Inhibition of LIGHT (a cellular ligand for herpes virus entry mediator and lymphotoxin receptor)/herpes simplex virus entry mediator (HVEM) and LIGHT/lymphotoxin β receptor (LTβR) interactions decreases mortality in MHC class I and II disparate graft-vs-host disease (GVHD). The present studies assessed the effects of these interactions on the generation of CD4+ T cell alloresponses in MHC class II-disparate MLC and GVHD. An inhibitor protein of LIGHT and LTαβ2 (LTβR-Ig) and an inhibitor protein of LIGHT (HVEM-Ig) caused similar decreases in alloresponses of control B6 or B6.129S1-IL12rb2tm1Jm (B6.IL12R−/−) spleen cells (SpC) in a MHC class II-disparate MLC. GVHD-induced wasting disease in MHC class II-disparate recipients of B6 CD4+ SpC who received either the LTβR-Ig-encoding adenovirus (LTβR-Ig Adv; 13.1 ± 10.9%; n = 10; p = 0.0004) or the HVEM-Ig-encoding adenovirus (HVEM-Ig Adv; 16.4 ± 9.9%; n = 13; p = 0.0008) was significantly reduced compared with that in recipients of a control adenovirus (30.4 ± 8.8%; n = 13). Furthermore, gut GVHD histologic scores of recipients of B6 CD4+ SpC who received the LTβR-Ig Adv (0.8 ± 0.8; n = 5; p = 0.0007) or the HVEM-Ig Adv (1.4 ± 0.5; n = 5; p = 0.008) were reduced compared with scores of recipients of a control adenovirus (2.5 ± 0.75; n = 11). In the intestine, both LTβR-Ig Adv and HVEM-Ig Adv decreased CD4+ T cells (0.35 ± 0.4 × 106 (n = 6) vs 0.36 ± 0.02 × 106 (n = 9); p = 0.03 and p = 0.007) compared with control adenovirus (0.86 ± 0.42 × 106; n = 9). LIGHT is critical for optimal CD4+ T cell alloresponses in MHC class II-disparate MLC and GVHD.
Allogeneic bone marrow transplantation (BMT)3 is the treatment of choice for many malignant conditions. Acute graft-vs-host disease (GVHD) is a major obstacle to successful outcomes after allogeneic BMT (1). Gastrointestinal involvement is a major cause of morbidity and mortality in human GVHD. Animal models of GVHD have been used in determining the pathogenesis of intestinal GVHD. In many murine models of intestinal GHVD, including the DBA/2J*B6D2F1 and the B6*bm12 × B6F1 GVHD models, CD4+ T cells play a prominent role in the pathogenesis (2, 3, 4, 5).
Multiple clinical and experimental observations indicate that cytokine dysregulation occurs during acute GVHD (5). A general hypothesis offered as an explanation for this phenomenon is that donor T cells encounter allogeneic histocompatibility Ags on host tissues and, in the presence of IL-12, secrete the Th1 cytokines, IFN-γ and IL-2 (6). IFN-γ primes monocytes and macrophages to secrete large amounts of proinflammatory cytokines, including TNF. Prevention of acute GVHD while retaining the mature T cells in the bone marrow graft may be possible if the amplification of inflammatory cytokine effectors is disrupted. TNF appears to influence intestinal inflammation induced by GVHD by promoting a Th1 cytokine profile (IL-2 and IFN-γ) independently of IL-12-IL-12R interactions (7, 8). Furthermore, TNF/TNFR type 2 interactions are important for the development of intestinal inflammation and activation/differentiation of Th1 cytokine responses by intestinal lymphocytes in MHC class II-disparate GVHD (8, 9).
The TNF superfamily includes a related cytokine lymphotoxin (LTα) also called TNF-β, which is secreted as homotrimers that bind to the two TNF receptors (10). However, LTα is also found complexed with LTβ, forming the membrane LTαβ2 complex (11, 12). This heteromeric complex binds uniquely to the LTβ receptor (LTβR), suggesting a distinct biological function (13). Inhibition of the LT pathways in adult mice showed a dominant role for LT in the maintenance of splenic architecture and aspects of Ig formation and B cell follicular structure (14, 15, 16, 17). When the LTαβ2/LTβR pathway has been disrupted by genetic deletion, the animals lacked Peyer’s patches (18, 19, 20, 21).
Recently, studies indicate that the LT system is interwoven with the LIGHT (a cellular ligand for herpes virus entry mediator and lymphotoxin receptor)/herpes simplex virus entry mediator (HVEM) receptor system because LIGHT can also bind effectively to the LTβR (22). The name LIGHT was derived from “homologous to lymphotoxins shows inducible expression and binds to HVEM.” Importantly, the HVEM receptor can be detected on T, B, and NK cells as well as endothelial cells. In contrast, LTβR is not found on T and B cells, but, rather, on monocytes. A mouse homologue of human LIGHT has recently been shown to have both in vitro and in vivo LIGHT costimulatory activity, leading to T cell growth and secretion of IFN-γ. Blockade of LIGHT by administration of soluble LTβR or an anti-LIGHT Ab leads to decreased cell-mediated immunity and reduces MHC class I- and II-disparate GVHD, leading to decreased mortality (23). In addition, other investigators have demonstrated, using LIGHT transgenic mice, that expression of LIGHT in T cells of the intestinal mucosa may regulate inflammation in the intestine (24).
Unlike models used by previous investigators, the proposed studies determined the roles of LTαβ2/LTβR, LIGHT/LTβR, and LIGHT/HVEM interactions in MHC class II-disparate MLC and MHC class II-only-disparate GVHD where the donor CD4+ T cells are responsible for the initiation and progression of the intestinal disease. Importantly, we had previously constructed an inhibitor protein with an extracellular domain of the human LTβR linked to the Fc portion of murine IgG (LTβR-Ig), which was subsequently subcloned into an adenoviral vector (LTβR-Ig Adv) (25, 26). This construct inhibits LIGHT/LTβR and LTαβ2/LTβR signaling in macrophages and LIGHT/HVEM signaling in T cells. The present studies examine the roles of these interactions using the LTβR-Ig and a new LIGHT inhibitor, which consisted of the extracellular domain of the murine HVEM linked to the Fc region of the murine IgG (HVEM-Ig). Importantly, HVEM-Ig blocks only LIGHT/HVEM interactions on T cells and LIGHT/LTβR interactions in macrophages. By using both inhibitors, similar effects on CD4+ T cell responses would support the hypothesis that LIGHT signaling alone is sufficient for optimal alloresponses.
In MHC class II-disparate recipients of B6 donor spleen cells (SpC), colonic GVHD histological scores are lower when the donor B6 SpC lack the IL-12R (8). The decreased scores were associated with lower IFN-γ levels in lamina propria lymphocytes (8). IL-12 signaling through the IL-12R on T cells induces high levels of IFN-γ production through Stat 4 signaling (27). By using responder T cells that lack the IL-12R β2 subunit of the IL-12R (28) in MHC class II-disparate MLC, the need for IL-12-IL-12R interactions for the optimal effect of LIGHT/LTβR and LTαβ2/LTβR signaling in macrophages and LIGHT/HVEM signaling in T cells could be assessed.
Importantly, our results revealed that both LTβR-Ig and HVEM-Ig decreased alloproliferation and IFN-γ responses in MHC class II-disparate MLC using B6 responder T cells. To assess whether IL-12 signaling was critical for LIGHT-mediated effects on alloresponses and IFN-γ production, LTβR-Ig or HVEM-Ig was used in MHC class II-disparate MLC with control B6 or B6.IL12R−/− responder SpC. Similar decreases in CD4+ T cell alloresponses were noted with both inhibitors with either control B6 responder SpC or IL-12R-deficient B6 responder SpC, supporting the hypothesis that the alloproliferation effects of LIGHT were independent of IL-12 signaling from macrophages to IL-12R on T cells. Furthermore, both HVEM-Ig and LTβR-Ig reduced wasting disease, histopathological scores, and infiltrating CD4+ T cells during MHC class II-disparate gut GVHD, supporting the hypothesis that LIGHT signaling is critical for the development of gut GVHD and proliferation of disease-producing CD4+ T cells.
Materials and Methods
C57BL/6J (B6; H-2b), B6.C-H-2bm12KhEg (bm12; H-2bm12), B6.129S1-IL12rb2 tm1Jm (B6.IL12R−/−; H-2b), and B6.129S2 Ltα(tm1Dch) (B6.LTα−/−; H-2b) cells were obtained from The Jackson Laboratory and were maintained in a specific pathogen-free environment. B6 males and bm12 females were bred to produce an F1 strain (bm12 × B6F1; H-2b/bm12).
Construction of the HVEM inhibitor construct and insertion into a recombinant adenoviral vector
A vector containing the extracellular domain of the murine HVEM receptor with specific cut sites of EcoRI and BclII was provided by Drs. Wang and Fu at the University of Chicago (29). Initially, the M1 FLAG polypeptide sequence (Asp-Tyr-Lys-Asp-Asp-Asp-Asp-Lys) was inserted before the extracellular domain of the mouse receptor for HVEM. Specifically, the flag peptide and preferred restriction sites (KpnI and BamHI) were sewed into the HVEM sequence using PCR amplification. Previously established restriction sites (EcoRI and BclII) were used to link the sequences. The oligonucleotides used to produce this linker had the sequences: KPNI FLAG EcoRI, GATggtaccGACTACAAGGACGACGATGACAGGgaattc; and BclII BamHI, GAactagtTCTTcctaggGCGC. The construct was cloned into a Topo TA vector (Invitrogen Life Technologies), then cut out with KpnI and BamHI. The FLAG-extracellular domain of the mouse receptor for HVEM (FLAG-HVEM-ED) sequence with the correct restriction sites was verified by sequencing. FLAG-HVEM-ED was then inserted before the mouse IgG H chain, previously cloned into a PACCMV vector (30, 31), creating a FLAG-HVEM-Ig fusion protein. PACCMV containing the fusion protein was cotransfected together with the vector pJM17 into 293 cells to permit recombination and the generation of infectious particles. Plaque-purified adenovirus (Adv) was then grown to a titer of 1010 PFU/ml, purified by a cesium chloride gradient, and stored at −80°C until use.
Western blot detection of recombinant HVEM inhibitor protein
A B6 mouse, injected with 100 μl of purified Adv expressing the HVEM-Ig fusion protein with a titer of 1010 PFU/ml, was bled 48, 72, and 96 h after injection. One microliter of plasma was subjected to electrophoresis. Standard samples of the FLAG-BAP protein were run in parallel to the experimental samples to quantitate the amount of the FLAG peptide present. Serum from another Adv encoding the chimeric inhibitor protein consisting of the extracellular domain of the human 55-kDa TNF receptor linked to murine Fc region of IgG1 (not including a FLAG protein) was used for the negative control. The proteins were transferred to a nitrocellulose membrane by electroblotting. The membrane was blocked with 5% dry milk, and the peroxidase conjugated anti-FLAG Ab was added at a concentration of 1 μg/ml. After incubation and three cycles of washing, the membrane was treated with ECL reagent (Amersham Biosciences) and exposed to film. In some experiments, quantitation of inhibitor proteins were estimated by the density of the bands.
In other B6 mice that received the purified Adv expressing the HVEM-Ig fusion protein (HVEM-Ig Adv), 1 ml of plasma was purified by affinity chromatography using Sepharose G beads (Amersham Biosciences). Briefly, the columns were washed with Dulbecco’s PBS, and the protein was eluted with glycine-HCl, pH 2.5, and immediately neutralized with Tris buffer, pH 9.5. A BSA protein assay was performed on each fraction of the eluant. Fractions containing IgG, as indicated by the protein assay, were pooled, dialyzed, and concentrated. A Western blot was performed to detect FLAG-HVEM-Ig in the purified fractions.
Adenoviral vectors containing β-galactosidase (β-gal Adv) and the LTβR-Ig fusion protein (LTβR-Ig Adv) were also used in experiments in addition to the HVEM-Ig Adv (25, 32). Briefly, 100 μl of β-gal Adv or LTβR-Ig Adv with a titer of 1010 PFU/ml was injected into BMT recipients at the time of transplantation. In some experiments another Adv encoding an extracellular domain of another member of the TNF receptor family (CD27) linked to FLAG and murine IgG1 Fc region (CD27-Ig Adv) was used as a Ig control.
Isolation of SpC
SpC were harvested by surgically removing the spleen, mincing and disrupting the spleen over a funnel covered with nylon mesh, and washing repeatedly with HBSS into a 50-ml conical tube (33). After the suspension was centrifuged at 600 × g for 10 min, the cell pellet was resuspended in complete medium and counted.
Control B6, B6.IL12R−/−, or B6.LTα−/− responder SpC were cultured with irradiated bm12 or syngeneic stimulator cells. In triplicate wells, the cells were cultured in the presence of 1 μl of plasma from a B6 mouse expressing the HVEM-Ig, 1 μl of plasma from a B6 mouse expressing the LTβR-Ig, or 1 μl of plasma from a B6 mouse infected with the control Adv and then incubated for 72 h. In other experiments the cells were cultured in the presence of 0.5 μl of murine anti-LIGHT Ab (1 mg/ml; R&D Systems; AF 1794) or control IgG (1 mg/ml; R&D Systems; AB 108C) and incubated for 72 h. One hundred microliters of supernatant was collected from each well, and 20 μl of tritiated thymidine was added to each well. After 18–21 h of incubation, the plate was frozen for 1 h to lyse the cells, then cells were harvested on a scintillation counter.
CD4-purified proliferation assays
Purified B6 or B6.IL12R−/− CD4+ responder SpC were cultured with irradiated MHC class II-disparate bm12 or syngeneic stimulator cells. Responder B6 or B6.IL12R−/− SpC were first separated using Lympholyte-M (Cedarlane Laboratories), a density separation medium (34). The isolated lymphocytes were then added to Mouse CD4 Cell Recovery Column (Cedarlane Laboratories) according to the protocol, and CD4+ SpC were collected. The responder and irradiated stimulator SpC were then cultured in triplicate wells as previously described.
IFN-γ levels in the MLC supernatants were assessed by ELISA. Briefly, supernatants from MLC triplicate wells were removed and assayed using ELISA for IFN-γ (XMG1.2; BD Pharmingen) 72 h after culture initiation. Briefly, plates were coated with 100 μl/well of the diluted capture Ab for the specific cytokine and incubated overnight at 4°C. After removal of excess capture Ab and washing, between 3 and 50 μl of the MLC supernatants were added to triplicate wells and incubated for 2 h. After washing, enzyme reagent was added to the wells and incubated at room temperature. Substrate solution was added and incubated for 30 min, then stop solution was added, and OD was read at 450 nm.
In all experiments, recipient bm12 × B6F1 mice (H-2b/bm12; 8–12 wk of age) were maintained on acidified (pH 2) antibiotic (100 mg/L neomycin and 10 mg/l polymyxin B) water for 2–3 days before transplantation and 10 days after transplantation. SpC and bone marrow cells (BMC) from B6 mice were used as a source of donor cells. On the day of transplantation, all recipient mice were irradiated (900 cGy) and 2 h later were injected via the lateral vein with donor cells and LTβR-Ig Adv, HVEM-Ig Adv, or control β-gal Adv as previously described (8). Briefly, 8.89 × 106 CD4+ B6 SpC and 4.6 × 106 T cell-depleted B6 BMC were injected into the lateral tail vein of lethally irradiated bm12 × B6F1 mice as previously described (8). The LTβR-Ig Adv and HVEM-Ig Adv generate an endogenous chimeric inhibitor protein. The soluble, secreted LTβR-Ig protein binds LIGHT and LTαβ2 at high affinity with specific neutralizing activity (25, 26). After infusion of the LTβR-Ig Adv, the inhibitor concentration was maintained at high levels for the duration of the studies (25, 26).
BMC from B6 mice were T cell depleted as previously described. Briefly, BMC were incubated with anti-HO13.4 Ab for 30 min at 4°C, then incubated with adsorbed rabbit complement at 37°C for 30 min (35).
Histological scoring of intestinal GVHD
Individual recipients of the B6 SpC and BMC transplants were evaluated according to the scoring system used by Snover (36). The large intestine was isolated by removing all the intestine distal to the appendix, then was stained with H&E. Peyer’s patches were removed from the small intestine, and 2-cm sections from the entire small intestine were cut from the duodenum to the ileum, fixed in formalin, and stained with H&E. Specifically, colonic GVHD was reported as grade 1 with evidence of increased apoptosis, grade 2 with evidence of cystically dilated crypts containing necrotic debris with individual crypt loss, grade 3 with loss of multiple crypts with preservation of surface epithelium, and grade 4 with complete loss of epithelium. The pathologist reviewing the slides was blinded as to which animals were assigned to each of the experimental protocols.
Flow cytometry of the intestinal lamina propria lymphocytes
Lamina propria lymphocytes were isolated from the intestines of the BMT recipients; the HVEM-Ig Adv, LTβR-Ig Adv, or β-gal Adv; or syngeneic control animals as previously described (35). Cells were labeled with FITC-labeled CD4+ or control Abs, then fixed with 4% paraformaldehyde at room temperature for 10 min. The cells were analyzed by fluorescence-activated flow cytometry on a FACScan (BD Biosciences).
GVHD-induced weight loss and survival data were analyzed by the Mann-Whitney rank-sum nonparametric test; the cellular experiments were analyzed by Student’s t test.
Recombinant HVEM-Ig and LTβR-Ig were detected by Western blot analysis in immunocompetent B6 and lethally irradiated bm12 × B6F1 recipients of B6 BMC and SpC mouse plasma
Mice were bled daily for 3 days after injection of the FLAG-HVEM-Ig-encoding adenoviral vector (HVEM-Ig Adv). A Western blot was performed on concentrated fractions of the purified HVEM-Ig protein to assess the expression of the FLAG-HVEM-Ig fusion protein in Sepharose G-purified mouse serum. Detection by ECL of 66-kDa bands indicated that the FLAG-HVEM-Ig fused protein was expressed in the plasma (Fig. 1 A). HVEM-Ig increased to detectable levels 48–72 h after adenoviral injection and was present up to 13 days after adenoviral injection. The peak HVEM-Ig chimeric protein concentration in mouse plasma was 0.6 mg/ml, as assessed by a BSA protein assay of Sepharose G-purified serum.
In BMT experiments, mice were bled repeatedly to determine the level of recombinant protein in the plasma. Detection by ECL of 66-kDa bands indicated that the FLAG-HVEM-Ig and the FLAG-LTβR-Ig fusion protein were expressed in plasma (Fig. 1 B). Hence, HVEM-Ig and LTβR-Ig were expressed in plasma for the entire course of the experiments, days 2–14.
Alloproliferation in MHC class II-disparate MLC using control B6 or B6.IL12R−/− SpC is decreased with LTβR-Ig or HVEM-Ig
Previous investigators noted that the LIGHT/HVEM and LIGHT/LTβR interactions were involved in the development of MHC class I- and II-disparate GHVD (23). The present studies were designed to assess the importance of the LTαβ2/LTβR, LIGHT/HVEM, and LIGHT/LTβR interactions in MHC class II-only-disparate alloproliferative responses. Furthermore, the present studies were designed to evaluate whether the effects of LTαβ2 or LIGHT on the generation of in vitro CD4+ T cell alloresponses in MHC class II-disparate MLC were IL-12 dependent. Control B6 or B6.IL12R−/− responder cells were cultured with irradiated MHC class II-disparate bm12 or irradiated syngeneic stimulator cells in the presence or the absence of LTβR-Ig or HVEM-Ig inhibitor protein or control mouse Ig for 3 days before assay of [3H]thymidine incorporation. Importantly, T cell proliferative responses of B6.IL12R−/− responder SpC had been previously shown to be comparable to those of control B6 responder SpC (8).
Both LTβR-Ig and HVEM-Ig reduced alloproliferation in MLCs using control B6 SpC (36,882 ± 13,681 vs 16,041 ± 9,687 vs 18,689 ± 7,038 cpm; p = 0.01 and p = 0.04, respectively; n = 5) with irradiated MHC class II-disparate bm12 stimulator SpC (Fig. 2,a) In addition, LTβR-Ig and HVEM-Ig inhibition decreased alloproliferation similarly in B6.IL12R−/− SpC in MHC class II-disparate MLC (29,835 ± 10,911 vs 11,570 ± 2,521 vs 12,733 ± 3,451 cpm; p = 0.006 and p = 0.02, respectively; n = 5; Fig. 2 b). Furthermore, no additive effects were noted with the addition of both inhibitors (B6 control, 34,483 ± 2,421 cpm; B6-LTβR-Ig and HVEM-Ig, 17,675 ± 7,468 cpm (n = 3); B6.IL12R−/− control, 35,520 ± 5,788 cpm; B6.IL12-LTβR-Ig and HVEM-Ig, 22,028 ± 9,132 cpm (n = 3)). In additional experiments, a reduction in proliferation was noted with a murine anti-LIGHT mAb vs a control Ab (15,524 ± 553 vs 9,793 ± 291 cpm; n = 3).
Supernatants from MHC class II-disparate MLC were assayed for IFN-γ production 72 h after the addition of a chimeric LTβR-Ig or HVEM-Ig inhibitor protein. IFN-γ production was reduced similarly in MLC using control B6 with the addition of LTβR-Ig or HVEM-Ig (B6 control, 6499 ± 1795 pg/ml (n = 10); B6-LTβR-Ig, 2522 ± 766 pg/ml (n = 9); p = 0.0000006; B6-HVEM-Ig, 2332 ± 388 pg/ml (n = 9), p = 0.000002; Fig. 2,c). In MLC using B6.IL12R−/− cells, IFN-γ levels were also decreased with both LTβR-Ig, an inhibitor of both LIGHT and LTαβ2, and with HVEM-Ig, an inhibitor of LIGHT (B6.IL12R−/− control, 3272 ± 830 pg/ml (n = 10); B6.IL12R−/− and LTβR-Ig, 1905 ± 300 pg/ml (n = 7); p = 0.0008; B6.IL12R−/− and HVEM-Ig, 720 ± 91 pg/ml (n = 5); p = 0.0001; Fig. 2 d). Thus, LIGHT inhibition alone is sufficient to reduce CD4+ T cell alloproliferation and IFN-γ levels in MHC class II-disparate MLC. Furthermore, the effect on Th1 polarization is IL-12 independent, suggesting that indirect activation of the CD4+ T cell by the macrophage through IL-12-IL-12R interactions is not essential.
B6 and B6.IL12R−/− CD4+ T cells were purified and used in MHC class II-disparate MLC with HVEM-Ig or LTβR-Ig or control Ig. LTbR-Ig and HVEM-Ig decreased proliferation similarly in purified CD4+ B6 SpC (132,382 ± 30,104 vs 76,437 ± 21,642 vs 54,542 ± 38,274 cpm; n = 6; p = 0.001 and p = 0.0002, respectively; Fig. 3,a). In addition, LTbR-Ig and HVEM-Ig decreased proliferation similarly in purified CD4+ B6.IL12R−/− SpC (160,006 ± 25,047 vs 98,645 ± 17,334 vs 88,429 ± 34,476 cpm; n = 9; p = 0.000007 and p = 0.0002, respectively; Fig. 3,b). Supernatants from purified CD4+ T cell MHC class II-disparate MLC were assayed for IFN-γ production 72 h after the addition of the LTβR-Ig or HVEM-Ig inhibitor protein. IFN-γ levels were decreased by LTβR-Ig and HVEM-Ig in the supernatants of MLC using CD4+ purified B6 (control, 22,880 ± 5,378 vs 11,364 ± 2,741 vs 9,897 ± 4,295 pg/ml; n = 6; p = 0.0003 and p = 0.00008, respectively; Fig. 3,c). IFN-γ levels were decreased in MLCs using purified CD4+ B6.IL12−/− SpC (control, 11,230 ± 2,335 pg/ml) with LTβR-Ig (5,940 ± 830 pg/ml; n = 3; p = 0.014) and HVEM-Ig (5,707 ± 1,657 pg/ml; n = 3, p = 0.018; Fig. 3 d). These data confirm that LIGHT signaling is essential in purified CD4+ responder T cells that lack the IL-12R for optimum alloproliferation and IFN-γ production.
HVEM-Ig and LTβR-Ig were also added in MHC class II-disparate MLC using B6.LTα−/− responder SpC. Similar decreases in alloproliferation occurred with LTβR-Ig and HVEM-Ig in MLC using B6.LTα−/− responder SpC (LTα-control, 15,150 ± 4,284 cpm (n = 9); LTα and LTβR-Ig, 10,702 ± 602 cpm (n = 5); p = 0.04); LTα and HVEM-Ig, 7,371 ± 2,592 cpm (n = 9; p = 0.0002); Fig. 4,a). Syngeneic cpm ranged from 20–30% of allogeneic cpm. IFN-γ levels were decreased in MLC using B6.LTα−/− SpC with the addition of LTβR-Ig or HVEM-Ig (LTα-control, 2,312 ± 958 pg/ml (n = 5); LTα and LTβR-Ig, 368 ± 348 pg/ml (n = 4; p = 0.006); LTα and HVEM-Ig, 884 ± 274 pg/ml (n = 4; p = 0.02; Fig. 4 b).
LTβR-Ig Adv and HVEM-Ig Adv reduced GVHD-induced weight loss in MHC class II-disparate GVHD
To assess the roles of LIGHT and LTαβ2 in MHC class II-disparate GVHD, experiments used B6 donor SpC and BMC transfer into MHC class II-disparate, bm12 × B6F1 mice. Lethally irradiated bm12 × B6F1 recipients mice were infused with B6 SpC and BMC. LTβR-Ig Adv, HVEM-Ig Adv, or control (β-gal Adv) cells were injected into the lateral tail vein of the BMT recipients at the time of transplantation. Eight to 12 days after transplant, bm12 × B6F1 recipients of B6 CD4+ Spc and BMC that received either LTβR-Ig-Adv or HVEM-Ig Adv exhibited less weight loss than the recipients of the control Adv (13.14 ± 10.9% (n = 10; p = 0.0004) or 16.42 ± 9.88% (n = 13; p = 0.0008) vs 30.43 ± 8.79% (n = 13); Fig. 5 a).
In separate experiments, BMT recipients of CD27-Ig Adv (Adv encoding FLAG fused to the extracellular domain of CD27 (TNF receptor family member) fused to a murine IgG1 FcR) demonstrated similar weight loss as recipients of Adv-β-gal (17.7 ± 4.7% weight loss (n = 4) vs 20.0 ± 7.1% weight loss (n = 3); p = 0.5888). With similar weight differences observed with both LTβR-Ig and HVEM-Ig and no difference with the FLAG-CD27-Ig construct, LIGHT signaling appears to be the critical component of these interactions in the development of MHC class II-disparate intestinal GVHD induced by B6 CD4+ T cells.
LTβR-Ig Adv and HVEM-Ig Adv reduced the histological features of intestinal GVHD in bm12 × B6F1 recipients of B6 CD4+ T cells
In other experiments, bm12 × B6F1 mice that received the HVEM-Ig and the LTβR-Ig or control Adv were harvested 8–12 days post-transplantation, and sections of the large intestines were stained with H&E. Histological features of apoptosis of crypt cells and cystically dilated crypts containing necrotic debris were observed in colons obtained 8–12 days after transplantation in recipients of B6 CD4+ T cells and control Adv, but only grade 1 GVHD with apoptosis of crypt cells was noted in the recipients of B6 CD4+ T cells and LTβR-Ig Adv or HVEM-Ig Adv.
In additional studies, H&E-stained sections of colon were obtained 8–12 days after bm12 × B6F1 were injected with the B6 donor CD4+ splenic T cells, and histopathology was graded as previously described by Snover (36). Colonic tissue was harvested 8–12 days after transplantation from 11 recipients of B6 CD4+ splenic T cells and control Adv, from 10 recipients of B6 CD4+ splenic T cells and either LTβR-Ig Adv or HVEM-Ig Adv, and from five recipients of syngeneic donor cells (B6 × bm12F1). The colonic tissue was fixed, sectioned, stained with H&E, and scored as detailed in Materials and Methods. The large intestines isolated from bm12 × B6F1 recipients of the B6 CD4+ T cells and the control Adv had evidence of a higher grade of colonic GVHD (2.4 ± 0.63; n = 11) than the five recipients of the B6 cells and the LTβR-Ig Adv or the five recipients of the HVEM-Ig Adv (0.84 ± 0.8; p = 0.0007) and (1.4 ± 0.5; p = 0.008). The histological scores of the recipients of the syngeneic bm12 × B6F1 BMC were low (0.2 ± 0.4; n = 5; Fig. 5, left panel). Thus, inhibition of LIGHT signaling reduces, but does not completely eliminate intestinal GVHD induced by the transfer of B6 donor T cells into MHC class II-disparate bm12 × B6F1 recipients.
LTβR-Ig Adv and HVEM-Ig Adv decreases the number of infiltrating donor T cells in the intestine in recipients of the B6 CD4+ T cells
T cells were isolated from the intestines of the recipients of the B6 CD4+ T cells and either the control Adv or LTβR-Ig Adv or HVEM-Ig Adv. Intestinal lymphocytes from bm12 × B6F1 recipients that had received HVEM-Ig Adv, LTβR-Ig Adv, or the control Adv were assessed for absolute numbers of CD4+ T cells. In the intestine, both LTβR-Ig Adv and HVEM-Ig Adv decreased number of intestinal CD4+ T cells (0.35 ± 0.4 × 106 (n = 6) vs 0.36 ± 0.02 × 106 (n = 9); p = 0.03 and p = 0.007) compared with numbers of intestinal CD4+ T cells in recipients of control Adv (0.86 ± 0.42 × 106; n = 9; Fig. 5, right panel).
These studies report that LIGHT/HVEM interactions on CD4+ T cells promote optimal stimulation of IFN-γ production and alloresponses during MHC class II-disparate MLC and GVHD. This conclusion is based on results of studies using two different inhibitors. These inhibitors are a chimeric inhibitor protein LTβR-Ig (25, 26), which inhibits LIGHT/LTβR and LTαβ2/LTβR interactions on macrophages and LIGHT/HVEM interactions on T cells, and another chimeric inhibitor protein, HVEM-Ig, which inhibits only the LIGHT/LTβR and LIGHT/HVEM interactions. We observed similar effects on CD4+ T cell alloresponses with addition of either HVEM-Ig or LTβR-Ig, which supported the hypothesis that LIGHT signaling alone was sufficient for optimal CD4+ T cell alloresponses. Because similar results were observed in experiments using anti-LIGHT Abs, our studies suggest that the inhibitors are interacting with LIGHT and not another unknown ligand.
Importantly, understanding the effect of LIGHT in this MHC class II-disparate system is found by reviewing LIGHT signaling pathways. LIGHT, by binding to LTβR, can cause activation of NFκB in the macrophage, which, in turn, may up-regulate Th1 cytokines indirectly (37). Therefore, engagement of LIGHT/LTβR by inducing additional macrophage activation and up-regulation of NFκB and cytokine secretion (IL-12) may be critical for the optimal alloresponses in an MHC class II-disparate MLC. However, in studies using responder T cells that lack the IL-12R β2 subunit of the IL-12R (28), IL-12 signaling to IL-12R on T cells was found not to be essential for the optimal effect of these inhibitors. Hence, indirect activation and Th1 polarization of the CD4+ T cell by the macrophage through IL-12-IL-12R interactions was not essential for the effects of these inhibitors. Furthermore, by using LTα−/− responder SpC, the ligand LTαβ was also found not to be essential for optimal MHC class II-disparate MLC alloresponses. Hence, by assessing alloproliferation and IFN-γ responses of B6.IL12R−/− responder SpC after LIGHT blockade, the studies have demonstrated that IL-12 (from macrophages)-IL-12R (on T cells) interactions are not critical for optimal effects of LIGHT.
Alternatively, LIGHT signaling through HVEM can induce activation of NF-κB in the T cell, which, in turn, may up-regulate Th1 cytokines directly (37), which is the mechanism that probably occurs within the MHC class II-disparate MLC. Our studies support other investigators’ findings that T-T cell interaction through LIGHT is required for optimal T cell activation and expansion in an APC-deficient environment (38); they also demonstrate the importance of LIGHT’s costimulatory activity in an MHC class II-disparate CD4+ T cell alloproliferative response.
Although previous studies have demonstrated that blockade of LIGHT ameliorates acute MHC class I- and II-disparate GVHD by anergizing host-specific CTL (23), our studies demonstrate that LIGHT’s costimulatory effect reduces MHC class II-disparate GVHD by reducing CD4+ T cell alloproliferation. The decrease in absolute number of intestinal CD4+ T cells is associated with an improvement in the histological intestinal GVHD score. Although some investigators report that LIGHT-HVEM is important only in the generation and maintenance of CD8+ T cell responses, others have indicated its importance in CD4+ T cell responses and Th1 cytokine responses (24, 39, 40). Our studies differ from previous reports in the use of an MHC class II-only-disparate system.
MHC class II-disparate intestinal GVHD induced by B6 CD4+ SpC has been described previously (7, 8). The severity of colonic GVHD has differed depending upon the use of IL12R−/− donor cells (score, 1.1 ± 1.5) to those scores of colons isolated from the bm12 × B6F1 recipients of the control B6 CD4+ T cells, where seven of 10 BMT recipients had evidence of grade II or higher GVHD (7, 8). The close correlation between the severity of colonic disease and the level of IFN-γ expression in vivo in previous studies suggests that colonic disease is directly associated with IFN-γ production. In all strain combinations examined to date (6, 7, 8), the most impressive effect of the TNF receptor family has been its effects on colonic GVHD and associated IFN-γ responses. These findings suggest that the effects of TNF family members on Th1 cytokine responses play a major role in the evolution of intestinal GVHD.
Both HVEM-Ig and LTβR-Ig decrease not only the MHC class II-disparate GVHD-induced weight losses and gut GVHD, but also intestinal CD4+ T cells. Similar down-regulation of intestinal GVHD by both inhibitors suggests that LIGHT signaling is sufficient for CD4+ T cell alloresponses, which are responsible for gut GVHD. Like other investigators that have shown that LTα and LTαβ were not required for the induction phase of MHC class I- and II-disparate GVHD by using LTα−/− donor SpC and have supported the involvement of LIGHT in the induction of GVHD (23), we have shown the importance of LIGHT in the induction of MHC class II-disparate GVHD.
With regard to mucosal lymphocytes, LIGHT was shown to be important in the number of donor T cells in the intestine during GVHD. Other investigators have demonstrated that LIGHT transgenic mice spontaneously developed severe chronic colitis with infiltration of the intestine over 5–6 mo, and LIGHT blockade ameliorated the severity of T cell-mediated disease (3). In contrast, our studies showed that in an acute GVHD process of 1- to 2-wk duration, T cell expansion within the intestines was dependent on LIGHT signaling. Thus, expression of LIGHT in T cells of the mucosa is essential in both acute and chronic intestinal inflammatory processes. LIGHT may also alter donor T cell migratory patterns and thus account for the T cell expansion in the intestine. The lower number of intestinal T cells, induced by the inhibitors, may have been associated with fewer cells undergoing apoptosis or a decrease in proliferation. In vitro experiments at least in part support the inhibitor’s effect on proliferation.
Of interest, the specific gut GVHD scoring system developed by Snover (36) for human GVHD directly correlated with number of T cells in the lamina propria in murine GVHD after LIGHT inhibition, implicating a mechanism for gut GVHD involving T cell numbers. Although numerous costimulatory molecules have been shown to improve GVHD, LIGHT has been shown to improve a specific aspect of gut GVHD with a lower number of intestinal lamina propria T cells. The mechanism is likely a T-T cell interaction, because IL-12 responses are not required for optimal alloproliferative responses.
In conclusion, similar decreases in CD4+ T cell alloresponses were noted with both LTβR-Ig and HVEM-Ig with both control B6 responder SpC and IL-12R-deficient B6 responder SpC, supporting the hypothesis that LIGHT effects were independent of IL-12 signaling from macrophages to IL-12R on T cells. Furthermore, both HVEM-Ig and LTβR-Ig reduced wasting disease, histopathological scores, and gut-infiltrating CD4+ T cells during MHC class II-disparate GVHD, supporting the hypothesis that LIGHT signaling is critical for the development of gut GVHD and proliferation of disease-producing CD4+ T cells.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by Veterans Affairs Merit Grant 65 (to G.B.) and National Institutes of Health Grant R01HL69006-01A1 (to G.B.).
Abbreviations used in this paper: BMT, bone marrow transplantation; Adv, adenovirus; BMC, bone marrow cell; β-gal, β-galactosidase; GVHD, graft-vs-host disease; HVEM, herpes simplex virus entry mediator; LIGHT, a cellular ligand for herpes virus entry mediator and lymphotoxin receptor; LT, lymphotoxin; LTβR, LTβ receptor; SpC, spleen cell.