crucial role in lymphoid cell development.¹⁷ In addition to transcriptional programming, the hematopoietic cytokine fms-like tyrosine kinase 3 ligand (Flt3L) is critical for pDC propagation. pDCs function both as innate anti-viral immune effectors, and as inducers and regulators of adaptive immunity,¹⁸ including hepatic T cell responses.¹⁹ They drive natural regulatory T cell (Treg) development,²⁰ promote central tolerance²¹ and induce²² and maintain²³ Ag-specific Treg.

There is also evidence that pDCs regulate the induction and/or maintenance of tolerance to hematopoietic stem cell or organ allografts (reviewed in [15]). Thus, Abe et al²⁴ first showed that adoptive transfer of pDCs could prolong cardiac allograft survival, whereas Ochando et al²⁵ reported that pDCs were essential for the induction of tolerance to heart grafts through Ag acquisition and induction of alloAg-specific Tregs. More recently, Oh et al²⁶ have found that pDC-driven induction of Treg correlates with the spontaneous acceptance of mouse renal allografts.

Non-lymphoid tissue pDCs, such as those that reside in the airways, gut and liver, play a significant role in regulating mucosal immunity and are critical for the development of tolerance to inhaled or ingested/dietary Ags.²⁷ The liver is a site of oral Ag presentation and compared to secondary lymphoid tissue, is comparatively rich in pDCs²⁸ that appear to rapidly induce anergy or deletion of Ag-specific T cells.²⁹ However, the role of pDCs in liver transplant tolerance has not been examined. Here we show for the first time, that hepatic pDCs of donor origin, that express high levels of DAP12, TREM2 and high ratios of T cell coinhibitory PD-L1:costimulatory CD80/86 compared with lymphoid tissue pDCs, play a key role in attenuating graft-infiltrating T effector cell responses, enhancing forkhead box p3 (Foxp3)⁺ Tregs, and promoting spontaneous acceptance of mouse liver allografts.

2 |. MATERIALS AND METHODS

2.1 |. Mice

Male C57BL/6 (B6; H2^b) and C3H/HeJ (C3H; H2^k) mice were purchased from The Jackson Laboratory.

2.2 |. Isolation of mouse liver non-parenchymal cells (NPC)

Mouse liver NPC were isolated by collagenase digestion and centrifugal elutriation as described,³⁰ with minor modifications.

2.3 |. Purification of liver and spleen pDCs

pDCs were positively selected from total liver NPC or splenocytes using pDC-Ag-1 (PDCA-1) magnetic microbeads (Miltenyi Biotec), as described.³¹ The purity of the sorted cells was routinely >85%, where the majority of contaminating cells were B220⁻CD11c⁺ cDCs, with negligible contamination by B220⁺CD11c⁻ B cells, as reported previously.³² In some cases, mice were injected with 10 μg of the endogenous DC poietin Flt3L for 10 consecutive days to expand DC in liver and spleen, as described.³³

2.4 |. Adoptive transfer of pDCs

C3H mice received 3 × 10⁶ B6 liver pDCs in 200 μl via portal vein injection. Liver, spleen, and mesenteric lymph nodes were isolated 10 days after cell transfer. Control mice received PBS alone.

2.5 |. Flow cytometry

Cells were stained with Zombie Aqua^™ dye (Zombie; Biolegend) for 30 min at room temperature (RT), then incubated with FcγR-blocking rat anti (α)-mouse CD16/32 Ab (93; Biolegend) for 20 min at RT to prevent non-specific Ab binding. For cell surface staining, the cells were incubated for 20 min at 4°C with different combinations of fluorochrome-conjugated antibodies (Abs) against mouse CD3 (clone #: 17A2), CD4 (H129.19), CD8α (53–6.7), CD11c (HL3), B220/CD45R (RA3–6B2), CD11b (M1/70), H-2 Kb (AF6–88.5), CD80 (16–10A1), CD86 (GL1), I-A^b (AF6–120.1), NK1.1 (PK136), PDCA-1 (927), programmed cell death-1 (PD-1) (J43), PD-L1 (10F.9G2), CD25 (PC61.5), CD40L (SA047C3), CD134 (OX-86), CD69 (H1.2F3), Neuropilin-1 (3E12) or T cell Ig and mucin domain-containing 3 (Tim-3; 8B.2C12). After staining, cells were fixed with 4% v/v paraformaldehyde. For intracellular staining, cells were fixed and permeabilized using Fix and Perm reagent (eBioscience) and then stained with the following fluorochrome-conjugated Abs against: cytotoxic T lymphocyte Ag 4 (CTLA-4; UC10–4B9), Foxp3 (FJK-16s), Helios (22F6), Ki-67 (SolA15), interferon (IFN)γ (XMG1.2), tumor necrosis factor α (TNFα; MP6-XT22), IL-2 (JES6–5H4), or IL-6 (MB5–20F3). Appropriate fluorochrome-conjugated, isotype-matched IgG was used as a negative control. Before cytokine staining, the cells were maintained for 4 h at 10⁶ cells /ml in 15 ml tubes with phorbol myristate acetate (PMA; 20 ng/ml, Sigma) and ionomycin (100 mmol/L, Sigma). GolgiStop^™ (0.66 μl/ml; BD) was added 30 min after PMA/ionomycin. An Annexin-V Apoptosis Detection Kit was purchased from eBioscience and used according to the manufacturer’s instructions. Flow data were acquired on a LSR Fortessa flow cytometer (BD) and analyzed using FlowJo version 10 software (Tree Star).

2.6 |. Cytokine quantitation

Cytokine levels were determined using mouse Th1, Th2, and Th17 cytokine bead array (CBA) kits (BD).

2.7 |. Liver transplantation

Orthotopic liver transplantation was performed as described in detail,⁵ without immunosuppressive therapy. For Ab-mediated pDC depletion in vivo,³⁴ donor B6 mice were injected for 3 consecutive days with anti-mPDCA-1 mAb (clone 927, 500 μg/mouse, intraperitoneally; BioXCell). For in vivo PD-L1 neutralization, anti-B7-H1 (clone MIH5; BioXCell) mAb was administered intraperitoneally to recipient C3H mice on day 0 (500 μg/mouse) and day 3 (250 μg/mouse). Serum aspartate aminotransferase (AST) and alanine aminotransferase (ALT) were quantified as described.³³ Allograft rejection was determined by host survival and confirmed/graded by “blinded” histological evaluation by a transplant pathologist (MIM) using Banff schema for acute liver rejection.³⁵ Liver and spleen tissues were fixed in 10% vol/vol formalin, embedded in paraffin, then sectioned and stained with H&E.

2.8 |. Statistical analysis

GraphPad Prism (version 7.00; Graphpad Software Inc.) was used for statistical analyses. Results are expressed as means ± SEM. Significances of differences between means were determined using Student’s ‘t’ test or the log-rank test. Multiple comparisons on a single data set were performed by ANOVA. In all experiments, p < .05 was considered significant.

2.9 |. Additional methods

Additional methods, including statistical analyses are provided in Data S1.

3 |. RESULTS

3.1 |. Liver pDCs exhibit a more regulatory phenotype than secondary lymphoid tissue pDCs

We first examined the cell surface phenotype of liver and spleen pDCs freshly isolated from WT B6 mice. As shown in Figure 1A, flow cytometric analysis revealed that mouse steady-state liver pDCs expressed lower levels (mean fluorescence intensity; MFI) of MHC class II (IA^b) and costimulatory molecules (CD80 and 86), but higher levels of co-regulatory PD-L1 than spleen pDCs. Moreover, strikingly higher ratios of PD-L1/CD80 and PD-L1/CD86 were expressed by the liver pDCs. The same pattern of differences in expression of MHC II and costimulatory and co-regulatory molecules was observed following stimulation of liver and spleen pDCs with the Toll-like receptor 9 (TLR9) ligand CpG (Figure 1A). RT-PCR analysis further revealed that liver pDCs expressed much higher levels of DAP12 and its co-receptor TREM2 (that we have shown are associated with negative regulation of liver DC stimulatory function)^13,36 and that expression of IL-10 mRNA by liver pDCs was also much higher than that by spleen pDCs. There was also a trend toward higher levels of IDO and Arg-1 expression by liver pDCs (Figure 1B). Consistent with their low levels of cell surface MHC and costimulatory molecules, liver pDCs elicited only low levels of allogeneic (C3H) CD4⁺ and CD8⁺ T cell proliferation, that were significantly inferior to those induced by splenic pDCs (Figure 1C,D). Moreover, when stimulated with CpG, liver pDCs secreted significantly lower levels of proinflammatory IL-6, TNFα and IFNγ than spleen pDCs (Figure 1E).

FIGURE 1.

Comparison of the phenotype and function of mouse liver plasmacytoid (p)DCs and spleen pDCs. (A) PDCA-1–purified pDCs isolated from normal WT B6 liver non-parenchymal cells or splenocytes were either unstimulated (upper row) or cultured in the presence of 1 mg/ml CpG A ODN for 18 h and analyzed by flow cytometry for MHC, costimulatory and coinhibitory molecule expression. (B) DAP12, TREM2, and immunoregulatory molecule mRNA expression determined by real-time RT-PCR for liver pDCs and spleen pDCs. (C) Representative and (D) aggregate data, showing allogeneic T cell stimulatory activity of unstimulated or CpG-stimulated liver or spleen pDCs determined in CFSE-MLR. (E) Levels of specific cytokines quantified in MLR supernatants. Data shown are from 4 to 5 independent experiments; *p < .05, **p < .01, ***p < .001

3.2 |. Liver pDCs suppress αCD3/CD28- and alloAg-induced T cell proliferation

Next, we determined the ability of liver pDCs to regulate T cell responses to αCD3/CD28− or alloAg-induced proliferation. When B6 liver pDCs were added at the start of C3H T cell cultures stimulated by αCD3/CD28, unlike splenic pDCs, they significantly reduced both CD4⁺ and CD8⁺ proliferative responses (Figure 2A,B). This effect was accompanied by elevated incidences of apoptotic CD4⁺ and CD8⁺ T cells; however, these increases in apoptosis were reversed by PD-L1 blockade (Figure 2C,D). B6 liver pDCs but not splenic pDCs also potently inhibited alloAg (B6)-induced C3H CD4⁺ and CD8⁺ T cell proliferation and increased the death of both CD4⁺ and CD8⁺ T cells (Figure S1A,B).

FIGURE 2.

Liver pDCs inhibit αCD3/CD28-induced T cell proliferation and enhance apoptosis of responder T cells via PD-L1 expression. (A) Representative flow data showing the ability of freshly isolated WT mouse liver pDCs to suppress CD4+ or CD8+ T cell proliferation induced by αCD3/CD28 stimulation. C3H spleen T cells were plated as responders, α mouse CD3/CD28 Ab beads used as stimulators, and B6 liver or spleen pDCs tested as regulators. The ratio of pDC:beads: T cells was 1:10:10. Cultures were harvested on day 5. (B) Histograms showing aggregate data from n = 5 independent experiments. (C) Representative flow data showing the ability of freshly isolated WT mouse liver pDCs to promote CD4⁺ and CD8⁺ T cell apoptosis induced by αCD3/CD28 stimulation that was reversed by neutralizing PD-L1 mAb (20 μg/ml). The ratio of pDC:beads: T cells was 1:10:10. Cultures were harvested on day 5. (D) Histograms showing aggregate data from n = 6 independent experiments. *p < .05; **p < .01; ***p < .001

3.3 |. Liver pDCs increase T regs in allogeneic recipients

The ability of pDCs to drive the development of CD4⁺CD25⁺Foxp3⁺ Treg has been documented.^20,26 Thus, to further investigate the potential immune regulatory effects of liver pDCs on allogeneic T cells, we ascertained whether the adoptive transfer of B6 liver pDCs into normal (non-transplanted) allogeneic (C3H) recipients could enhance Treg in vivo. Following liver pDC injection (3 × 10⁶) via the portal vein, significant increases in liver and spleen weights relative to body weight (Figure 3A,B) were observed, accompanied by increases in liver, spleen, and lymph node T cell and myeloid cell populations (Figure S2A). Histological examination revealed a few small clusters of lymphoid cells in the liver, and lymphoid aggregates in the spleen, 10 days after pDC injection (Figure S2B), although there were no increases in AST and ALT plasma levels (Figure S2C). Mice given allogeneic B6 liver pDCs showed significantly increased incidences and absolute numbers of Treg in liver and secondary lymphoid tissue (Figure 3C,D). We also assessed the phenotype of CD4⁺Foxp3⁺ cells (Treg) and CD4⁺Foxp3⁻ cells (non-Treg) in liver, spleen, and lymph nodes. PD-1, Tim-3, CD40L, and CD69 expression by Tregs from allogeneic pDC-injected C3H livers were each increased significantly (Figure 3E; Figure S3A). Helios and Neuropilin-1 expression by Tregs was similar between the liver pDC-injected and control groups (Figure S3B). Both Tregs from pDC-injected C3H liver and C3H WT liver suppressed alloAg-induced CD4⁺ and CD8⁺ T cell proliferation to similar degrees on a per cell basis (Figure S4).

FIGURE 3.

Liver pDCs enhance numbers of allogeneic CD4⁺ Treg in vivo. Following injection of B6 WT liver pDCs (3 × 10⁶) or PBS (control) into C3H recipients via the portal vein, livers, spleens, and mesenteric lymph nodes were harvested and T cell populations enumerated. (A) Macroscopic findings of liver (upper panel) and spleen (lower panel) after liver pDC injection. (B) Liver and spleen weights in relation to body weight after liver pDC injection. (C,D) Incidences (% CD4⁺ cells) and absolute numbers of CD25⁺ Foxp3⁺ cells in livers, spleens, and lymph nodes. (E) Expression of PD-1, CTLA-4, Tim-3, and CD40L by CD4⁺ Foxp3⁺ cells (upper panels) and CD4⁺ Foxp3⁻ (lower panels) in livers, spleens, and lymph nodes. *p < .05; **p < .01; ***p < .001; n = 5 mice in each group

3.4 |. pDC deficiency in donor livers increases graft inflammation/injury and abrogates transplant tolerance

To examine the role of donor liver pDCs in orthotopic liver transplantation, untreated or pDC-depleted WT B6 (H2^b) livers were transplanted into normal allogeneic C3H (H2^k) or syngeneic recipients, without immunosuppressive therapy. Following pDC depletion of the donor, no significant changes in liver or spleen histology, liver function tests, cytokine gene expression, or immune cell subsets in livers or spleens were observed (Figure S5A–D). In keeping with previous reports,^6,33,37,38 WT liver allografts survived indefinitely (median survival time [MST]: >100 days; n = 6), whereas pDC-depleted allografts were rejected acutely (MST: 25 days; n = 5; p < .05) (Figure 4A). pDC-depleted syngeneic liver grafts transplanted into B6 recipients also survived indefinitely (MST: >100 days; n = 3). To evaluate graft function, we euthanized liver allograft recipients on day 4 posttransplant and determined serum AST and ALT levels and examined graft histology using Banff schema rejection criteria for acute T cell–mediated rejection. Serum ALT levels were elevated markedly in recipients of pDC-depleted liver allografts compared with those given untreated grafts (Figure 4B). Consistent with elevated liver enzyme levels, there was evidence of enhanced lymphocytic cholangitis and venulitis and elevated rejection indices in pDC-depleted allografts compared with untreated grafts (Figure 4C,D).

FIGURE 4.

pDC deficiency in donor livers increases graft inflammation/injury and abrogates transplant tolerance. (A) Actuarial graft survival curves showing that whereas normal WT B6 livers transplanted into allogeneic C3H recipients were accepted indefinitely (●; n = 6; median survival time [MST] >100 days), those from pDC-depleted B6 donors were rejected acutely (▲; n = 5; MST = 25 days), **p < .01. MST of pDC-depleted B6 liver grafts in syngeneic recipients (n = 3) was >100 days. (B) Serum ALT and AST levels 4 days posttransplant; *p < .05; n = 5 mice in each group. (C) Histopathological appearance of the liver allografts showing lymphocytic cholangitis and venulitis (arrowheads) in pDC-depleted donor allografts. (D) Grading of acute liver allograft rejection using Banff criteria rejection activity index. n = 4 mice in each group

3.5 |. Rejection of liver allografts from pDC-depleted donors correlates with enhanced anti-donor effector T cell responses

Liver transplantation in mice is associated with the migration of immature donor interstitial cDCs and their precursors to host lymphoid tissues, an event that has been implicated in the induction of liver transplant tolerance.³⁹ As shown in Figure 5A,B, donor pDCs (approx. 4% of total pDCs) could be detected in WT B6 liver allografts 4 days posttransplant, as determined by co-staining for donor MHC Ag (H2^b). At the same time, similar low incidences of donor pDCs could also be observed in the recipients’ spleens and lymph nodes. By contrast, pDCs were virtually absent from the graft or secondary lymphoid tissues of recipients of pDC-depleted donor livers. Seven days posttransplant, very low incidences of donor pDCs could still be detected in WT B6 liver allografts (Figure S6). To measure anti-donor immune reactivity, we recovered splenocytes from allograft recipients on day 4 posttransplant. Recipient T cells were re-stimulated with T cell–depleted donor (B6) splenocytes and anti-donor proliferation determined in 5-day CFSE-MLR. As shown in Figure 5C,D, compared with animals given normal WT livers, those given pDC-depleted liver allografts displayed significantly enhanced anti-donor CD4⁺ and CD8⁺ T cell proliferation. Moreover, pDC-depleted liver recipients’ T cells produced significantly higher levels of pro-inflammatory cytokines (IL-6, TNFα, IFNγ) in response to ex vivo donor Ag stimulation in CFSE-MLR (Figure 5E).

FIGURE 5.

Donor pDCs can be identified in host lymphoid tissue and pDC-depleted liver allografts elicit enhanced T cell proliferative responses to donor. (A,B) Incidences of donor-derived (H2^b) B220⁺ CD11c⁺ PDCA-1⁺ pDCs in the allograft, spleen, and lymph nodes, determined by flow cytometry 4 days after transplantation of WT B6 or pDC-depleted B6 livers into WT C3H recipients. Representative data are shown in the upper panel (A), and aggregate data in the lower panel (B). (C,D) Splenocytes isolated from normal, untreated WT B6 liver or pDC-depleted B6 liver allograft recipients (C3H) 4 days posttransplant were stimulated with T cell–depleted donor (B6) splenocytes in 5-day CFSE-MLR. Recipient T cell proliferation was determined by flow cytometry. Representative flow data, showing the proliferative responses of host CD4⁺ and CD8⁺ T cells (C), and aggregate data from each group (D). (E) Specific cytokine levels determined in MLR supernatants (day 5) using mouse CBA kits. *p < .05; n = 4 mice in each group

3.6 |. Donor liver pDCs promote alloreactive T cell exhaustion, while Tregs are reduced in recipients of pDC-depleted livers

Given the enhanced anti-donor T cell responses we observed in recipients of pDC-depleted livers that rejected their grafts, we assessed the phenotype of T cells infiltrating the graft, and in spleen and lymph nodes on day 4 posttransplant. As shown in Figure 6A,B a significant proportion of CD4⁺ T cells, and a much higher fraction of the CD8⁺ T cells infiltrating untreated donor liver grafts, exhibited an exhausted phenotype (PD-1⁺TIM-3⁺).⁴⁰ Levels of exhausted cells were significantly lower in grafts from pDC-depleted donors (Figure 6B,C). Notably, although in addition, a much lower proportion of lymph node T cells exhibited an exhausted phenotype compared to graft-infiltrating T cells, a significantly higher level of exhausted CD8⁺ T cells was observed in lymph nodes of the untreated donor group compared with the pDC-depleted donor group (Figure 6A–C). Consistent with these observations, a significant proportion of CD8⁺ T cells in grafts and lymph nodes of the untreated donor group exhibited the phenotype PD-1⁺CTLA-4⁺⁴¹ (Figure 6D). In support of attenuated T cell reactivity, a lower incidence of graft-infiltrating and lymphoid tissue CD4⁺ T cells in the untreated donor group produced IL-2, while a lower incidence of CD8⁺ T cells in the untreated donor group produced TNF-α (Figure 6E).

FIGURE 6.

Donor liver pDCs promote host alloreactive T cell exhaustion. (A) Representative flow data showing the incidences of PD-1 and Tim-3 double-positive (i.e., exhausted) CD4⁺ and CD8⁺ cells in liver allografts, and host spleens and lymph nodes of the untreated donor group and the pDC-depleted donor group on day 4 posttransplant. (B,C) Incidences and the absolute numbers of PD-1 and Tim-3 double-positive CD4⁺ and CD8⁺ T cells in liver allografts and host spleens and lymph nodes. (D) Incidences of PD-1 and CTLA-4 double-positive CD4⁺(upper) and CD8⁺(lower) T cells in liver grafts, spleens, and lymph nodes in the untreated donor group and the pDC-depleted donor group on posttransplant day 4. (E) Compilation of cytokine (IL-2, IL-6, IFNγ, and TNFα) expression by liver grafts, spleens, and lymph nodes CD4⁺ (upper) and CD8⁺ (lower) T cells in the untreated donor group and the pDC-depleted donor group on day 4 posttransplant. *p < .05; n = 3–4 mice in each group

There is evidence that mouse liver transplant tolerance is dependent on host CD4⁺Foxp3⁺ Treg.³⁸ We observed that the incidence and absolute numbers of CD4⁺CD25⁺Foxp3⁺ Treg were significantly higher in the liver grafts and lymph nodes of the untreated donor group compared with the pDC-depleted donor group on day 4 posttransplant (Figure 7A–C). Furthermore, pDC-depleted donor allografts expressed significantly higher granzyme B and perforin levels, and lower coinhibitory PD-L1 levels compared with untreated allografts, 4 days posttransplant (Figure 7D).

FIGURE 7.

CD4⁺CD25⁺Foxp3⁺ T cells in liver allografts and lymph nodes are reduced in recipients of pDC-depleted livers. (A) Representative data showing the incidences ofCD25⁺ Foxp3⁺ cells (% of CD4⁺ T cells) in recipients of normal untreated WT B6 or pDC-depleted liver allografts determined by flow cytometry. (B,C) Incidences of CD25⁺Foxp3⁺ cells (% of CD4⁺ T cells) and the absolute numbers of CD4⁺CD25⁺ Foxp3⁺ T cells. (D) Expression of IFNγ, granzyme B, perforin, PD-L1, and IDO determined by real-time RT-PCR in allograft tissue on posttransplant day 4. *p < .05; **p < .01; n = 4 mice in each group

3.7 |. PD-L1 blockade suppresses CD8⁺ T cell exhaustion in graft recipients

The PD-1-PD-L1 pathway acts as an important negative regulator of immune reactivity, and PD-L1 upregulation in liver allografts appears to play a critical role in the induction of mouse liver spontaneous transplantation tolerance.³⁷ Since we observed much higher levels of co-regulatory PD-L1: CD80 or PD-L1: CD86 ratios on liver pDCs and PD-L1 upregulation in liver allografts, we investigated the role of PD-L1 expression in regulation of graft-infiltrating T cell exhaustion and numbers of Treg. To achieve this, we used blocking Ab to neutralize PD-L1 in recipients of WT liver allografts. Significantly higher levels of PD1⁺Tim3⁺ exhausted CD4⁺ and CD8⁺ T cells were observed in the graft, spleen, and lymph nodes of the control group compared with recipients given anti-PD-L1 mAb (Figure 8A–C). PD1-PD-L1 pathway blockade also significantly reduced the incidence of CD4⁺Foxp3⁺ Treg within the graft (Figure 8D).

FIGURE 8.

In vivo PD-L1 blockade suppresses host CD8⁺ T cell exhaustion. (A) Representative flow data for PD-1 and Tim-3 double-positive (i.e., exhausted) CD4⁺ and CD8⁺ cells in liver allografts, spleens, and lymph nodes of control and anti-PD-L1 mAb-injected recipients on posttransplant day 4. (B,C) Incidences and absolute numbers of PD-1 and Tim-3 double-positive CD4⁺ and CD8⁺ T cells in liver allografts, spleens, and lymph nodes on day 4 posttransplant. (D) Incidence of CD25⁺Foxp3⁺ cells (% of CD4⁺ T cells) in liver grafts, spleens, and lymph nodes of control and anti-PD-L1 mAb-injected recipients on posttransplant day 4. *p < .05; **p < .01; n = 4 mice in each group

4 |. DISCUSSION

In rodents, the hematopoietic activity of the liver⁴² and donor-derived hematopoietic cells have been implicated in the regulation of host T cell responses and the promotion of liver transplant tolerance.^6,11,43,44 Among the liver’s unique constituency of innate immune cells, that includes conventional and non-conventional DCs, liver-resident macrophages (Kupffer cells) and innate lymphoid cells (ILCs: NK cells and non-NK ILCs), each with tolerogenic properties,^12,45 DCs are especially well-equipped APCs. They have ability to migrate to lymphoid tissues and instigate, integrate and regulate innate and adaptive immunity.⁴⁶

Within the liver microenvironment, hematopoietic progenitors are programed to differentiate into regulatory DCs that maintain liver tolerance.⁴⁷ Furthermore, in mouse liver transplantation, depletion of intrahepatic cDCs before transplant prevents the induction of spontaneous allograft tolerance.⁸ Moreover, donor-derived cDCs can be generated ex vivo from progenitors within the lymphoid tissue of untreated mouse recipients of liver, but not heart allografts from the same donor strain that are rejected acutely.³⁹ In addition, when adoptively transferred to prospective pancreatic islet allograft recipients, donor liver-derived cDCs prolong graft survival.⁴⁸ Collectively, these and other observations have implicated donor-derived liver cDCs in the promotion of liver transplant tolerance.¹² In addition, our recent studies³⁰ suggest that graft-infiltrating host cDC that have acquired donor MHC Ag via cross-dressing, regulate anti-donor T cell responses and promote mouse liver allograft tolerance. By contrast, the role of non-conventional pDCs, that are comparatively abundant in mouse liver,²⁸ and that have well-described tolerogenic properties,^15,25,34 has not been examined in the context of (spontaneous) liver transplant tolerance.

pDCs derived from the liver, but not the spleen, are reportedly endowed with a high intrinsic tolerogenic potential,^14,15 and interactions between pDCs and CD4⁺ T cells, particularly within LNs, have been proposed to be instrumental for peripheral Ag-specific Treg development and subsequent heart graft survival under tolerogenic conditions.²⁵ In this study, we show that freshly isolated liver pDCs, that express high levels of DAP12 (shown previously to negatively regulate liver cDC maturation)¹³ and its co-receptor TREM2, and also high T cell coinhibitory PD-L1:costimulatory CD80/86 ratios compared with lymphoid tissue pDCs, potently suppress donor alloreactive T cell responses. Notably, in studies of clinically tolerant liver transplant patients, we have reported⁴⁹ that high PD-L1:CD86 ratios on circulating pDCs correlate with elevated Treg. Our current observations further show that, unlike transplantation of normal, fully MHC-mismatched livers, grafting of pDC-depleted donor liver allografts is associated with reduced levels of both host Tregs in grafts and LNs and exhausted, graft-infiltrating PD-1⁺Tim3⁺ CD8⁺ T effector cells, that was reversed by in vivo blockade of the PD1-PD-L1 pathway. These findings implicate donor pDCs in the regulation of anti-donor T cell reactivity and the promotion of “spontaneous” liver transplant tolerance. Like liver graft-infiltrating host cDCs that acquire donor MHC Ag via cross-dressing³⁰ and express comparatively high ratios of PD-L1:CD80/86,³⁰ donor-derived pDCs appear to augment the exhaustion of effector T cells early posttransplant. The findings are also in keeping with earlier accounts of abortive T cell responses (apoptotic cell death/deletion) early (3–5 days) posttransplant within liver allografts tolerated in MHC-mismatched mouse or rat recipients.^9–11

Based on the data obtained in this study, the role of pDCs in promotion of spontaneous liver transplant tolerance can be ascribed both to their capacity to enhance Treg and their ability to promote alloreactive CD8⁺ T cell exhaustion. Indeed, in mice, Tregs can induce a dysfunctional state resembling T cell exhaustion in tumor-infiltrating CTLs that is characterized by low expression of effector cytokines, inefficient cytotoxic granule release, and coexpression of coinhibitory PD-1 and Tim-3.⁵⁰

These novel observations are consistent with the immune regulatory activity of pDCs (reviewed in [16]) and in particular, with previous reports from our laboratory of the regulatory function of mouse liver pDCs, that we have shown,^29,31,32 can promote alloreactive T cell apoptosis, dependent on CD4⁺ Tregs in vitro. They are also in agreement with the ability of adoptively transferred liver pDCs loaded with oral Ag to induce Ag-specific suppression of CD4⁺ and CD8⁺ T cell responses,²⁷ and with the ability of pDCs of donor origin to prolong vascularized heart allograft survival when administered alone or in combination with costimulation blockade in the same strain combination as used in the present study.^24,51 In addition, our findings are consistent with the recently reported observation that Foxp3 induction in naïve T cells by allogeneic pDCs in vitro correlates with mouse strain combinations in which spontaneous acceptance of kidney allografts is observed.²⁶

Engagement of PD-1 by PD-L1 negatively regulates lymphocyte activation,⁵² whereas blockade of PD-1 ligands on DCs enhances T cell activation and cytokine production.⁵³ PD-L1/PD-1 interaction plays a pivotal role during CD8⁺ T cell exhaustion from chronic Ag stimulation, and blockade of the PD-L1/PD-1 pathway reverses exhaustion.⁵⁴ Several liver cell types can express PD-L1¹² and Morita et al³⁷ have reported that mesenchyme-mediated immune control, in particular PD-L1 expression by the graft, is important for the development of mouse liver transplant tolerance. They have further shown³⁷ that PD1-PD-L1 pathway blockade, or transplantation of PD-L1^−/− liver allografts, leads to pronounced leukocyte infiltration and abrogation of transplant tolerance. The present data, showing that transplantation of pDC-depleted donor livers results in reduced levels of exhausted, graft-infiltrating PD-1⁺TIM3⁺CD8⁺ T effector cells, reversed by PD1-PD-L1 pathway blockade, links for the first time, the immunoregulatory function of donor liver interstitial pDCs expressing high levels of PD-L1 relative to costimulatory molecules, host alloreactive T cell exhaustion, Treg development and the induction of mouse spontaneous liver transplant tolerance.

Abbreviations:

Ab

antibody

Ag

antigen

ALT

alanine aminotransferase

APC

antigen presenting cell(s)

CBA

cytokine bead array

CFSE

carboxyfluorescein succinimidyl ester

DAP12

DNAX-activating protein of 12 kDa

DC

dendritic cell(s)

GAPDH

glyceraldehyde 3-phosphate dehydrogenase

MFI

mean fluorescence intensity

MHC

major histocompatibility complex

MLR

mixed leukocyte reaction

mRNA

messenger RNA

NPC

non-parenchymal cell(s)

PD1

programed cell death protein 1

pDC(s)

plasmacytoid dendritic cell(s)

PD-L1

programed death ligand 1

RT

room temperature

Tim-3

T cell immunoglobulin and mucin domain-containing protein 3

Treg(s)

regulatory T cell(s)

TREM

triggering receptor expressed on myeloid cells

WT

wild-type

DATA AVAILABILITY STATEMENT

The data that support the findings are available from the corresponding author upon reasonable request.