Induction of tissue-specific experimental autoimmune diseases involves an obligatory adjuvant effect to trigger an innate response of a type that will drive a Th1-biased adaptive response. This is achieved by use of CFA containing mycobacteria (Mycobacterium tuberculosis), whose recognition by cells of the innate immune system depends on TLRs that signal through the adaptor molecule MyD88. We examined the role of selected components of the MyD88 pathway in promoting experimental autoimmune uveitis (EAU). Mice deficient in MyD88, TLR2, TLR4, or TLR9 were immunized with the retinal Ag interphotoreceptor retinoid-binding protein in CFA, and their EAU scores and associated immunological responses were examined. MyD88−/− mice were completely resistant to EAU and had a profound defect in Th1, but not Th2, responses to autoantigen challenge. Surprisingly, TLR2−/−, TLR4−/−, and TLR9−/− mice were fully susceptible to EAU and had unaltered adaptive responses to interphotoreceptor retinoid-binding protein. Examination of IL-1R family members, which share the common adaptor MyD88 with the TLR family, revealed that IL-1R-deficient mice, but not IL-18-deficient mice, are resistant to EAU and have profoundly reduced Th1 and Th2 responses. These data are compatible with the interpretation that TLR9, TLR4, and TLR2 signaling is either not needed, or, more likely, redundant in the adjuvant effect needed to induce EAU. In contrast, signaling through the IL-1R plays a necessary and nonredundant role in EAU and can by itself account for the lack of EAU development in MyD88 mice.
Toll-like receptors are a family of pattern recognition receptors which recognize distinct molecular patterns associated with microbial pathogens (1, 2). Thus far, 11 TLRs have been identified and many of their ligands are known. TLR2 is involved in the response to microbial lipoproteins and peptidoglycans such as those in Gram-positive bacteria and yeasts, whereas TLR4 is required for the recognition of endotoxin of Gram-negative bacteria (1). TLR9 is the receptor for bacterial CpG-DNA (3). These receptors constitute the first line of defense against many pathogens and play a crucial role in the function of the innate immune system by activating NF-κB and other signaling pathways to produce inflammatory cytokines (4). TLRs share homology with other members of the IL-1R/TLR superfamily, which includes the receptors for IL-1 and IL-18. Signal transduction by members of this family requires an adapter molecule, MyD88 (5, 6).
The activation of the innate immune system by TLRs in turn influences the development of adaptive immune responses. MyD88−/− animals fail to generate proinflammatory and Th1 responses when stimulated with TLR ligands (6, 7). These animals are highly susceptible to infection with a wide variety of different pathogens, including Staphylococcus aureus (8), Listeria monocytogenes (9, 10), Toxoplasma gondii (11), and Mycobacterium tuberculosis (MTB)2 (12, 13). MyD88-deficient mice are also highly susceptible to Leishmania major infection associated with a polarized Th2 response (14). These findings have led to the conclusion that MyD88 plays a critical role in host resistance to microbial infection as well as adaptive immunity.
Induction of experimental autoimmune diseases, such as experimental autoimmune encephalomyelitis (EAE), experimental autoimmune orchitis, and experimental autoimmune uveitis (EAU) by immunization of susceptible animals with the respective organ-specific autoantigens involves an obligatory adjuvant effect. Traditionally, this has been achieved by emulsifying the Ag in CFA that contains MTB (15, 16, 17, 18). MTB is the causative agent of human tuberculosis. It drives a predominantly cell-mediated immune response, characterized by the production of IFN-γ and other proinflammatory cytokines. TLR2 and TLR4 have been implicated in the immune response to MTB (19, 20, 21). In addition, Segal et al. (22) reported that CpG oligonucleotides, such as are present in bacterial DNA, are potent adjuvants for activation of autoreactive encephalitogenic T cells in vivo. Many, but not all, strains of mice and rats also require pertussis toxin (PT) as an adjuvant. We showed previously that PT enhances the Th1 response when used as part of a disease-inducing immunization regimen (23, 24, 25) and that it binds to TLR4 on dendritic cells (Z. Y. Wang, D. Yang, Q. Chen, C. A. Leifer, D. M. Segal, S. B. Su, R. R. Caspin, Z. O. M. Howard, and J. J. Oppenheim, submitted for publication). It is therefore reasonable to hypothesize that components of MTB and PT acting as ligands to stimulate TLRs are the source of the adjuvant effect necessary to facilitate induction of experimental autoimmune diseases.
The present study was undertaken to assess the role of the TLRs in the adjuvant effect involved in induction of EAU. Toward that end, we challenged animals deficient in MyD88, TLR2, TLR4, or TLR9 with a uveitogenic regimen of the retinal Ag interphotoreceptor retinoid-binding protein (IRBP) in adjuvant to induce a Th1-mediated adaptive response leading to EAU. Because both IL-1R and IL-18R are the members of the IL-1R/TLR family and share the adaptor molecule MyD88 for their signaling by interacting with the Toll/IL-1R domain (6), we also examined the susceptibility to the EAU induction challenge of mice deficient in IL-1R α-chain (IL-1RI) and IL-18. We found that the mice deficient in MyD88 or IL-1RI were completely resistant to EAU and have a profound defect in Th1 response to autoantigen challenge. In contrast, mice deficient in TLR2, TLR4, TLR9, or IL-18 were fully susceptible to EAU and had essentially unaltered immunological responses. Our data led to the conclusion that signaling through TLR2, TLR4, TLR9, and IL-18 receptors are either unnecessary, or redundant, for EAU induction, whereas IL-1R signaling is both required and nonredundant and can by itself account for the resistance of MyD88−/− mice to EAU.
Materials and Methods
MyD88−/−, TLR2−/−, TLR4−/−, or TLR9−/− mice generated by Dr. S. Akira (Osaka University, Osaka, Japan) (26) were backcrossed eight or more generations onto the C57BL/6 background, and were then intercrossed to obtain the knockout (KO) genotypes. Littermates of both sexes between 8 and 12 wk old were used in all experiments. IL-18−/− mice (on C57BL/6 background) were provided by Dr. K. Heeg (Munich, Germany) (27). Mice deficient in IL-1R (IL-1RI−/− on B6/129 background) and wild-type (WT) C57BL/6 mice were purchased from The Jackson Laboratory. WT B6/129S1 mice (controls for IL-1RI−/−) were obtained from Taconic Farms. Animals were kept in a specific pathogen-free facility and given water and standard laboratory chow ad libitum. Animal care and use were in compliance with institutional guidelines and with the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research.
Ag and reagents
CFA was purchased from Difco. MTB strain H37RA was purchased from Sigma-Aldrich. Purified Bordetella PT was purchased from Sigma-Aldrich. IRBP was isolated from bovine retinas, as described previously, using Con A-Sepharose affinity chromatography and fast performance liquid chromatography (28). IRBP preparations were aliquoted and stored at −70°C.
EAU induction and scoring
EAU was induced by active immunization with 150 μg of IRBP in PBS emulsified 1:1 v/v in CFA that had been supplemented with MTB to 2.5 mg/ml. A total of 200 μl of emulsion was injected s.c., divided among the base of the tail and both thighs. PT (0.5 μg/mouse) in PBS containing 2% normal mouse serum was given by i.p. injection concurrently with immunization. Clinical EAU was evaluated by fundoscopy under a binocular microscope after dilation of the pupil and was graded on a scale of 0–4 using criteria based on the extent of inflammatory lesions, as described in detail elsewhere (29). Eyes harvested 21 days after immunization were prefixed in 4% phosphate-buffered glutaraldehyde for 1 h (to prevent artifactual detachment of the retina) and then transferred to 10% phosphate-buffered formaldehyde until processing. Fixed and dehydrated tissue was embedded in methacrylate, and 4- to 6-μm sections were stained with standard H&E. Eye sections cut through pupillary-optic nerve planes were scored in a masked fashion. Severity of EAU was graded on a scale of 0–4 in half-point increments using the criteria described previously, based on the type, number, and size of lesions (29, 30).
To assess delayed-type hypersensitivity (DTH), 10 μg of IRBP in 10 μl of PBS was injected into the ear pinna. Ear thickness increment was measured 48 h later using a spring-loaded micrometer. The specific response was calculated as the difference between ear thickness after the IRBP injection minus ear thickness before the IRBP injection.
Lymphocyte proliferation assay
Draining (inguinal and iliac) lymph nodes were collected 21 days after immunization and were pooled within the group. Triplicate 0.2-ml cultures containing 5 × 105 cells were seeded in round-bottom 96-well microtiter plates. The RPMI 1640 (BioWhittaker) was supplemented with mouse serum, mercaptoethanol, antibiotics, glutamine, and nonessential amino acids, as described (31), and contained gradient IRBP started at 20 μg/ml as stimulant. The cultures were incubated for a total of 60 h. Tritiated thymidine (1 μCi/well) was added during the last 18 h. The data are shown as cpm.
For detection of Ag-induced cytokines in cell culture supernatants, inguinal and iliac lymph node cells from immunized mice were cultured in 24-well flat-bottom plates (5 × 106 cells/1 ml of culture medium per well) either alone or with IRBP at 20 μg/ml. Supernatants were collected after 48 h and were kept frozen in small aliquots at −70°C. Cytokines were measured by multiplex ELISA using the Pierce SearchLight technology (Pierce Boston Technology) (32) (http://www.searchlightonline.com). For intracellular cytokine detection, lymph node cells or splenocytes from immunized mice were cultured in 6-well-plates with IRBP at 20 μg/ml for 24 h, and Golgistop (BD Biosciences Cytofix/Cytoperm kit) was added during the last 10 h. The cells were surface stained with FITC-labeled CD4, CD8, CD11b, CD11c, NK1.1, or Gr-1 or allophycocyanin-labeled CD19, and then fixed, permeabilized, and stained with anti-IL-4, anti-IL-5, or anti-IL10 Abs using the BD Biosciences Cytofix/Cytoperm kit per the manufacturer’s instructions. The samples were analyzed by flow cytometry.
Determination of IRBP-specific IgG Ab isotypes
Serum levels of anti-IRBP IgG2a and IgG1 subclasses in IRBP-immunized mice were determined by ELISA as previously described (29). Briefly, 96-well microtiter plates (Costar) were coated with IRBP (1 μg/ml). After blocking the plates with BSA and overnight incubation with samples of the tested sera, the plates were developed using HRP-conjugated goat anti-mouse IgG1 or goat anti-mouse IgG2a Abs (Southern Biotechnology Associates). The amount of each isotype bound to the IRBP-coated wells was estimated from standard curves constructed by coating wells with the same goat anti-mouse IgG1 or goat anti-mouse IgG2a Abs and adding dilutions of Ig standards of the pertinent isotype.
Reproducibility and statistical analysis
Experiments were repeated at least twice. Results were highly reproducible. Figures show pooled data from repeat experiments, or representative experiments, as indicated. Each point is one mouse (average of both eyes). Statistical significance of differences in disease scores was calculated using Snedecor and Cochran’s test (33) for linear trend in proportions, with each mouse (average of both eyes) as one statistical event. This is a nonparametric test that generates its p values by frequency analysis of the number of individuals at each possible score, thus taking into account both severity and incidence of disease. Delayed hypersensitivity and lymphocyte proliferation data were analyzed using an independent t test. Probability values of p ≤ 0.05 were considered to be significant.
MyD88−/−, but not TLR2−/−, TLR4−/− and TLR9−/−, mice are resistant to EAU induction
To investigate the role of TLR/MyD88 signaling in the induction of EAU, MyD88−/−, TLR2−/−, TLR4−/−, and TLR9 −/− mice on the C57BL/6 background and their phenotypically normal littermates were immunized with the retinal Ag IRBP in CFA plus PT to induce EAU. Evaluation of eyes collected for histopathology on day 21 showed that MyD88−/− mice failed to develop EAU (Fig. 1,A). Surprisingly, the disease scores of mice deficient in TLR2, TLR4, or TLR9 were similar to those of control normal homozygous and heterozygous littermates (Fig. 1, B–D).
Mice on the C57BL/6 background typically develop fairly moderate EAU pathology, with scores of 0.5–1. Nevertheless, this level of disease is easily discernible on clinical observation as well as by histopathology. Eyes harvested 21 days after immunization showed characteristic EAU histopathology in WT littermates from the MyD88 line (Fig. 2,A), with retinal vasculitis, vitritis, choroiditis, and minimal or absent photoreceptor cell damage (Fig. 2,A). In contrast, the histopathology of eyes from homozygous MyD88−/− mice was completely normal (Fig. 2,B). The eyes of TLR2-, TLR4-, and TLR9-deficient mice, as well as those of their respective WT littermates, were indistinguishable from the EAU pathology depicted in Fig. 2 A (data not shown).
Adaptive immune responses in TLR−/− and MyD88−/− mice
DTH responses of TLR2-, TLR4-, TLR9-, MyD88-deficient mice and their normal littermates that had been immunized for EAU were elicited by ear challenge 19 days after immunization as described in Materials and Methods, and the responses were read on day 21. The specific increment in ear thickness showed that the DTH reaction of MyD88−/− mice in response to IRBP ear challenge was strongly reduced in comparison to WT littermates (Fig. 3,A). In addition, a gene-dose effect was apparent in that MyD88+/− littermates had an intermediate DTH response. TLR2-, TLR4-, and TLR9-deficient mice and their WT littermates had normal DTH responses to IRBP, similar to those of MyD88 WT controls. Responses of TLR9−/− mice and their controls are depicted in Fig. 3 B; responses of TLR2−/− and TLR4−/− showed the same pattern (data not shown).
In keeping with the DTH responses, proliferation to IRBP of draining lymph node cells collected 21 days after immunization was greatly decreased in MyD88−/− mice as compared with their WT littermate controls (Fig. 3,C), but was similar in both WT and deficient littermates of the TLR9 (Fig. 3 D), TLR2, and TLR4 lines (data not shown).
To evaluate the Ag-specific cytokine responses, supernatants were collected from IRBP-stimulated lymph node cell cultures and were assayed for cytokine content by multiplex ELISA as described in Materials and Methods. Production of Ag-specific IL-6, IFN-γ, and TNF-α by MyD88-deficient mice was profoundly impaired and IL-1α and IL-18 were reduced. Conversely, IL-4, IL-5, and IL-10 were significantly increased in MyD88−/− mice (Fig. 4,A). In contrast, Ag-specific production of IL-1α, IL-2, IL-4, IL-5, IL-6, IL-10, IL-12, IL-18, IFN-γ, and TNF-α were not significantly changed among null and normal littermates of TLR9 (Fig. 4 B), TLR2, and TLR4 lines (data not shown).
To identify the cells producing the elevated Th2-type cytokines, we performed immunofluorescent staining of Ag-stimulated lymph node and spleen cell cultures for intracellular IL-4, IL-5, or IL-10, combined with surface staining for T cells (CD4, CD8), B cells (CD19), monocyte/macrophages (CD11b), dendritic cells (CD11c), granulocytes (Gr-1), and NK cells (NK1.1). Flow cytometric analysis revealed that, interestingly, the cells with detectable cytoplasmic IL-4, IL-5, or IL-10 were not T cells. Both in MyD88−/− and in WT mice the majority of cells producing these cytokines were B cells (Fig. 5). The remaining cytokine-positive cells were distributed among the monocyte, granulocyte, and NK populations, whereas dendritic cells did not stain (data not shown). Notably, although the identity of cytokine-producing cells did not appear altered, MyD88−/− mice had more cytokine-positive cells and a more intense cytokine stain per cell compared with WT mice for all three cytokines, in line with their enhanced titers in the culture supernatants by ELISA.
Because IFN-γ and IL-4 promote Ab isotype switching to IgG2a and IgG1, respectively, the relative titers of these Ab isotypes are a good indicator of the type of response that develops to the Ag in vivo. We therefore assayed the anti-IRBP Abs of each isotype in sera of MyD88 KO and WT mice (Fig. 6). The results showed that WT mice made approximately equivalent amount of both isotypes, yielding a ratio of IgG1:IgG2a close to 1. MyD88-deficient mice made extremely low titers of serum IgG2a and similar amounts of IgG1 compared to their WT controls, confirming a predominantly Th2-like response (Fig. 6).
EAU and adaptive responses of IL-1RI- and IL-18-deficient mice
Because MyD88 is required not only for TLR, but also for IL-1 and IL-18 signaling, MyD88 deletion may affect EAU development by compromising the functions of IL-1 and IL-18. To test the role of these cytokines in the pathogenesis of EAU, we examined EAU induction and adaptive immune responses in IL-1R-deficient and IL-18-deficient mice. As shown in Fig. 7,A, IL-1RI-deficient mice were completely resistant to EAU. In contrast, IL-18 deficiency did not interfere with EAU development (Fig. 7 B).
DTH reaction to IRBP by IL-1RI-deficient mice was extremely low (Fig. 7,C) and proliferation to IRBP of explanted lymph node cells collected 21 days after immunization was greatly decreased (Fig. 7,D). DTH responses and Ag-specific lymphocyte proliferation of IL-18−/− mice were similar to their WT controls (data not shown). In keeping with the DTH and IRBP-specific proliferation data, production of Ag-specific IL-4, IL-5, IL-6, IL-10, IL-18, IFN-γ, and TNF-α by lymph node cells from IL-1RI-deficient mice was profoundly impaired, with a defect in both Th1 and Th2 responses (Fig. 8). IL-1α and IL-12p70 were not significantly changed from controls (Fig. 8). In contrast, production of Ag-specific cytokines was not significantly different between WT and IL-18-deficient mice (data not shown). These data suggest that unlike TLR2, TLR4, TLR9, and IL-18, signaling through the IL-1R has a necessary and nonredundant role in EAU development.
The discovery of mammalian TLRs has provided new insights into the association between innate responses to microbes and adaptive responses and consequently their impact on autoimmune diseases. Activation of TLRs leads to the induction of antimicrobial pathways central to innate defense as well as the up-regulation of Ag presentation molecules and secretion of cytokines that influence the nature of the adaptive immune response. MTB and PT as necessary adjuvants for the induction of experimental autoimmune diseases is an excellent model of the relationship of microbial infection and immunologically driven disease.
In this study, we used the EAU model to study the role of TLRs in the induction of autoimmunity. Because all TLRs signal through MyD88, we first examined whether MyD88−/− mice have an altered susceptibility to EAU induced by the standard immunization protocol involving IRBP in CFA plus PT. The data showed that MyD88 deficiency confers resistance to EAU. Interestingly, although the adaptive response in MyD88−/− mice exhibited strongly depressed IFN-γ and skewing toward production Th2-type cytokines, reminiscent of responses seen in IFN-γ-deficient mice, MyD88−/− mice did not exhibit signs of “Th2 disease” such as is seen in IFN-γ−/− mice and in immunodeficient mice transferred with Th2 effector cells (34, 35). A defect in Th1, but not Th2, response to IRBP in MyD88−/−mice is in line with published reports in other experimental systems (6, 7). Interestingly, intracytoplasmic staining revealed that the cells producing the Th2 cytokines IL-4, IL-5, and IL-10 in response to an in vitro IRBP recall were not T cells. Most of the producers were found in the B cell population, with a minority distributed among the monocyte/macrophage, granulocyte, and NK lineages. Whether this finding and lack of a “Th2-type” pathology in MyD88 KO mice are connected remains to be determined.
Since the MyD88 pathway appeared necessary for adjuvanticity, we reasoned that the Toll receptors involved in responses to MTB and PT, the two prototypic adjuvants that promote tissue-specific autoimmunity, should be the ones involved in their adjuvant effects. This would include TLR2 (peptidoglycans), TLR4 (LPS), TLR9 (methylated CpG DNA motifs) for MTB (19, 20, 21, 22), and TLR4 for PT (Z. Y. Wang, S. B. Su, R. R. Caspin, and J. J. Oppenheim, unpublished observations, and Ref.36). In contrast to the recent report by Waldner et al. (37), who demonstrated that T cell tolerance in a TCR-transgenic model of EAE could be broken by activating TLR9 or TLR4 with their cognate ligands, none of our single TLR-deficient strains, namely, TLR2−/−, TLR4−/−, and TLR9−/− showed inhibited EAU induction or changes in the associated immunological responses: DTH, proliferation, and cytokine production. Double KO mice deficient in TLR2 and TLR4, or TLR4 and TLR9, were also susceptible to EAU (data not shown). These findings lead to the conclusion that either these receptors are not involved, or, more likely, that signaling through TLR2, TLR4, and TLR9 is redundant with each other and/or with other receptors that utilize the MyD88 adapter molecule.
We next proceeded to examine the roles of IL-1 and IL-18, whose receptors are also known to utilize the MyD88 pathway in their signaling process (6), and may therefore contribute to the MyD88−/− phenotype independently of TLR-mediated effects. Thus, a number of investigators reported that disruption of the MyD88 signaling pathway by mutation, dominant negative forms, or targeted disruption attenuates IL-1R-mediated and IL-18-mediated NF-kB activation, with a consequent loss of functions that are dependent on signaling by these cytokines (6, 38, 39, 40). Thus, MyD88−/− mice display defects in T cell proliferation as well as induction of acute phase proteins and cytokines in response to IL-1, whereas loss of IL-18 signaling may contribute to defects in the Th1 response. IL-18 synergizes with IL-12 in inducing IFN-γ production from T cells and plays a role in polarization toward the Th1 pathway (6, 41). IL-18-deficient mice display defective Th1 response development after injection of Propionibacterium acnes and bacillus Calmette-Guérin in physiological responses (41). Interestingly, although IL-18 is constitutively expressed in the epithelial cells of the iris, ciliary body, and retina of the eye (42), IL-18 signaling did not seem to affect development of autoimmune uveitis. Earlier data from our laboratory (J. L. Wahlsten, L. M. Bagenstose, and R. R. Caspi, unpublished observations) indicated that IL-18 treatment of mice immunized for EAU failed to enhance disease, and IL-18-transgenic mice (developed by Hoshino et al. (43)) did not display enhanced susceptibility to EAU. Our present data, showing that IL-18-/ mice are fully susceptible to EAU induction, are consistent with these results as well as with a recent report by Jiang et al. (42) showing unaltered susceptibility to EAU in an independently derived line of IL-18−/− mice.
In sharp contrast to IL-18−/−, TLR2−/−, TLR4−/−, and TLR9−/− strains, the IL-1R-deficient animals immunized with IRBP/CFA/PT remained free of EAU, similarly to MyD88−/− mice. Thus, IL-1R signaling serves a necessary and nonredundant function in EAU induction and can by itself account for lack of EAU development in MyD88−/− mice. However, judging by their Ag-specific cytokine profile, IL-1RI−/− mice displayed a defect not only in Th1, but also in Th2 effector function. This could suggest that IL-1R may also utilize a MyD88-independent signaling pathway, since its disruption compromises EAU-associated responses in a more profound way than disruption of MyD88 signaling alone. Our observations underscore the central role of IL-1 in EAU and are consistent with studies in other autoimmune diseases (44, 45, 46, 47). IL-1 promotes collagen-induced arthritis in mice (44) and plays an important role in the pathogenesis of rheumatoid arthritis in humans (45). Mice deficient in IL-1-type I receptor were resistant to EAE (46). Eriksson et al. (47) showed that IL-1 receptor I on dendritic cells plays an important role in the induction of autoimmune myocarditis. This cytokine is a multifunctional player and stimulates the acute phase response, the secretion of matrix metalloproteinases, chemokines, and other proinflammatory cytokines. These diverse activities of IL-1 may help explain its central role in induction of autoimmunity. In view of its central importance for pathogenesis of autoimmune and inflammatory diseases, IL-1 constitutes a potential therapeutic target. Studies in the rat model of EAE (48, 49), and more recently also in EAU (50), confirm that treatment with the rIL-1R antagonist (IL-1RA) can suppress induction of disease, supporting the possible clinical utility of this approach.
In summary, although TLR2, TLR4, and TLR9 play an important role in responses to cell wall components of MTB to their methylated CpG DNA and to PT, our data in single TLR gene KO mice do not point to a unique role of these receptors in the induction of EAU. The data indicate that TLR2, TLR4, and TLR9 are either unnecessary or have redundant roles in EAU induction. Our findings do not negate the possibility that TLR signaling plays a role, however, double and triple TLR KO may need to be generated to establish the role of TLRs in induction of autoimmune diseases such as EAU. IL-1R signaling, on the other hand, is necessary and nonredundant and can by itself account for the resistance of MyD88−/− mice to EAU.
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.
Abbreviations used in this paper: EAE: experimental autoimmune encephalomyelitis; EAU, experimental autoimmune uveitis; MTB, Mycobacterium tuberculosis; PT, pertussis toxin; IRBP, interphotoreceptor retinoid-binding protein; KO, knockout; WT, wild type; DTH, delayed-type hypersensitivity.