PD-1 and CTLA-4 up regulation on donor T cells is insufficient to prevent GvHD in allo-HSCT recipients

The expression of checkpoint blockade molecules PD-1, PD-L1, CTLA-4, and foxp3+CD25+CD4+ T cells (Tregs) regulate donor T cell activation and graft-vs-host disease (GvHD) in allogeneic hematopoietic stem cell transplant (allo-HSCT). Detailed kinetics of PD-1-, CTLA-4-, and PD-L1 expression on donor and host cells in GvHD target organs have not been well studied. Using an established GvHD model of allo-HSCT (B6 → CB6F1), we noted transient increases of PD-1- and CTLA-4-expressing donor CD4+ and CD8+ T cells on day 10 post transplant in spleens of allo-HSCT recipients compared with syngeneic HSCT (syn-HSCT) recipients. In contrast, expression of PD-1- and CTLA-4 on donor T cells was persistently increased in bone marrow (BM) of allo-HSCT recipients compared with syn-HSCT recipients. Similar differential patterns of donor T cell immune response were observed in a minor histocompatibility (miHA) mismatched transplant model of GvHD. Despite higher PD-1 and CTLA-4 expression in BM, numbers of foxp3+ T cells and Tregs were much lower in allo-HSCT recipients compared with syn-HSCT recipients. PD-L1-expressing host cells were markedly decreased concomitant with elimination of residual host hematopoietic elements in spleens of allo-HSCT recipients. Allo-HSCT recipients lacking PD-L1 rapidly developed increased serum inflammatory cytokines and lethal acute GvHD compared with wild-type (WT) B6 allo-HSCT recipients. These data suggest that increased expression of checkpoint blockade molecules PD-1 and CTLA-4 on donor T cells is not sufficient to prevent GvHD, and that cooperation between checkpoint blockade signaling by host cells and donor Tregs is necessary to limit GvHD in allo-HSCT recipients.


Introduction
Donor T-lymphocyte infusion can be an effective form of adoptive immunotherapy in the context of allo-HSCT, but life threatening complications related to GvHD limit its clinical application. Removal of donor T cells from the graft reduces GvHD but increases the incidences of graft failure, opportunistic infection, and tumor relapse [1][2][3]. Immunosuppressive drugs are commonly used to control GvHD, but have incomplete efficacy, and are frequently associated with drug-related toxicities and mortality [4]. Therefore, modulating donor T cell activity to enhance immune response against opportunistic infection and against tumor relapse in allo-HSCT recipients without increasing GvHD remains a long-standing goal. Programmed death-1 (PD-1) and cytotoxic T-lymphocyte antigen-4 (CTLA-4) expression negatively regulate T cell activity and lack of their expression leads to autoimmune diseases [5][6][7][8][9]. Therefore, in vivo immune modulation of donor T cells through PD-1 and CTLA-4 signaling pathways may play an important role in controlling GvHD in allo-HSCT recipients. Detailed kinetic studies of PD-1 and or CTLA-4 expression on donor CD4+ and CD8+ T cells and the kinetics of inducible PD-L1 expression on host cells and donor cells in GvHD target organs have not been done.
PD-1 has two known ligands, PD-L1 and PD-L2. PD-L1 binding to PD1 has a greater negative regulatory effect on T cells than does its interaction with PD-L2 [10][11][12][13][14]. PD-L1 is widely expressed on hematopoietic and non-hematopoietic cells and its expression is upregulated in the presence of immune stimuli [15]. Functional signaling through PD-1 is critical to limit GvHD in allo-HSCT, since blockade of PD-1 or PD-L1 with monoclonal antibody (mAb) increases GvHD toxicity and PD-L1 expression in host tissues is required to reduce GvHD [16,17]. In vivo PD-L1 blockade increases CD4+ T cell-mediated skin allograft rejection and prevents apoptosis of alloantigen-specific effector T cells [18]. In addition to PD-1 signaling, inhibitory signaling through CTLA-4 binding to CD80/CD86 limits GvHD. CTLA-4 blocks the CD28-B7 mediated co-stimulatory pathway and inhibits donor T cell allo-reactivity in allo-HSCT recipients [7]. Deficiency of CTLA-4 in Tregs causes spontaneous development of systemic lymphoproliferation, fatal T cell-mediated autoimmune disease, and hyperproduction of immunoglobulin E in mice [19]. Administration of antibodies against CTLA-4 immediately after allo-HSCT was shown to increase acute GvHD, while the blocking of CTLA-4 at late phases of allo-HSCT did not increase GvHD but resulted in lethal lympho-splenomegaly, autoimmune hepatitis and increased circulating anti-DNA auto-antibodies [9]. In addition to the role of CTLA-4 and PD-1 in regulating allo-reactivity and GvHD, foxp3-expressing inducible donor Tregs and/or the addition of in vitro expanded donor Tregs to the graft have shown promising therapeutic benefit in the prevention of GvHD [20,21].
In this study, we have explored the kinetics of expression of PD-1 and CTLA-4 by donor CD4+ T and CD8+ T cells in the spleen and BM, two hematolymphoid organs that are a target for GvHD. We show herein that, despite the abundant expression of inducible PD-1 and CTLA-4 by donor CD4+ and CD8+ T cells, and inducible PD-L1 expression by both host and donor cells in lethally irradiated allo-HSCT recipients, expression of these molecules failed to prevent the development of severe GvHD in allo-HSCT recipients. Importantly, up-regulation of PD-1 and CTLA-4 on donor CD4+ and CD8+ T cells were differentially regulated in spleen versus the BM of allo-HSCT recipients. The numbers of PD-L1 expressing host spleen cells markedly decreased over time, concomitant with the induction of GvHD in allo-HSCT recipients. PD-L1 KO allo-HSCT recipients had lethal acute GvHD associated with increased serum inflammatory cytokines, increased numbers of activated donor CD4+ and CD8+ T cells expressing ICOS-1, CD69 and PD-1 and decreased numbers of Tregs in the spleen compared with WT B6 allo-HSCT recipients. Our data highlight the differential kinetics of PD-1 and CTLA-4 expression by donor T cells in hematolymphoid tissues of allo-HSCT recipients and their role as potential targets for therapeutic intervention.

Preparation of donor splenocytes and BM from donor mice
For B6 ! CB6F1 allo-and syn-HSCT models, splenocytes were obtained from PepBoy B6 congenic mice and cultured overnight at 10 7 cells/ml in complete media at 37˚C in 5% CO 2 [22] to deplete adherent cells, particularly monocytes. Live cells were counted under fluorescence microscopy using ethidium bromide mixed with acridine orange dye [23]. Following overnight culture the cells were washed once and transplanted via intravenous injection. Bone marrow (BM) was flushed from femora and tibia of naïve CD45.2+Thy1.1+ B6 congenic donor mice and CD3+ T-cells were depleted (TCD) as described [24]. For the B10. BR ! PD-L1 KO B6, PD-L2 KO B6 or WT B6 allo-HSCT models, splenocytes from naïve B10.BR mice were harvested, cultured overnight and T cells were further enriched for hematopoietic stem cells by depleting CD11b+CD11c+CD19+ cells using a Macs column. BM cells were harvested from naïve BA.B10 mice and hematopoietic stem cells were enriched by depleting CD3+ CD11b+ CD11c+ CD19+ cells using a MACS separation column as described previously [24].

Lymphocyte isolation from the spleen and BM of allo-and syn-HSCT recipients
Splenocytes were harvested from allo-and syn-HSCT recipients by gently crushing tissue between frosted glass slides (Fisher Scientific, Pittsburgh, PA, USA) [25]. BM cells were harvested from the tibia and femur of one hind leg as described before [27]. Single cell suspensions were prepared by passing the harvested cell suspension through a cell strainer (Becton Dickinson, Franklin Lakes, NJ, USA). Live lymphocytes were quantitated by fluorescence microscopy as previously described [24].

Serum cytokines measured by Luminex assay
Serum was harvested on day 8 after transplantation from PD-L1 KO and B6 allo-HSCT recipient mice. Mouse 26-plex kits were purchased from Affymetrix Inc (Santa Clara, CA) and the Luminex assay (Luminex Corp., Austin, Texas) was performed in a blinded fashion at the Immunology Core Laboratory at Stanford University (Stanford, CA) according to the manufacturer's recommendations. All samples were assayed in a single batch, and each sample was measured in duplicate. Plates were read using a Luminex 200 instrument (Luminex Corp) as previously described [28].

Statistical analyses
Mean values obtained from 4 to 6 mice per group were compared using two-sided Student's ttest. Differences were considered significant when p values below the 0.05 levels were obtained.

Increased expression PD-1 and CTLA-4 on donor CD4+ and CD8+ T cells is inadequate to reduce GvHD in allo-HSCT recipients
Previous work using in vivo imaging has shown that donor T cells migrate to almost all body compartments (GvHD target organs and other tissues) within 72 hours of transplantation and signaling through PD-1 and CTLA-4 play important roles in the regulation of allo-reactive donor T cells that cause GvHD [7] [29]. Using an established model of sub-acute GvHD (B6 ! CB6F1) in which allo-HSCT recipients experienced 30% mortality and 20% weight loss by 100 days post transplant (S1 Fig To explore the basis for the differences in inflammation seen in the various hematolymphoid organs we studied the kinetics of PD-1 expression on donor T cells during sub-acute GvHD. Donor CD4+ and CD8+ T cells recovered from the spleen and BM of allo-HSCT recipients were phenotyped by flow cytometry and compared with T cells from syn-HSCT recipients. Numbers of CD45.1+ donor CD4+ and CD8+ donor T cells were significantly higher in the spleens of allo-HSCT recipients on day 10 after transplant but were lower at later time points compared with syn-HSCT recipients ( Fig 1A). On the other hand, the numbers of CD45.1+ donor CD4+ T cells (but not CD8+ T cells) were higher at all time points in the BM of allo-HSCT recipients compared with syn-HSCT recipients ( Fig 1B). These data suggested the hypothesis that local inflammation caused by GvHD in hematolymphoid organs is differentially regulated in the spleen versus the BM according to the immunological state of donor T cells at each site. To address this question, we measured the kinetics of PD-1 expression on CD4+ and CD8+ T cells in the spleen and BM of allo-and syn-HSCT recipients. Higher percentages of CD4+ T cells expressing PD-1 were detected in the spleen and BM on days 10 and 25 (but not on day 103) in allo-HSCT recipients compared with syn-HSCT recipients ( Fig 1C). The percentages of PD-1 expressing CD8+ donor T cells in both spleen and BM were higher at all times post transplant in allo-HSCT recipients compared with syn-HSCT recipients ( Fig  1D). Accordingly, the absolute numbers of PD-1 expressing donor CD4+ and CD8+ T cells were significantly increased on day 10 post transplant in spleen of allo-HSCT recipients but equivalent or lower at subsequent time points compared with syn-HSCT recipients (Fig 1E  and 1F). In contrast, the numbers of PD-1 expressing donor CD4+ T cells were persistently higher at all time points in the BM of allo-HSCT recipients compared with syn-HSCT ( Fig 1G) and numbers of PD-1 expressing donor CD8+ T cells were higher only on days 10 and 25 (but not on day 103) after transplant ( Fig 1H). These data suggest that the kinetics of PD-1 expression on donor CD4+ and CD8+ T cells is differentially regulated in spleen versus BM.
We next determined the kinetics of CTLA-4 expression on donor CD4+ and CD8+ T cells in the spleen and BM of allo-and syn-HSCT recipients. Similar percentages of CTLA-4 expression were detected in donor CD4+ (Fig 2A) and CD8+ (Fig 2B) T cells in the spleen and BM of allo-and syn-HSCT recipients at all three time points post transplant. However, a transient increase in the absolute numbers of CTLA-4 expressing CD4+ and CD8+ T cells was seen on day 10 after transplant in the spleen of allo-HSCT compared with syn-HSCT recipients ( Fig  2C). On the other hand, absolute numbers of CTLA-4 expressing CD4+ T cells (but not CD8+ T cells) in the BM of allo-HSCT recipients were significantly increased at all time points after transplant compared with syn-HSCT ( Fig 2D). These data suggest that higher levels of CTLA4-expressing donor T cells may be responsible for limiting GvHD-associated inflammation in the BM (but not the spleen) of allo-HSCT recipients.
Next, we investigated whether similar increased expression of PD-1 and CTLA-4 by donor T cells also occurs in a minor histocompatibility (miHA)-mismatched mouse GVHD model  Fig 3A) which may more closely resemble clinical GvHD seen in human allo-HSCT patients [26]. To address this we used the CD8+ T cell dependent C3H.SW!B6 miHA GvHD model which has minimal early lethality allowing collection of data from mice of multiple time points on after BMT [26,30]. As anticipated, all recipient mice receiving 5 million donor splenocytes from C3H.SW donor mice survived with only slight delays in bodyweight gain and without significantly increased GvHD clinical scores compared with the syngeneic control recipients (Fig 3B and 3D). Total numbers of leukocytes were significantly lower (p<0.05) on day 10 in spleen and on day 26 in BM in miHA recipients compared with the syngenic control recipients (Fig 4A). Consistent with CD8+ T-cell mediated GVHD in this model, numbers of CD8+ T cells were higher in spleen on day 26 after allo-HSCT (Fig 3B) while numbers of donor spleenderived CD4+ T cells in the spleen were significantly lower on day 10 compared with syn-HSCT recipients. In contrast, examination of BM showed higher numbers of donor spleenderived CD4+ and CD8+ T cells T cells on day 10 post-transplant compared with syngenic control recipients (Fig 4B). Similarly, the numbers of donor BM-derived CD4+ and CD8+ T cells were significantly higher in BM of miHA recipients versus syngenic controls but lower in spleen on day 10 and 26 post-transplant (Fig 4B). PD-1 expressing donor spleen-derived CD4 + T cells were significantly higher in spleen only whereas PD-1 expressing donor spleen- derived CD8+ T cells were also significantly higher in the spleen and BM of miHA recipients on day 10 post-transplant compared with the syngenic control recipients (Fig 4C). PD-1 expressing CD4+ and CD8+ T cells, generated de novo from donor BM, were also significantly higher in the BM of miHA transplant recipient on day 10 compared with the syngenic control recipients (Fig 4C). We could not detect differences in the expression of CTLA-4 by donor spleen-and BM-derived CD4+ and CD8+ T cells in spleen or BM comparing miHA with Since CTLA-4 causes activation-induced cell death on T cells and PD-1 functions as a cell death inducer [31,32] we reasoned that expression of one or both molecules could play an important role in the contraction phase of donor cells in the spleen on day 25 after transplant. We next measured the kinetics of both PD-1 and CTLA-4 expressing CD4+ and CD8+ T cells in the spleen and BM of allo-and syn-HSCT recipients. The percentage of PD-1 and CTLA-4 expressing CD4+ and CD8+ donor T cells decreased considerably on day 25 but rebounded by day 103 after transplant in the spleen of both allo-HSCT and syn-HSCT recipients (Fig 2E). In contrast, the percentage of CD8+ T cells (not CD4+ T cells) expressing PD-1 and CTLA-4 decreased considerably on day 25 after transplant in the BM of both allo-and syn-HSCT recipients ( Fig 2E). These data suggest that the activation status and levels of PD-1 and CTLA-4 expression on donor CD4+ and CD8+ T cells in the spleen and BM of allo-HSCT recipients GvHD are differentially regulated and overall data are summarized in Table 1.

GvHD causes reduced expression of foxp3 transcription factor by donor T cells
The transcription factor foxp3 is expressed by subsets of CD4+ and CD8+ T cells and induces negative regulatory effects on donor allo-reactive T cells. Increased foxp3 expression on donor T cells is associated with reduced levels of GvHD in both mice and in human [21,[33][34][35][36][37][38]. To explore the role of foxp3 transcription factor expressing donor T cells in the induction of GvHD, we measured foxp3 expression in CD4+ T cells and Tregs in the spleen on day 10 after transplantation in allo-HSCT recipients. The proportion of CD4+ and CD4-T cells expressing foxp3 was significantly lower (p<0.005) on day 10 post-transplant in the spleen of allo-HSCT recipients compared with syn-HSCT recipients (Fig 5A). Accordingly, the percentages of T cells that were foxp3+ and the absolute numbers of Tregs per spleen were significantly lower in allo-HSCT recipients compared with syn-HSCT recipients (Fig 5B and 5C). These data show that lower levels of foxp3 expression by donor T cells was associated with the peak expansion of activated donor T cells on day 10 post allo-HSCT.
Inducible PD-L1 expression on host and donor tissues was insufficient to prevent GvHD in allo-HSCT recipients PD-L1 is expressed by various tissues of hematopoietic and non-hematopoietic origins [15] and Saha et al showed that expression of PD-L1 by host tissues is required to limit GvHD in allo-HSCT recipients [17]. To explore PD-L1 expression on donor and host cells in allo-HSCT recipients, we sacrificed allo-HSCT recipients on days 4, 10 and 36 after transplant and measured PD-L1 expression on donor spleen-, donor BM-and host-derived cells in the spleen of allo-HSCT recipients. The data were compared with similar data obtained from the syn-HSCT recipients. The percentages of donor spleen-derived (CD45.1+H-2 d -) and host-derived (CD45.1-H-2 d +) cells decreased over time after transplant in allo-HSCT recipients (Fig 6A,  upper panel). More CD45.1+ donor cells expressed PD-L1 at all three time points in allo-HSCT recipients than syn-HSCT recipients (Fig 6B, upper panel). While most of the hostderived CD3+ T cells and CD3-non-T cells also expressed PD-L1 on day 4 post transplant, host non-T cells expressing PD-L1 eventually disappeared over time in allo-HSCT recipients (Fig 6C), with the remaining PD-L1 positive cells being donor BM-derived cells (Fig 6D). These data suggest that PD-L1 expression on host and donor hematopoietic cells was insufficient to prevent GvHD in target organs of allo-HSCT recipients.

Days post-BMT
Comparison of absolute numbers of PD-1 and CTLA-4 expressing donor spleen-derived CD4+ and CD8+ T cells, and Tregs harvested on day 10, 25 or 26 and 103 from the spleen and BM of the major and minor MHC mis-matched allo-HSCT and syn-HSCT recipients. The symbols ", "", and """ represent the relative significance of the increase, and #, ##, and ### represent the relative significance of the decrease (corresponding to p<0.05, p<0.005 and p<0.001 values (Student's t-Test), respectively) in the mean numbers of PD-1+ and Tregs and CTLA-4+ donor spleen-derived CD4+ and CD8+ T cells in the spleen and BM of allo-HSCT recipients compared to data obtained from the spleen and BM of B6! B6 syngeneic transplants. The "-"is for data for which there was no significant difference noted between the allo-HSCT and syn-HSCT recipients. The "ND" is for the experiments not done.
https://doi.org/10.1371/journal.pone.0184254.t001  (Fig 7A and 7B), while donor CD4+ and CD8+ T cells increased~6-fold in the spleen of PD-L1 KO allo-HSCT recipients compared with PD-L2 KO and WT B6 allo-HSCT recipients, with >95% of all T cells expressing PD-1 in allo-HSCT recipients (Fig 7C). Accordingly, total numbers of donor CD4+ ( Fig 7D) and CD8+ (Fig 7E) T cells per spleen increased significantly in PD-L1 KO allo-HSCT recipients compared with PD-L2 KO and WT B6 allo-HSCT recipients. In addition to increased PD-1 expression in donor CD4+ and CD8+ T cells in the spleen (Fig 7F and 7G), ICOS-1 expression on donor CD4+ and CD8+ T cells was also significantly increased in PD-L1 KO allo-HSCT recipients (Fig 7H and 7I). Moreover, significantly higher percentages of PD-1+ CD69+ double-positive CD4+ T cells with an activated phenotype (but not double-positive CD8+ T cells) were detected in the spleen of PD-L1 KO allo-HSCT recipients compared with WT B6 allo-HSCT recipients (Fig 7J). While the numbers of CD25 +CD4+ activated T cells were similar in the spleens of PD-L1 KO allo-HSCT and WT allo- HSCT recipients (Fig 7K, left panel), significantly lower numbers of foxp3+ Tregs were seen in the spleen of PD-L1 KO allo-HSCT recipients compared with WT B6 allo-HSCT recipients ( Fig  7K, right panel). Collectively, these data suggest that increased GvHD in PD-L1 KO allo-HSCT recipients is associated with increased proliferation and activation of donor CD4+ and CD8+ T cells, but not non-T cells and donor Tregs, and that up-regulation of PD-1 expression on donor T cells is independent of the expression of PD-L1 in allo-HSCT recipients.
Severe acute GvHD in PD-L1 KO allo-HSCT recipients is associated with increased production of inflammatory cytokines To determine the effect of the absence of PD-L1 expression on host cells on the immunological milieu in allo-HSCT recipients we transplanted B10.BR cells into PD-L1 KO mice and control WT B6 mice, and measured serum cytokine and chemokine levels on day 8 after transplant (S3 Fig). PD-L1 KO allo-HSCT recipients had significantly higher levels of 13 inflammatory cytokines/chemokines compared with WT B6 recipients ( Fig 8A). Next, we determined the acute GvHD pathological score from formalin fixed histological tissue sections of liver, small and large intestines harvested on day 8 post-transplant from WT B6, PD-L1 KO and PD-L2 KO allo-HSCT recipients (as described in Materials and methods). PD-L1 KO allo-HSCT recipients had higher GvHD pathology as determined through their clinical GvHD scores (weight loss, posture, activity, skin integrity) and histological GvHD scores compared with WT or PD-L2 KO recipients (Fig 8B and 8C). The markedly increased acute GvHD in the absence of PD-L1 expression on host tissues compared with WT recipients indicate that PD-L1 expression by host tissue is required to limit allo-activation of donor T cells, presumably through PD1-signaling.

Discussion
In this study, we have explored the kinetics of expression of three inhibitory immune checkpoint pathways, PD-1/PD-L1, CTLA-4, and foxp3+ Tregs, that are relevant to the control of GvHD in allo-HSCT recipients. Our data indicate that the expression levels of immune checkpoint blockade molecules and Tregs varies dynamically over time and across different hematolymphoid tissues and organs in transplant recipients. Among transplant recipients with established GvHD, the expression levels and numbers of PD-1 and CTLA-4 positive donor T cells were higher in the BM than the spleen, consistent with a recent publication showing that the BM microenvironment is relatively immune-suppressive [39]. Host APCs surviving in the spleen and BM after lethal irradiation have been shown to play a pivotal role in activating alloreactive donor T cells that cause GvHD [40][41][42]. The larger numbers of mature APC in the spleen versus BM may also explain the increased susceptibility of the spleen to the effects of GvHD. The presence of GvHD in the spleen leads to marked atrophy and loss of splenic white pulp. Our data also support the data published by Zhang et al, showing that the recruitment and activation of donor allo-reactive CD8+ T cells in the liver and spleen by host DCs and macrophages was differentially regulated compared with that in lymph nodes [40]. The GvHD pathology seen in miHA mismatched allo-HSCT model has resemblance to the GvHD pathology seen in human allo-HSCT recipients [43], and the data presented in this study could provide important clues for designing therapeutic interventions to treat or prevent clinical GvHD. Significantly increased numbers of donor C3H.SW CD8+ T cells were seen in recipient BM on day 10 and in spleen on day 26 in the C3H.SW ! B6 miHA mismatched allo-HSCT recipients presumably because donor CD8+ T cells mediate GvHD pathogenicity in this model (Fig 4B) [30]. Of note, in miHA mismatched model, we were unable to quantify significant numbers of CTLA-4 expressing CD4+ and CD8+ T cells in the spleen and BM through flow cytometry analysis on both day 10 and 26 after transplant, suggesting that increased PD-1 expression on donor CD8+ T cells in spleen and BM in the miHA mismatched allo-HSCT model may be sufficient to reduce GvHD pathogenicity in miHA versus MHC mis-matched allo-HSCT recipients may (Fig 4C).
PD-1 expression on T cells significantly increases in the presence of viral infection and after transplantation [17,24,44]. Our data confirm that PD-1 expression on CD4+ and CD8+ T cells persists in allo-HSCT recipients until 100+ days post transplant in spite of active GvHD. In the absence of PD-L1-expressing host cells the numbers of donor Tregs are decreased, inflammatory cytokine increase leading to expansion of PD-1+ donor effector T-cells and acute GvHD damage (Figs 5 and 6). Our findings in allo-HSCT recipients are consistent with those in acute LCMV-infected mice, in which PD-1+ T cells are cytolytically active [44]. Our data from allo-HSCT recipients lacking PD-L1 with severe acute GvHD and mortality are also consistent with the observation that chronically (but not acutely) LCMV-infected PD-L1 KO mice develop severe immunopathologic tissue damage and die within a few days of infection [44]. However, the effect of PD-1/PD-L1 mediated inhibitory signaling is more noticeable in tumor models as it helps tumors to escape immune surveillance by inducing apoptosis of tumor specific T cells [45]. Inhibition of either PD-1 or PD-L1 with mAbs restores the effector function of viral specific and tumor specific T cell responses, and clinical trials are currently testing PD-1 or PD-L1 blockade in cancer patients [6,12]. Although PD-1/PD-L1 blockade can increase anti-tumor activity of donor T cells [44,45], antibodies to PD-1/PD-L1 increase GvHD lethality and mortality in allo-HSCT recipient mice [16]. Administration of engineered soluble PDL1-Ig protein, which reduces the proliferation of T cells during transplantation, may be an alternative approach to reduce GvHD [46] although autoimmune complications will be an issue [47].
Our data showing decreased numbers of donor Tregs associated with significantly reduced expression of the intracellular foxp3 transcription factor in CD4+ T cells in the spleen of allo-HSCT recipients (Fig 5) is consistent with previously published clinical data demonstrating that allo-HSCT patients with GvHD had decreased Tregs in blood compared with patients treated with either allo-HSCT recipients without GvHD or auto-HSCT recipients [48].
Although the mechanism behind the reduced generation of Tregs in allo-HSCT recipients with GvHD is not well documented, it is conceivable that the increased production of inflammatory cytokines in allo-HSCT recipients with GvHD may prevent the generation of Tregs.
In conclusion, our data provide important information showing that expression of inducible PD-1 and CTLA-4 on donor T cells, and PD-L1 expression on both donor and host cells in allo-HSCT recipients are insufficient to prevent donor T cell mediated allo-reactivity and local GVHD-mediated inflammation in hematolymphoid organs. Although allo-HSCT recipients lacking PD-L1 had increased numbers of activated donor CD4+ and CD8+ T cells and increased GvHD severity, induction of donor Tregs did not occur. Our data suggest that activation of donor Treg through PD1 signaling by PD-L1-expressing host cells necessary to prevent severe acute GvHD. Novel approaches to regulate GvHD should focus on the preservation of PD-L1 expression on host hematopoietic and non-hematopoietic host cells of GvHD target organs alone or in combination with adoptive therapy using ex vivo expanded Tregs in the graft, with the goal of providing potential therapeutic benefit against GvHD.