IκB-ζ Expression Requires Both TYK2/STAT3 Activity and IL-17–Regulated mRNA Stabilization

Psoriasis is an immune-related chronic skin disease characterized by erythematous, scaly, sharply demarcated plaques. On lesions, excessive proliferation of epidermal keratinocytes is observed with infiltration of neutrophils, macrophages, and activated T cells (1). IL-17A (IL-17) is a key inflammatory cytokine present within psoriasis lesions (2–4) and is mainly produced from immune cells, such as activated T cells. IL-17 has a direct effect on the induction of genes expressed by keratinocytes involved in innate immune defense and a range of CXC chemokines that regulate immune cell trafficking (5–7). This immune cell feedback has been considered to amplify and prolong psoriatic inflammation (2, 8). Biologic agents neutralizing IL-17 (i.e., secukinumab and ixekizumab) or antagonizing its receptor (i.e., brodalumab) are used as therapeutic drugs and show a high clinical efficacy for treating psoriasis (9). Elucidating the signaling mechanism of IL-17–induced keratinocyte activation is important for understanding psoriasis pathogenesis.

Tyrosine kinase-2 (TYK2) is a JAK family kinase member and is activated in response to various cytokines, including type I IFNs (IFN-α/β), IL-12, and IL-23 (10–12). We previously reported that TYK2 deficiency attenuates imiquimod (IMQ)- and IL-23–induced ear thickening resulting from epidermal hyperplasia with inflammatory cell infiltration, suggesting that signals through TYK2 are required for psoriasis-like skin inflammation in mice (13, 14). Genome-wide association studies identified TYK2 as a psoriasis susceptibility gene (15–17). Because TYK2 is involved in IL-12/IL-23 signaling that activates IFN-γ/IL-17 production in T cells (12), suppression of psoriasis-like pathologic condition due to TYK2 deficiency has been mainly attributed to suppressed T cell activity (14, 18). It has not been sufficiently verified that TYK2 in keratinocytes may have a role in the IL-17 response, despite TYK2 protein expression in keratinocytes. In this study, we aimed to elucidate the role of TYK2 in the keratinocyte IL-17 response.

The IκB-ζ protein (also known as INAP or MAIL, encoded by NFKBIZ) is highly expressed in the epidermal keratinocytes of psoriatic lesions and is considered to be involved in psoriasis pathogenesis (19). The NFKBIZ gene is located in a psoriasis susceptibility locus on 3q12.3 (17). We have reported that IκB-ζ is an IL-17–inducible protein and mediates IL-17–induced gene expression (20). IκB-ζ is a nuclear IκB family protein that positively or negatively modulates NF-κB–dependent transcription depending on the cellular context (21). The contribution of IκB-ζ in the pathogenesis of IMQ-induced skin inflammation in mice seemed to be larger than the contribution of IL-17 (19), presumably because IκB-ζ has multiple induction pathways in addition to IL-17 signaling, as exemplified by other psoriasis-associated cytokines such as IL-17F, IL-1β, or IL-36 (22, 23). The transcriptional activator of IκB-ζ expression has been suggested to be transcription factors NF-κB (23, 24) and/or STAT3 (23, 25).

Signaling pathways activated by IL-17 control mRNA stability (26). The importance of IL-17–induced stabilization of mRNAs in inflammation, as well as its multiple mechanisms, has become increasingly evident (27–30). Regnase-1 (also known as MCPIP1, encoded by ZC3H12A) is the IL-17–inducible protein with endoribonuclease activity that acts as a negative feedback regulator for inflammatory signaling induced by TLR ligands or various cytokines, including IL-17 (31–35). Regnase-1 has been shown to be degraded rapidly through the ubiquitin–proteasome pathway in LPS-treated macrophages and activated T cells, as well as IL-17–, IL-1β–, or IL-36α–treated keratinocytes (33, 36–38). Stimulus-induced decrease of Regnase-1 protein levels has been considered to release a brake on mRNA expression. Recent studies have shown that the 3ʹ-untranslated region (3ʹUTR) of IκB-ζ mRNA is required for the recognition, degradation, and/or translational suppression by Regnase-1 (39, 40). Altogether, these findings suggested that IL-17 signaling upregulates IκB-ζ through quantitative and/or qualitative inhibition of Regnase-1; however, this has not been sufficiently investigated.

In this study, we showed that TYK2 positively regulates IL-17–induced IκB-ζ expression in keratinocytes. TYK2-deficient mice showed decreased inflammation and concomitant reduction of IκB-ζ mRNA compared with wild-type (WT) mice in the IMQ-induced skin inflammation model. Small interfering RNA (siRNA)–mediated TYK2 knockdown reduced IL-17–induced expression of IκB-ζ and its target genes in keratinocytes. The TYK2–STAT3 pathway enhanced IκB-ζ promoter activity in a TYK2 kinase activity–dependent manner, but STAT3 phosphorylation was not enhanced by IL-17 stimulation. IL-17 stimulation posttranscriptionally stabilized IκB-ζ mRNA through Regnase-1 inhibition independently of TYK2. Our data showed that the activity of the TYK2–STAT3 pathway and IL-17–induced inhibition of Regnase-1 are regulated independently of each other, and both of them are required for IL-17–induced IκB-ζ upregulation.

Human TYK2 cDNA was a gift from Dr. J. Ihle (St. Jude Children’s Research Hospital, Memphis, TN). Full-length cDNA of TYK2, kinase-negative TYK2 (harboring a K930R mutation in the ATP-binding site), and the kinase domain–deleted TYK2 (ΔKD: aa 1–867) were generated by PCR and were subcloned into the pCS2-MT vector for mammalian expression. The plasmid expressing the TYK2 kinase domain (KD; aa 833–1187) was described previously (41). Expression vector for FLAG-tagged STAT3C, a constitutively active form of STAT3, has been described previously (42). The firefly luciferase reporter pGL3-mIκB-ζ (mNfkbiz)-3ʹUTR-Full (1–1353 nt after the stop codon), pFLAG-CMV2-mZc3h12a (mRegnase-1)-WT, and pFLAG-mZc3h12a (mRegnase-1)-D141N were kindly provided by Dr. O. Takeuchi (Kyoto University, Kyoto, Japan) (39). The mutant forms of Regnase-1 (C157A and D225/226A) were generated by site-directed mutagenesis. The pFN21A clone expressing an N-terminal HaloTag fusion of human full-length ACT1 (HaloTag-ACT1) was obtained from the Kazusa DNA Research Institute (Kisarazu, Japan). A fluorescent reporter construct, pCAG-mVenus-Nfkbiz-3ʹUTR, was generated in two steps: first, pCAG-mVenus was constructed from pCAG-YFP (43) (gift from C. Cepko; Addgene plasmid number 11180) by replacing YFP with the monomeric Venus coding sequence, which was obtained by cutting mVenus N1 (44) (gift from S. Vogel; Addgene plasmid number 27793) with restriction enzymes EcoRI and NotI. The 150-nt portion of mouse Nfkbiz 3ʹUTR after the stop codon, which contains two Regnase1-recognition stem-loops and plays a crucial role in mRNA stability (39), was inserted in the pCAG-mVenus downstream of the Venus gene using the In-Fusion HD cloning kit (Clontech). Stem-loop–disrupting mutations in the reporter plasmid were inserted by PCR. The secondary structures of the intact and mutated Nfkbiz 3ʹUTR mRNA were predicted by the CentroidFold Web server (45) (http://www.ncrna.org). Human NFKBIZ promoter–luciferase reporter (IκB-ζ-promoter-Luc; HPRM11632-PG04) was purchased from GeneCopoeia (Rockville, MD). This promoter clone simultaneously expresses naturally secreted Gaussia luciferase (Gluc) under the control of the promoter sequence (−1148 to +98) of NFKBIZ transcript variant 1 (NM_031419) and secreted alkaline phosphatase under the control of the CMV promoter, which serves as the internal control. The mutated promoter clones for each of two putative STAT-binding sites in the NFKBIZ promoter region were generated by PCR. All primers used for the plasmid constructions are listed in Supplemental Table I. Recombinant human IL-17 was purchased from R&D Systems (Minneapolis, MN). Cycloheximide (CHX), actinomycin D, and pyridone-6 were purchased from Wako Pure Chemical Industries (Osaka, Japan). Cerdulatinib was obtained from MedChem Express (Monmouth Junction, NJ). Tofacitinib/CP-690550 was obtained from Phoenix Pharmaceuticals (Burlingame, CA).

Tyk2^−/− mice, BALB/c background, were described previously (13). Mice were kept under specific pathogen-free conditions and provided with food and water ad libitum. All experiments were performed according to the guidelines of the Institutional Animal Care and Use Committee of Hokkaido University. At 8–10 wk of age, mice received a daily topical dose of 10 mg of commercially available IMQ cream (5%, Beselna Cream; Mochida Pharmaceuticals, Tokyo, Japan) on the side of the ears for three consecutive days. To evaluate the severity of ear skin inflammation, affected ear thickness was measured. At the days indicated, ear thickness was measured using the thickness gauge (Mitutoyo, Kawasaki, Japan) and averaged. After application for three consecutive days, ears were collected for quantitative PCR analysis.

Primary normal human epidermal keratinocytes (NHEKs) were purchased from KURABO (Tokyo, Japan) and cultured using the DermaLife K Comp Kit (KURABO). The human keratinocyte cell line HaCaT and the human cervix carcinoma cell line HeLa were maintained in DMEM containing 10% FCS. For siRNA transfection, the Lipofectamine 2000 reagent (Life Technologies, Carlsbad, CA) was used. The siRNAs targeting human TYK2, ZC3H12A, TRAF3IP2, and STAT3 used in this study are purchased from Shanghai GenePharma (Shanghai, China) and are listed in Supplemental Table II.

Cells were harvested and total RNA was prepared using the TRI Reagent (Molecular Research Center, Cincinnati, OH). First-strand cDNA was synthesized from 1 μg of total RNA with ReverTra Ace (TOYOBO, Osaka, Japan). Quantitative real-time PCR (qPCR) analysis of transcripts was carried out using a combination of a KAPA SYBR FAST qPCR master mix (KAPA Biosystems, Woburn, MA) with a Mx3005P real-time PCR system (Stratagene, Santa Clara, CA). qPCR primers used in this study are listed in Supplemental Table I.

IL-19 protein levels in culture supernatants of IL-17–treated cells were assayed by the Quantikine Human IL-19 Immunoassay kit according to the manufacturer’s instructions (R&D Systems).

HeLa cells were transfected with IκB-ζ-promoter-Luc, together with each expression vector for human TYK2, its kinase mutants (K930R, ΔKD, and KD), or a constitutively active form of STAT3 (STAT3C). At 24 h posttransfection, cells were treated with 100 ng/ml IL-17 and incubated for an additional 1.5 h. The culture supernatants were collected, and Gluc activities were measured using a Secrete-Pair Dual Luminescence Assay kit (GeneCopoeia) according to the manufacturer’s instructions.

HeLa cells were transfected with luciferase reporter plasmid pGL3 containing the mouse Nfkbiz 3ʹUTR (full length: 1353 nt after the stop codon), together with the expression plasmid for human TYK2 or empty (control) plasmid. The gene encoding Renilla luciferase was transfected simultaneously as an internal control. After 24 h of cultivation, treatment with 100 ng/ml IL-17, and a subsequent 1.5 h of incubation, cells were lysed and luciferase activity in lysates was determined with the Dual-Luciferase Reporter Assay System (Promega, Madison, WI). The Venus reporter assay was conducted for further analysis. pCAG-mVenus-Nfkbiz-3ʹUTR (150 nt) or its stem-loop–disrupted mutant in the 3ʹUTR, together with each expression vector for Regnase-1, Regnase-1 mutants (D141N, C157A, or D225/226A), or ACT1, was introduced to HeLa cells. At 24 h posttransfection, cells were treated with 100 ng/ml IL-17 and incubated for an additional 1.5 h. The number of Venus-expressing cells, which is considered to reflect mRNA stabilization/degradation control through the 3ʹUTR, was counted using flow cytometry (FACSCalibur; Becton Dickinson), and the IL-17–induced fold change of the number of Venus-positive cells was calculated.

Immunoblotting was performed as described previously (46). Briefly, cell lysates were resolved on SDS-PAGE and transferred to PVDF transfer membranes (PerkinElmer, Boston, MA). The filters were then immunoblotted with each Ab. Immunoreactive proteins were visualized using an ECL detection system (Millipore, Bedford, MA). Anti–IκB-ζ (number 9244) and anti-TYK2 (number 9312) Abs were from Cell Signaling Technology (Beverly, MA); anti–β-defensin 2 (FL-64) and anti-STAT3 (C-20) Abs were from Santa Cruz Biotechnology (Santa Cruz, CA). Anti–Myc-tag (9E10) and anti-actin (AC15) Abs were from Sigma-Aldrich (St. Louis, MO); anti–phospho-STAT3 (Y705) Ab (EP2147Y) was from Abcam (Cambridge, U.K.).

HaCaT cells in a 6-cm dish were exposed to 10 μM cerdulatinib or DMSO for 1 h. Chromatin immunoprecipitation (ChIP) assays were performed as described previously (47), with some modifications. The Abs used in this study for immunoprecipitations were rabbit monoclonal anti-STAT3 Ab (79D7; Cell Signaling Technology) and normal rabbit IgG (negative control). After incubation with the Ab, DNA–protein–Ab complexes were collected using Dynabeads M-280 Sheep Anti-Rabbit IgG (Thermo Fisher Scientific). Eluted DNA was analyzed by qPCR. The following sequence-specific ChIP qPCR primers were used to amplify two different NFKBIZ promoter regions: the promoter region upstream of transcriptional start site (TSS) 1 for NFKBIZ isoform 1 (NM_031419) and the TSS2 for NFKBIZ isoform 2 (NM_001005474) (Supplemental Table I). The promoter regions for TSS1 and TSS2 harbor putative STAT3-binding motifs. The negative-control primer was designed in an intronic region of the NFKBIZ gene lacking putative STAT3-binding motifs. Input DNA (5% of sample) was used as a control in each reaction. The relative ChIP amplification levels of each fragment were presented as percentages of total inputs in three experiments.

Data are expressed as mean ± SEM. For statistical comparison of two groups, a Student t test was performed. For multiple comparisons, an ANOVA and Tukey test were conducted. A p value <0.05 was chosen as an indication of statistical significance.

We investigated the role of TYK2 in IκB-ζ gene expression in skin inflammation using TYK2-deficient mice. In psoriasis-affected skin, IL-17 produced by T cells induces neutrophil infiltration and production of antimicrobial peptides (such as β-defensins) from keratinocytes, thereby exacerbating psoriasis symptoms (48). We used a psoriasis-like skin inflammation model induced by topical application of an imidazoquinoline derivative, IMQ (49). As we reported earlier, by applying IMQ to the ear pinna of WT mice for three consecutive days (at days 0, 1, and 2), a significant auricular thickening was observed at days 2 and 3 (Fig. 1A). In contrast, application of IMQ to TYK2-deficient mice resulted in significant auricular thickening at day 3, but the degree was significantly milder than that of WT mice. The influence of TYK2 deficiency on the mRNA expression of IL-17–responsive genes IκB-ζ (Nfkbiz), Il19, and Defb2 in the ear skin was examined. In IMQ-treated WT mice, expression of these IL-17–responsive genes was enhanced. In TYK2-deficient mice, IMQ application did not significantly increase IL-17–responsive genes (Fig. 1B). Of note, the level of Il17a mRNA in the IMQ-treated ears of TYK2-deficient mice was lower than in those of WT mice but was still firmly detected (Fig. 1B). These results suggested that TYK2 deficiency resulted in the decreased induction of IL-17–responding as well as IL-17 genes in the IMQ-treated lesions.

IκB-ζ is important for IL-17–induced gene expression and inflammation, and we examined the role of TYK2 in NFKBIZ (encoding IκB-ζ) expression. The human NFKBIZ gene has two transcript variants driven by two distinct promoters (Fig. 2A). This feature and the overall gene structure of NFKBIZ are conserved between humans and mice (Supplemental Fig. 1). NFKBIZ transcript variant 1 (NM_031419), which is transcribed from the TSS1, was upregulated at 1.5 h after IL-17 addition in control siRNA-transfected HaCaT cells. The induction was significantly attenuated in TYK2 siRNA-treated cells (Fig. 2A). NFKBIZ transcript variant 2 (NM_001005474), which is transcribed from the alternative TSS2 and has a slightly distinct mRNA sequence because of differential exon usage, was also induced by ∼3-fold at 1.5 h after IL-17 addition, and it was not affected by TYK2 knockdown (Fig. 2A). These results suggested that transcript variant specificity exists in TYK2-mediated regulation of NFKBIZ expression. We have reported that IL-17–induced expression of DEFB4A and IL19 is mediated by IκB-ζ (20). In TYK2-knockdown HaCaT cells, the IL-17–induced expression of these targets was reduced compared with control siRNA-treated cells (Fig. 2B). IL-17–induced secretion of IL-19 protein was significantly decreased by siRNA knockdown of TYK2 (Fig. 2C). Similar results showing TYK2 involvement in the IL-17 response were obtained in NHEKs (Fig. 2D) and HeLa cells (data not shown). Thus, TYK2 may have an effect on gene expression in keratinocytes and other cells responsive to IL-17.

The mechanism of TYK2-mediated upregulation of IκB-ζ was investigated. First, a luciferase reporter plasmid expressed under the control of the promoter region sequence (−1198 to +98) upstream of the TSS1 of the IκB-ζ gene (IκB-ζ-promoter-Luc) was transiently introduced into HeLa cells (Fig. 3A). We investigated the effects of transient overexpression of TYK2 and IL-17 stimulation on transcriptional induction of IκB-ζ (Fig. 3B, 3C). The luciferase activity was significantly enhanced by TYK2 overexpression. In contrast, IL-17 did not activate the IκB-ζ-promoter-Luc. Next, we introduced plasmids expressing WT TYK2, kinase-inactive point mutant (K930R), kinase domain–deleted mutant (ΔKD), or the kinase domain (KD) into HeLa cells (Fig. 3D, 3E, 3F). In cells transfected with TYK2 K930R and TYK2 ΔKD, which do not have catalytic activity, IκB-ζ promoter activation was not induced (Fig. 3E). Overexpression of TYK2 KD was sufficient for enhancing IκB-ζ promoter activity. These results indicated that TYK2-mediated IκB-ζ-promoter activation depends on its kinase activity.

The 3′UTR downstream of the protein coding region of IκB-ζ mRNA affects IκB-ζ mRNA stability via regulatory stem-loop elements conserved to mammalian species including humans and mice (39). A reporter plasmid (IκB-ζ-3′UTR-Luc), in which a 3′UTR sequence derived from murine IκB-ζ mRNA was inserted downstream of a luciferase gene expressed under the constitutively active CMV promoter (39), was used to determine the effects of TYK2 overexpression and IL-17 stimulation (Fig. 3G, 3H, 3I). TYK2 overexpression did not affect 3ʹUTR-regulated reporter activity, whereas IL-17 stimulation showed a considerable enhancing effect (Fig. 3H). These data suggested that the TYK2-mediated signaling pathway is involved in the transcriptional regulation of IκB-ζ expression. IL-17 was found to primarily enhance the stability of IκB-ζ mRNA posttranscriptionally.

Regarding the posttranscriptional mRNA stabilizing effect of IL-17 via the 3ʹUTR IκB-ζ mRNA sequence, our flow cytometric assay using the fluorescent protein (Venus)–based reporter fused to the truncated IκB-ζ 3ʹUTR (1–150 nt from stop codon) showed that this region of the IκB-ζ 3ʹUTR can respond to IL-17 stimulation (Fig. 4A, 4B). This region reportedly has two characteristic highly conserved stem-loop structures (denoted as “e” and “f” in Fig. 4A) that are required for the recognition, degradation, and/or translational suppression by an RNase, Regnase-1/MCPIP1 (encoded by ZC3H12A) (39, 40). The two stem-loops in the Venus-IκB-ζ 3ʹUTR reporter were mutated to disrupt these secondary structures (Fig. 4A). The loop-mutant reporter showed enhanced expression compared with the control intact reporter under unstimulated conditions, whereas it did not respond to IL-17 (Fig. 4B). Knockdown of endogenous Regnase-1 in HaCaTs or NHEKs by siRNA resulted in an accumulation of basal IκB-ζ mRNA levels, but it did not significantly affect IκB-ζ mRNA levels in IL-17–treated conditions (Fig. 4C). A recent study showing that IL-17 stimulation caused Regnase-1 protein degradation (33) has implied that the decrease of functional Regnase-1 protein could be a common mechanism between IL-17 stimulation and siZC3H12A for target mRNA stabilization. This overlapping mechanism would account for the nonsignificant effect of siZC3H12A on the accumulation of IκB-ζ mRNA, especially after IL-17 stimulation. Overexpression of either WT Regnase-1 or the C157A mutant, which lacks a deubiquitinase function but retains RNase activity (50), significantly suppressed Venus-IκB-ζ-3ʹUTR reporter expression (Fig. 4D), suggesting that the 3ʹUTR-mediated suppression is deubiquitinase-independent. The introduction of RNase-defective mutants of Regnase-1 (D141N and D225/226A) (36, 50) showed a possible dominant-negative effect on endogenous Regnase-1 and resulted in enhanced basal expression of the IκB-ζ-3ʹUTR reporter and unresponsiveness to IL-17 (Fig. 4D). These results suggested that Regnase-1–mediated degradation of IκB-ζ mRNA may occur constitutively and that the IL-17–induced signal may counteract Regnase-1–mediated mRNA degradation.

An IL-17 signaling adaptor protein, ACT1 (encoded by TRAF3IP2) (51), has been shown to mediate IL-17–induced upregulation of IκB-ζ mRNA in keratinocytes (17), although it is unclear whether ACT1 is required for IκB-ζ 3ʹUTR-mediated mRNA stabilization in IL-17–treated cells. We found that IL-17–induced expression of Venus-IκB-ζ-3ʹUTR reporter was abrogated by ACT1 knockdown (Fig. 4E). In addition, ACT1 overexpression counteracted the suppressing function of Regnase-1 on Venus-IκB-ζ-3ʹUTR reporter expression (Fig. 4F). These results suggested that IL-17/ACT1 signaling counteracts the constitutively occurring Regnase-1–mediated degradation of IκB-ζ mRNA.

Regnase-1 is considered to destabilize target mRNAs in a protein translation–dependent manner, as the translation inhibitor CHX can block mRNA decay mediated by Regnase-1 (39). Consistently, inhibition of the constitutive activity of Regnase-1 by CHX resulted in IκB-ζ mRNA accumulation in cells not treated with IL-17 (Fig. 4G). This CHX-induced accumulation of IκB-ζ mRNA was not significantly increased by combined stimulation with IL-17 (Fig. 4G). The treatment with transcription inhibitor actinomycin D completely abrogated CHX-induced accumulation of IκB-ζ mRNA in HaCaT cells (Fig. 4G), suggesting that CHX-induced accumulation of IκB-ζ mRNA is probably the reflection of the rate of IL-17–unrelated constitutive transcription of the IκB-ζ gene.

We speculated that TYK2-mediated STAT3 activation may be required for constitutive transcription of IκB-ζ because STAT3 reportedly plays a critical role in maintaining epithelial cell survival via IκB-ζ induction (25). siRNA-mediated knockdown of STAT3 or TYK2 showed significant suppression of CHX-induced IκB-ζ accumulation (Fig. 5A). Western blot analysis showed that tyrosine phosphorylation of STAT3, which is a hallmark for STAT3 protein activation, constitutively occurs in HaCaT cells, and it was attenuated by TYK2 siRNA introduction (Fig. 5B). Treatment with IL-17 did not increase STAT3 phosphorylation levels (Fig. 5B), which seemed to be in accordance with data showing that IL-17 did not activate IκB-ζ promoter activity (Fig. 3B). ChIP experiments using the anti-STAT3 Ab showed that STAT3 constitutively binds to the genomic promoter region of IκB-ζ TSS1 in nontreated HaCaT cells (Fig. 5C). STAT3 did not show specific binding to the promoter region of IκB-ζ TSS2 (Fig. 5C), the induction of which was unrelated to TYK2 expression (Fig. 2A), suggesting that the constitutive transcription of IκB-ζ is mediated by the TYK2–STAT3 pathway in a promoter region–specific manner. The promoter region of IκB-ζ TSS1 contains two putative STAT-binding sites (Fig. 5D). The IκB-ζ-promoter-Luc responded to the plasmid overexpression of STAT3C, a constitutively active mutant of STAT3 (Fig. 5D). We mutated each of the putative STAT-binding sites in the IκB-ζ-promoter-Luc and performed luciferase assays to test the significance of STAT3 activity. As shown in Fig. 5D, the deletion of site 1 (−133 to −125) abolished the STAT3C-mediated promoter activation. In contrast, the deletion of site 2 (−462 to −453) did not affect the promoter activity. These results indicated that the TYK2–STAT3 pathway drives IκB-ζ gene transcription with no particular requirement of IL-17 signaling.

Similar to TYK2 knockdown (Fig. 2A), siRNA-mediated knockdown of STAT3 resulted in the suppression of IL-17–induced upregulation of IκB-ζ mRNA in HaCaT cells (Fig. 6A), suggesting that the transcription of IκB-ζ by STAT3 is required for IL-17–induced upregulation of IκB-ζ. However, IL-17 did not stimulate STAT3 phosphorylation (Fig. 5B). To further verify the significance of TYK2–STAT3 pathway in the induction of IκB-ζ, we pharmacologically inhibited STAT3 activity using three small-molecule JAK inhibitors (cerdulatinib, pyridone-6, and tofacitinib), each of which has been reported to have distinct JAK selectivity (52–54) (summarized in Fig. 6B). Cerdulatinib and pyridone-6, compared with tofacitinib, have relatively low IC₅₀ values for TYK2 (52). IκB-ζ mRNA induction by IL-17 was suppressed by each of these three JAK inhibitors in a dose-dependent manner, with a relatively low effectiveness of tofacitinib (Fig. 6C). These JAK inhibitors efficiently suppressed the constitutive phosphorylation level of STAT3 (Fig. 6D). ChIP analysis showed that cerdulatinib addition abrogated constitutive STAT3 binding to the IκB-ζ TSS1 promoter region (Fig. 6E). Moreover, cerdulatinib treatment suppressed IL-17–induced increase of IκB-ζ and β-defensin 2 proteins (Fig. 6F). These results suggested that the activity of the TYK2–STAT3 pathway defines the strength of IL-17–induced expression of IκB-ζ in keratinocytes.

JAK/TYK2 inhibitors have been developed and intended to be used for psoriasis therapy (55–57). The basis of efficacy of TYK2 inhibition has been considered to be the suppression of lymphocyte activation, including IL-12/IFN-γ and IL-23/IL-17 axes (13, 18). In the current study, we investigated the role of TYK2 in IL-17 responsiveness in epithelial (nonlymphocyte) cells and demonstrated that TYK2-mediated maintenance of STAT3 phosphorylation is a substantial driver for the keratinocyte IL-17 response via IκB-ζ gene transcription. This finding expands our understanding of the mechanism of action of clinically used TYK2 selective inhibitors. We speculated that TYK2 inhibition can suppress IL-17–induced keratinocyte activation in addition to IL-12/IFN-γ and IL-23/IL-17 axes, thus showing high clinical efficacy.

The transcription factor STAT3 is a well-known kinase substrate of JAK/TYK2 and has a major function in the signal transduction of psoriatic keratinocytes in response to cytokine/growth factors (23, 58–61). In the current study, we showed that the TYK2-mediated STAT3 phosphorylation is a rate determinant for the IL-17–unrelated transcription of the IκB-ζ gene and that pharmacological inhibitors of JAKs could be used to suppress the transcription. In psoriatic lesions, the persistently enhanced phosphorylation of STAT3 in keratinocytes occurs more as the effects of various cytokines become pronounced, such as IL-19, IL-21, IL-22, IL-25, and IL-36, which are upregulated in lesions (23, 60, 61). As a result, enhanced transcription of the IκB-ζ gene by STAT3 occurs, and this STAT3 action is the therapeutic target of JAK/TYK2 inhibitors. Our previous study (14) demonstrated that TYK2-deficient mice possessed diminished skin inflammation and psoriasis-like pathologic condition after direct IL-22 injection into ear pinna. We showed that TYK2 plays a role in STAT3 phosphorylation in IL-22–treated keratinocytes. These previous observations supported the idea that TYK2 is required not only for T cell–dependent inflammation but also for non–T cell reactions in a dermatitis pathologic condition, such as keratinocyte activation. The promoter activation effect by WT TYK2 overexpression was only ∼2-fold in HeLa cells (Fig. 3B). We speculate that background STAT3 activity had an effect to worsen the experimentally observed signal/noise ratio in the reporter system that is transcriptionally activated by STAT3. Indeed, in HeLa (data not shown) and HaCaT (Fig. 5B) cells, phosphorylated STAT3 is constantly detected even without introducing exogenous TYK2. Nevertheless, the result that overexpression of WT but not kinase-inactive mutant TYK2 enhanced IκB-ζ promoter activity (Fig. 3E) supported the contribution of catalytic activity of TYK2.

We demonstrated that the TYK2–STAT3 pathway regulated promoter-selective induction of human IκB-ζ. Publicly available genome and transcripts data (Gene identification number 64332 for human IκB-ζ and 80859 for mouse) in the National Center for Biotechnology Information Gene database (https://www.ncbi.nlm.nih.gov/gene/) show that the overall gene structure, as well as the two major TSS (described as TSS1 and TSS2 in the present paper), are conserved between human and mouse IκB-ζ genes (Supplemental Fig. 1). The promoter activity for upstream-located TSS2 has been investigated and revealed to be regulated by NF-κB transcription factor (24). However, the transcriptional regulatory mechanism at the downstream-located TSS1 was unclear. In the current study, we demonstrated the role of STAT3 in the transcription from TSS1. We identified site 1 (5′-TTACTGGAA-3′) at the position upstream (−133 to −125) of TSS1 as a STAT-binding sequence (Fig. 5D). Comparative genomic analysis showed that site 1 is well conserved among multiple mammalian species (Supplemental Fig. 2), suggesting the importance of this STAT-binding site for IκB-ζ gene transcription. A recent study using RNA sequencing analysis of IL-36–stimulated keratinocytes has provided a clear example that the downstream-located promoter region (corresponding to TSS1 in the present paper) is the dominant promoter used for IκB-ζ transcription in keratinocytes (23). Also, in that study, STAT3 involvement in the TSS1 promoter regulation has been shown. Our data are consistent with these previous results and, thus, reasonably support the possibility of the use of JAK/TYK2 inhibitors for targeting keratinocytes.

It should be noted that no difference in Nfkbiz mRNA levels was found between WT and TYK2 knockout (KO) under the conditions without IMQ application (Fig. 1B). Also, TYK2 knockdown in HaCaT cells did not decrease NFKBIZ mRNA under nonstimulated conditions (Fig. 2A). These results implied that IκB-ζ mRNA levels were determined not solely by the activity of the TYK2–STAT3 pathway and that without IL-17, the effects of TYK2 knockdown on IκB-ζ mRNA levels do not emerge. Thus, a specific IL-17–elicited event may be required for efficient IκB-ζ induction. In the current study, we focused on Regnase-1–mediated posttranscriptional stability control of mRNA. Constitutive activity of endogenous Regnase-1 appeared to play a more important role than the TYK2–STAT3 pathway in terms of determining the basal expression level of IκB-ζ (Fig. 4C). In addition, our results showed that IL-17 signaling seemed to act as the key for the deactivation of posttranscriptional mRNA degradation mediated by Regnase-1. The transient inhibition of Regnase-1 in IL-17–treated keratinocytes can be achieved promptly through proteasome-mediated degradation, as reported recently (33). This brake-releasing mechanism may allow for the accumulation of IL-17–induced mRNA. TYK2–STAT3–mediated transcription is considered to have a substantial (visible) role in driving IκB-ζ expression, especially when combined with the stimulation that suppresses Regnase-1 activity, such as IL-17 (Fig. 2) and CHX (Fig. 5A). Altogether, the combined activation of STAT3 and IL-17 signaling is important to elicit successful expression of the IκB-ζ protein (Fig. 6G).

To inhibit posttranscriptional mRNA destabilization and experimentally monitor the rate of IL-17–unrelated constitutive transcription of the IκB-ζ gene, we used protein translation inhibitor CHX. CHX has been shown to block Regnase-1–mediated destabilization of target mRNAs, the mechanism of which is translation-coupled (39). In the current study, we showed that CHX treatment can actually accumulate IκB-ζ mRNA in HaCaT cells (Fig. 4G), supporting the idea that Regnase-1 constitutively degrades IκB-ζ mRNA. The mRNA stabilizing effect of CHX is consistent with results from earlier studies that have shown that CHX treatment increases IκB-ζ and CXCL8 mRNA (62, 63), both of which have been reported to be degraded by Regnase-1 (39, 64). We showed that CHX-induced accumulation of IκB-ζ mRNA was not enhanced by combined stimulation with IL-17 (Fig. 4G). These observations supported the idea that the pathway targeted by IL-17 signaling may correspond to Regnase-1 activity.

The persistently high expression of IκB-ζ is considered to be pathogenic and can be a candidate therapeutic target in psoriasis (19). In addition, the complete absence of the IκB-ζ protein in murine epithelial cells exhibited elevated apoptosis, which triggers lymphocyte-mediated autoimmune inflammation resembling Sjögren syndrome (25), suggesting the indispensable function of epithelial IκB-ζ in the maintenance of cell viability and immune homeostasis. These findings indicated that IκB-ζ protein levels should be under tight regulation at an adequate level and should avoid persistent aberrant upregulation. Regnase-1 has been shown as an IL-17–inducible protein and a negative feedback inhibitor of IL-17–induced inflammation (31); it would be reasonable to consider that this feedback mechanism accounts for the prevention of persistent upregulation of the IκB-ζ protein. However, the mechanisms of constitutively regulated expression and IL-17–induced rapid induction of IκB-ζ were elusive. In the current study, we demonstrated that the regulated balance between the supply of IκB-ζ mRNA by STAT3 activity and constitutive degradation of IκB-ζ mRNA by Regnase-1 determines the steady-state expression level of IκB-ζ. It should be emphasized that Regnase-1 seems to suppress IκB-ζ expression not just as an IL-17–inducible negative feedback inhibitor but as a constitutive inhibitor. We speculated that IL-17–induced rapid IκB-ζ upregulation is attributed to the STAT3-driven transcription of IκB-ζ mRNA, which presumably omits the latency time required for multiple processes, including signal transduction, transcription factor activation, and mRNA synthesis. Overall, this regulatory system might enable cells both to express IκB-ζ protein at a level pertinent for cell viability and to quickly exert boosted induction if there is a need to fight against extrinsic fungal and bacterial infections.

In summary, we demonstrated that TYK2-mediated phosphorylation of STAT3 underlies the transcription of IκB-ζ. This IL-17–independent supply of IκB-ζ mRNA may act as a determinant for the strength of IL-17–induced responses in keratinocytes. This finding suggested that JAK/TYK2 inhibition can bring simultaneous inhibition of the IL-12/IL-23 responses in lymphocytes and the IL-17 response in keratinocytes of psoriatic lesions. Thus, JAK/TYK2 inhibition may be an effective method to treat IL-17–mediated diseases.

We express our gratitude to Dr. James N. Ihle (St. Jude Children’s Research Hospital, Memphis, TN), Dr. Osamu Takeuchi (Kyoto University, Kyoto, Japan), Dr. Connie Cepko (Harvard Medical School, Boston, MA), and Dr. Steven Vogel (National Institutes of Health, Rockville, MD) for providing plasmids. We gratefully acknowledge the generous support of the Hokkaido University Global Facility Center Pharma Science Open Unit, which is funded by the Ministry of Education, Culture, Sports, Science and Technology–Japan under the Support Program for Implementation of New Equipment Sharing System.

The authors have no financial conflicts of interest.