NLRP3 inflammasome activation contributes to development of alopecia areata in C3H/HeJ mice

Abstract Alopecia areata (AA) is an autoimmune non‐scarring hair loss disease. Recently, several reports have suggested that innate immune systems such as interferon‐α (IFN‐α)‐producing plasmacytoid dendritic cells and NOD‐like receptor family pyrin domain‐containing protein 3 (NLRP3) inflammasomes play a role in the pathogenesis of AA. However, critical studies about their involvement in the initiation of AA have not yet been reported. Therefore, we investigated the expression of innate immune cytokines in serum and skin, and examined the effect of a selective NLRP3 inhibitor, MCC950, on AA in C3H/HeJ mice, induced by transferring cultured skin‐draining lymph node cells. IFN‐α production was upregulated in lesions of AA‐affected mice, and interleukin‐1β in serum and skin was highly expressed before onset as well as postonset. Furthermore, MCC950 treatment prevented AA development and promoted hair growth in AA mouse models by reducing NLRP3 signalling and Th1/Tc1 chemokines and cytokines in the skin. These results suggest that NLRP3 inflammasome contributes to AA onset and chronicity, and NLRP3 inhibitor may be a potential therapeutic agent for AA.


| INTRODUC TI ON
Alopecia areata (AA) is an autoimmune disease characterized by non-scarring hair loss. Although AA is not life-threatening, it has a detrimental impact on quality of life. It has been suggested that AA develops in anagen hair follicles (HF) through collapse of constitutive immune privilege (IP). This hypothesis is supported by infiltration of lymphocytes, natural killer (NK) cells, Langerhans cells, dendritic cells (DCs) and macrophages into the peribulbar area of anagen HFs. It has also been demonstrated that MHC class I and II expression increases and IP guardians decrease in AA lesions. [1][2][3] What triggers IP collapse is still under debate. Several studies have indicated that the innate immune system plays an important role in the pathogenesis of AA, based on clinical reports concerning associations with viral infection, 4-6 psychological stress 7 and oxidative stress. 8 In innate immunity, pattern recognition receptors (PRRs), such as toll-like receptors (TLRs) and NOD-like receptor family pyrin domain-containing protein 3 (NLRP3), are key regulators.
Activation of PRR signalling is essential for host defense against infection, whereas overactivation of PRRs often causes uncontrolled inflammation and autoimmune diseases. 9,10 TLRs have been identified from TLR1 to TLR11 in humans. Several TLRs are localized in the endosome to detect nucleic acids derived from bacteria, viruses and damaged cells. Endogenous nucleic acids, which are not recognized by the endosomal TLRs under homeostatic conditions, activate the TLRs and trigger development of autoimmune diseases 9 such as psoriasis, 11 systemic lupus erythematosus (SLE), 12 rheumatoid arthritis (RA), 13 diabetes mellitus 14 and systemic sclerosis (SSc). 15 In AA patients, expressions of TLRs 3, 7, 8 and 9 are reported to be upregulated in peripheral blood mononuclear cells 16,17 and/or around lesional HFs. 17 Stimulation of TLRs 7 and 9 mediates interferon (IFN)α secretion from plasmacytoid DCs (pDCs). pDCs are present in the peribulbar area of AA patients, 18 and type 1 interferon-inducible mexovirun protein A is expressed in AA lesions. 19 A recent study showed that IFNα-producing pDCs were infiltrated around HFs not only in AA lesions but also in non-lesional areas of AA-affected C3H/HeJ mice, and that intradermal injection of pDCs induced AA lesions in normal C3H/ HeJ mice. 20 IFNα-producing pDCs are known to play a role in scarring alopecia such as Lupus erythematosus-associated alopecia. 21 NLRP3 is activated via the P2X7 receptor, assembling with apoptosis-associated speck-like protein containing a caspase recruitment domain (ASC), and pro-caspase-1 into inflammasomes.
The NLRP3 inflammasome promotes secretion of interleukin (IL)-1β and IL-18. Several studies have revealed their significance in autoimmune disorders such as SLE, RA and SSc. 10 In AA patients, NLRP3 inflammasome components and IL-1β are reported to be highly expressed in the outer root sheath of HF. 22 Serum levels of IL-18 in patients with extensive AA are significantly higher than in healthy controls, 23 and single-nucleotide polymorphisms in IL-18 (re187238 and rs549908) are related to the susceptibility to AA in Koreans. 24 Thus, several studies have suggested a contribution of IFNαproducing pDCs and NLRP3 inflammasome activation to AA pathogenesis. However, their roles in the initiation of AA are yet to be clarified.
To define their roles in the onset of AA, we investigated the expression of IFNα and NLRP3-related factors, comparing before and after AA onset, using an AA mouse model reflecting AA pathogenesis. 25 Furthermore, we studied the therapeutic effect of an NLRP3 inhibitor on AA. This is the first report to demonstrate that inhibition of NLRP3 inflammasome may be a promising therapeutic approach to AA.

| Mice and induction of AA
Female 7-week-old C3H/HeJ mice were obtained from Japan SLC (Shizuoka, Japan). All mice were housed in specific pathogen-free conditions. They received food pellets for long-term breeding (CR-LPF; Oriental Yeast) and ultraviolet sterile water ad libitum. All animal experimental procedures were approved by the Ethics Committee naive mice (without transferred LN cells); preonset mice (with transferred LN cells but without hair loss); AA mice at initial phase (with hair loss localized in less than half of the abdominal area for less than 2 weeks); and AA mice at severe phase (with hair loss in whole abdominal area and less than half of dorsal area for more than 4 weeks).

| MCC950 treatment
Among AA-affected mice at initial phase (hair growth index between 50 and 100), mice were randomized into two groups: MCC950-

| Evaluation of hair growth
Hair growth was assessed in the square area (2 × 3 cm) on abdominal skin. Hair growth was evaluated using the 4-point score with definition as follows: 0 (no hair); 1 (non-dense short hairs); 2 (non-dense intermediate length hairs); and 3 (normal length and density hairs).
The percentage of the area with each score point in 2 × 3 cm area was measured using ImageJ software (National Institutes of Health).
Hair growth index was calculated as the sum of multiplying each hair growth score by the percentage of scored area.

| Real-time polymerase chain reaction analysis
To determine mRNA expressions, total RNA was isolated from homogenized skin samples (lesions, hair-loss areas; non-lesions, normal hairy areas remote from lesions) using RNeasy Fibrous Tissue Mini Kit (Qiagen). cDNA was synthesized by reverse transcription of total RNA using High Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific). Amplification was performed using TaqMan ® Gene Expression Master Mix and TaqMan Gene Expression Assay (Applied Biosystems ™ ). The primer sets purchased from Thermo Fisher Scientific were as follows: Ager (Mm01134790_

| Histological analysis
Skin samples from MCC950-and PBS-treated mice were collected and fixed in 10% formalin. Paraffin-embedded sections (3 μm) were stained with haematoxylin and eosin (HE). The stained slides were subsequently evaluated using the SLIDEVIEW VS200 digital slide scanner (Olympus).

| Statistical analysis
Data are expressed as mean ± standard error of the mean (SE).
Statistical analysis was conducted using Student's t test or Dunnett's multiple comparison test, implemented using EXSUS software (CAC Croit). A p-value less than 0.05 was considered statistically significant. and induces IFNα production. 28 HSP70 potentiates DNA-induced IFNα production. 29 To elucidate the trigger of pDC activation and IFNα production, we determined mRNA expressions of HMGB1, RAGE and HSP70 ( Figure 1F-H). The expressions of HMGB1, RAGE and HSP70 increased in lesions of initial phase AA mice, which was similar to IFNα. Serum level of HSP70 was increased in initial phase AA mice ( Figure 1I), and the expression of mmu-miR-30b-3p, which is reported to inhibit HSP70 expression, was higher in blood of AA mice, and particularly elevated in severe AA mice (Figure 1 J). Taken together, HMGB1/RAGE and HSP70 signal pathways were activated and IFNα production was increased in lesions at the initial phase.

| NLRP3 Inflammasome signal pathway contributes to development and exacerbation of AA
Next, to investigate whether NLRP3 inflammasome contributes to AA development, we evaluated the production of IL-1β and IL-18 in serum and skin by comparing naive, preonset, initial AA and severe AA mice. Serum IL-1β level was subtly higher in preonset and AA mice than in naive mice (Figure 2A). Serum IL-18 level was increased in initial AA mice and decreased in preonset and severe AA mice ( Figure 2B). Serum level of IFNγ, that is reported to play a crucial role in the pathogenesis of AA, was significantly increased in severe AA mice ( Figure 2C). The blood level of mmu-miR-150-5p, which is reported to inhibit IL-18 expression, was lower in preonset, initial AA and severe AA mice than in naive mice ( Figure 2D). IL-1β production in skin was increased in preonset and AA mice, and the level was elevated depending on the progression of AA ( Figure 2E). IL-18 production was increased in non-lesional areas and lesions of AA mice, and the level was similar between initial AA and severe AA ( Figure 2F).
Comparing IL-1β and IL-18 production, IL-18 level was significantly higher than IL-1β level both in serum and in skin. Furthermore, IL-1β production, but not IL-18 production, was increased in preonset mice. These data demonstrate that IL-1β production increased before onset of AA, and IL-1β and IL-18 production increased in non-lesional areas and in lesions after onset. IFNγ production in skin tended to increase in mice injected with lymph node cells compared to naïve mice ( Figure 2G). mRNA expressions of NLRP3 were increased both in non-lesional areas and in lesions of initial and severe AA mice, particularly in lesions of initial AA mice ( Figure 2H). mRNA expressions of NLRP3 inflammasome components, ASC (Pycard) and caspase-1, were increased in lesions of initial AA mice and non-lesional areas and lesions of severe AA mice, and the level was similar between non-lesional areas and lesions at the severe phase ( Figure 2,I J). These data demonstrate that NLRP3 inflammasome signalling in skin was activated at the initial phase and remained activated at the severe phase, suggesting that NLRP3 inflammasome contributes to aggravation as well as onset of AA. As shown in Figure 3B, the increase in hair growth index was observed even 7 days after the first MCC950 administration and was maintained up to 42 days. Therefore, to elucidate the effect of MCC950 on cytokines, chemokines and NLRP3 inflammasome signalling in AA mice, we investigated pathological characteristics, mRNA expressions in the skin and cytokine production in serum on day 7. HE staining revealed that MCC950 treatment decreased infiltration of inflammatory cells such as neutrophils and lymphocytes ( Figure 4A). MCC950 treatment decreased mRNA expression of Th1/Tc1 chemokines (CXCL9/10/11), Th1/Tc1 cytokine (IFNγ) and cytotoxic marker (granzyme B) in the skin ( Figure 4B). mRNA expressions of NLRP3 inflammasome-related factors (NLRP3, ASC and caspase-1) were reduced by MCC950 treatment (Figure 4C).

| DISCUSS ION
AA is thought to arise as a consequence of the collapse of local IP with cytotoxicity induced by CD8 + T cells and NK cells. 1 Although several studies imply that AA results primarily from dysregulation of innate immunity, similar to other autoimmune diseases, 3 the trigger of IP collapse has not been clarified. Here, we compared serum and skin levels of innate immune-related factors between naive mice, preonset mice and AA-affected mice (initial phase and severe phase). This is the first study to demonstrate that NLRP3 inflammasome activation contributes to development and aggravation of AA and that NLRP3 inhibition may have potential for AA therapy.
In this study, we focused on IFNα-producing pDCs and NLRP3 inflammasome in innate immunity, and investigated their contributions to AA onset using a C3H/HeJ AA mouse model, which is the most popular and well-defined model for AA. First, we found that IFNα in serum was upregulated in preonset mice, whereas its expression in skin was unchanged ( Figure 1A, B). Skin IFNα level was upregulated in lesions of initial AA and severe AA mice ( Figure 1B).
Our data demonstrate that IFNα (all subtypes) protein level was higher in lesions than in non-lesional areas, whereas a previous report showed that IFN-α2 and IFN-α4 mRNA expressions were higher in non-lesional areas than in lesions. 20 This discrepancy could be caused by differences of IFNα subtype, definition of non-lesions (normal hairy areas remote from lesions in our study; the vicinity of lesions in the previous report), AA mouse model (induced by celltransferring or spontaneously affected) and immune status due to ageing (our AA-affected mice were younger than 32 weeks old, but the mice in the previous report were 1 year old). IFNα production is activated by TLR7/9 signalling that requires the formation of a complex consisting of MyD88 (myeloid differentiation factor 88), TRAF3/6, IRAK1/4 and IRF7. [30][31][32] We found that these adaptor molecules and transcriptional factors were also highly expressed in skin of AA-affected mice, particularly in lesions of initial AA mice ( Figure 1C-E). These data indicate that the IFNα signal pathway is locally activated in non-lesional areas at the initial phase. A number of studies have suggested the relevance of TLRs to autoimmune diseases and to triggering IFNα production. HMGB1-DNA immune complex activates TLR9 and induces pDC activation followed by IFNα secretion. 28 HMGB1 is released from pDCs after stimulation F I G U R E 1 IFNα signal pathway is activated in lesions in the initial phase of alopecia areata. IFNα levels in serum (A) and skin (B) from naive, preonset, initial AA and severe AA mice were measured by ELISA. mRNA expressions of IRF7 (C), IRAK4 (D), TRAF3 (E), HMGB1 (F), RAGE (G) and HSP70 (H) were assessed by RT-PCR. (I) Serum HSP70 level was measured by ELISA. (J) Expression of mmu-miR-30b-3p in blood was assessed using RT-PCR. Skin samples were collected from non-lesional areas (NL; normal hairy area) and lesions (L; hair loss area). Data are expressed as mean ± SE (N = 4). ‡ , p < 0.05. ‡ ‡ , p < 0.01. vs. naive (Dunnett's multiple comparison test). Abbreviations: AA, alopecia areata; ELISA, enzyme-linked immunosorbent assay; RT-PCR, reverse transcription polymerase chain reaction; SE, standard error; and N.D., not detected and regulates IFNα production via RAGE in an autocrine way. 33 HMGB1 is also shown to be released from damaged cells, based on in vitro study indicating that outer root sheath cells, which are nonprofessional immune cells, secrete HMGB1 by poly(I:C)-mediated TLR3 activation. 22 In AA patients, HMGB1 levels in scalp and serum have been reported to be upregulated, compared with normal controls. 34 HSP70 has also been reported to induce pDC activation and IFNα production. 29 Moreover, the previous study for investigating the effect of heat treatment on AA development has suggested that induction of HSP70 may precipitate the development of AA in C3H/ HeJ mice. 35 We investigated whether these factors activate pDCs, and showed that mRNA expressions of HMGB1, RAGE and HSP70 were increased in lesions at the initial phase ( Figure 1F-H). Serum level of HSP70 was also increased at the initial phase ( Figure 1I).
These results indicate that these factors might induce IFNα production in lesions at the initial phase of AA.
Although NLRP3 inflammasome has been reported to be associated with AA pathogenesis, little is known about its role in AA onset. Our data showed that IL-1β level was increased in both serum and skin before and after AA onset (Figure 2A, E), whereas IFNα level was increased in lesions after AA onset ( Figure 1B).
Thus, the increase of IL-1β preceded that of IFNα, which suggests that increase of IL-1β production results in initiation of innate immune response around immune-privileged HF. As well as IL-1β production, IL-18 production is also known to be regulated by NLRP3 inflammasome. Our data showed that IL-18 production was increased in both lesions and non-lesional areas of AA mice ( Figure 2F). IL-18 has been reported to induce Th1 and NK cell activation, IFNγ production, 36 and type 1 chemokines. 37 The blood level of mmu-miR-150-5p, which downregulates IL-18 expression, was lower before and after AA onset than in naive mice ( Figure 2D), which is in concordance with the clinical report that the blood level of hsa-miR-150-5p was decreased in AA patients. 38 These data support the involvement of IL-18 secretion in AA pathogenesis.
The expressions of NLRP3 inflammasome components (NLRP3, ASC and caspase-1) were highly increased in lesions at the initial phase and also upregulated in non-lesional areas and lesions at the severe phase ( Figure 2H MCC950 is a potent, selective small molecule inhibitor of NLRP3, and its therapeutic effects on various autoimmune diseases have recently been anticipated. 39 In the AA mouse model, MCC950 treatment prevented AA development and promoted hair growth ( Figure 3A, B). The effect of MCC950 was observed just 7 days after first administration, and we investigated pathological characteristics on day 7. HE staining revealed that MCC950 treatment reduced infiltration of inflammatory cells such as neutrophils and lymphocytes ( Figure 4A). The treatment also reduced expressions of type 1 chemokines, type 1 cytokine and cytotoxic marker, which were increased in AA-affected mice, as well as NLRP3 inflammasome-related factors ( Figure 4B, C). Serum levels of IFNγ, TNFα and IL-17A, which were increased in both AA patients and mice (data not shown), were decreased following MCC950 treatment ( Figure 4D). IL-1β is shown to inhibit hair elongation in HF culture. 40 IL-18 induces IFNγ production and, vice versa, IFNγ induces IL-1β and IL-18 production via NLRP3 inflammasome activation. 41,42 We speculate that MCC950 inhibits the negative loop of IFNγ and NLRP3 inflammasome and reduces Th1/Tc1 activation, consequently exerting a hair growth effect in AA mice. Our study sheds light on pharmacotherapy targeting the innate immune systems for AA, unlike the conventional pharmacological strategy, which targets T cells. Considering our finding that NLRP3 inflammasome activation contributes to not only AA exacerbation but also AA initiation, MCC950 might have not only have therapeutic effect but also preventive effect on AA relapse. Further studies will be needed to address whether NLRP3 inhibitor has a preventive effect against AA.
In conclusion, we reported that the IFNα-producing pathway was activated in lesions at the initial phase, whereas the NLRP3 inflammasome pathway was activated before as well as after AA onset. Furthermore, administration of NLRP3 inflammasome inhibitor MCC950 induced hair regrowth and inhibited infiltration of inflammatory cells, expression of type 1 chemokines and cytokine in skin, and inflammatory cytokines in serum. The inhibitors of NLRP3 inflammasome may be candidates for novel therapeutic agents for AA.