Rapid immunosurveillance by recirculating lymphocytes in the rat intestine: critical role of unsulfated sialyl-Lewis X on high endothelial venules of the Peyer's patches.

Naive lymphocytes systemically recirculate for immunosurveillance inspecting foreign antigens and pathogens in the body. Trafficking behavior such as the migration pathway and transit time within the gastrointestinal tract, however, remains to be elucidated. Rat thoracic duct lymphocytes (TDLs) were transferred to a congeneic host that had undergone mesenteric lymphadenectomy. The migration pathway was investigated using newly developed four-color immunohistochemistry and immunofluorescence. Donor TDLs showed rapid transition in gut tissues from which they emerged in mesenteric lymph around 4 h after intravenous injection. Immunohistochemistry showed that donor TDLs predominantly transmigrated across high endothelial venules (HEVs) at the interfollicular area of the Peyer's patches (PPs), then exited into the LYVE-1+ efferent lymphatics, that were close to the venules. The rapid recirculation depended largely on the local expression of unsulfated sialyl-Lewis X on these venules where putative dendritic cells (DCs) were associated underneath. Recruited naive T cells briefly made contact with resident DCs before exiting to the lymphatics in the steady state. In some transplant settings, however, the T cells retained contact with DCs and were sensitized and differentiated into activated T cells. In conclusion, we directly demonstrated that lymphocyte recirculation within the gut is a very rapid process. The interfollicular area of PPs functions as a strategically central site for rapid immunosurveillance where HEVs, efferent lymphatics and resident DCs converge. PPs can, however, generate alloreactive T cells, leading to exacerbation of graft-versus-host disease or gut allograft rejection.

Introduction

All immune cells constantly monitor for the intrusion of harmful factors as a way to maintain organismal homeostasis. To respond systemically to local inflammatory stimuli, lymphocytes successively patrol tissues, known as recirculation, entering from the blood vessels, scanning the tissue and then exiting to the lymphatics (1). A large number of non-self factors, such as food antigens, ions and sometimes harmful pathogens, are ordinarily internalized through the absorptive epithelium of the intestine. The intestines are equipped with the highest number of immune cells, mostly resident innate cells and migratory lymphocytes (2). What remains unclear, however, are the specifics of the trafficking of recirculating lymphocytes to the intestine, including aspects related to the migration pathway, dwell time within the tissue and relevant molecules. Previously, we reported the migration kinetics and routes of recirculating lymphocytes in the rat liver (3). Their kinetics involve a rapid process in which the transit time of recruited lymphocytes is 4–6 h at steady state. This rapid transit may enable an efficient surveillance of the liver by the recirculating lymphocytes.

In this study, we first estimated the transit time of recirculating lymphocytes in the intestine by directly counting gut-derived cells in the mesenteric lymph. Second, we traced the migration route in the gut tissue using a newly developed four-color immunohistochemistry approach and then identified relevant molecules crucial for the lymphocyte trafficking. Finally, we addressed the significance of these findings in terms of homeostatic immunosurveillance as well as pathogenesis of transplant immunity in the gut.

Methods

Animals

Inbred PvG/c (RT1.AcBc), Lewis (RT1.AlBl) and DA (RT1.AaBa) rats were supplied by SLC Co. (Shizuoka, Japan). Congeneic PvG-RT.7b/OlaHsd rats (RT.1AcBc, RT.7b) and (PvG/c x Lewis) F1 hybrid rats (RT1.AclBcl) were bred and maintained in the laboratory Animal Research Center (Dokkyo Medical University). All animals were reared under specific pathogen-free conditions. Animal handling and care was approved by the Dokkyo Medical University Committee, and in accordance with the Dokkyo University’s Regulations for Animal Experiments and with Japanese Government Law (No. 105). Congeneic PvG-RT.7b male rats were used as donors, unless otherwise specified. General anesthesia was used for surgery, and euthanasia. Anesthesia was administrated using isoflurane (Mylan Inc., Tokyo, Japan), given with an isoflurane vaporizer (SN-487-OT, Shinano Manufacturing, Tokyo, Japan).

Antibodies and reagents

Mouse hybridoma cells against the congeneic marker RT.7b, clone HIS41, were kindly donated by Dr F. G. M. Kroese (Groningen University, the Netherlands). Monoclonal antibodies (mAbs), polyclonal antibodies (pAbs), labeled secondary antibodies and substrates used are listed in Supplementary Table S1 (available at International Immunology Online). Some mAbs were purified from culture supernatants and coupled to FITC, biotin (Dojindo, Kumamoto, Japan) and Alexa Flour conjugates (Thermo Fisher Scientific) in house.

Experimental design

In the first experiment, the intestinal blood-lymph transit assay, the transit time of gut-derived recirculating lymphocytes was estimated by counting the number of donor cells in the thoracic duct lymph of recipient rats that had undergone mesenteric lymphadenectomy (MLNx) 6 weeks earlier, which resulted in the direct influx of the gut lymph into the thoracic duct after regeneration of the lymphatics (Fig. 1A). In the second experiment, multicolor immunohistochemistry or immunofluorescence was performed to analyze the spatiotemporal distribution of donor cells in the intestinal tissue. We also explored the molecules involved in the rapid blood-lymph transition in the gut in an immunohistological study of gut endothelium, flow cytometry of donor lymphocytes and an in vivo short-homing and blood-lymph transit assay with anti-selectin ligand antibody. Finally, we focused on T-cell behavior in the Peyer’s patches (PPs), especially an interaction with tissue dendritic cells (DCs) in terms of immunosurveillance at steady state and their significance in transplant immunity.

Fig. 1.

Kinetics of recirculating lymphocytes in rat intestine. (A) Schematic summary of lymphocyte trafficking in the MLNx group and control group. In the MLNx group, congeneic donor lymphocytes (filled circle) from intestine directly flowed into the thoracic duct without being trapped in the MLN. Open circles indicate host lymphocytes. (B, C) Intestinal blood-lymph transit assay: time kinetics of total donor cell output in the thoracic duct lymph in the MLNx group and control group. An early-stage variant of (B) is more precisely shown in (C). (B) More donor recirculating lymphocytes appeared in the MLNx group (filled circle) than in the control group (closed circle) through the experiment. (C) An early time scale (~12 h) of (B). Donor cell output in the MLNx group was significantly increased at 4 h after transfer. Each bar in (B) and (C) represents means ± SD (n = 5). *P < 0.05, versus control group.

Animal studies

MLNx of 5- to 7-week-old PvG/c rats was performed as described previously, with minor modification (3). The rats were allowed to recover more than 6 weeks to ensure anastomosis of the afferent and efferent lymphatics of the excised nodes (Fig. 1A). The thoracic duct lymphocytes (TDLs) of PvG-RT.7b rats were obtained by routine thoracic duct cannulation and were collected aseptically overnight at 4°C. The TDLs were labeled with 10 µM CFSE (Thermo Fisher Scientific) for 20 min at 37°C. The viability of labeled TDLs was consistently >95% as assessed by the trypan blue dye exclusion method. A total of 1 × 108 cells per rat of TDLs were injected intravenously (i.v.) into host PvG/c rats that had received thoracic duct cannulation immediately before cell transfer.

In the intestinal blood-lymph transit assay, thoracic duct lymph was collected every hour up to 12 h, then at 15, 18, 21, 24, 30, 36, 42, 48 and 72 h after transfer. To avoid imposing great stress, the subject rats were cared for dedicatedly during cannulation. An actual body weight (BW) loss after 72-h cannulation was 13.4 ± 2.4% (n = 6), which was much less than those of wasting conditions such as in rats developing acute graft-versus-host disease (GvHD), where their BW rapidly dropped >30% in 3 days (4). After counting the hourly output of total lymphocytes, cytospot smears were prepared. The proportion of CFSE+ cells to total cells for each cytospot smear was counted under a fluorescent microscope with a differential interference contrast aid (Axioskop2 plus, Carl Zeiss, Jena, Germany), and the total donor cell output per hour was calculated. In some cases, the host PvG/c rat was injected i.v. with 1 mg of anti-sialyl-Lewis X mAb, F2 1 h before congeneic TDL transfer. Donor TDLs in thoracic duct lymph were counted and analyzed by FACS.

In the short-homing assay, donor TDLs were prepared from PvG-RT.7b male rats by thoracic duct cannulation as above. After two washes with PBS, 1 × 108 cells were i.v. transferred to congeneic PvG/c rats that 1 h before had received F2 mAb (500 µg per rat). Two hours after the donor cell transfer, single cell suspensions were prepared from the host PPs, small intestines and peripheral lymph nodes (LNs), and the proportion of donor cells was determined by FACS.

Rat small intestinal transplantation was performed orthotopically between allogenic (DA rat to Lewis rat) or syngenic (Lewis rat to Lewis rat) combinations as previously described (5).

Immunohistological analysis

To analyze the spatiotemporal distribution of donor lymphocytes in the gut tissue, double or triple immunohistochemistry was performed as described previously (6). In brief, tissues were embedded in Tissue-Teck O.C.T. compound (Sakura Fineteck, Tokyo, Japan) and fresh cryosections 4 µm thick prepared. For labeling proliferating cells, rats received an i.v. injection of BrdU (6 mg per 200 g body weight, for immunohistochemistry) or equivalent moles of 5-ethynil-2′-deoxyuridine (EdU) (for immunofluorescence (7)) in sterile PBS 1 h before sacrifice. Immunofluorescent images were captured by a fluorescent microscope equipped with an AxioCam MRm camera (Carl Zeiss). To depict the tissue framework more clearly, original blue images from type IV collagen staining were converted to pseudocolor white with Axiovision software (Carl Zeiss). To quantitatively analyze the positional relationship between MHCII cells and MAdCAM-1+ high endothelial venules (HEVs) with or without F2 antigen, the sub-endothelial zone was defined as a continuous belt with 10-µm width in the peripheral margin of the basement membrane of the MAdCAM-1+ HEV by Axiovision software, and then the number of MHCII+ cells inside this zone was counted. In some cases, fluorescent images were captured by the confocal laser microscopy LSM780 (Carl Zeiss).

For analyzing the in situ kinetics of donor cells in detail, donor cells and multiple tissue components and/or interacting cells were simultaneously visualized in the same section by newly established four-color immunohistochemistry. We first incubated a cocktail of antibodies consisting of a biotinylated mAb against recipient class II MHC (MHC-II), anti-LYVE1 pAb and a fluorescein-conjugated mAb against anti-congeneic marker (HIS41). After three washes with PBS, streptavidin-conjugated β-galactosidase (β-gal) was incubated then light blue color was developed by using X-gal (Tokyo Chemical Industry Co. Ltd, Tokyo, Japan) as a substrate for β-gal. Sequential incubation with enzyme-linked secondary antibody followed by the substrate reaction was performed step by step, which resulted in red (NewFuchsin, Dako, Carpinteria, CA, USA) and blue (Vector Blue, VECTOR, Burlingame, CA, USA) staining for lymphatics and donor lymphocytes, respectively. As a last step, the tissue framework was stained with anti-type IV collagen pAb and visualized in brown with diaminobenzidine. In some cases, biotinylated anti-MAdCAM-1 mAb was used instead, followed by streptavidin-conjugated peroxidase. Black color was developed by using ImmPACT™ SG (VECTOR) as a substrate. Enzyme inhibition was conducted after color development when the same enzyme–substrate system was used at a later step. A schematic of the protocol is shown in Supplementary Figure S1 (available at International Immunology Online).

Flow cytometry

Cervical LNs, mesenteric LNs, the PPs and intestinal tract without the PPs were aseptically excised and then incubated with Liberase TM (Roche Diagnostics, Indianapolis, IN, USA) for 5 min at 37°C. Before enzymatic digestion, the PPs and gut were pretreated for 30 min at 37°C with stirring in EDTA solution to remove mucus. After quenching the enzyme digestion by further incubating for 5 min with 5 mM EDTA, the cell suspension was washed three times with PBS containing 2 mM EDTA, 5% FCS and 100 µg ml−1 DNase I (Roche). After FcγII receptor blocking (BD Bioscience, Franklin Lakes, NJ, USA), the cells were stained with fluorochrome-conjugated antibodies. Stained cells were acquired on an Attune NxT flow cytometer (Thermo Fisher Scientific Inc.) or FACSCalibur (BD Biosciences) and data were analyzed with FlowJo ver. 9.2 (FlowJo LLC, Ashland, OR, USA). T cells and B cells were defined as TCRαβ+B220− and TCRαβ−B220+ cells, respectively.

Statistical analysis

Statistical analysis was performed using the Student’s t-test.

Results

Kinetics of recirculating lymphocytes from the intestine

In the MLNx group, recirculating lymphocytes that had egressed from the intestine could directly enter the thoracic duct lymph without being trapped in the regional LNs of the intestine, the MLN (Fig. 1A). In the intestinal blood-lymph transit assay, donor cells appeared in the thoracic duct lymph more rapidly in the MLNx group when compared with controls (Fig. 1B). Of note, the donor cell emergence in the thoracic duct lymph was significantly higher at 4 h after cell transfer in the MLNx group (Fig. 1C).

PPs as the rapid blood-lymph transition site in the intestine

Immunohistology of intestinal tissue 3 h after congeneic lymphocyte transfer revealed that the cells were distributed predominantly at the interfollicular areas of the PPs (Fig. 2A–C). In FACS analysis for donor cells in the intestinal tissue, about 7 × 106 of donor cells migrated to the PPs, whereas a negligible number of cells did so to the small and large intestine (Fig. 2D). Thus, these studies suggested that donor lymphocytes preferentially entered gut tissues from HEVs in the PPs.

Fig. 2.

Naive lymphocytes recirculate the gut by transiting at PPs. (A–C) Donor cell distribution in the host intestinal tissue 2 h after transfer; double stained with congeneic donor cells (blue) and type IV collagen (brown). (B, C) Interfollicular area (B) and lamina propria (C) encircled in dotted lines in (A). (D) Total number of donor cells in the tissue 2 h after transfer (1 × 108 cells). SI, small intestine; LI, large intestine. (E–G) Triple staining of MAdCAM-1 (black), LYVE-1 (red) and type IV collagen (brown). (F, G) Interfollicluar area and villus lamina propria encircled in dotted lines in (E). (H, I, K, L) Four-color immunohistochemistry for donor cells (blue), MAdCAM-1 (black), LYVE-1 (red) and type IV collagen (brown) at 0.5 h (J), 2 h (K) and 3 h (H, L; encircled area in (H)) after donor TDL transfer (5 × 108 cells). Note some donor cells in the LYVE-1+ lymphatics at 3 h (L). (I) Three-dimensional aspects of donor cells (green), MAdCAM-1 (light blue) and LYVE-1 (red) at the interfollicular area of PPs 3 h after donor cell transfer (5 × 108 cells). IF, interfollicular area; LF, lymph follicle. Scale bars: (A) 400 µm; (B) 20 µm; (C) 50 µm; (E) 300 µm; (F, G) 50 µm; (H) 200 µm; (I) 100 µm; (J) 50 µm; (K, L) 20 µm.

Next, we investigated the expression of MAdCAM-1 (mucosal addressin cell adhesion molecule-1) on blood vessels and the vasculature of LYVE-1+ lymphatics in the gut tissues. Triple immunohistochemistry showed that MAdCAM-1+ venules as well as LYVE-1+ lymphatics were distributed both in the interfollicular area of the PPs and in the lamina propria of intestinal villi. It should be noted that these vasculatures run side by side in close proximity (Fig. 2E–G) in the gut tissues. To analyze the interrelationship between donor cells and these vessels in the PPs, four-color immunohistochemistry was newly developed in which donor cells, MAdCAM-1+ HEV and LYVE-1+ lymphatics were visualized simultaneously using brightfield microscopy (Fig. 2H, J–L). At 3 h after transfer, donor lymphocytes that had transmigrated across the MAdCAM-1+ HEV were readily found near to LYVE-1+ lymphatics at the interfollicular area (Fig. 2J and K) and some had begun to egress into LYVE-1+ vessels (Fig. 2H, I and L). This cellular kinetics was consistent with the appearance of gut-derived donor cells in the thoracic duct lymph of the MLNx group at 4 h after cell transfer (Fig. 1C).

Site-specific expression of unsulfated sialyl-Lewis X in MAdCAM-1+ blood vessels in the intestine and its role for lymphocyte trafficking

To clarify the mechanisms for the preferential homing for donor lymphocytes to the PPs, we examined the expression of cell migration-associated molecules in both lymphocytes and blood vessels in the gut. FACS analysis showed that although transferred donor T and B cells were strongly positive for L-selectin, they were weakly positive for β7 integrin and negative for CCR9, which are involved in the selective migration for gut-homing lymphocytes (Fig. 3A) (8). Immunohistochemical analysis for their corresponding ligands in the gut tissues showed uniform expression of ICAM-1, P-selectin (both not shown) and MAdCAM-1 (Fig. 2E–G) on vascular endothelial cells of HEVs in the PPs and venules in the intestinal villi. In contrast, 6-sulfo sialyl-Lewis X (6-sulfo sLeX), known as a major ligand for l-selectin and recognized by S2 mAb in the peripheral LNs (9), was expressed on neither the PP HEVs (Fig. 3B and C) nor venules in the intestinal villi (not shown). Of note, the vessels in the PPs, but not in the villi, were selectively stained with another recently developed mAb, F2, which reacts with the N-acetyl and N-glycolyl forms of sLeX structure (Fig. 3C) (10), in a PP-dependent manner (Fig. 3D–I).

Fig. 3.

Local expression of unsulfated sialyl-Lewis X and its role in lymphocyte recirculation. (A) Expression of gut-homing molecules (filled histogram) on naive T cells and B cells. Open histogram: isotype control. (B) Double staining of S2 and type IV collagen in the interfollicular area of PPs. (C) Structures of O- and N-glycans modified with the 6-sulfo sialyl-Lewis X moiety, and recognition determinants S2 and F2. (D–I) Double staining for MAdCAM-1 (blue, D–F) or F2 (blue, G–I) and type IV collagen in the PPs (D, G) or intestinal villi (E, F, H and I). (F, I) Inset of (E) and (H), respectively. (J) Short-homing assay. Donor TDLs (1 × 108 cells) were injected intravenously into a host rat that had received 0.5 mg of F2 1 h before. Two hours after cell transfer, CFSE+ donor cells in single cell suspensions from target tissues were quantified by FACS. *P < 0.01, versus control group. (K) Scheme of inhibition assay for lymphocyte blood-lymph transition by F2. (L) Kinetics of donor cell proportion in the host thoracic duct lymph. CFSE+ donor cells in thoracic duct lymph cells collected hourly were quantified by FACS. Note a significant decrease in donor cell proportion in the F2-treated group (open square) compared with control group (filled circle). *P < 0.05, versus control group. IF, interfollicular area; LF, lymph follicle. Scale bars: (B) 100 µm; (D, G) 200 µm; (E, H) 100 µm; (F, I) 20 µm. Each bar in (J) and (L) represents means ± SD (n = 3).

We then examined the role of F2 glycotope in lymphocyte trafficking to the gut tissue by a short-term homing assay. With F2 pretreatment, migration of donor lymphocytes at 2 h after transfer was significantly suppressed not only in the peripheral LNs but also in the PPs when compared to isotype IgG (Fig. 3J). Donor cells in the villus lamina propria of the small intestine were rather slightly increased by the F2 pretreatment, but not statistically significantly (Fig. 3J). Furthermore, in the intestinal blood-lymph transit assay with MLNx rats, donor cell emergence in thoracic duct lymph was significantly inhibited by F2 pretreatment, when compared with non-pretreated MLNx rats (Fig. 3K and L). Control studies, in which normal mouse IgG1, an isotype IgG of F2 mAb, did not affect per se both transmigration of donor TDLs to host secondary lymphoid organs and the output of recirculating lymphocytes in the thoracic duct lymph (Supplementary Figure 2, available at International Immunology Online). These results confirm the validity of the inhibitory effect of F2 mAb in this assay. Of note, fluorescent immunohistochemistry revealed that the expression of the F2 glycotope was observed in the MAdCAM-1+ HEVs at the interfollicular area (Fig. 4A). Four-color immunofluorescence further revealed that unsulfated sLeX was selectively expressed at MAdCAM-1+ HEVs where MHC-II+ cells with dendritic morphology were significantly accumulated underneath (Fig. 4B–I).

Fig. 4.

Area-specific expression of F2 glycotope on MAdCAM-1+ HEVs. (A) Four-color immunofluorescence for F2 (green), MAdCAM-1 (red), MHC-II (purple) and type IV collagen (gray) in the PPs. Note that F2 glycotope was started to express at the IF area (arrowhead) but not the V area. (B–E) Four-color immunofluorescence at the boundary area of F2 glycotope expression. Note that local expression of F2 (green) on MAdCAM-1+ HEVs (blue) where MHC-II+ cells (red) accumulating underneath the vessels (D). (E) Defining the sub-endothelial zone, and (F) Number of MHC-II+ cells in the sub-endothelial zone. Note F2 glycotope was predominantly expressed on MAdCAM-1+ HEVs where more than four MHC-II+ cells distributed in 1000 µm2 of sub-endothelial zone surrounding HEVs. Bar indicates mean ± SD (n = 3). P < 0.01. (G–I) Same horizontal section of MAdCAM-1+ HEVs at interfollicular area. Yellow bar in (E) 10 µm; (B–D) 20 µm.

Rapid recirculation of naive T cells via PPs and its significance in immunosurveillance and transplantation immunity

FACS analysis for donor cells in the migration site revealed that >60% of migrant cells in the peripheral LNs were T cells as in the original TDLs, while B cells were dominant in the PPs at 2 h after cell transfer (Fig. 5A). However, the donor cell type in the thoracic duct lymph that had egressed from MLNx-treated gut up to 6 h post-transfer was exclusively T-cell dominant (Fig. 5A). Sphingosin-1-phosphate receptor, which is involved in T-cell exit from tissue to lymphatics, was expressed to the same degree in donor T cells in the PPs when compared with those of peripheral LNs (Fig. 5B). Four-color immunohistochemistry of the PPs, 3 h post-transfer, showed that many donor lymphocytes in the interfollicular area made contact with MHC-II+ cells with dendritic morphology, and some of the donors had already entered the lymphatics (Fig. 5C and D). These results suggested that migrated T cells first contacted with the PP DCs, forming a cellular cluster at the interfollicular area to inspect antigens presented by DCs; if not, they soon detached from the cluster and departed into lymphatics nearby to restart circulation at the steady state. In contrast, when donor T cells were transferred to a semi-allogenic recipient, they similarly entered the intestinal tissue and made contact with the PP DCs as well (Fig. 5E). Some of them became BrdU+ large lymphoblasts within the cluster (Fig. 5F). The number of proliferating cells at the interfollicular area was largely increased at day 2 (Fig. 5G), and ~10% of total donor cells were proliferating from days 2 to 4 (Fig. 5H).

Fig. 5.

Rapid recirculation of naive T cells in the PPs and the clinical significance for transplant immunity. (A) Proportions of a lymphocyte subset were analyzed in original donor TDLs, from donor cells in the single cell suspension from each host lymphoid organ 2 h after donor cell transfer, and from thoracic duct lymph up to 6 h by FACS. (B) Expression of Sphingosin-1-phosphate receptor (filled histogram) on donor T cells in the PPs 3 h after cell transfer. Open histogram: isotype control. CLN, cervical LN. (C, D) Four-color immunohistochemistry for donor cells (blue), MHC-II (light blue), LYVE-1 (red) and type IV collagen (brown). Note that some donor cells are contacting MHC-II cells in the interfollicular area (arrowhead in (D)). (D) Inset of (C). (E–H) Fate of donor T cell in the allogenic PPs. Donor T cells that contacted with MHC-II+ cells and reside in the host PPs were originally small and round cells at 3 h after transfer (E), but some became proliferating large lymphoblasts at day 3 (arrowheads (F)). (G, H) Kinetics of the proliferative response of donor cells. (G) The number of donor proliferating cells/mm2, (H) the proportion of proliferating donor cells. (I–N) Fate of host T cells in the donor PPs after small intestinal transplantation. (I) Graft PPs were immediately occupied by host MHC-I+ cells at day 2 after transplantation. (J, K) Massive proliferative responses at the interfollicular area in allogeneic graft (J), but not in syngeneic graft (K) at day 2. (L–N) Four-color immunofluorescence for host MHC-I (green), T-cell receptor αβ (red), donor MHC-II (white) and EdU (blue). Note that host MHC-I+TCRαβ+ T cells became enlarged and proliferated with forming of the cell cluster with donor MHC-II+ cells (purple arrowheads in (N)). Day 2 after transplantation. IF, interfollicular area; LF, lymph follicle; LV, lymphatic vessel; S1PR1, Sphingosin-1-receptor 1; EdU, 5-ethynil-2′-deoxyuridine. Scale bars: (C) 100 µm; (D) 20 µm; (E, F) 20 µm; (I–K) 200 µm; (L–N) 10 µm. Each bar in (A, G and H) represents means ± SD (n = 3).

In addition, in the allotransplantation settings of the small intestine, where a part of the donor intestinal tract including the PPs was transplanted to the host, graft PPs were subjected to enormous numbers of infiltrating host cells and were occupied by host MHC-I cells at day 2 (Fig. 5I). Also, vigorous proliferative responses were observed exclusively at the interfollicular area of the graft PPs when compared with those of the syngenic graft (Fig. 5J and K). Recently established four-color immunofluorescent microscopy using EdU (7) clearly showed that most of the proliferating cells were host MHC-I+TCRαβ+ T cells, which directly contacted with donor MHC-II+ cells at day 2 thereafter (Fig. 5L–N). These activated T cell–DC clusters may represent a site of the direct pathway of allosensitization as reported previously (11, 12).

Discussion

In this study, we directly showed the transition time of recirculating lymphocytes in the intestine by counting transferred cells in the mesenteric lymph of congeneic host rats that had been subjected to MLNx. The estimated minimal transit time was 4 h after i.v. transfer, which is very rapid similar to that of the liver (3) but shorter than that of the peripheral LNs (13). Using multicolor immunohistochemistry, we clearly demonstrated the close proximity of MAdCAM-1+ HEVs to LYVE-1+ lymphatics in the gut tissue. A previous report involving scanning electron microscopic observation of corrosion casts in rabbit PPs suggested a close association of lymphatics with HEVs (14). This association was also observed in mice (S. Simmons, unpublished data), suggesting that the close apposition of HEVs with the lymphatics is conserved across animal species at least in the PPs. These anatomical features would enable the migrated lymphocytes to exit more rapidly in starting recirculation.

Previous studies indicated that a predominant distribution of donor cells in the interfollicular area of the PPs and few cells in the small and large intestine (15). In this study, we confirmed these observations quantitatively by FACS, indicating that the PPs are the major site for intestinal blood-lymph transition in a steady state. In this respect, immunohistochemistry also revealed that MAdCAM-1+ HEVs express a unique type of l-selectin ligand in a PP-specific manner, as discussed below. Research involving intravital microscopy in rats has indicated that fluorochrome-labeled lymphocytes migrate into microlymphatics of the PPs, which were assigned by drainage of pre-injected vital dyes at 240 min after transfer (16). These results may support our findings. Antigen sampling in the small intestine has largely been performed in a specialized compartment of the PPs, the follicle-associated epithelium rather than the villus epithelium (17). In addition, although intestinal DCs constantly migrate to draining LN, MLN where they present cognate antigen to naive T cells, their turnover rate is much slower than that of naive lymphocytes (18, 19). Thus, in terms of defense against incoming pathogens and toxins landing at the mucous surface of the body, it would be strategically reasonable to selectively recruit naive lymphocytes at the PPs where foreign antigens are preferentially incorporated in order to survey them quickly and efficiently.

It is well known that modifications of HEV peripheral node addressin (PNAd), such as sialylation and fucosylation, are indispensable for L-selectin+ lymphocyte homing across the HEVs. Sulfation of sLeX is also important for effective homing to the peripheral LNs: however, lymphocyte migration toward the PPs is not impaired in knockout mice lacking sulfotransferases (GlcNAcST-1 and 2) (20). Recent transcriptomic analysis of mice HEVs showed that expression of the HEV-specific sulfotransferase, GlcNAcST-2 is less than 10-fold lower in the PPs than in the peripheral LNs (21). In addition, the S2 mAb, which recognizes the 6-sulfo sLeX structure, does not inhibit lymphocyte homing to mice PPs, although it reacts with PP HEV in situ (10). These results suggested a minor role of sulfation in the PP HEV homing in mice. In the present study, the rat PPs were not stained with S2 at all but stained with F2 instead. F2 recognizes the sLeX moiety on O- and N-glycan regardless of sulfation while S2 recognizes only the sulfated form (Fig. 3C), suggesting that sLeX on the rat PP HEVs should be unsulfated. The significant block of lymphocyte homing to the PPs and the delayed donor cell appearance in the mesenteric lymph (Fig. 3K and M) by F2 mAb indicated that unsulfated sLeX did play a critical role in the rapid transition of recirculating lymphocytes in the PPs. On the other hand, in agreement with the abovementioned mouse study (10), lymphocyte homing partly persisted in rat PPs when compared with other LNs by F2 pretreatment (Fig. 3J). The residual migrants after F2 pre-treatment were not only B cells but also T cells (H. Ueta, unpublished data), suggesting the presence of selectin-independent homing abilities in each cell type, such as CD22-mediated B-cell homing (22) and the involvement of MAdCAM-1:α4β7 integrin interaction, which compensates for the weaker avidity between lymphocyte l-selectin and unsulfated sLeX on the PP HEVs when compared with 6-sulfo sLeX (20). Also, this axis might be involved in naive lymphocyte homing to the intestinal lamina propria.

It is noteworthy that in gut tissues, unsulfated sLeX was expressed exclusively in the MAdCAM-1+ HEVs in the interfollicular area (Fig. 3J). According to a recent review of HEVs, MAdCAM-1+ immature HEVs develop into MAdCAM-1+PNAd+ cuboidal mature HEVs by non-canonical NF-κB signaling driven by continuous stimulation with lymphotoxin-β receptor on HEVs (23). CD11c+ DCs are thought to be among the candidates for lymphotoxin-providing cells and depletion of CD11c+ cells in CD11c-DTR transgenic mice results in the loss of PNAd expression and inhibition of lymphocyte recruitment (24). In our multicolor immunohistological study, some PP DCs were distributed in the vicinity of and made contact with MAdCAM-1+F2+ HEVs (Fig. 4). Therefore, local unsulfated sLeX expression may represent functional maturation of the HEVs, which might be regulated by a distinct type of DC subset located at the interfollicular area of the PPs (25).

Also of interest is that the cell types of recirculating lymphocytes in the PPs 2 h after transfer were relatively B-cell dominant, while those of mesenteric lymph up to 6 h were almost all T cells (Fig. 5A). Therefore, homeostatic naive T-cell kinetics are very fast; immediately after transmigrating HEVs, they formed cell clusters with resident DCs, and then rushed into neighboring lymphatics at steady state. Blast formation followed by transient proliferative response of T cells in an allogeneic host (Fig. 5F–H) suggested that the PPs act not only as a blood-lymph transition site but also as a major site of immunoactivation for naive T cells when encountering the cognate antigen presented by DCs, as well as spleen and peripheral LNs (26, 27). In other words, the PPs could be an essential site in generating alloreactive T cells in the transplant setting. Indeed, it was proposed in a mouse model of acute GvHD that donor naive T cells differentiated to effector cells in the PPs and that the disease did not develop in the PP-deficient host (28).

Furthermore, in the case of small intestinal transplantation, an enormous number of host T cells would quickly flood into the graft PPs soon after re-vascularization and make contact with donor resident DCs within the graft (Fig. 5I, J, L–N). This mode of interaction is a direct host-versus-graft reaction (HvGR), where host naive T cells developed into anti-donor effector T cells because of direct sensitization by donor DCs, leading to graft rejection (11, 29). Of note, unlike liver transplantation where intragraft proliferative responses are prominent around day 4 after transplantation (11, 29), much more severe BrdU responses were induced even at day 2 at the interfollicular area of the graft PPs in the current work (Fig. 5J). In terms of sensitization site, alloreactive host T cells against the hepatic allograft are mainly generated at host secondary lymphoid organs by ‘intrahost’ direct HvGR where passenger donor DC migration is a prerequisite (11), while those of intestine are generated within the graft by ‘intragraft’ direct HvGR. Thus, effector T cells from the latter would soon exert their cytotoxicity without trafficking and accumulate stronger damage against the graft. Preoperative donor DC depletion along with HEV blockade by each antibody ex vivo (4) might be an effective manipulation for preventing intragraft sensitization and prolonging graft survival. Taken together, as a mucosa-associated lymphoid tissue, the PPs are equipped with specialized structures that enable rapid transition and antigen presentation to naive T cells for effective immunosurveillance at steady state. However, these features also contribute to serious complications involving transplant immunity.

Methodologically, although immunofluorescence staining is useful for analyzing multiple parameters, it is difficult to gain a complete observation of the signals through the tissue in the dark field. In this study, we newly established four-color immunohistochemistry in which the particular cells, vessels and tissue framework were simultaneously visualized using brightfield (Supplementary Figure S1, available at International Immunology Online). The result was high visibility and made it possible to understand more deeply and precisely the positional relationship among migrated lymphocytes, HEVs and lymphatics in the PPs. This method would be applicable for other detailed analysis for in situ multicellular interactions in other investigations.

Conclusion

In conclusion, we directly demonstrated that lymphocyte recirculation within the gut is a rapid process comparable to that of LN and liver. Four-color immunohistochemistry revealed that PPs function as a strategically central site for rapid immunosurveillance and recirculation for T cells where unsulfated sLex-positive HEVs, efferent lymphatics and resident DCs converge. However, the PPs also may provide the opportunity to generate alloreactive T cells which can lead to complications for transplantation therapy.

Supplementary data

Supplementary data are available at International Immunology Online.

Click here for additional data file.

Click here for additional data file.

Click here for additional data file.

Funding

This study was partly supported by Dokkyo Medical University Young Investigator Award (2015-10 to T.U.), Dokkyo Medical University Project Research Grant (2012-02-1 to H.U.) and JSPS KAKENHI Grant Number (JP17K08518 to K.M.).