Regulation of Th2 responses by different cell types expressing the interleukin-31 receptor

C57BL/6 mice were purchased from Sankyo Laboratories (Tokyo, Japan) and were housed in our facilities under specific pathogen-free conditions. All experiments were performed following the Animal Experimentation Guidelines of The Jikei University School of Medicine and the RIKEN Kobe Branch. The IL-31RA mutant (Accession No. CDB1012K: http://​www2.​clst.​riken.​jp/​arg/​mutant%20​mice%20​list.​html) was established as follows. To generate IL-31RA deficient mice, homologous recombination in embryonic stem (ES) cells was used to create a mutant allele in which exon 4 of the IL-31RA gene was replaced with a cassette expressing the selective marker neomycin transferase. In brief, homologous regions (5′5 8001 bp, 3′: 3074 bp) were subcloned into a knock-in vector (DT-A-pA/lox71/LacZ-pA/frt/PGK-Neo/frt/loxP/pA: http://​www2.​clst.​riken.​jp/​arg/​cassette.​html) and electroporated into mouse ES cells. Two recombinant ES cells were found to be IL-31RA^tm1(LacZ). Homologously integrated ES cells were injected into 8-cell-stage zygotes (Fig. 1a). Chimeric mice were mated with C57BL/6 mice to produce mutant IL-31RA^{+/LacZ(+/−)} progeny. The generation of mutant IL-31RA^+/− mice was verified by Southern blot analysis (Fig. 1b). Digestion with ApaI followed by hybridization of membranes with probe 1 (a 900-bp genomic DNA fragment obtained by PCR: GenBank AC154767.2, nt 49651–50550) yielded a 15,638-bp fragment for the WT allele and a 21,281-bp fragment for the correctly targeted mutant allele. Similarly, digestion of genomic DNA with NheI followed by hybridization of membranes with probe 2 (a 800-bp genomic DNA fragment obtained by PCR: GenBank NC_000079.6, nt 112555274–112556073) yielded a 16,632-bp fragment for the WT allele and a 10,930-bp fragment for the correctly targeted mutant allele. To reduce heterozygosity, the IL-31RA^+/− allele was then backcrossed to C57BL/6 mice for 15 more generations using IL-31RA^+/− males. Disruption of IL-31RA was identified by PCR using the corresponding primers to give a 602-bp product (primer1: 5′-caaagcgccattcgccattcaggctgcgca-3′,primer2: 5′-tgtgcattgtgagtgggtgagtggtatgca-3′). The presence of wild-type IL-31RA was identified as a 377-bp product (primer 2 and primer 3: 5′-tgaatttgcagaggaaagagaatgcccaca-3′).

Homozygous IL-31RA^−/− and WT (IL-31RA^+/+) littermates were generated by intercrossing IL-31RA^+/− mice from the 15th generation of backcrossed mice. Heterozygous IL-31RA^+/− and homozygous IL-31RA^−/− littermates were generated by crossing IL-31RA^+/− mice with homozygous IL-31RA^−/− mice.

To examine the expression of IL-31RA or beta-galactosidase (β-Gal) from WT C57BL/6 mice (IL-31RA^+/+) and IL-31RA-deficient (IL-31RA^−/−) mice, skin specimens were fixed for 3 days in 10% normal buffered formalin and embedded in paraffin using standard techniques. Five-micrometer sections were heated at 60 °C for 30 min for tissue adhesion. Slides were subsequently dewaxed by incubation in xylene and then rehydrated in 100, 95, and 70% EtOH. Finally, the slides were rinsed with Tris-buffered saline Tween buffer, and prepared as recommended by the manufacturer. Endogenous peroxidase was blocked with 4.5% H₂O₂ in MeOH for 30 min at room temperature. A protein block (phosphate-buffered saline block containing 10% normal goat serum) was applied overnight at 4 °C.

To examine the expression of IL-31RA, we prepared rabbit anti-mouse IL-31RA polyclonal antibodies as follows. A partial peptide of murine IL-31RA, mGPL 65–86 (YSDNATEASYSFPRSCAMPPDI), was synthesized by referring to [8] and conjugated to keyhole limpet hemocyanin. Rabbits were immunized with the conjugates and Freund’s Complete Adjuvant. After booster shots, whole blood from immunized rabbits was collected and serum was separated. Serum antibodies that recognize IL-31RA were purified using a column containing Sepharose-bound mGPL 65–86 peptide and the antigen-binding activity of the purified antibodies to the peptide was checked by ELISA. To examine the expression of β-Gal, an anti-beta β-Gal antibody (Abcam PLC, Cambridge, US) was used as a primary antibody and biotin-conjugated goat anti-chicken IgY (H+L) polyclonal antibody (GeneTex Inc, CA, USA) was used as a secondary antibody for β-Gal staining.

To induce allergen-specific Th2 responses, mice were administered an allergen nasally or intraperitoneally. For nasal administration, 2.5 µg of Cry j 2 (Hayashibara Biochemical Laboratories, Okayama, Japan), one of the major allergens of Japanese cedar pollen, was dissolved in 4 μl of PBS, and simply administered intranasally 15 times (5 times per week). For intraperitoneal immunization, mice were intraperitoneally injected with 20 µg of Cry j 2 in 0.2 ml of PBS without any adjuvants once a week for 5 weeks.

Seven days after the last immunization, the spleen and lymph node cells were collected. Cry j 2-specific T-cell proliferative responses were determined by an in vitro [³H] thymidine incorporation assay. RPMI-1640 medium (Thermo Fisher Scientific Inc., MA, USA) supplemented with 1% normal mouse serum was used to suspend cells and 8 × 10⁵ cells were seeded into each well of 96-well plates (Nunc Microwell 96F; Thermo Fisher Scientific Inc.) and cultured with 2.5 µg/ml of Cry j 2 for 88 h. The cultures were pulsed with 0.5 µCi of [³H] thymidine (American Radiolabeled Chemicals Inc., USA) for the last 16 h in Nunc Maxisorp plates (Kamnstrup, Denmark). The cells were harvested using a Micro 96 Harvester (Skatron Instruments, Norway), and [³H] thymidine uptake by the cells was determined by measuring the radioactivity using a liquid scintillation counter (LSC-6000; ALOKA, Tokyo, Japan). T-cell proliferation was expressed as a stimulation index (SI), which represents the ratio of [³H] thymidine incorporation in cultures with and without antigen. Responses with SIs greater than 2.0 were considered positive.

Spleen or lymph node cells were stimulated with 2.5 µg/ml of Cry j 2 and supernatants were collected at 72 h and stored at −20 °C for cytokine detection assays. The concentrations of IL-5, IL-13, and IFN-γ were determined using standard sandwich ELISA protocols (IL-5: BD Biosciences, NJ, USA; IL-13: Thermo Fisher Scientific Inc.; IFN-γ: BD Biosciences).

For the assay of Cry j 2-specific IgE, IgG1, IgG2a, and IgG2b, we used serum collected from mice 7 days after immunization. Antigen-specific IgE antibody titers were measured by IgE capture ELISA as previously described [9], with several modifications. Microtiter plates were incubated with anti-mouse IgE (BD Biosciences, CA, USA) capture antibody overnight at 4 °C, washed, and blocked with PBS containing 10% FCS for 1 h. The plates were incubated with serially diluted serum samples for 2 h. After washing, 0.3 µg/ml of biotinylated Cry j 2 was added to each well for further incubation for 1 h at room temperature. The plates were washed and streptavidin-horseradish peroxidase (HRP) conjugates (BD Biosciences) was added and incubated for 30 min. After 5 times washing, tetramethylbenzidine (TMB) substrate solution was added to each well and incubated for 30 min in the dark, and 2 N H₂SO₄ solution was added to each well to stop the reactions. The absorbance in each well was measured with a microplate reader (Bio-Rad, CA, USA). To calculate the anti-Cry j 2 IgE antibody titer, sera of five C57 BL/6 mice immunized five times with Cry j 2 in the peritoneal cavity were pooled and a standard curve was prepared to calculate anti-Cry j 2 IgE antibody by measuring total IgE. There was a correlation between total IgE and the Cry j 2-specific IgE antibody titer in the diluted samples from pooled sera. On the other hand, anti-Cry j 2 IgE antibody could not be detected in non-immunized mice (data not shown).

Antigen-specific IgG, IgG1, IgG2a, and IgG2b antibody titers were measured by indirect ELISA. Microplates, which were coated with Cry j 2 (1 µg/mL), were blocked with FCS, and serum samples were added to the wells and incubated for 2 h at room temperature. After the plates were washed, HRP-conjugated isotype-specific antibody was added to the wells. The bound antibodies were detected by adding TMB substrate solution, and color development was assessed by measuring the absorbance at 450 nm.

To investigate the function of endogenous IL-31–IL-31R interactions, IL-31RA-deficient mice were generated as IL-31RA knock-in mice by homologous recombination, as shown in Fig. 1a. Correct homologous recombination was confirmed by Southern blot analysis (Fig. 1b). The phenotypes were not significantly different in the two IL-31RA^+/LacZ(−) knock-in mouse lines. Inheritance of WT and mutant alleles was followed by PCR analysis of genomic DNA obtained from tail biopsies with pairs of primers specific for WT or mutant alleles. Digestion with ApaI followed by hybridization of membranes with a 5′ probe (900-bp) of a genomic DNA fragment obtained by PCR with oligos and a 3′ probe (800-bp) yielded a 15,638-bp fragment for the WT allele and a 21,281-bp fragment for the correctly targeted mutant allele. The IL-31RA^+/− allele was then backcrossed to C57BL/6 mice for 15 more generations using males IL-31RA^+/−. Disruption of IL-31RA was identified by PCR with the corresponding primers.

Homozygous IL-31RA^−/− and WT (IL-31RA^+/+) littermates were generated by intercrossing IL-31RA^+/− mice from the 15th generation of backcrossed mice. Heterozygous IL-31RA^+/− and homozygous IL-31RA^−/− littermates were generated by crossing IL-31RA^+/− mice with homozygous IL-31RA^−/− mice. As shown in Fig. 1c, the absence of IL-31RA expression in the skin of IL-31RA-deficient mice was confirmed by immunohistological staining with antibodies against IL-31RA or β-Gal. Staining with the anti-IL-31RA antibody revealed that IL-31RA was expressed in the hair roots of the skin of the WT mice but not in that of IL-31RA^−/− mice. In contrast, β-Gal was found to be expressed in the skin hair roots of the IL-31RA^−/− mice but not in that of WT mice.

In order to clearly analyze the differential effects of tissue-specific IL-31R signaling on Th2 immune responses, mice were injected nasally or intraperitoneally with the Th2 antigen Cry j 2. First, to examine the positive role of IL-31–IL-31R interactions in the skin on Th2 immune responses, IL-31RA-deficient and WT mice were injected nasally with Cry j 2, one of the major allergens of Japanese cedar pollen, which induces Th2 polarized and allergic inflammation [10]. Following repeated nasal administration of Cry j 2, the Cry j 2-specific serum IgE and IgG1 levels in IL-31RA-deficient mice were significantly lower than those observed in the serum of WT mice injected nasally with Cry j 2 (Fig. 2a). To evaluate the effects of tissue-specific IL-31R signaling on Th2 immune responses more accurately, we used the Cry j 2 allergen as an immunogen without any adjuvants to induce Th2 immune responses, as Cry j 2 has protease activity similar to that of Th2 adjuvants [11]. However, when Cry j 2 was administered in a mixture with Alum, there was no difference in the level of induced Th2 immune response between IL-31RA-deficient and WT mice (data not shown), even though Alum adjuvants strongly induce antigen-specific Th2 immune responses [12]. To confirm our previous observations that IL-31–IL-31R interactions positively influence Th2 immune responses following nasal administration of Cry j 2 (Fig. 2a), heterozygous IL-31RA^+/− and homozygous IL-31RA^−/− littermates were generated by crossing IL-31RA^+/− with IL-31RA^−/− mice. The littermates were nasally administered Cry j 2 and examined for Cry j 2-specific Th2 immune responses. Consistent with previous findings, in IL-31RA^−/−, Cry j 2-specific IgE and IgG1 (Fig. 2b) serum levels were significantly lower than those in IL-31RA^+/− mice. In addition, serum levels of IgG2b (Fig. 2b), which are influenced by Th2 immune responses [13], were also significantly lower in IL-31RA^−/− mice than in IL-31RA^+/− mice. In contrast, serum IgG2a (Fig. 2b), which is influenced by IFN-γ secreted by Th1-cells [14], did not differ between the two strains. Furthermore, in IL-31RA-deficient mice, not only antigen-specific proliferative responses (Fig. 3a) of spleen cells but also secretion levels of the Th2 cytokines IL-5 (Fig. 3b) and IL-13 (Fig. 3c) were significantly lower than those in WT mice injected nasally with Cry j 2. These results demonstrated that IL-31R signaling following nasal administration of an allergen can positively regulate the Th2 immune response, which is consistent with the findings of numerous previous studies that have shown that IL-31R signaling exacerbates Th2-associated immune responses [1‐3, 15‐17].

Furthermore, spleen cells isolated from IL-31RA^−/− mice nasally administered Cry j 2 showed significantly lower levels of Cry j 2-specific T-cell proliferative responses (Fig. 4a) and Th2 cytokine IL-5 and IL-13 secretion than did IL31 RA^+/− mice nasally administered Cry j 2 (Fig. 4b, c). In addition, IL31 RA^−/− mice nasally administered Cry j 2 in which submandibular lymph node cells were draining showed significantly lower levels of Cry j 2-specific Th2 cytokine IL-5 and IL-13 secretion than IL31 RA^+/− mice nasally administered Cry j 2 (Fig. 5a, b), whereas IFN-γ secretion was not detected in either strain of immunized mice (Fig. 5c). Collectively, these results strongly demonstrated that IL-31R signaling following nasal administration of an allergen can positively regulate Th2-associated immune responses.

To investigate whether IL-31R signaling in different cells and tissues influences the patterns of Th2 immune responses, IL-31RA-deficient (IL-31RA^−/−) and IL-31RA^+/− littermate mice were intraperitoneally administered Cry j 2. Following a repeated cycle of intraperitoneal administration with Cry j 2, the level of Cry j 2-specific serum IgE (Fig. 6a), IgG1 (Fig. 6b), and IgG2a (Fig. 6c) in IL-31RA-deficient mice were significantly higher than IL-31RA^+/− mice intraperitoneally administered Cry j 2. On the other hand, the serum level of IgG2a (Fig. 6c) did not differ between the two strains. Furthermore, in IL-31RA-deficient mice, the secretion levels of Th2 cytokines IL-5 and IL-13 (Fig. 7a, b) were significantly higher than those in IL-31RA^+/− mice injected intraperitoneally with Cry j 2. In contrast, IFN-γ production by Th1-cells did not differ between the two strains. These data are consistent with the increase in Th2 cytokines observed in the draining lymph node of the lungs in IL-31RA^−/− mice following S. mansoni egg injection [5], and support a role for IL-31–IL-31R interactions in limiting intestinal Th2 cytokine production following Trichuris infection [6]. These results demonstrate that IL-31R signaling in mice intraperitoneally administered Cry j 2 negatively regulates the Th2-associated immune responses, even though IL-31R signaling following nasal administration positively regulates the Th2 responses. Collectively, these data demonstrated that IL-31R signaling following nasal administration of an allergen positively regulates, whereas that following intraperitoneal administration negatively regulates, antigen-specific Th2 responses. These results indicated that regulation of the Th2 immune response depends on the cell types expressing IL-31RA, in which IL-31R signaling plays an important role.