Differential migration of passenger leukocytes and rapid deletion of naive alloreactive CD8 T cells after mouse liver transplantation

Authors

  • Szun S. Tay,

    1. Liver Immunology Group, Centenary Institute, Newtown, Australia
    2. A. W. Morrow Gastroenterology and Liver Centre, Royal Prince Alfred Hospital, University of Sydney, Sydney, Australia
    Search for more papers by this author
    • These authors contributed equally to this work.These authors are equal last authors.

  • Bo Lu,

    1. Immunology Research Centre, St. Vincent's Hospital, Melbourne, Australia
    Search for more papers by this author
    • These authors contributed equally to this work.These authors are equal last authors.

  • Fred Sierro,

    1. Liver Immunology Group, Centenary Institute, Newtown, Australia
    2. A. W. Morrow Gastroenterology and Liver Centre, Royal Prince Alfred Hospital, University of Sydney, Sydney, Australia
    Search for more papers by this author
  • Volker Benseler,

    1. Liver Immunology Group, Centenary Institute, Newtown, Australia
    2. A. W. Morrow Gastroenterology and Liver Centre, Royal Prince Alfred Hospital, University of Sydney, Sydney, Australia
    Search for more papers by this author
  • Claire M. McGuffog,

    1. Liver Immunology Group, Centenary Institute, Newtown, Australia
    2. A. W. Morrow Gastroenterology and Liver Centre, Royal Prince Alfred Hospital, University of Sydney, Sydney, Australia
    Search for more papers by this author
  • G. Alex Bishop,

    1. Collaborative Transplantation Research Group, Bosch Institute, Royal Prince Alfred Hospital, University of Sydney, Sydney, Australia
    Search for more papers by this author
  • Peter J. Cowan,

    1. Immunology Research Centre, St. Vincent's Hospital, Melbourne, Australia
    Search for more papers by this author
  • Geoffrey W. McCaughan,

    1. Liver Immunology Group, Centenary Institute, Newtown, Australia
    2. A. W. Morrow Gastroenterology and Liver Centre, Royal Prince Alfred Hospital, University of Sydney, Sydney, Australia
    Search for more papers by this author
  • Karen M. Dwyer,

    1. Immunology Research Centre, St. Vincent's Hospital, Melbourne, Australia
    2. Department of Medicine, University of Melbourne, Melbourne, Australia
    Search for more papers by this author
  • David G. Bowen,

    Corresponding author
    1. Liver Immunology Group, Centenary Institute, Newtown, Australia
    2. A. W. Morrow Gastroenterology and Liver Centre, Royal Prince Alfred Hospital, University of Sydney, Sydney, Australia
    3. Collaborative Transplantation Research Group, Bosch Institute, Royal Prince Alfred Hospital, University of Sydney, Sydney, Australia
    • Address reprint requests to David G. Bowen, Ph.D., M.D., Liver Immunology Group, Centenary Institute, Locked Bag No. 6, Newtown, New South Wales 2042, Australia. Telephone: + 61 2 95656264; FAX: +61 2 95656101; E-mail: d.bowen@centenary.org.auAddress reprint requests to Patrick Bertolino, Ph.D., Liver Immunology Group, Centenary Institute, Locked Bag No. 6, Newtown, New South Wales 2042, Australia. Telephone: +61 2 95656186; FAX: +61 2 95656101; E-mail: p.bertolino@centenary.org.au

    Search for more papers by this author
    • These authors are equal last authors.

  • Patrick Bertolino

    Corresponding author
    1. Liver Immunology Group, Centenary Institute, Newtown, Australia
    2. A. W. Morrow Gastroenterology and Liver Centre, Royal Prince Alfred Hospital, University of Sydney, Sydney, Australia
    • Address reprint requests to David G. Bowen, Ph.D., M.D., Liver Immunology Group, Centenary Institute, Locked Bag No. 6, Newtown, New South Wales 2042, Australia. Telephone: + 61 2 95656264; FAX: +61 2 95656101; E-mail: d.bowen@centenary.org.auAddress reprint requests to Patrick Bertolino, Ph.D., Liver Immunology Group, Centenary Institute, Locked Bag No. 6, Newtown, New South Wales 2042, Australia. Telephone: +61 2 95656186; FAX: +61 2 95656101; E-mail: p.bertolino@centenary.org.au

    Search for more papers by this author
    • These authors are equal last authors.


Abstract

Donor passenger leukocytes (PLs) from transplanted livers migrate to recipient lymphoid tissues, where they are thought to induce the deletion of donor-specific T cells and tolerance. Difficulties in tracking alloreactive T cells and PLs in rats and in performing this complex surgery in mice have limited progress in identifying the contribution of PL subsets and sites and the kinetics of T cell deletion. Here we developed a mouse liver transplant model in which PLs, recipient cells, and a reporter population of transgenic CD8 T cells specific for the graft could be easily distinguished and quantified in allografts and recipient organs by flow cytometry. All PL subsets circulated rapidly via the blood as soon as 1.5 hours after transplantation. By 24 hours, PLs were distributed differently in the lymph nodes and spleen, whereas donor natural killer and natural killer T cells remained in the liver and blood. Reporter T cells were activated in both liver and lymphoid tissues, but their numbers dramatically decreased within the first 48 hours. These results provide the first unequivocal demonstration of the differential recirculation of liver PL subsets after transplantation, and show that alloreactive CD8 T cells are deleted more rapidly than initially reported. This model will be useful for dissecting early events leading to the spontaneous acceptance of liver transplants. Liver Transpl 19:1224–1235, 2013. © 2013 AASLD.

Abbreviations
APC

allophycocyanin

B6

Ly5.2+ C57BL/6 allogeneic

B10

Ly5.2+ B10.BR syngeneic

cDC

conventional dendritic cell

CFSE

carboxyfluorescein succinimidyl ester

Cy

Cyanin

DC

dendritic cell

EOS

eosinophil

FSC

forward scatter

LN

lymph node

MHC

major histocompatibility complex

ni

number of cells in the ith division peak

NK

natural killer

NKT

natural killer T

PE

phycoerythrin

PerCP

peridinin chlorophyll protein

PL

passenger leukocyte

RAG

recombinase activating gene

SSC

side scatter

TCR

T cell receptor

Treg

regulatory T cell

Liver allografts possess unique tolerogenic properties in comparison with other solid organ transplants. In the clinical setting, human leukocyte antigen matching is not required, chronic rejection is uncommon, and up to 20% of recipients can be weaned off immunosuppression.[1] In animal models, liver transplants are spontaneously accepted in outbred pigs and in most mouse and many rat strain combinations, and they are further able to induce donor-specific tolerance and prevent the rejection or prolong the survival of other organs from the same donor, including the kidneys, heart, pancreas, and skin.[1] Despite the observation of this phenomenon for more than 40 years, the mechanisms of liver allograft tolerance are not fully understood.

In comparison with other organs, the liver contains a large and diverse population of bone marrow–derived hematopoietic cells, which include lymphocytes, myeloid cells, and progenitor cells. A syngeneic liver transplant is able to reconstitute the majority of the hematological lineages in a lethally irradiated host.[2] Collectively termed passenger leukocytes (PLs), these donor liver leukocytes play important roles in liver transplant tolerance. Depleting PLs by irradiation[3] or replacing donor PLs[4, 5] prevents spontaneous liver graft acceptance in rats, which can be re-established by the infusion of purified donor leukocytes.[6]

Which PLs induce liver transplant tolerance and how and where their effects are mediated are unclear. In rats, tolerance is associated with the rapid migration of PLs and immune activation in the recipient draining lymph nodes (LNs) and spleen.[7] High rates of apoptosis in the spleen[8] and liver[8, 9] are also associated with tolerance. It has been suggested that the inappropriate timing and magnitude of allograft-reactive T cell activation in lymphoid organs, mediated by PLs, are followed by their death by neglect in the liver because of insufficient cytokine signaling, which results in the depletion of donor-reactive T cells.[10] However, the LNs and spleen are not equivalent in the priming of skin and heart transplant rejection.[11] The spleen is sometimes considered more tolerogenic,[12] although most studies have not distinguished the contributions of these organs. Studies using donor livers reconstituted with PLs from the recipient strain have also shown that instead of PLs, the liver parenchyma and/or liver-resident cells are key determinants of graft outcomes.[13] The liver parenchyma has been suggested to play an important role in tolerance by secreting soluble donor major histocompatibility complex (MHC) class I molecules[4] or by inducing the abortive activation of graft-reactive T cells.[13]

Although studies in rats have demonstrated the early migration of PLs into lymphoid tissues followed by the apoptosis of alloreactive T cells, their conclusions have been limited: they have not addressed whether all PL subsets migrate into lymphoid tissues or some remain in the liver. It is not also clear whether the spleen and LNs are seeded by distinct PL subsets. Finally, tracking alloreactive T cells in rats is difficult because of insufficient reagents for identifying specific lineages. Few studies have been performed in mice, for which leukocyte lineages are better defined and transgenic strains are available, because of technical challenges in performing this surgery. A few studies in mice have indicated roles for the apoptosis of recipient and donor cells[9, 14] and for natural killer T (NKT) cells,[15] which are enriched in the liver. However, none of these studies have systematically investigated the migration of all PL subsets. In agreement with the clonal exhaustion/deletion hypotheses, mouse liver allografts were shown to deplete a large population of donor-specific T cell receptor (TCR)–transgenic CD8 T cells in 5 days.[16]

In this study, we established a mouse liver transplant model in which donor and recipient leukocytes could be distinguished by allelic differences in CD45. To follow the fate of CD8 T cells activated via the direct allorecognition pathway, naive Des CD8 T cells expressing a transgenic TCR specific for the donor MHC molecule H-2Kb were adoptively transferred into recipient mice before transplantation. Using multicolor flow cytometry, we were able to quantify PLs, recipient leukocytes, and directly allograft-specific CD8 T cells in allografts and recipient organs after transplantation. This study confirms that PL migration occurs very early after transplantation. However, by systematically distinguishing all main PL subsets, it demonstrates for the first time that all subsets do not migrate equally in the recipient. The compositions of the PL subsets in the LNs and the spleen were different, and this suggests that these compartments can induce distinct pathways of alloreactive T cell activation and play different roles in tolerance induction. We report for the first time that although some natural killer (NK) and NKT cells recirculate in the blood, a significant proportion of these cells are retained in the liver. Finally, this study reveals that the deletion of graft-reactive T cells is evident by 24 hours, an earlier time point than previously reported.

MATERIALS AND METHODS

Animals

Male C57BL/6 and B10.BR mice were purchased from the Animal Resources Centre (Perth, Australia). Ly5.1 B10.BR mice (H-2k, Ly5.1+/CD45.1+) were derived via the crossing of C57BL/6-Ptprca mice (H-2b, Ly5.1+/CD45.1+) for more than 10 generations into a B10.BR background. Des and Des–recombinase activating gene−/− (RAG−/−) mice have been described previously.[17] The mice were housed at the Centenary Institute Animal Facility under specified pathogen-free conditions. All experimental procedures were approved by the animal ethics committee of Sydney University.

Adoptive Transfer and Orthotopic Mouse Liver Transplantation

LN cells were isolated from Des-RAG−/− mice and were carboxyfluorescein succinimidyl ester (CFSE)–labeled (5 μM) as previously described.[18] Cells (2.5-5.0 × 106) were injected into recipients via the lateral tail vein 24 hours before liver transplantation. Donor livers were perfused and preserved in University of Wisconsin solution before orthotopic transplantation according to protocols described previously[19, 20] without hepatic re-arterialization. The infrahepatic vena cava, portal vein, and common bile duct were anastomosed. This procedure resulted in indefinite survival for more than 90% of the recipients of syngeneic grafts in the absence of immunosuppression and antibiotics (data not shown).

Preparation of Leukocytes From Different Organs and Flow Cytometry Analysis

The liver, spleen, LNs, and blood were harvested and the leukocytes were prepared as previously described.[18] The antibodies, which were obtained from BD Pharmingen unless otherwise stated, were CD45.1 (Horizon V450), CD45.2-biotin, streptavidin (Pacific Orange, Invitrogen), CD19-phycoerythrin (PE), CD3–peridinin chlorophyll protein (PerCP)–cyanine 5.5 (Cy5.5), NK1.1-allophycocyanin (APC), CD4 (Alexa Fluor 700), CD8-APC-Cy7, F4/80-PE (AbD Serotec), B220-PerCP, CD11c-PE-Cy7, CD11b-APC, Gr1-APC-Cy7, Des-biotin (in house), CD69-PE, CD25-APC, CD44-APC-Cy7 (Biolegend), and CD8 (Pacific Blue). Cells were stained with 4′,6-diamidino-2-phenylindole (Invitrogen) before their acquisition on an LSR-II flow cytometer (BD Biosciences) and were analyzed with FlowJo 9.4.11 (TreeStar, Inc., Ashland, OR).

Gating Strategy for Leukocyte Counting and Identification

Total leukocyte counts were calculated for each sample by flow cytometry with Sphero AccuCount particles as a reference population. Two panels of antibodies were used for the analysis. For panel 1, 4′,6-diamidino-2-phenylindole–negative events were analyzed for lymphocytes: NK (CD3NK1.1+) and NKT cells (CD3+NK1.1+) were first distinguished on the basis of the expression of CD3 and NK1.1. B cells were identified as CD19+CD3NK1.1 cells. T cells were further subgated into CD4+ and CD8+ populations. Panel 2 was used to identify myeloid cells after the exclusion of CD3+ and B220+ cells. Eosinophils (EOSs; CD3B220side scatter (SSC)-Ahigh), macrophages (CD3B220SSC-AlowF4/80+CD11bint), and conventional dendritic cells (cDCs; CD3B220SSC-AlowCD11chiCD11bhi) were distinguished. F4/80+ monocytes were further subgated into Gr1+ inflammatory monocytes and F4/80int Gr1 patrolling monocytes. F4/80Gr1+ cells were identified as neutrophils.

Calculation of the Number of Cell Precursors Using CFSE

The number of cells in each cell division CFSE peak was corrected for the increase due to the division itself via the summation of the corrected numbers in each peak with the following formula[21]:

display math(1)

where ni is the number of cells in the ith division peak.

Histology and Staining of Liver Sections

Frozen sections from the spleen or LNs collected 5 hours after transplantation were stained with biotin-conjugated anti–H-2Kb (BD Biosciences), fluorescein isothiocyanate–conjugated anti-CD4 (BD Biosciences), and Alexa Fluor 555–conjugated streptavidin (Invitrogen).

RESULTS

Rapid Migration of Donor PLs From the Liver to Lymphoid Tissues After Transplantation

To characterize PL migration from the liver after transplantation in the absence of alloreactivity, we first transplanted syngeneic livers from CD45.2+ B10.BR donors into CD45.1+ B10.BR recipients. CFSE-labeled H-2Kb–specific Des-transgenic T cells (also CD45.2+) were adoptively transferred into recipient mice 1 day before transplantation (Supporting Fig. 1). This experimental protocol allowed us to easily distinguish donor PLs (CFSECD45.2+), directly graft-reactive CD8 Des T cells (CFSE+CD45.2+), and recipient leukocytes (CFSECD45.2) by flow cytometry. Most PLs exited the liver and were rapidly replaced by recipient leukocytes as early as 1.5 hours after transplantation (Fig. 1). PLs were found in the blood at 1.5 hours and in the seeded spleen and LNs 1.5 to 24 hours after the completion of the transplant. The percentages of PLs found in the spleen and LNs at 24 and 48 hours were similar (Fig. 1). There was no statistically significant difference between the numbers of donor PLs detected in the recipient LNs and spleen after syngeneic transplants and allogeneic transplants (data not shown). These results demonstrated that PLs migrated very rapidly into recipient lymphoid tissues after liver transplantation, irrespective of host MHC.

Figure 1.

Rapid exit of donor PLs from the liver and migration to lymphoid organs after syngeneic transplantation. Ly5.2+ B10.BR livers were transplanted into Ly5.1+ B10.BR recipients. Livers, spleens, LNs, and blood were harvested at 1.5, 24, and 48 hours from syngeneic transplant recipients. Lymphocytes were stained and analyzed for CD45.1 and CD45.2 expression after transgenic CFSE+ Des T cells were gated out. Recip indicates recipient leukocytes.

Leukocyte Composition in the Donor Liver Before Transplantation

Before analysing the migration of donor leukocytes after the transplantation of a liver allograft, it was important to characterize intrahepatic leukocytes present in the donor organ because this would determine the PL subsets migrating out of the graft after transplantation. Flow cytometry analysis demonstrated that the liver contained 1.7 ± 0.4 × 106 total leukocytes (mean  ±  standard error of the mean; n = 9), and this was less than the number in the LNs (2.8 ± 0.1 × 107) and the spleen (3.5 ± 0.4 × 107; Fig. 2A). The blood contained 7.9 ± 1.4 × 106 cells per 2 mL (estimated volume of blood per mouse; Supporting Fig. 2). The numbers of each leukocyte subset analyzed in various organs are shown in Fig. 2A. NK and NKT cells were enriched within the liver in comparison with the other organs analyzed (Fig. 2A), and this was consistent with previous reports.[22, 23] This composition was unique to the liver because the LNs contained mostly T and B cells, and the spleen contained predominantly B cells with a smaller proportion of T cells. In comparison with the LNs, the spleen also contained more myeloid cells, including large proportions of macrophages and cDCs, populations that are rarely detected in the blood, LNs, and liver (Fig. 2A and Supporting Fig. 3).

Figure 2.

Migration of different liver leukocyte and myeloid cell subsets in the blood. (A) Normal distributions and numbers of leukocyte subsets in the livers, blood, LNs, and spleens of B10.BR mice before transplantation. Leukocytes are represented in the left panels, whereas myeloid subsets are shown in the middle and right panels. To better demonstrate the distributions of the different myeloid subsets, the y axes of the middle panels have been magnified and are shown in the right panels. Counts were collected from 9 separate experiments. Donor PLs in (B) liver grafts and (C) blood were analyzed 24 and 48 hours after the grafting of livers from B10 or B6 donors into Ly5.1+ B10.BR recipients. The absolute number of the various PL subsets in each compartment was quantified as described. Bars show the means of 3 samples. A statistical analysis performed via an analysis of variance followed by Bonferroni post hoc tests showed significant differences between groups (P < 0.05). Asterisks show statistically significant differences between groups. Mye, Mac, Mono, and Neutro indicate myeloid cells, macrophages, monocytes, and neutrophils, respectively.

Specific Retention and/or Preferential Homing of Donor NKT Cells to the Liver Graft

The PL composition was then quantified in the liver grafts, lymphoid tissues, and blood of recipient animals after syngeneic and allogeneic transplantation. Donor PL counts 24 and 48 hours after transplantation are shown in the same graphs to highlight the differences in their fates. Figure 2B shows that with the exception of NKT cells, most PL subsets had exited the graft by 24 hours. Donor NKT cells represented more than 80% of intrahepatic donor PLs at 24 and 48 hours (Supporting Fig. 3). This was significantly increased in comparison with the NKT representation (30%) before transplantation (Fig. 2A and Supporting Fig. 3), and this suggested that donor NKT cells either had not left the graft or, as their presence in the blood might suggest (Fig. 2C), had preferentially returned to the liver. Donor NK, NKT, and myeloid cells were also better represented in the blood in comparison with lymphocytes, and this suggested that these cell types might not have homed to lymphoid tissues as efficiently as T and B cells. Recirculating donor myeloid cells in the blood were largely Gr1 monocytes (not shown), which is a phenotype characteristic of patrolling monocytes. Interestingly, the number of donor NK cells in the blood 48 hours after transplantation was significantly higher for allogeneic recipients versus syngeneic recipients. However, because the total number of NK cells retrieved from allogeneic recipients did not exceed pretransplant numbers (not shown), it was unclear whether selective expansion after allogeneic transplantation had occurred.

Differential Migration Patterns of Donor PLs to Lymphoid Tissues Reflect Their Steady-State Distribution

To address whether PL subsets were distributed equally to the spleen and LNs, the total number and composition of PL subsets were analyzed in these compartments. Donor PLs migrating into the LNs and spleen 24 and 48 hours after transplantation were almost exclusively T and B cells (Fig. 3 and Supporting Fig. 3). However, the distribution of T and B cells in these tissues was different: in the LNs, more donor T cells than B cells were found, whereas in the spleen, donor B cells predominated (Fig. 3 and Supporting Fig. 3). In addition, although the LNs contained very few donor myeloid cells (Fig. 3A,B), the spleen contained a significant number of donor NK cells, NKT cells, and myeloid cells (composed largely of patrolling Gr1 monocytes; not shown). The differential distribution in the LNs and spleen reflected the original distribution of the cells in these compartments (Fig. 2A). Interestingly, a significant loss of donor T cells was observed at 48 hours in both the LNs and the spleens of allograft recipients (Fig. 3A), and this raises the possibility that donor lymphocytes were rejected between 24 and 48 hours. In contrast, NK and NKT cells persisted in the blood and liver at 48 hours (Fig. 2C), and this suggested that they did not undergo rejection as rapidly. There was a significant increase in donor B cell numbers in the recipient spleen 24 hours after allogeneic transplantation (Fig. 3A).

Figure 3.

Differential migration of donor PLs to recipient spleens and LNs. (A) Absolute numbers of various PL subsets in LNs and spleens 24 and 48 hours after the grafting of livers from B10 or B6 donors into Ly5.1+ B10.BR recipients. Bars show the means of 3 samples. A statistical analysis performed via an analysis of variance followed by Bonferroni post hoc tests showed significant differences between groups (P < 0.05). Asterisks show statistically significant differences between groups. (B) Proportions of the different PL subsets in the LNs and spleens of syngeneic and allogeneic transplant recipients 24 hours after transplantation. An overrepresentation of T cells and B cells in the LNs and spleens, respectively, is shown. Mye indicates myeloid cells.

Increased Recipient T and B Cell Infiltrates in Allograft Grafts Versus Syngeneic Grafts

After transplantation, recipient leukocytes infiltrated the liver and rapidly replaced donor PLs within 1.5 hours (Fig. 1), although the infiltrating subsets did not necessarily reach pretransplant numbers (Fig. 4). Recipient myeloid cells were significantly increased in the liver 24 and 48 hours after transplantation, regardless of whether donors were allogeneic or syngeneic, and this suggested that this was a nonspecific response to transplant surgery and associated reperfusion injury (Fig. 4). The myeloid subsets were predominantly neutrophils and Gr1+ inflammatory monocytes (not shown). The number of recipient T and B cells increased in the allografts and blood of allogeneic recipients between 24 and 48 hours (Fig. 4). However, only the increase in recipient B cells in the blood of allogeneic recipients reached statistical significance.

Figure 4.

Recipient cell subsets in different compartments 24 and 48 hours after the grafting of livers from B10 or B6 donors into Ly5.1+ B10.BR recipients. The absolute numbers of the various leukocyte subsets in each compartment were quantified. The bars show the means of 3 samples. Pretransplant values of leukocyte subsets were collected from 9 separate experiments. Asterisks show statistically significant differences between groups. Mye indicates myeloid cells.

These data indicate that the mobilization of some leukocyte subsets was associated with allograft-specific responses and that they could be distinguished from nonspecific responses.

Primary Activation of Transgenic Alloreactive CD8 T Cells in the Graft Followed by Rapid Deletion

To analyze the site of activation and the fate of CD8 T cells specific for donor H-2Kb via the direct allorecognition pathway, we assessed activation markers and cell division in reporter Des T cells isolated from recipient mice. Five hours after the receipt of an allogeneic donor liver, all Des T cells in the liver, spleen, blood, and LNs had up-regulated the early activation marker CD69, which was indicative of parallel activation in these compartments (Fig. 5A). Two hours after transplantation, CD69+ CD8 Des cells were found in the liver and blood and were mostly absent from lymphoid tissues (Fig. 5A), and this suggested that intrahepatic activation occurred independently of lymphoid tissues. Des T cell activation was a result of cognate recognition and not surgery because activation was never observed in recipients of syngeneic B10.BR livers (Fig. 5B). Des T cells responding to C57BL/6 donor livers also increased in size (in agreement with blast formation) and up-regulated CD25 and CD44 in all compartments by 24 hours (Fig. 5B). Together, these results suggest that naive CD8 T cells were directly activated simultaneously by PLs in lymphoid tissues and in the graft by liver resident cells and/or intrahepatic PLs that did not exit the graft. To visualize where allograft-reactive transgenic T cells interacted with donor cells and were activated in lymphoid tissues, Des T cells were labeled with CellTrace Violet before adoptive transfer into transplant recipients. Spleens and LNs were harvested 5 hours after the T cell transfer. The immunostaining of liver sections revealed Des-TCR cell migration almost exclusively into the T cell zone in both lymphoid tissues (Fig. 5C), with interactions between allograft-reactive transgenic T cells and donor H-2Kb+ PLs observed (Fig. 5C).

Figure 5.

Early activation of alloreactive reporter Des T cells by donor PLs followed by their rapid loss from analyzed organs. Des-RAG−/− T cells were adoptively transferred into B10.BR mice subjected to transplantation with a C57BL/6 or B10.BR liver. Livers, LNs, and spleens were harvested 2 or 5 hours after transplantation, and lymphocytes were recovered and analyzed. (A) Representative flow histograms of CD69 expression by Des-transgenic T cells in recipients of allogeneic C57BL/6 grafts (black lines) versus untreated B10.BR controls (shading in top panels) and recipients of syngeneic B10.BR grafts (shading in bottom panels). (B) Representative histograms of FSC, CD69, CD25, and CD44 expression (at 24 hours) and CFSE fluorescence intensity (at 48 hours) of CD8+Des+ T cells in the spleens, LNs, and livers of recipients that received allogeneic C57BL/6 grafts (black lines in right panels) or syngeneic B10.BR grafts (black lines in left panels) versus Des T cells transferred into nontransplanted B10.BR controls (shading in both sets of panels) within the same experiment. (C) LN cells (10 × 106) were isolated from Des-RAG−/− mice, labeled with CellTrace Violet (5 μM), and transferred into B10.BR recipients 24 hours before the grafting of C57BL/6 livers. Frozen sections from spleens (top panels) or LNs (bottom panel), collected 5 hours after transplantation, were stained for H-2Kb to label PLs (red) and CD4 (green) to reveal T cell zones. The bars represent 50 μm. Arrowheads indicate interactions between donor H-2Kb+ PLs (red) and CellTrace Violet+ Des T cells (blue). (D) Calculated numbers of Des T cell precursors retrieved from LNs, spleens, blood, and livers 24 and 48 hours after transplantation. Bars represent means and standard deviations for 4 mice per group. Statistically significant differences between groups (P < 0.05), found with an analysis of variance and Bonferroni post hoc tests, are indicated. Asterisks show statistically significant differences between groups.

By 48 hours, all Des T cells transferred into recipients of allografts had proliferated, as shown by CFSE dilution (bottom panels in Fig. 5B). Despite the similar extent of Des T cell proliferation observed in lymphoid tissues and livers, the total number of Des T cells retrieved from allograft recipients was significantly lower than that retrieved from recipients of syngeneic grafts or nontransplant controls, in which proliferation was not observed. To account for cell division, we calculated the number of precursors from which the daughter Des T cells at 48 hours were derived. These results indicated a profound loss of transferred Des T cells even before division (within the first 24 hours; Fig. 5D) and suggested early deletion or redistribution to other tissues.

DISCUSSION

Since the first report of the spontaneous acceptance of liver transplants by Calne et al. in 1969,[24] research on this phenomenon has been hampered by a lack of transgenic tools and reagents for phenotyping rat leukocytes, by the difficulty in transplanting mouse livers, and by the problems faced in tracking complex polyclonal responses, which have limited our progress in identifying how alloreactive T cells are tolerized. Using mouse strains with congenic markers in combination with multicolor flow cytometry, we have identified and tracked multiple PL subsets and monitored the fate of a reporter population of directly allograft-reactive CD8 T cells, and we have shown for the first time that (1) most liver PLs except for NKT cells exit the graft rapidly and migrate to recipient lymphoid organs as early as 1.5 hours after transplantation, (2) PL subsets are not distributed equally to the spleen and LNs but rather reflect steady-state leukocyte population frequencies in these organs, (3) donor-specific naive CD8 T cells are activated within both the graft and the lymphoid tissues within 2 hours but are deleted within 24 hours, and (4) recipient cell subsets infiltrating the liver graft differ in response to allogeneic and syngeneic transplants.

The early migration of PLs to recipient lymphoid tissues is thought to be a critical event in the spontaneous acceptance of liver allografts.[7] Most studies investigating the fate of alloreactive T cells have focused on lymphoid tissues, in which activation by PLs presenting the donor MHC on the surface of donor antigen-presenting cells (direct pathway) or by allograft-derived peptides associated with the recipient MHC (indirect pathway) occurs. The presence of donor PLs in the recipient draining LNs and spleen after transplantation has been shown to be associated with high levels of interleukin-2 and interferon γ messenger RNA production and with increased apoptosis, presumably of recipient T cells[10]; this suggests that PLs induce abortive activation or exhaustion of alloreactive T cells and reduce the repertoire that mediates rejection.[16] Although this is an attractive hypothesis, it remains unclear why liver PLs induce tolerance when PLs from kidney, skin, and heart grafts prime recipient T cells for rejection.[1] It has been hypothesized that the large number of PLs in the liver versus other organs could be responsible for this liver property. However, the inability of large numbers of donor splenocytes to induce long-term acceptance of fully mismatched heart and kidney grafts[1] argues against this possibility. Alternatively, it has been proposed that intrahepatic PLs might differ qualitatively from splenocytes and contain subsets more potent in inducing tolerance. Although this cannot be discounted, donor splenocytes are equivalent to purified donor liver leukocytes in their ability to reconstitute the acceptance of PL-depleted transplanted livers.[6] Finally, it is possible that the combination of PLs and liver parenchyma contributes to the induction of tolerance by liver allografts. Characterizing the type of PLs seeding lymphoid tissues and the site of recipient T cell activation was thus among the aims of this study.

We confirmed here the results from rat models, which showed that donor PLs exited the graft rapidly and that some migrated to the recipient LNs and spleen. However, by systematically analyzing all the main liver PL subsets, we demonstrated that all PL subsets did not migrate equally in the recipients. First, PL migration to the LNs and spleen was not equivalent because myeloid cells were hardly recovered from the LNs. Twenty-four hours after transplantation, the majority of PLs in the LNs were T and B cells, whereas in the spleen, B cells predominated, and they were followed by T cells, donor NK cells, and myeloid cells. This differential migration reflects the distribution of the various cellular subsets in these organs. It further suggests that alloreactive T cells might encounter different types of antigen-presenting cells at these 2 sites, and this could potentially lead to different outcomes. Antigen presentation by B cells has been shown to turn off naive T cells.[25, 26] In a fully MHC-mismatched kidney transplant model, the transfer of donor B cells, but not T cells, led to prolongation of graft survival, and this suggests that there are inherent differences in their ability to induce tolerance.[27] Interestingly, at 48 hours, there was a significant loss of donor T cells in the LNs and spleens of allograft recipients, and this suggested that they had been recirculated or rejected. We could not detect any significant differences between the numbers of recipient CD4 and CD8 T cells in allogeneic and syngeneic transplant recipients (data not shown). This was expected because of the low percentage of alloreactive cells in the recipient polyclonal T cell population (1%-10%). Deletion by the apoptosis of recipient allogeneic T cells after liver transplantation has been suggested in other studies,[8, 9] although it was not clear whether apoptosis occurred in recipient T cell subsets.

The ability of myeloid cells to home preferentially to the spleen suggests that liver myeloid cells, particularly dendritic cells (DCs), can exert their immunomodulatory properties (T helper 2/tolerogenic induction[28] or veto function[29]) at this site. The role of liver DCs in liver transplantation is controversial: when they were infused, purified liver DCs prolonged allograft survival, but livers containing FMS-like tyrosine kinase 3 ligand–mobilized DCs were rapidly rejected.[30, 31] Because DCs mature upon culture and recirculation, infusion or mobilization approaches might not recapitulate the endogenous response. There are also different subsets of DCs that might exert different functions in transplantation tolerance. Higher rates of circulating plasmacytoid DCs with respect to myeloid or conventional ones have indeed been reported in operationally tolerant pediatric liver allograft recipients and in patients on low-dose immunosuppressive therapy undergoing prospective drug weaning, and they have been correlated with increased frequencies of CD4+CD25hi forkhead box P3+ regulatory T cells (Tregs).[32] Plasmacytoid DCs are also known to promote tolerance in several animal models of transplantation (reviewed by Matta et al.[33]). Donor DCs have been detected in recipients after liver transplantation in rats and humans,[34, 35] although their role in the induction of liver transplant tolerance has never been directly demonstrated. Unfortunately, we did not include the plasmacytoid DC markers mPDCA1 and B220 in our panels to distinguish plasmacytoid DCs. More in-depth experiments that include a proper enzymatic digestion protocol are thus required to address the phenotype and migration of the different DC subsets. Our study also did not address whether donor CD4 or CD8 Tregs migrated differently in the different compartments after liver transplantation and this resulted in a different effector cell/Treg ratio. Although Tregs might play a role in liver transplantation, we excluded an analysis of this cell subset from our panels because we were investigating events of tolerance induction occurring within the first 48 hours after transplantation. Tregs take several days to differentiate from naive T cells, and their involvement in the spontaneous acceptance of liver transplants is in long-term maintenance rather than the induction of tolerance (reviewed by Benseler et al.[1]).

The second important finding of this study was the enrichment of donor NK and NKT cells in the liver and blood after transplantation, and this suggested that these cells failed to migrate to lymphoid organs and/or homed preferentially back to the liver. It is also possible that some donor NKT cells never left the graft; the residence of NKT cells in the liver would be consistent with recent observations in parabiotic mice in which these cells failed to recirculate despite their location in the sinusoids and hence direct access to the circulation.[36]

In response to liver transplantation, the number of nontransgenic recipient cells also increased in lymphoid tissues. An increase in recipient myeloid cells (mainly neutrophils and inflammatory monocytes) was observed after both allogeneic and syngeneic transplantation, and this indicated that non–antigen-specific responses related to the inflammation and reperfusion injury had occurred.

A significant advantage of the mouse liver transplant model over previous studies was that it allowed the fate of a reporter population of naive donor-specific transgenic CD8 T cells to be monitored after transplantation, whereas studies analyzing the polyclonal T cell response in rats did not reveal significant differences in infiltrates between tolerant and rejecting livers and did not distinguish between naive and activated/memory T cells. We were able to demonstrate that naive CD8 T cell activation occurred within hours in both recipient lymphoid tissues and hepatic allografts. Activation in the LNs and spleen seems to have occurred in the T cell zones, where the transgenic T cells were predominantly located and contacted PLs (Fig. 5C). In the liver, primary activation occurred in the sinusoids (data not shown), as we have previously reported.[18] We also found that Des T cells started to be deleted within the first 24 hours, and most were deleted within 48 hours. This suggests that alloreactive T cell deletion occurs at a significantly earlier time point than the one described by Steger et al.,[16] who reported the deletion of a large cohort of donor-specific naive CD8 T cells following T cell proliferation by day 5 after mouse liver transplantation. This suggests that deletion occurs even before T cells start to divide. This type of rapid allo-T cell deletion is consistent with recent studies performed in a nontransplant setting, which showed that a large proportion of naive CD8 T cells activated within livers expressing high levels of antigen underwent deletion by a nonapoptotic process because of invasion and degradation inside hepatocytes within hours,[37] and the residual cells were further deleted because of subsequent Bim-dependent apoptosis.[38] This is in contrast to a previous study of mouse liver transplantation, which reported that naive CD8 T cells developed cytotoxic function 7 days after intrahepatic activation in the liver graft.[39] Importantly, however, in that particular study, the cognate antigen was expressed only transiently in contradistinction to the continuous expression of donor antigen occurring in allografts, as reflected in the current work.

It is unclear from the current study which of the 3 compartments supporting naive CD8 T cell activation is also essential for their early deletion. We recently observed Des T cell deletion in splenectomized recipients (S. S. Tay 2011, unpublished data), and this suggests that the spleen is not essential for this process. The nature of the antigen-presenting cells leading to the activation and deletion of Des T cells is as yet undefined, but because of the preferential migration patterns of the different PL subsets, activation in the separate compartments could involve distinct antigen-presenting cells. In the liver, the ability to tolerize CD8 T cell responses has largely been attributed to hepatocytes expressing cognate antigen. However, the unique hepatic architecture and slow blood flow would also permit prolonged interactions in the sinusoids with Kupffer cells and liver sinusoidal endothelial cells and, as discussed previously, NK and NKT cells, which could act as antigen-presenting cells. Kupffer cells and liver sinusoidal endothelial cells have reported abilities to tolerize CD8 and CD4 T cells and, in the transplant setting,[40] might play key roles in modulating both subsets.

In summary, we have developed a mouse model of liver transplantation to elucidate a question that has fascinated transplant immunologists for 4 decades. This new approach has yielded new insights into the general mechanisms leading to the spontaneous acceptance of liver grafts and will be further developed in the future to properly address the sequence of events leading to this intriguing phenomenon.

ACKNOWLEDGMENTS

The authors thank the Centenary Institute Animal Facility and the Advanced Cytometry Facility for their technical support and Ben Roediger for his help with the flow cytometry experimental design and for helpful discussions.

Ancillary

Advertisement