Two immunogenic passenger dendritic cell subsets in the rat liver have distinct trafficking patterns and radiosensitivities

Authors


  • Potential conflict of interest: Nothing to report.

  • The authors thank Satoshi Ueha and Yoshikazu Sado for their kind donation of polyclonal antibodies.

Abstract

The aim of this study was to investigate the trafficking patterns, radiation sensitivities, and functions of conventional dendritic cell (DC) subsets in the rat liver in an allotransplantation setting. We examined DCs in the liver, hepatic lymph, and graft tissues and recipient secondary lymphoid organs after liver transplantation from rats treated or untreated by sublethal irradiation. We identified two distinct immunogenic DC subsets. One was a previously reported population that underwent blood-borne migration to the recipient's secondary lymphoid organs, inducing systemic CD8+ T-cell responses; these DCs are a radiosensitive class II major histocompatibility complex (MHCII)+CD103+CD172a+CD11bCD86+ subset. Another was a relatively radioresistant MHCII+CD103+CD172a+CD11b+CD86+ subset that steadily appeared in the hepatic lymph. After transplantation, the second subset migrated to the parathymic lymph nodes (LNs), regional peritoneal cavity nodes, or persisted in the graft. Irradiation completely eliminated the migration and immunogenicity of the first subset, but only partly suppressed the migration of the second subset and the CD8+ T-cell response in the parathymic LNs. The grafts were acutely rejected, and intragraft CD8+ T-cell and FoxP3+ regulatory T-cell responses were unchanged. The radioresistant second subset up-regulated CD25 and had high allostimulating activity in the mixed leukocyte reaction, suggesting that this subset induced CD8+ T-cell responses in the parathymic LNs and in the graft by the direct allorecognition pathway, leading to the rejection. Conclusion: Conventional rat liver DCs contain at least two distinct immunogenic passenger subsets: a radiosensitive blood-borne migrant and a relatively radioresistant lymph-borne migrant. LNs draining the peritoneal cavity should be recognized as a major site of the intrahost T-cell response by the lymph-borne migrant. This study provides key insights into liver graft rejection and highlights the clinical implications of immunogenic DC subsets. (HEPATOLOGY 2012)

The trafficking of dendritic cell (DC) subsets is important because appropriate site-specific antigen delivery is essential for immune regulation, both in the healthy state and in various immune-mediated diseases.1 However, information regarding trafficking patterns of the specific DC subsets remains incomplete.2-5 We demonstrated previously that after rat liver transplantation (LT), a small, but notable number of graft DCs systemically migrate to the recipient's secondary lymphoid organs through the bloodstream; these cells form clusters with the recipient's T cells and induce diffuse CD8+ T-cell responses that may promote graft rejection.6 T-cell proliferative responses originate within the clusters, which thus represent sites for the intrahost direct allorecognition pathway in which migrated donor DCs sensitize the recipient's T cells through cognate interaction within the cluster.7 Because these DCs actively transmigrate through the blood-vessel wall, whereas lymph DCs at the antigen-transporting stage do not,8 they presumably constitute a distinct DC subset. Although these cells are class II MHC antigen positive (MHCII+) and either CD11c+ or CD103+, other phenotypes and radiosensitivities have not been examined.6

The hepatic lymph contains a constant large efflux of liver DCs9, 10 and lymphocytes,11 even in the absence of invading pathogens. In healthy rat hepatic lymph, this DC output is ∼1 × 106 cells/overnight collection.10 In steady-state rat intestinal and hepatic lymph, DCs are mostly MHCIIhigh αE2 integrin (CD103)high3 and include three distinct subsets (i.e., CD172ahigh, CD172aint, and CD172alow) at various ratios in both lymphs.12 Notably, CD172a is another term for signal-regulatory protein-alpha (SIRP-α). However, the role of hepatic lymph DCs and the role of specific subsets in transplantation immunity remain unknown.

At steady state, hepatic lymph DCs usually migrate to regional liver lymph nodes (LNs), which are the celiac LNs in rats and hepatic LNs in humans.9 In LT, graft lymph ducts are unavoidably injured during surgery and all of the donor DCs entering the hepatic lymph leak into the peritoneal cavity. In rats, the parathymic LNs and posterior mediastinal LNs drain the peritoneal cavity through the diaphragmatic lymphatics,13-15 and peritoneal exudate cells migrate to these LNs in acute gastrointestinal inflammation.16 We define these LNs as parathymic LNs. We suspected that many donor DCs in the peritoneal cavity might further migrate to these LNs. There were relatively higher proliferative responses in the parathymic LNs than in other secondary lymphoid organs,6 with extensive cluster formation between donor MHCII+ cells and recipient proliferating cells after rat LT (Ueta, unpublished observation). This finding suggests that LNs that drain the peritoneal cavity comprise the special secondary lymphoid organ where donor DCs accumulate not only through the blood, but also through the lymph, resulting in the highest allostimulation among the recipient lymphoid organs. However, this hypothesis awaits experimental validation.

Concerning the sessile population, conventional DCs in the liver mainly localize to the portal area, with many fewer conventional DCs, within the hepatic lobule at steady state in rats, mice, and humans.9 Sun et al.17 reported that some donor MHCII+ cells with dendritic morphology persist in the graft liver after preoperative lethal irradiation of the donor rat, suggesting the presence of radioresistant DCs. Surprisingly, the irradiated livers were acutely rejected when transplanted to allogeneic recipients, even when there were fewer DCs. The role of these remaining DCs, as well as other factors that are important in the rejection process, merits investigation.

Here, we show that rat liver conventional DCs contain at least two immunogenic subsets that have distinct trafficking patterns and radiosensitivities. We also found a novel migration pathway of passenger DCs that involves lymph-borne migration to the peritoneal cavity and then to regional LNs through diaphragmatic lymphatics. Even after DC depletion by graft irradiation, the remaining radioresistant DC subset induced an intense alloresponse in vitro that was probably responsible for the rejection.

Abbreviations

BrdU, 5-bromo-2'-deoxyuridine; DC, dendritic cell; FACS, fluorescence-activated cell sorting; IL-2, interleukin-2; Irr(+)/Irr(−), irradiated/nonirradiated; LN, lymph node; LT, liver transplantation; MHCI, class I major histocompatibility complex; MHCII, class II major histocompatibility complex; PALS, periarterial lymphoid sheath; SIRP-α, signal-regulatory protein-alpha.

Materials and Methods

Additional Materials and Methods information can be found in the online Supporting Information (Supporting Materials).

Whole-Body Irradiation.

Half of the donor DA rats received total-body sublethal split X-irradiation (filter: 0.5 mm aluminum + 0.1 mm copper, Hitachi MBR-1505R; Hitahci, Tokyo Japan). Animals were dosed twice at 3 Gy with a 4-hour intermission18 5 days before LT. We used a 5-day interval between irradiation and LT after conducting a preliminary kinetic study to determine how long it took to achieve a significant reduction of donor MHCII+ or CD103+ cells in the graft (not shown). Recipient Lewis rats that received LT with or without irradiation were designated as the irradiated [Irr(+)] or nonirradiated [Irr(−)] groups.

Immunohistological Analysis.

To investigate the recipient's immune response, cryosections were triple-immunostained for CD8β, FoxP3, recipient class I MHC (MHCI) (I169.1+)(19), or donor MHCII (alkaline phosphatase-blue), type IV collagen (peroxidase brown), and 5-bromo-2′-deoxyuridine (BrdU) (alkaline phosphatase red).

Graft tissues were divided into three anatomical compartments: the sinusoidal; portal; and hepatic vein areas.2 Because the loss of the sinusoidal area should be parallel to the loss of hepatocytes (i.e., liver function), we assumed that compression of the sinusoidal area by the expanding portal and hepatic vein areas reflected the degree of rejection of the graft liver. Accordingly, the percentage of the sinusoidal area relative to the total surface area of stained sections was estimated by image analysis. The number of donor MHCII+ cells that formed clusters with more than two recipient BrdU+ cells in the Irr(+) group was counted in 20 continuous fields of the portal and hepatic vein areas in each rat using the ×20 objective. Similarly, the number of BrdU+CD8β+ cells, BrdU+FoxP3+ cells, or total FoxP3+ cells in the portal area were counted. All counting was performed in a blinded fashion. The phenotypes of donor MHCII+ cluster-forming cells were analyzed, as reported previously,6 in fresh serial 2-μm cryosections of the parathymic LNs and graft liver.

Results

DC Subsets in the Liver at Steady State and After Irradiation.

DCs in the liver nonparenchymal cells and hepatic lymph were defined as the MHCIIhighCD103high population, based on fluorescence-activated cell-sorting (FACS) data (Fig. 1A,D) and in accord with earlier studies.3, 12 Those DCs in the healthy rat liver could be subdivided into three phenotypically different groups: CD172a+CD11b+ DCs (∼44%); CD172a+CD11b DCs (∼20%); and CD172aCD11b+ DCs (∼28%) (Fig. 1A,B). Five days after sublethal irradiation, the total number of liver DCs decreased from ∼2.8 × 105 to ∼5.1 × 104 (Fig. 1C), and the percentages of the three subsets changed to ∼64%, ∼28%, and ∼5%, respectively (Fig. 1A,B). Notably, the CD172aCD11b+ subset was radiosensitive and decreased dramatically after irradiation, but ∼25% of the other two subsets remained.

Figure 1.

Cell yields and phenotypes of DC subsets in the rat liver and lymph at steady state (0 Gy) or after sublethal irradiation (3 + 3 Gy). Low-density liver nonparenchymal cell (NPC) fractions or whole hepatic lymph cells were used for FACS analysis, with MHCIIhighCD103high cells gated as the DC fraction (A and D). These DCs were subdivided into CD172a+CD11b+, CD172a+CD11b, and CD172aCD11b+ subsets in NPCs (A and B) and lymph (D and E), with the total yield shown in (C) and (F), respectively. The numbers in (A) and (D) represent percentages. Each panel shows representative data from three separate experiments.

DC Subsets in the Hepatic Lymph at Steady State and After Irradiation.

MHCIIhighCD103high DCs in the hepatic lymph could also be subdivided into three phenotypically different groups: CD172a+CD11b+ (∼80%); CD172a+CD11b (∼15%); and CD172aCD11b+ DCs (∼3%) (Fig. 1D,E). Five days after sublethal irradiation, the total number of lymph DCs decreased from ∼1.3 × 105 to ∼1.4 × 104 (Fig. 1F), and the percentages of the three subsets changed to ∼90%, ∼8%, and ∼1%, respectively (Fig. 1D,E). Thus, among lymph DCs, the CD172a+CD11b+ subset was relatively radioresistant, with ∼13% remaining after irradiation, whereas the other two subsets were very radiosensitive and were almost abolished.

The CD172a+CD11b DC Subset Migrates Through Blood to the Recipient's Secondary Lymphoid Organs After LT.

As in our previous study,6 donor MHCII+ and donor MHCI+ cells readily migrated to the recipient's secondary lymphoid organs (i.e., the spleen, skin LNs, and Peyer's patches), and donor MHCII+ DCs formed clusters with recipient BrdU+ cells in the T-cell areas on days 1-3 after LT in the Irr(−) group (Supporting Fig. 1A,C). Because Peyer's patches do not possess afferent lymphatics, DC entry should be through the blood, presumably through the high endothelial venules.6 FACS analysis of skin LNs (Fig. 2A) and Peyer's patches (not shown) revealed that more than 90% of the migrated donor MHCIIhighCD103high DCs were CD172a+CD11b. The exception was the parathymic LNs, in which both CD172a+CD11b and CD172a+CD11b+ donor DCs were found (Fig. 2A); the CD172a+CD11b+ subset constituted ∼70% of all DCs.

Figure 2.

Migration of donor DC subsets to the recipient's secondary lymphoid organs after LT. (A) Whereas only the CD172a+CD11b subset (c) was detected in skin LNs, both CD172a+CD11b+ (b) and CD172a+CD11b (a) subsets were detected in the parathymic LNs (PTN: upper part of panel A). In the Irr(+) group on day 2 after LT, only the CD172a+CD11b+ (b') subset remained in the parathymic LNs, but the CD172a+ CD11b subset (a') disappeared in both LNs (middle part of panel A). n.d., not detectable. The lower part of (A) also shows the absence of donor MHCI+MHCII+ large cells in the skin LNs (d). (B) The CD172a+CD11b+ subset expressed high levels of CD25 in both the Irr(+) and Irr(−) groups. However, the CD172a+CD11b subset expressed a lower level of CD25. (A and B) Representative data from two separate experiments. (C and D) Triple immunostaining of serial sections of the PTN in the Irr(+) group on day 3 after LT showed that donor MHCII+ cells (blue cell, arrowhead) clustering with BrdU+ cells (red) were CD103+, CD11c+, CD25+, or CD86+.

Donor MHCII+ Cells in the Hepatic Lymph Migrate Through Lymph to the Recipient's Parathymic LNs After LT.

In the Irr(−) group, donor MHCII+ and MHCI+ cells appeared in the peritoneal cavity on days 1-3 after LT. There were comparable numbers of donor cells in the Irr(+) group (Fig. 3A). Furthermore, donor MHCII+ DC-like cells were found both in the omentum and diaphragm spreads and in the subcapsular sinus of the parathymic LNs not only in the Irr(−) group (not shown), but also in the Irr(+) group (Fig. 3B-D). Because the subcapsular sinus is the portal for afferent lymph entry,20 this result confirms that donor MHCII+ cells in the injured hepatic lymphatics migrate to the parathymic LNs through the peritoneal cavity and diaphragmatic lymphatics. The donor MHCII+ cells in the subcapsular sinus in the Irr(+) group might represent a radioresistant lymph DC subset, because irradiation eliminates lymphocytes, including B cells, which are constitutively MHCII+.17

Figure 3.

Appearance of donor MHCI+ and MHCII+ cells in the peritoneal cavity and parathymic lymph nodes (PTN) in both the Irr(+) and Irr(–) groups. (A) Number of donor MHCI+ or MHCII+ cells/rat 1-3 days after LT. Mean ± standard deviation (n = 3). (B-D) Presence of donor MHCII+ DC-like cells (blue) in the recipient omentum spread (B) and diaphragm spread (C) on day 2 after LT and in the subcapsular sinus of the PTN on day 3 after LT in the Irr(+) group. (E) Schematic showing of the migration of CD172a+CD11b+ and CD172a+CD11b DC subsets in the Irr(−) and Irr(+) groups.

Blood-borne, but Not Lymph-Borne, Migration of DC Subsets Is Inhibited in the Irr(+) Group.

Donor MHCII+ and MHCI+ cells migrating to the host secondary lymphoid organs were almost completely abolished after irradiation (Fig. 2A and Supporting Fig. 1B,D), demonstrating that the blood-borne migrating CD172a+CD11b DC subset and the donor lymphocytes were radiosensitive. The exception was the parathymic LNs, where donor MHCII+ DC-like cells remained (Fig. 2C,D and Supporting Fig. 1F). These cells were mostly CD172a+CD11b+ (Fig. 2A), confirming that these cells were the radioresistant DC subset that migrated through the lymphatics through the peritoneal cavity. This CD172a+CD11b+ population expressed high levels of CD25 (interleukin-2 [IL-2] receptor alpha) (Fig. 2B) in both the Irr(+) and Irr(−) groups. Additionally, abdominal LNs (i.e., the celiac and mesenteric LNs) contained very few donor MHCII+ DC-like cells with a weak T-cell response, suggesting that these cells were also the CD172a+CD11b+ DC subset that migrated from the peritoneal cavity (not shown).

Intrahost T-cell Response Is Restricted to the Parathymic LNs in the Irr(+) Group.

In the Irr(−) group, a proliferative response in the T-cell areas of the recipient's secondary lymphoid organs was observed, as reported previously.6 As expected, the proliferative response in the T-cell area of the parathymic LNs was considerably higher than that in other secondary lymphoid organs tested (Fig. 4A,C-E and Supporting Fig. 1E). The CD8+ T-cell proliferative response was clear in splenic periarterial lymphoid sheath (PALS) (Fig. 4B and Supporting Fig. 2A) and even more intense in the T-cell area of the parathymic LNs (Fig. 4F and Supporting Fig. 2C).

Figure 4.

Proliferative response in the T-cell area of the recipient's secondary lymphoid organs, determined by the number of total BrdU+ cells/mm2 (A, C, D, and E), and the CD8+ T-cell response in the T-cell area of the spleen (PALS; B) and parathymic lymph nodes (PTN; F), determined by the number of BrdU+CD8β+ cells/mm2. PP, Peyer's patches. Mean ± standard deviation (n = 3). *P < 0.05 versus the Irr(−) group.

In contrast, in the Irr(+) group, the T-cell proliferative response in the splenic PALS and T-cell areas of the cervical LNs and Peyer's patches was significantly suppressed (Fig. 4A,C,D). The CD8+ T-cell response was also significantly suppressed in the splenic PALS (Fig. 4B and Supporting Fig. 2B). These results indicate that suppression of the T-cell response was the result of impairment of the direct allorecognition pathway through inhibition of blood-borne migration of the CD172a+CD11b subset. One exception to this was observed in the parathymic LNs. Here, there was a CD8+ T-cell proliferative response that became comparable with the response in the Irr(−) group by day 3 (Fig. 4F and Supporting Fig. 2D). As described above, the T-cell area in the parathymic LNs contained a small, but notable, number of donor MHCII+ DC-like cells that clustered with BrdU+ cells (Supporting Fig. 1F). These cells appeared on day 2, peaked on day 3, and disappeared on day 4 after LT (not shown). Serial immunostaining showed that ∼43% of these cells were CD103+ and that ∼63% were CD11c+ (Table 1; Fig. 2C); some were also CD25+ and/or CD86+ (Fig. 2D).

Table 1. Phenotype of CFCs
Area Irr(−)Irr(+)
  1. Proportion of CD103+ or CD11c+ donor MHCII+ CFC/total donor MHCII+ CFC that contains two or more BrdU+ cells. More than 36 and 29 CFCs/rat are counted in recipient parathymic LNs at day 3 and portal and hepatic vein area at day 2 after LT, respectively. Mean ± standard deviation.

  2. Abbreviation: CFCs, cluster-forming cells.

Parathymic LN,Mean % CD103+ CFC20.0, 20.942.9 ± 6.1 (n = 3)
T-cell areas, 3d after LTxMean % CD11c+ CFC29.5, 35.163.4 ± 26.8 (n = 3)
Portal and hepatic-vein areas, 2d after LTxMean % CD103+ CFCnot tested64.7 ± 1.3
Portal and hepatic-vein area, 2d after LTxMean % CD11c+ CFCnot tested82.3 ± 8.1

Comparable Mean Survival Time and Intragraft Response Between the Irr(+) and Irr(−) Groups.

The mean survival time of the Irr(−) and Irr(+) groups were 10.3 ± 1.6 (n = 9) and 8.8 ± 1.0 days (n = 4) after LT, respectively (Fig. 5A), indicating that irradiation enhanced rejection slightly, but not significantly. The ratio of the sinusoidal area to the total surface area was significantly higher in the Irr(+) group than in the Irr(−) group's on day 5, but became comparable by day 7 (Fig. 5B). The CD8+ T-cell responses were comparable in the Irr(+) and Irr(−) groups, as shown by the kinetics of BrdU+CD8β+ cell numbers in the graft portal and sinusoidal areas (Fig. 5C-F).

Figure 5.

Survival rate and intragraft response. (A) Survival rate in the Irr(−) and Irr(+) groups after LT. (B) Percentage of the sinusoidal area relative to total surface area on days 1-7 after LT. *P < 0.05 versus the Irr(−) group. (C and D) Triple immunostaining of the portal area in the Irr(−) (C) and Irr(+) groups (D) on day 3 after LT: CD8β (blue), type IV collagen (brown), and BrdU (red). (E and F) CD8+ T-cell response in the portal area (E) and sinusoidal area (F). The number of BrdU+CD8β+ cells in 20 continuous fields of each area/rat was counted using the ×20 objective. Mean ± standard deviation (n = 3).

Radioresistant Sessile DCs Persist in the Grafts the Irr(+) Group.

In both the Irr(−) and Irr(+) groups, donor MHCII+ DC-like cells were observed in clusters with BrdU+ cells that were found in the graft portal and hepatic vein areas on days 2-4 after LT (Fig. 6A). FACS analysis to detect nonparenchymal cells on day 3 after LT showed that the sessile donor DCs were mainly in the CD172a+CD11b+ population in both groups and that they expressed similar levels of CD25 and CD86 (Fig. 6B-D). Immunostained serial sections showed that of these donor MHCII+ cluster-forming cells, ∼65% were CD103+ and ∼82% were CD11c+ (Table 1; Fig. 6E,F). Furthermore, some also coexpressed CD86 (Fig. 6F).

Figure 6.

Presence of sessile DCs and their cluster formation within the graft after LT. (A) Kinetics of donor MHCII+ cluster-forming cells (CFC) in the portal and hepatic vein areas. (B-D) Persistence of donor MHCIIhighCD103high cells in the low-density liver nonparenchymal cell (NPC) fraction (B). Both groups had mostly CD172a+CD11b+ cells (C) with similar expression levels of CD25 and CD86 (D). (B) The numbers of red circles (open arrowheads), red squares (closed arrowheads), and dotted black circles indicate the percentage of donor MHCII+ cells, MHCIIhighCD103high cells, and recipient MHCII+ cells, respectively. (C) Numbers indicate cell percentages. (D) Numbers indicate the mean fluorescent intensities (MFI) with the isotype control in parenthesis. (B-D) Representative data from three separate experiments. (E and F) Triple immunostaining of serial sections of the graft portal areas show that donor MHCII+ cells (blue cells, arrowheads) cluster with BrdU+ cells (red) and are CD103+, CD11c+, or CD86+.

In Vitro Allosimulating Activity in Mixed Leukocyte Reaction.

Cytosmears of FACS-sorted liver DC subsets showed their DC cytology (Fig. 7A). The positive stimulator control of the donor splenic DCs induced a dose-dependent proliferation of responder T cells. The CD172a+CD11b+ DCs (3 × 103/well) that were isolated from the donor liver with or without irradiation and from the irradiated donor hepatic lymph induced high proliferation comparable to the control splenic DCs (3 × 103/well) (Fig. 7B). In contrast, CD172aCD11b+ DCs isolated from the nonirradiated donor liver (3 × 103/well) showed a lower stimulation index (Fig. 7B). The CD172a+CD11b+ DCs formed huge clusters in vitro that were larger than clusters formed by the CD172aCD11b+ DCs (Fig. 7C).

Figure 7.

Mixed leukocyte reaction using isolated donor DC subsets and recipient T cells. (A) Cytosmears of three DC subsets separated from the low-density fraction of healthy liver nonparenchymal cells using a FACSAria (May-Giemsa stain). (B) Isolated donor DC fractions (3 × 103 cells/well) were cultured with (+) or without (−) recipient T cells (Resp. T; 1 × 105 cells/well) for 4 days, and the proliferative response was estimated as relative light units/sec (rlu/s). As a positive stimulation control, we used varying amounts of donor splenic DCs (allo Spl. DC). The CD172a+CD11b+ DCs isolated from the donor liver after irradiation (3 + 3 Gy) or without irradiation (0 Gy) and irradiated hepatic lymph DCs induced proliferation comparable to the control splenic DCs. Representative data are shown from three separate experiments. Data are reported as mean ± standard deviation (n = 3). (C) Photomicrograph of culture wells showing the cluster formation in vitro.

The Regulatory T-Cell Responses in the Spleen and Graft Is Not Suppressed in the Irr(+) Group.

The number of BrdU+FoxP3+ regulatory T cells was suppressed slightly on day 2 in both the spleen and graft portal areas in the Irr(+) group, compared to the Irr(−) group (Supporting Fig. 3A,B); however, suppression was not significantly different over the entire examination period. The total number of FoxP3+ cells in the portal areas was also comparable (Supporting Fig. 3C).

DNA Microarray Analysis.

The 35,129-element oligonucleotide microarray of graft tissues used to analyze the Irr(+) group identified 117 up-regulated and 79 down-regulated genes on day 2 and 95 up-regulated and 79 down-regulated genes on day 3 after LT, compared to the Irr(−) group. Among these, several genes were related to immune responses. Although we analyzed the relevant genes further by quantitative reverse-transcription polymerase chain reaction analysis and found a significant difference in the expression of macrophage inhibitory factor genes in the two groups, we could not confirm the readable presence of radiosensitive suppressive genes in the graft liver (Supporting Fig. 4). This discrepancy may be the result, in part, of the low expression of the listed genes, to sample variability, or to use of pooled samples for the DNA microarray.

Discussion

There were several novel findings in this study of the rat liver DC subsets. First, three DC subsets with distinct phenotypes were identified in rat liver nonparenchymal cells. The largest subset comprised CD172a+CD11b+ cells that were relatively radioresistant, compared to the other two subsets found in liver tissues and lymph. Second, the CD172a+CD11b DC subset migrated from the liver transplant through the blood to the recipient's secondary lymphoid organs. This migration was completely inhibited in the Irr(+) group, and the intrahost T-cell response was also suppressed, except in the parathymic LNs, where a radioresistant CD172a+CD11b+ DC subset migrated through the lymphatics and formed clusters with proliferating cells (shown schematically in Fig. 3E). Third, the intragraft CD8+ T-cell responses as well as the mean survival time were comparable between the Irr(+) and Irr(−) groups, and FoxP3+ regulatory T-cell responses in the spleen and graft were also similar between the two groups. Fourth, the radioresistant CD172a+CD11b+ DC subset persisted in the graft in the Irr(+) group and formed clusters with proliferating cells from the recipient. When isolated from the liver or hepatic lymph, this subset showed very strong allostimulation activity in the mixed leukocyte reaction.

Phenotype and Radiosensitivity of Liver and Hepatic Lymph DC Subsets.

The donor MHCIIhighCD103high cells were defined as the DC fraction. Although CD103low DCs are found in liver sections,9 we could not detect them in the low-density fraction of liver nonparenchymal cells by the method used here (Fig. 1A). We identified three subsets with distinct phenotypes in the liver tissues and lymph, as reported on previously in rat intestinal and hepatic lymph.12 The CD172a+CD11b+ subset was the largest subset, and these cells were relatively radioresistant: After irradiation, ∼25% of these cells persisted in the liver and ∼13 % persisted in the lymph. In contrast, the CD172a+CD11b and CD172aCD11b+ subsets were more radiosensitive and were almost eliminated in the hepatic lymph after irradiation. Yrlid et al.12 reported that the CD172a+CD11b subset was very small in untreated hepatic lymph. This might be the result of differences in FACS settings and especially in gating positive cells: Whereas Yrlid et al. analyzed the low-density fraction, we analyzed all lymph cells because the yield of DCs in lymph after irradiation was especially low.

The CD172a+CD11b Subset Migrates Through Blood.

The blood-borne migrating DC population reported on previously6 was defined as a distinct radiosensitive DC subset. The DCs migrating to the skin LNs (Fig. 2A) and Peyer's patches (not shown) were mostly MHCII+CD103+CD172a+CD11b. These cells thus belonged to the second subset of liver DCs, and they almost disappeared after irradiation. This migration was observed in different rat strains, that is, ACI6 and Lewis rats (Yu, unpublished data), indicating the essential nature of this subset. Because crude bone marrow cells also contain blood-borne migrating DCs,6 this subset can be isolated easily and might have potential for use in a DC vaccine.

Observations in Parathymic LNs.

As expected, the proliferative response in the parathymic-LN T-cell area after LT was considerably higher than in other secondary lymphoid organs. The maximal response in the LNs8 was as high as ∼2,500 BrdU+ cells/mm2 in the T-cell area (Fig. 4E), reflecting the additive effect of the CD172a+CD11b and CD172a+CD11b+ migrating subsets through the direct allorecognition pathway.

The diaphragmatic lymphatics provide major drainage for fluids or cells from the peritoneal cavity in many animals, including humans,21 by connecting to the regional LNs. In dogs, cats, rabbits, guinea pigs, and sheep, the mediastinal and/or parasternal LNs are draining LNs.14 In humans, the diaphragmatic lymphatics are connected to the anterior, right and left lateral, and posterior diaphragmatic LN groups, which drain into the parasternal, posterior mediastinal, and brachiocephalic LNs.22 Therefore, after LT in both experimental and clinical settings, the LNs that drain the peritoneal cavity should be recognized as major sites of the intrahost T-cell response by immunogenic passenger DCs that migrate through the lymph, rather than the ordinary regional liver LNs.

The CD172a+CD11b+ Subset Induces T-Cell Response In Vitro and Probably In Vivo, Even After Irradiation.

In the Irr(+) group, the intrahost T-cell response was restricted to the parathymic LNs, where the CD172a+CD11b+ subset formed clusters with recipient proliferating cells from the beginning of the response. In the graft liver, the CD172a+CD11b+ subset made up the majority of DCs (Fig. 6C) and formed clusters similar to those involving the parathymic LNs (Supporting Fig. 1F). Therefore, we suggest that this subset induces a CD8+ T-cell response in vivo in the parathymic LNs and probably also in the graft through the direct allorecognition pathway in the Irr(+) group.

Notably, in vitro experiments involving the mixed leukocyte reaction showed that the radioresistant CD172a+CD11b+ subset in the liver and hepatic lymph of donor rats induced an intense T-cell proliferative response comparable to the control splenic DCs (Fig. 7B). This subset constitutively expressed the B7-2 costimulatory molecules (CD86) and had up-regulated IL-2 receptor alpha (CD25) expression when isolated from the parathymic LNs (Fig. 2B) and the graft liver (Fig. 6D). Expression of a functional IL-2 receptor can be induced in mouse splenic and lung DCs as well as in Langerhans cells during maturation, and a synergistic effect of IL-2 on interferon-gamma production by DCs has been reported on.23 Taken together, these data demonstrate that this subset is functionally mature and possesses the strong allostimulating activity in vitro. These data support the idea that this subset induces CD8+ T-cell responses in parathymic LNs and in the graft through the direct allorecognition pathway.

Rejection in the Irr(+) Group.

Because of the large reduction in the number of donor DCs after elimination of the blood-migrating DC subset and inhibition of systemic alloreactive T-cell generation in the recipient's lymphoid organs, one might expect rejection to be delayed in the Irr(+) group. The question then arises: Why doesn't preoperative irradiation of the graft liver suppress rejection? Some of the possible effects of irradiation on the graft that promote rejection include the persistence of immunostimulatory factors and/or down-regulation of immunosuppressive factors. As for persistent stimulatory cells, the CD172a+CD11b+ DC subset is likely to be a central player. Other stimulatory factors may also be present that were not detected in our analysis.

With respect to suppressive factors, there is a combination of fully MHC-incompatible strains that allows a rat liver to be spontaneously accepted. However, this tolerance is abrogated by donor irradiation17 in different rat strain combinations after different doses of irradiation (Supporting Table 2).24 These reports suggest the presence of some radiosensitive factors that promote liver graft-induced tolerance. Possible factors include regulatory T cells,25 tolerogenic passenger cells,3, 26 and an apoptosis-inducing microenvironment in the liver27 that includes negative costimulatory molecules, such as B7-H1.28

In the present study, FoxP3+ regulatory T-cell responses were not different between the Irr(−) and Irr(+) groups (Supporting Fig. 3), indicating that graft irradiation did not down-regulate the recipient's regulatory T-cell response. Another study suggested that in the LT tolerant group, passenger leukocytes in the recipient's lymphoid organs play a suppressive role by causing apoptosis of the recipient's T cells in the graft.26 Accordingly, MHCIIlowCD11c+ immature DCs in mice29 have been suggested to be tolerogenic. We also found immature DCs in the liver of some rat strains. These DCs were MHCIIlow (Yu, unpublished data) and were not examined in this study. Concerning the liver microenvironment, the liver may facilitate CD8+ T-cell proliferation, leading to apoptosis30; however, we were not able to confirm changes in the expression of radiosensitive genes in the graft liver in this study.

Our findings suggest that preoperative irradiation of the graft liver did not suppress rejection, because a radioresistant CD172a+CD11b+ DC subset generated a sufficient number of effector T cells. Immunosuppressive factors other than regulatory T cells might be down-regulated, but we did not observe them.

Conclusion

Rat liver conventional DCs contain at least two immunogenic subsets that have distinct trafficking patterns and radiosensitivities. One subset comprises radiosensitive MHCII+CD103+CD172a+CD11bCD86+ cells. We previously described this subset as a DC population that readily undergoes transmigration to the recipient's secondary lymphoid organs through the blood and induces a systemic CD8+ T-cell response there. The second subset comprises relatively radioresistant MHCII+CD103+CD172a+CD11b+CD86+ cells that steadily undergo lymph-borne migration to the regional hepatic LNs. When freshly isolated from the liver and hepatic lymph of donor rats after irradiation, these cells have strong allostimulating activity in vitro. After LT, the cells further migrate to the regional LNs of the peritoneal cavity (i.e., the parathymic LNs). These cells up-regulate CD25 (the IL-2 receptor) and are probably responsible for T-cell responses in the parathymic LNs and in the graft through the direct allorecognition pathway as they form clusters with recipient T cells. The LNs that drain the peritoneal cavity, rather than ordinary regional liver LNs, should be recognized as major sites of the intrahost T-cell response because of these immunogenic passenger DCs that migrate through the lymph. Irradiation completely eliminated the migration and immunogenicity of the first subset of DCs, but did not suppress rejection. However, the remaining second subset may generate a sufficient number of intragraft CD8+ T cells. Other immunosuppressive factors might be down-regulated as well. This study provides key insights that help shed light on the mechanisms underlying liver graft rejection. The findings also have clinical implications for the manipulation of immunogenic DC subsets.

Acknowledgements

The authors are grateful to the late professor Ralph Steinman and to Drs. Xiao-Kang Li, Atsushi Sugioka, Kouji Matsushima, and Hiroyuki Yoneyama for their valuable discussions and suggestions. The authors appreciate the excellent technical support provided by Junko Sakumoto and Yasuko Nonaka.

Ancillary

Advertisement