Contact hypersensitivity (CHS) is a typical inflammatory response against contact allergens. Inflammatory cytokines, including IL‐1 and tumor necrosis factor (TNF)‐α, are implicated in the reaction, although the precise roles of each cytokine have not been completely elucidated. In this report, we dissected the functional roles of IL‐1 and TNF‐α during CHS. CHS induced by 2,4,6‐trinitorochlorobenzene as well as oxazolone was suppressed in both IL‐1α/β–/– and TNF‐α–/– mice. Hapten‐specific T cell activation, as examined by T cell proliferation, OX40 expression and IL‐17 production, was reduced in IL‐1α/β–/– mice, but not in TNF‐α–/– mice, suggesting that IL‐1 but not TNF‐α is required for hapten‐specific T cell priming in the sensitization phase. On the other hand, TNF‐α, induced by IL‐1, was necessary for the induction of local inflammation during the elicitation phase. We also found that the expression of IFN‐γ‐inducible protein 10 (IP‐10) was augmented at the inflammatory site. Although IP‐10 mRNA expression was abrogated in TNF‐α–/– mice, both CHS development and TNF‐α mRNA expression occurred normally in IFN‐γ–/– mice, indicating that the induction of IP‐10 during CHS was primarily controlled by TNF‐α. Interestingly, CHS was suppressed by treatment with anti‐IP‐10 mAb, suggesting a critical role for IP‐10 in CHS. Reduced CHS in TNF‐α–/– mice was reversed by IP‐10 injection during the elicitation phase. Thus, it was shown that the roles for IL‐1 and TNF‐α are different, although both cytokines are crucial for the development of CHS.
The development of contact hypersensitivity (CHS) requires several distinctive events including the capture of contact allergens by Langerhans cells (LC), LC migration into lymph nodes (LN) and maturation, hapten‐specific T cell activation within the LN, and leukocyte infiltration into the skin requiring a series of adhesion events: rolling, firm adhesion and eventual transmigration into inflamed regions (1,2). Various cytokines are produced by keratinocytes (KC) in the skin and LC in LN (2), and they are involved in both the sensitization and elicitation phases of CHS (1,2). The pro‐inflammatory cytokines IL‐1 and tumor necrosis factor (TNF)‐α may play crucial roles in this reaction, as CHS is greatly suppressed in mice deficient for these molecules (3,4). These cytokines promote LC migration (4–8) by regulating E‐cadherin expression on KC and LC (9,10), CCR7 expression on immature dendritic cells (11,12), and the ligands for CCR7, SLC and ELC, expression in LN (13). Adhesion molecules, ICAM‐1, VCAM‐1, and E‐ and P‐selectin on vascular endothelium, are up‐regulated by IL‐1 and TNF‐α during the elicitation phase to facilitate leukocyte rolling and firm adhesion in inflammatory sites (14,15). Although these cytokines are critical, various additional chemokines produced by KC, LC and lymphatic endothelial cells are also involved in leukocyte migration during CHS (2,16).
In earlier studies, it was reported that CHS was suppressed by the treatment with anti‐IL‐1β antibody, but not by anti‐IL‐1α antibody (17). CHS induced by low‐dose 2,4,6‐trinitorochlorobenzene (TNCB) sensitization, but not by high‐dose sensitization, was reduced in IL‐1β–/– mice on a 129 × B6 background (18), suggesting the involvement of IL‐1β in CHS. On the other hand, Zheng et al. reported that oxazolone‐induced CHS was normal in IL‐1β–/– mice (19), and we also showed that TNCB‐induced CHS is suppressed only in IL‐1α–/– and IL‐1α/β–/– mice, but not in IL‐1β–/– mice, on a C57BL/6 background using both low‐ and high‐dose TNCB, indicating that IL‐1α is responsible for the development of this reaction (4). Reduced CHS in IL‐1α/β–/– mice is caused by a defect in antigen‐specific T cell activation rather than impairment of LC migration into LN (4). IL‐1 potentiates T cell priming through the CD40 ligand (CD40L) and OX40 induction on CD4+ T cells (20), and specifically IL‐1α, which is produced by LC, plays a crucial role in the hapten‐specific T cell activation during CHS (4). The possible role for IL‐1 in the elicitation phase, however, remains to be elucidated.
CHS is suppressed by the injection with anti‐TNF‐α antibody (21). Likewise, oxazolone‐induced CHS is reduced in TNF‐α–/– and TNF receptor II–/– mice (3,6), suggesting the involvement of TNF‐α in the development of CHS—although, again, the mechanisms remain unclear. Furthermore, CHS suppression in these single cytokine‐deficient mice was incomplete, suggesting that additional cytokines or these cytokines mutually compensate for the deficiency. Therefore, although IL‐1 and TNF‐α are clearly involved in CHS, their individual functional roles in CHS pathogenesis, including a functional discrimination of each and their relationship, remain to be elucidated. The current study sought to address these points using IL‐1α/β–/–, TNF‐α–/– and IL‐1α/β–/– × TNF‐α–/– mice. Our results demonstrate that IL‐1 and TNF‐α play important roles in the sensitization phase and elicitation phase respectively. During the sensitization phase, both cytokines are implicated in the migration of LC from the skin to LN, although IL‐1 alone is required for hapten‐specific T cell activation. In the elicitation phase, however, IL‐1 induces TNF‐α expression that then activates the expression of IFN‐γ‐inducible protein 10 (IP‐10), a chemokine crucial for the recruitment of inflammatory cells into inflammatory sites.
IL‐1α–/–, IL‐1β–/–, IL‐1α/β–/–, IL‐1Ra–/– and IFN‐γ–/– mice were generated by homologous recombination as described. Mice from each group were backcrossed to C57BL/6J and BALB/cA mice respectively for eight generations (22,23). TNF‐α–/– mice were backcrossed to C57BL/6J and BALB/cA mice for 10 and eight generations respectively (24). IL‐1α/β–/– × TNF‐α–/– mice and IL‐1Ra–/– × TNF‐α–/– mice were obtained by intercrossing IL‐1α/β–/– mice and TNF‐α–/– mice, and IL‐1Ra–/– mice and TNF‐α–/– mice respectively. These mice were housed under specific pathogen‐free conditions in an environmentally controlled clean room at the Center for Experimental Medicine, Institute of Medical Science, University of Tokyo. The experiments were conducted according to the institutional ethical guidelines for animal experiments and the safety guidelines for gene manipulation experiments. Sex‐ and age‐matched mice of 8–12 week in age were used for experimentation.
TNCB (Tokyo Kasei, Tokyo, Japan)‐induced CHS was assayed as described previously (4,19). Briefly, the abdomens of mice were shaved and sensitized epicutaneously with 25 µl of 3.0% TNCB dissolved in acetone mixed with olive oil (4:1). Five days after sensitization, the outside of one ear (auricle) was challenged with 25 µl of 1.0% TNCB while the other ear was given 25 µl of vehicle alone. Mice were euthanized 24 h after TNCB challenge. A disk of ear tissue from both ears of each mouse was removed using a 6‐mm biopsy punch. Each disk was weighed and ear swelling was calculated as follows: increase of ear swelling = [(weight of TNCB challenged ear) – (weight of vehicle‐treated ear)]. For reconstitution or neutralization of CHS, either 25 ng of mouse rTNF‐α (Peprotech, Rocky Hill, NJ), 50 ng of mouse rIL‐1α/β mixture (Peprotech), 1 µg of mouse rIP‐10 (Peprotech) or 100 µg of anti‐mouse IP‐10 mAb (25) was injected intradermally into the ear skin immediately after TNCB challenge. At 24 h after TNCB challenge, ear swelling was then measured as described above.
Migration and maturation of LC
Measurement of LC migration and maturation was carried out as described (4). The backs and abdomens of mice were shaved and painted with 50 µl of 0.5% FITC isomer I (Sigma, St Louis, MO) dissolved in an equal volume mixture of acetone and dibutylphthalate. After 24 h, inguinal, axillary and brachial LN were harvested. Single‐cell suspensions were prepared from collagenase‐treated LN, and stained with biotinylated anti‐mouse CD11c mAb (HL3; PharMingen, San Diego, CA) and PerCP–streptavidin (PharMingen). The content of FITC+ cells within the CD11c+ population was analyzed to estimate migration of LC (5000 cells) from the skin to the LN using flow cytometry on a FACScan cytometer (Becton Dickinson, Mountain View, CA). To examine maturation state of the LC, CD11c+FITC+ cells in LN were analyzed for staining with either phycoerythrin (PE)–anti‐mouse CD40 mAb (3.23; Immunotech, Marseille, France) or PE–anti‐mouse CD86 mAb (RMMP‐1; Immunotech).
T cell response
Five days after 3.0% TNCB sensitization, single‐cell suspensions were prepared from inguinal, axillary and brachial LN. T cells were purified by the depletion of B220+ cells and Mac‐1+ cells using a MACS system (Miltenyi Biotec, Bergish Gladbach, Germany). To prepare TNP‐conjugated spleen cells, Thy‐1.2+ cells were also depleted by MACS columns. Cells were incubated with 100 mM trinitrobenzene sulfonate (Wako, Osaka, Japan) in PBS for 5 min at 37°C and irradiated at 35 Gy with 137Cs. T cells derived from TNCB‐sensitized mice (5 × 105 cells/well) were cultured with TNP‐conjugated spleen cells (2 × 105 cells/well) in 200 µl RPMI 1640 (Sigma) containing 50 µM 2‐mercaptoethanol (Gibco/BRL, Gaithersburg, MD), 50 µg/ml streptomycin (Meiji, Tokyo, Japan), 50 U/ml penicillin (Meiji) and 10% heat‐inactivated FCS (Sigma) in a 96‐well flat‐bottom plate for 72 h. Following a 72‐h incubation, cultures were pulsed with [3H]thymidine (0.25 µCi/ml) (Amersham, Little Chalfont, UK) for 6 h. Cells were harvested with a Micro 96 cell harvester (Skatron, Lier, Norway); incorporated [3H]thymidine was measured with a Micro Beta counter (Pharmacia Biotech, Piscataway, NJ). To measure OX40 expression, dinitrobenzene sulfonate (DNBS)‐sensitized LN cells were cultured for 72 h in the absence or presence of 40 µg/ml DNBS as described elsewhere (26). Cells were then incubated with FITC–anti‐mouse CD4 mAb (BD PharMingen) and PE–anti‐mouse OX40 (OX86) (Immunotech), and analyzed by flow cytometry. IL‐17 levels in culture supernatants were measured by ELISA using monoclonal rat anti‐mouse IL‐17 and polyclonal biotinylated goat anti‐mouse IL‐17 antibody (Dako, Carpinteria, CA) as capture and detection antibody respectively. Horseradish peroxidase–avidin was obtained from BD PharMingen and TMB substrate was purchased from Dako.
CHS of mice transferred with T cells
T cells (2×107 cells/mouse) derived from TNCB‐sensitized mice were suspended in PBS and injected i.v. into an unchallenged mouse. After 12 h, the outside of one ear was challenged with 25 µl of 1.0% TNCB, while the other was given vehicle alone. Then, 24 h after challenge with TNCB, ear swelling was measured as described above.
Induction of IP‐10 mRNA expression by injection with rIL‐1 and rTNF‐α
rIL‐1α/β (100 ng) was injected into the ear with or without anti‐TNF‐α mAb (XT22; Endogen). rTNF‐α (100 ng) was also injected into the ear. Total RNA was prepared from ear tissue 3 h after the injection. IP‐10 mRNA expression was then detected by Northern blot hybridization.
Northern blot hybridization analysis
Total RNA was prepared from the ear 3 and 24 h after either TNCB challenge or cytokine injection. Northern blot hybridization analysis was carried out as described (23).
Student’s t‐test was used for statistical evaluation of results.
Effects of IL‐1α/β and TNF‐α deficiency on the development of CHS
To elucidate the role of IL‐1 and TNF‐α in CHS development, the ear swelling response against TNCB was assessed in IL‐1α/β–/– and TNF‐α–/– mice of the C57BL/6J background. The CHS response was significantly suppressed in both TNCB‐sensitized IL‐1α/β–/– mice and TNF‐α–/– mice, indicating that both IL‐1 and TNF‐α are involved in the reaction (Fig. 1A). Without sensitization, no ear swelling was observed. The suppression, however, was not complete, indicating that additional cytokines may compensate for the deficiency. We then examined the possible mutual compensation between IL‐1 and TNF‐α using IL‐1α/β–/– × TNF‐α–/– mice. The suppression of ear swelling was significantly greater in double cytokine knockouts than in IL‐1α/β–/– or TNF‐α–/– mice, clearly indicating that IL‐1 and TNF‐α have distinct roles in the CHS response.
As the genetic backgrounds of mice can affect the CHS response (27), we also examined the effects of cytokine deficiency on mice of the BALB/cA background. We obtained similar results to those seen for the C57BL/6J background. Suppression was significantly greater in IL‐1α/β–/– × TNF‐α–/– mice than in either IL‐1α/β–/– or TNF‐α–/– mice (Fig. 1B).
We also examined effects of these deficiencies on the development of CHS using oxazolone treatment. We obtained similar results (data not shown), demonstrating that both IL‐1 and TNF‐α play important roles in CHS pathogenicity irrespective of allergens.
Effect of IL‐1α/β and TNF‐α deficiency on LC migration and maturation
LC of the skin are the major antigen‐presenting cells (APC) involved in CHS. Following painting of the skin with FITC, we examined the effect of cytokine deficiency on LC migration by measuring the migration of FITC‐labeled LC to regional LN by flow cytometry. Migration of FITC+ LC from the skin to the draining LN was impaired in IL‐1α/β–/– mice, consistent with previous observations (4); while 44% of CD11c+ LN cells from wild‐type C57BL/6J mice were labeled with FITC, only 27% of CD11c+ LN cells from IL‐1α/β–/– mice were labeled with FITC (Fig. 2A). LC migration was also suppressed in TNF‐α–/– mice, indicating that TNF‐α also plays a role in the migration of LC. Interestingly, the phenotype in triple cytokine knockout mice was more severe; <3% of CD11c+ cells in draining LN of IL‐1α/β–/– × TNF‐α–/– mice were labeled with FITC. Similar results were obtained in BALB/cA mice (data not shown). These observations suggest that these cytokines have partially redundant functions in the induction of LC migration.
CD40 and CD86 expression on FITC+ CD11c+ cells was similar among IL‐1α/β–/–, TNF‐α–/–, IL‐1α/β–/– × TNF‐α–/– and wild‐type mice on either the C57BL/6J (Fig. 2B) or BALB/cA background (data not shown). The expression of CD80 and ICAM‐1 was also similar among these mice (data not shown), indicating that LC in mutant mice mature normally, measured by cell surface molecule expression.
Effects of IL‐1α/β and TNF‐α deficiency on hapten‐specific T cell activation
To address the defects in CHS responses in TNF‐α–/– mice, we analyzed the hapten‐specific T cell response. Purified T cells from TNF‐α–/– and IL‐1α/β–/– mice, sensitized with TNCB, were co‐cultured with TNP‐conjugated splenocytes to examine their proliferative responses. The IL‐1α/β–/– T cell response was lower than the wild‐type response, in agreement with previous findings (4). The response of T cells from TNCB‐sensitized TNF‐α–/– mice, however, was comparable to that of wild‐type T cells (Fig. 3A), demonstrating normal T cell priming in TNF‐α–/– mice. We also examined OX40 expression on CD4+ T cells, as OX40 induction by IL‐1 is crucial for T cell activation (20). OX40–OX40 ligand co‐signaling is required for hapten‐ specific T cell priming and the development of CHS (26). Although OX40 expression on IL‐1α/β–/– T cells was lower than that on wild‐type T cells, expression on TNF‐α–/– T cells was comparable to wild‐type, consistently with normal hapten‐specific T cell proliferative responses (Fig. 3B). Moreover, we found that IL‐17 production, which reflects T cell activation in the sensitization phase of the CHS response (28), was also reduced in IL‐1α/β–/– mice, whereas that in TNF‐α–/– mice was comparable with that in wild‐type mice (Fig. 3C). These results suggest that TNF‐α is not involved in antigen‐specific T cell priming in CHS.
These results were further confirmed by T cell transfer experiments. Non‐sensitized wild‐type mice were given LN T cells from TNCB‐sensitized wild‐type, IL‐1α/β–/– or TNF‐α–/– mice, and then challenged with TNCB. Ear swelling was impaired in mice given transferred TNCB‐sensitized IL‐1α/β–/– T cells, but not TNF‐α–/– T cells (Fig. 3D). The impairment was observed in mice receiving sensitized IL‐1α/β–/– T cells from mice on either the C57BL/6J or BALB/cA backgrounds (Fig. 3D; data not shown for the BALB/cA background mice). These results indicate that IL‐1, but not TNF‐α, is required for T cell activation during the sensitization phase of CHS.
Roles of IL‐1 and TNF‐α in the elicitation phase of CHS
Next, we examined the role of IL‐1 and TNF‐α in the elicitation phase. We transferred TNCB‐sensitized wild‐type T cells into IL‐1α/β–/–, TNF‐α–/– or IL‐1α/β–/– × TNF‐α–/– mice to avoid effects on the sensitization phase, as IL‐1 deficiency affects T cell priming. Ear swelling responses upon challenge with TNCB were significantly impaired in mutant mice receiving wild‐type T cells sensitized with TNCB compared with wild‐type mice (Fig. 4A). Under these experimental conditions, mice that received non‐sensitized T cells did not develop the CHS response. The suppression was observed on both the C57BL/6J (Fig. 4A) and BALB/cA (data not shown) backgrounds. Thus, both IL‐1 and TNF‐α play important roles in the elicitation phase of CHS. These findings also suggest that both TNF‐α and IL‐1 mediate local inflammation during CHS through common mechanisms as the suppression in IL‐1α/β–/–, TNF‐α–/– and IL‐1α/β–/– × TNF‐α–/– mice was similar. Consistent with this possibility, mRNAs encoding TNF‐α and IL‐1α/β were reduced in IL‐1α/β–/– and TNF‐α–/– mice respectively in comparison with wild‐type mouse tissue (Fig. 4B and C). Thus, the results indicate that the expression of IL‐1α and β mRNA is induced by TNF‐α, and TNF‐α expression is induced by IL‐1α/β respectively in the elicitation phase of CHS, suggesting that these cytokines amplify the expression of each other.
We next examined the possibility that TNF‐α and IL‐1 act within a cascade. After transplantation with TNCB‐sensitized wild‐type T cells, we injected rTNF‐α or rIL‐1α/β into IL‐1α/β–/– mice or TNF‐α–/– mice respectively. IL‐1α/β–/– mice receiving rTNF‐α demonstrated strong responses comparable to those seen in wild‐type mice injected with rTNF‐α, although the wild‐type response was also enhanced by rTNF‐α (Fig. 4D). The response in TNF‐α–/– mice, however, was not affected by rIL‐1α/β administration (Fig. 4E). Similarly, rTNF‐α administration rescued the defects observed in IL‐1α/β–/– × TNF‐α–/– mice, while rIL‐1α/β could not compensate (data not shown). We also examined CHS using IL‐1 receptor antagonist (IL‐1Ra)–/– × TNF‐α–/– mice on the BALB/cA background. IL‐1Ra–/– mice exhibited exacerbated CHS compared with wild‐type mice, while the CHS response in IL‐1Ra–/– × TNF‐α–/– mice was suppressed to levels seen in TNF‐α–/– mice (Fig. 4F). Although H‐2 loci are b/b in the TNF‐α–/– mice that we used, the CHS response in wild‐type BALB/cA (H‐2d/d) was comparable with that in wild‐type BALB.B (H‐2b/b), indicating that the difference in H‐2 loci did not affect the results (data not shown). Taken together, these results suggest that IL‐1‐induced TNF‐α plays an important role in induction of local inflammation during the elicitation phase of CHS, although TNF‐α also can induce IL‐1α/β.
Effects of IL‐1α/β and TNF‐α deficiency on the expression of genes for adhesion molecules and chemokines in the inflammatory sites of CHS
We next investigated the expression of various adhesion molecule and chemokine genes suspected to play roles in the development of inflammation. Ear tissues from IL‐1α/β–/–, TNF‐α–/– and IL‐1α/β–/– × TNF‐α–/– mice were examined following the induction of CHS with TNCB. The expression of ICAM‐1 mRNA, but not E‐cadherin and E‐selectin mRNA, was severely reduced in IL‐1α/β–/–, TNF‐α–/– and IL‐1α/β–/– × TNF‐α–/– mice (Fig. 5A and B). Of the examined chemokine mRNAs, MCP‐1 mRNA expression was not impaired in any of the deficient mice. KC mRNA was reduced in only IL‐1α/β–/– × TNF‐α–/– mice, while the mRNA expression of MIP‐1α, MIP‐1β and IP‐10 was impaired in all mice deficient in either IL‐1α/β or TNF‐α (Fig. 5C and D). The suppression of IP‐10 mRNA expression was more severe in TNF‐α–/– mice than in IL‐1α/β–/– mice. As IFN‐γ is a potent inducer of IP‐10 (29), we also examined IFN‐γ mRNA expression in mutant mice; although IP‐10 mRNA levels in IL‐1α/β–/– mice were ∼50% of the wild‐type level, the expression of IFN‐γ was more severely affected in the mutant mice (Fig. 5E and F). Similar results were obtained using both the C57BL/6J and BALB/cA backgrounds (data not shown for BALB/cA mice). Although CHS was impaired only in IL‐1α–/– mice (4), the expression of ICAM‐1, MIP‐1α and MIP‐1β mRNA at the inflammatory site was reduced in both IL‐1α–/– and IL‐1β–/– mice (Fig. 6A and B). IP‐10 mRNA expression, however, was reduced in only IL‐1α–/– mice, demonstrating a correlation between IP‐10 expression and CHS sensitivity (Fig. 6A and B). Furthermore, IP‐10 mRNA expression in IL‐1α/β–/– mice receiving TNCB‐sensitized wild‐type T cells was lower than that seen in wild‐type mice (Fig. 6C). This expression, however, recovered to wild‐type levels following treatment with rTNF‐α after transplantation. These results suggest that, during the elicitation phase, TNF‐α induces IP‐10 expression, resulting in ear inflammation.
Critical role of IP‐10 in the elicitation of CHS
We examined the possibility that TNF‐α induces IP‐10 mRNA expression by injecting rTNF‐α intradermally into the ears of wild‐type mice. rTNF‐α injection increased IP‐10 mRNA expression in a dose‐dependent manner (Fig. 7A). IP‐10 mRNA expression was also strongly induced following the injection of rTNF‐α into IL‐1α/β–/– × TNF‐α–/– mice, consistent with the possibility that TNF‐α is induced by IL‐1 (Fig. 7B).
IP‐10 mRNA expression was analyzed in IFN‐γ–/– mice to evaluate the possible involvement of IFN‐γ in IP‐10 induction and CHS (Fig. 7C). IP‐10 mRNA was induced by either TNF‐α or IL‐1 even in the absence of IFN‐γ. Furthermore, IP‐10 mRNA expression, induced by rIL‐1, was suppressed by anti‐TNF‐α administration (Fig. 7C), suggesting that IL‐1 induces IP‐10 through TNF‐α. To exclude the possible involvement of IFN‐γ in CHS, we examined the sensitivity of IFN‐γ–/– mice against TNCB. IFN‐γ–/– mice on either the C57BL/6J or BALB/cA background demonstrated a similar sensitivity to CHS as wild‐type mice (Fig. 7D). Although the expression levels were slightly reduced, IP‐10 mRNA expression was clearly shown in the mutant mice and TNF‐α mRNA expression was normal in their ear tissues (Fig. 7E). These results suggest that TNF‐α directly induces IP‐10 mRNA expression in the ear through an IFN‐γ‐independent mechanism.
To address the role of IP‐10 in the development of CHS, we examined the effect of a neutralizing anti‐IP‐10 mAb. TNCB‐induced CHS in mice treated with antibody was significantly reduced in comparison with mice treated with control Ig (Fig. 8A), indicating IP‐10 plays an important role in CHS. Moreover, the ear swelling of TNF‐α–/– mice after challenge with TNCB was restored to the levels of wild‐type mice following the injection of rIP‐10 into the ear. This is not due to the inflammatory activity of IP‐10, as ear swelling was not observed without TNCB challenge (Fig. 8B). Similar observations was obtained using IL‐1α/β–/– × TNF‐α–/– mice injected with rIP‐10 (data not shown). Collectively, these observations indicate that TNF‐α induces IL‐1α first, then the IL‐1α induces TNF‐α, resulting in the induction of IP‐10 that plays a critical role in CHS elicitation in an IFN‐γ‐independent manner.
IL‐1α/β and TNF‐α are believed to exert similar activities in CHS responses to contact allergens, such as LC migration, LC maturation and the expression of adhesion molecules. The precise relationship between these cytokines and their distinctive roles in CHS, however, has not been well elucidated. In this report, we demonstrated that IL‐1 is required for hapten‐specific T cell priming during the sensitization phase of CHS, consistent with our previous report indicating that IL‐1 plays a crucial role in antigen‐specific T cell activation through the induction of CD40L and OX40 on T cells (4,20). In contrast, TNF‐α deficiency did not affect antigen‐specific T cell priming. Furthermore, TNF‐α‐deficient T cells sensitized with TNCB were capable of efficiently transferring CHS to wild‐type recipient mice, indicating that TNF‐α is not involved in the sensitization phase of CHS. We showed that TNF‐α plays an important role in the elicitation phase, because TNCB‐sensitized wild‐type T cells cannot evoke CHS in TNF‐α–/– mice upon transplantation. IL‐1 is also involved in the elicitation phase. This is because TNF‐α is induced by the action of IL‐1 and the deficiency of CHS in IL‐1–/– mice at the elicitation phase is recovered by the addition of TNF‐α. Thus, it was clearly shown that IL‐1 and TNF‐α play distinctive roles during CHS.
Although the involvement of other cytokines such as IL‐4 in CHS development is only obvious under conditions using specific allergens or specific genetic backgrounds (27), the effects of IL‐1 or TNF‐α deficiency were clearly observed using both TNCB and oxazolone as contact allergens on both the C57BL/6J and BALB/cA genetic backgrounds. Thus, the requirement for IL‐1 and TNF‐α may be critical for CHS development irrespective of the nature of the allergen and the genetic backgrounds.
LC are the major APC in hapten‐specific T cell activation during the sensitization phase (1). Although both IL‐1 and TNF‐α are involved in the migration of LC from the skin to LN, we demonstrated in a recent report that the suppression of LC migration in IL‐1α/β–/– mice was only transient; migration recovered gradually to levels similar to wild‐type mice by 36 h after stimulation (4). We obtained similar results in TNF‐α–/– mice (data not shown). Thus, the reduced CHS in IL‐1α/β–/– and TNF‐α–/– mice cannot be completely explained by a deficiency in LC migration.
Recently, it has been reported that the CD40–CD40L interaction is important for the induction of CHS (30). It was suggested that TNF‐α, elicited by the CD40–CD40L interaction, plays a critical role in LC migration and maturation. Our results, however, demonstrated that LC maturation, as examined by the expression of activation surface marker and T cell priming activity, was normal in TNF‐α–/– and IL‐1α/β–/– × TNF‐α–/– mice, indicating TNF‐α is not essential for these processes. It is likely that similar cytokines, such as granulocyte macrophage colony stimulating factor, may compensate the activity.
TNF‐α is produced by various cell types including KC and LC in the skin (2). The specific cells producing this cytokine during the elicitation phase of CHS, however, have not been identified. Although TNF‐α is known to be produced by CD4+ T cells (31), we demonstrated that T cell‐derived TNF‐α is not directly involved in the development of CHS (Fig. 3D). In regard to this, it is reported that mast cell‐derived TNF‐α plays an important role in CHS induction (32). CHS is induced normally in mast cell‐deficient Kitw/Kitw–v mice receiving wild‐type mast cells, while the response could not be restored in Kitw/Kitw–v mice receiving TNF‐α–/– mast cells (32). Thus, mast cells, in addition to KC and LC, may contribute to the production of TNF‐α.
ICAM‐1 (14,33), MCP‐1 (1,16), MIP‐1α and MIP‐1β (34) are involved in ear swelling during the elicitation phase. We showed that the expression of these genes was reduced in both IL‐1α/β–/– and TNF‐α–/– mice. The expression of these cytokines was also reduced in IL‐1β–/– mice, in which ear swelling develops normally (4,19). In addition, the CHS response was exacerbated in CCR5 (a receptor of MIP‐1β)‐deficient mice (35) and that in MCP‐1–/– mice developed normally (36). Therefore, these molecules do not appear to be involved in the suppression of ear swelling in IL‐1α/β–/– and TNF‐α–/– mice.
IP‐10 mRNA is expressed by KC in the skin (2,37) and IP‐10 plays an important role in the trafficking of effector Th1 cells to inflammatory sites (16,38–41). We found that the levels of IP‐10 mRNA correlated well with the severity of ear swelling in IL‐1α–/–, IL‐1β–/–, IL‐1α/β–/–, TNF‐α–/– and IL‐1α/β–/– × TNF‐α–/– mice. Furthermore, anti‐IP‐10 mAb significantly suppressed the development of ear swelling and the reduced CHS in TNF‐α–/– mice recovered to wild‐type levels following injection with rIP‐10. Thus, we conclude that IP‐10 plays a critical role in the development of CHS at the elicitation phase. In agreement with our notion, recently it was shown that CHS is reduced in IP‐10–/– mice (42). It should be noted, however, that IP‐10 mRNA expression was much more severely impaired in TNF‐α–/– mice than in IL‐1α/β–/– mice (Fig. 5C and D), although the suppression of ear swelling in IL‐1α/β–/– mice was similar to that in TNF‐α–/– mice (Fig. 1). Moreover, the CHS response was suppressed only partially by anti‐IP‐10 mAb treatment (Fig. 8). Thus, in addition to the IP‐10 induction by inducing TNF‐α, IL‐1 may also contribute to the local inflammation through IP‐10‐independent pathways during CHS response.
It is well known that IFN‐γ is a potent inducer of IP‐10 (29) and actually it was shown that IP‐10 is induced by IFN‐γ, but not by TNF‐α, upon infection with Trypanosoma cruzi (43). Recently, however, it was shown that TNF‐α could potentiate IP‐10 mRNA expression in hepatocytes more effectively than IL‐1β and IFN‐γ, while the induction of IP‐10 mRNA in Kupffer and endothelial cells by these cytokines was equivalent (44). We demonstrate here that IP‐10 in the skin is mainly induced by TNF‐α, not by IFN‐γ. Consistent with this observation, we showed that TNCB‐induced CHS was normal in IFN‐γ–/– mice, in agreement with the normal development of CHS in IFN‐γRI–/– mice (45). On the other hand, FITC‐induced CHS was reduced in IFN‐γRII–/– mice (46). This apparent discrepancy may be caused by the difference of the inducers (47). Normal expression of TNF‐α is observed in IFN‐γ–/– mice at inflammatory sites, supporting our claim that TNF‐α‐induced IP‐10 is responsible for the induction of inflammation.
In summary, we have demonstrated specific functions for IL‐1 and TNF‐α in the development of CHS. These findings may provide a critical cue allowing for the development of novel therapeutics to treat inflammatory diseases of the skin.
We would like to thank Dr Hideo Nariuchi (Institute of Medical Science, University of Tokyo) for his critical reading of the manuscript and valuable discussion. We would also like to thank Drs Shinobu Saijo and Jun Tanaka for their technical support, and all the members of our laboratories for excellent animal care. This work was supported by grants from the Ministry of Education, Science, Sport and Culture of Japan, the Ministry of Health and Welfare of Japan, CREST, and Pioneering Research Project in Biotechnology.
IP‐10—IFN‐γ‐inducible protein 10
IL‐1Ra—IL‐1 receptor antagonist
TNF—tumor necrosis factor
1Center for Experimental Medicine, Institute of Medical Science, University of Tokyo, 4‐6‐1 Shirokanedai, Minato‐ku, Tokyo 108‐8639, Japan 2Department of Molecular Preventive Medicine, School of Medicine, University of Tokyo, 7‐3‐1 Hongo, Bunkyo‐ku, Tokyo 113 0033, Japan 3Department of Immunology, National Institute of Animal Health, 3‐1‐1 Kannonndai, Tsukuba City 305‐0856, Japan 4Present address: Institute of Experimental Animals, Shinshu University School of Medicine, 3‐1‐1 Asahi, Matsumoto, Nagano 390‐8621, Japan 5Present address: Institute for Experimental Animals, School of Medicine, Kanazawa University, 13‐1 Takaramachi, Kanazawa 920‐8640, Japan