Author response: Involvement of ILC1-like innate lymphocytes in human autoimmunity, lessons from alopecia areata

Abstract

Article Figures and data Abstract Editor's evaluation Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract Here, we have explored the involvement of innate lymphoid cells-type 1 (ILC1) in the pathogenesis of alopecia areata (AA), because we found them to be significantly increased around lesional and non-lesional HFs of AA patients. To further explore these unexpected findings, we first co-cultured autologous circulating ILC1-like cells (ILC1lc) with healthy, but stressed, organ-cultured human scalp hair follicles (HFs). ILClc induced all hallmarks of AA ex vivo: they significantly promoted premature, apoptosis-driven HF regression (catagen), HF cytotoxicity/dystrophy, and most important for AA pathogenesis, the collapse of the HFs physiological immune privilege. NKG2D-blocking or IFNγ-neutralizing antibodies antagonized this. In vivo, intradermal injection of autologous activated, NKG2D+/IFNγ-secreting ILC1lc into healthy human scalp skin xenotransplanted onto SCID/beige mice sufficed to rapidly induce characteristic AA lesions. This provides the first evidence that ILC1lc, which are positive for the ILC1 phenotype and negative for the classical NK markers, suffice to induce AA in previously healthy human HFs ex vivo and in vivo, and further questions the conventional wisdom that AA is always an autoantigen-dependent, CD8 +T cell-driven autoimmune disease. Editor's evaluation This manuscript provides fundamental data that implicate ILC1-like cells in alopeia areata. The data are solid in the use of cultured human hair follicles co-cultured with ILC-1-like cells and demonstration that alopecia phenotypes emerge. The authors also provide compelling evidence that injection of ILC1-like cells induces alopecia in a mouse model grafted with human hair follicle-containing skin. This work will be of interest to immunologists, skin biologists, and scientists interested in autoimmune disorders. https://doi.org/10.7554/eLife.80768.sa0 Decision letter eLife's review process Introduction Alopecia areata (AA) is both the most common inflammatory hair loss disorder and one of the most common human autoimmune diseases and exerts a major negative impact on quality of life (Gilhar et al., 2012; Gilhar et al., 2019a; Korta et al., 2018; Pratt et al., 2017). Despite major recent advances in AA therapy, a causal therapy does not yet exist, and disease relapse after therapy discontinuation is the rule, not the exception in long-standing AA (Meah et al., 2020; Gilhar et al., 2019a). Thus, the currently available, purely symptomatic AA therapy, including JAK inhibitors (Gilhar et al., 2019b), remains unsatisfactory. Since the exact pathobiology of AA and its clinical variants remains to be fully characterized, the – likely diverse – disease-initiating factors that ultimately result in the characteristic AA hair loss pattern shared by all AA variants, require more comprehensive dissection for optimal, personalized therapeutic targeting (Bertolini et al., 2020; Paus et al., 2018). Specifically, there is increasing awareness that a classical, autoantigen- and CD8 +T cell-dependent autoimmune variant of AA (AAA) and a possibly autoantigen-independent non-autoimmune variant (NAIAA) may have to be distinguished from each other (Gilhar et al., 2019a; Bertolini et al., 2020; Paus et al., 2018; Paus, 2020). This is in line with the long-standing, but often under-appreciated clinical recognition that AA shows a wide spectrum of phenotypes and sub-forms (Gilhar et al., 2012; Ikeda, 1965; Meah et al., 2021; King et al., 2022). One reason why the currently available AA therapy is not entirely satisfactory may be related to as yet insufficient therapeutic targeting of innate immunocytes in the immunopathogenesis of human AA, namely in NAIAA, even though these are now recognized as major players in AA pathobiology (Ghraieb et al., 2018; Ito et al., 2008; Li et al., 2016; Uchida et al., 2020; Uchida et al., 2021). Previously, we had demonstrated that AA lesions are associated with a massive increase in the number of perifollicular NKG2D+NK cells (Gilhar et al., 2013a), which recognize the activating NKG2D ligand MICA, a ‘danger’ signal that is greatly overexpressed by the epithelium of lesional AA HFs (Ito et al., 2008; Li et al., 2016; Connell and Jabbari, 2022). Subsequent work has confirmed the key role of NKG2D and its activating ligands in human and murine AA (Xing et al., 2014; Petukhova et al., 2010). In fact, AA lesions can be induced experimentally in healthy human scalp skin in vivo by the transfer of interleukin 2 (IL-2)-activated NKG2D+ cells (Gilhar et al., 2013a), most of which had NK cell characteristics, with only a small minority of CD8 +T cells being present, that is the best-recognized pathogenic lymphocyte population in AA (Gilhar et al., 2012; Gilhar et al., 2013a; Pratt et al., 2017; Bertolini et al., 2020; de Jong et al., 2018). Moreover, pro-inflammatory mast cells (Bertolini et al., 2014) and (likely autoantigen-non-specific) γδ T-cells are also increased around/in lesional human AA HFs (Uchida et al., 2020). Finally, these ‘intermediate immunity’ protagonists suffice to induce the hallmarks of AA ex vivo (Uchida et al., 2021). Taken together, this questions whether pathogenic, autoreactive CD8 +T cells are the only drivers of disease, and that all cases of AA, represent a genuine, autoantigen-dependent autoimmune disease (Bertolini et al., 2020; Paus et al., 2018) in the strictly defined sense of this term (Rose and Bona, 1993). In our ongoing exploration of the role of innate/transitional immunity in the pathobiology of AA (Paus, 2020; Uchida et al., 2020; Uchida et al., 2021; Gilhar et al., 2019a; Bertolini et al., 2014), we, therefore, have asked in the current study whether innate lymphoid cells type 1 (ILC1 cells) (Zhou et al., 2020; Nabekura and Shibuya, 2021a; Colonna, 2018) can initiate human AA lesions. We were interested in these immunocytes since human ILC1 cells secrete large amounts of interferon-γ (IFN-γ) (Ebbo et al., 2017), the crucial AA pathogenesis-promoting cytokine (Gilhar et al., 2012; Gilhar et al., 2019a; Paus et al., 2018), and this notably independent of classical autoantigen-specific CD8 +T cell activities. These ‘unconventional’ T-cells are placed in strategic tissue locations (Collins et al., 2017; Jiao et al., 2016; Kim et al., 2021) and represent an important link between innate and adaptive immunity (Vivier et al., 2018). While ILC1s play an essential role in human inflammatory bowel disease (IBD) (Ebbo et al., 2017; Luo et al., 2022; Clottu et al., 2021), their role in the pathophysiology of autoimmune hepatitis and rheumatoid arthritis requires further investigation (Ebbo et al., 2017; Fang et al., 2020; Yang et al., 2015), and their role in human autoimmune diseases overall remains insufficiently understood. We hypothesized that AA might offer a good model disease for interrogating this role. ILC1 cells are classified as a component of type 1 immunity (Shannon et al., 2021), express NKG2D, recognize conserved phosphoantigens (Nabekura and Shibuya, 2021a), and contribute to immunity against tumor cells, for example through NKG2D activation (Dadi et al., 2016). The activating receptor NKG2D and its ligands (MICA, ULBP3) play an important role in innate (NK, ILC1), ‘translational’ (γδ T-cells) and CD8 T-cell-mediated immune responses to tumors and in several autoimmune diseases (Frazao et al., 2019; Babic and Romagnani, 2018). Given that ILC1 cells produce TH1-type cytokines (such as IFN-γ) and share several phenotypic markers with NK cells, namely NKG2D (Spits et al., 2016), it is challenging to distinguish NKs and ILC1 cells (Tulic et al., 2019; Zhang et al., 2018; Seillet et al., 2021; Conlon et al., 2021). In fact, how to reliably discriminate between NK cells and ILC1s and unraveling the shared and distinct functions of these cell populations remains an important open quest (Seillet et al., 2021; Lopes et al., 2023; Cheng et al., 2023; Taggenbrock and van Gisbergen, 2023). For example, Eomeshi T-betlo liver-resident NK cells have been described in humans and mice (Park et al., 2019; Harmon et al., 2016), while ILC1s from human tonsil and blood was also found to be Eomes+ (Cella et al., 2019). Therefore, the distinction between NK cells and ILC1s remains provisional – which is exactly why we have cautiously labeled the latter as ‘ILC1lc.’. The transcriptional and functional identity of ILC1 cells in humans is still a matter of debate, given that in contrast to other ILC subsets ILC1 cells seem to lack robust markers that enable their unequivocal identification and isolation (Bennstein et al., 2020). However, although integrin α1 (CD49a) is upregulated on activated NK cells (Albini et al., 2021; Zheng et al., 2016), CD49a and integrin α2 (CD49b) are used as two mutually exclusive markers for distinguishing between NK and ILC1 cells, with NK cells being defined as CD49b+CD49a- and ILC1 as CD49b-CD49a+ (Gao et al., 2017; Vienne et al., 2021; Flommersfeld et al., 2021; Krzywinska et al., 2022) In the current study, we have accepted and employed this consensus. Also, in contrast to ILC1 and ILC1lc, classical NK cells demonstrate high T-bet and Eomes expression (T-bethi /Eomeshi) (Verma et al., 2020). Therefore, for the purpose of this study, we define ILC1lc as CD49a+CD49b- (Verma et al., 2020) and as lin-/CD127+/CD117-/CRTH2-phenotype, which are typical to classical ILC1 cells (Bennstein et al., 2020; Krabbendam et al., 2021), and also as T-betlo/ Eomeshi (Bennstein et al., 2020) (in contrast to classical T-bethi /Eomeslo ILC1 cells Verma et al., 2020). Specifically, we have asked whether (a) their number is increased in lesional AA skin, (b) they can damage human HFs ex vivo in a manner that mimics the AA phenotype, and finally (c) whether ILC1lc alone suffice to induce AA in previously healthy human scalp skin in vivo. To address these questions, we first analyzed the abundance, distribution, and phenotype of ILC1lc in human AA skin lesions compared to healthy human control skin. We then co-cultured autologous ILC1lc with freshly organ-cultured scalp HFs from the same patient, that is under conditions where the epithelium of these HFs transiently undergo an acute stress response and overexpresses MICA (Uchida et al., 2021), to check whether these innate lymphocytes exert any HFs cytotoxicity and/or impact on the physiological immune privilege (IP) of HFs (Bertolini et al., 2020; Paus et al., 2005; Ito et al., 2004; Peters et al., 2007; Bertolini et al., 2016). Finally, we injected autologous ILC1lc intradermally into healthy human scalp skin xenotransplants from the same human volunteers on SCID/beige mice to probe whether this suffices to induce classical AA hair loss lesions in vivo. Taken together, our data show that ILC1lc is increased in AA lesions and suffice to induce an AA phenotype in healthy human HFs ex vivo and in vivo. This provides the first functional evidence of a key role of ILC1lc innate lymphocytes in a model human autoimmune disease (Colonna, 2018; Seillet et al., 2021; Conlon et al., 2021; Flommersfeld et al., 2021; Daussy et al., 2014; Park et al., 2019) - but also questions whether AA always a classical autoimmune disease is and underscore the role of innate immune cells in AA pathobiology. Results Peri- and intrafollicular infiltrates of ILC1lc are seen in both lesional and non-lesional AA skin First, we investigated whether healthy and AA-affected human skin differs in their content and/or distribution of ILC1lc, using a comprehensive set of triple-immunofluorescence (IF) staining best suited to identify these immunocytes (Seillet et al., 2021; Bennstein et al., 2020; Gao et al., 2017). This revealed the presence of only extremely few ILC1lc in healthy control skin with all three staining settings employed (Eomes+, CD49a+, NKG2D+ [Figure 1A and Figure 1—figure supplement 1A], Eomes+, c-KIT-, CD49a+ [Figure 1B and Figure 1—figure supplement 1A], or NKp44+, CD103+, T-bet- cells [Figure 1C and D and Figure 1—figure supplement 1A; Kim, 2015; Fuchs et al., 2013; Salimi and Ogg, 2014]). These cells appeared to be preferentially scattered along the papillary dermis of healthy scalp skin biopsies and around the HFs (Figure 1C). This is reminiscent of the few Vδ1+T cells detectable in healthy human skin that also have a preferential perifollicular location and may ‘police’ the skin for molecular indications of tissue stress, namely of HFs (Uchida et al., 2020; Uchida et al., 2021). Figure 1 with 1 supplement see all Download asset Open asset Immunofluorescence microscopy analyses of ILC1lc and CD8+/NKG2D+ cells in alopecia areata (AA) scalp skin. (A) ILC1lc (EOMES+, CD49a+, and NKG2D+) around HF in normal scalp skin, intrafollicular and perifollicular ILC1lc infiltrates in lesional and in non-lesional AA scalp patient. (B) EOMES+, c-KIT-,CD49a+, and (C) NKp44+, CD103+, T-bet- ILC1lc. For each panel, yellow staining indicates double staining A-EOMES+, NKG2D+; B- EOMES+, CD49a+; C- NKp44+,CD103+ (D) Quantitative immunohistomorphometry (qIHM) shows an increased number of ILC1lc in AA patients as compared to normal volunteers and increased number of the cells in lesional versus non-lesional areas of the patients. There is a significant increased perifollicular than intrafollicular ILC1lc in the lesional and non lesional areas. (E) CD8+/NKG2D+ cells around HF in AA scalp patient and absence of these cells in normal scalp skin of normal scalp skin. (F) There is an increased number of CD8+/NKG2D+ cells in HFs of AA patients compared to normal scalp skin and a significant lower number of ILC1lc versus CD8+/NKG2D+ cells in AA scalp skin. N=6 biopsies /AA patients and six biopsies /healthy donors from six independent donors, three areas were evaluated per section, and three sections per biopsy. Following Shapiro-Wilk test, Student’s t-test: *p<0.05, **p<0.01 or Mann Whitney U test: #p<0.05. Scale bars, 50 µm. CTS- connective tissue sheath, DP - dermal papilla, HM - hair matrix, White arrow- c-KIT stained melanocyte. Figure 1—source data 1 Quantitative data for immunofluorescence microscopy analyses of ILC1lc and CD8+/NKG2D+ cells in AA scalp skin. https://cdn.elifesciences.org/articles/80768/elife-80768-fig1-data1-v1.xlsx Download elife-80768-fig1-data1-v1.xlsx Instead, intra and peri-follicular infiltrates of ILC1lc were frequently present in lesional AA HFs (Figure 1A, B, C and D and Figure 1—figure supplement 1A), typically in conjunction with a dominant infiltrate of CD8+/NKG2D+ cells around the hair bulb (p<0.05) (Figure 1E and F). Importantly, the number of ILC1lc was already significantly increased in/around non-lesional AA HFs compared to healthy scalp skin (p<0.01) (Figure 1A, B, C and D and Figure 1—figure supplement 1A). This may indicate that ILC1lc may actually have arrived around the HFs before the CD8 cells and may have contributed to attracting the CD8 cells into the perifollicular space. This strongly suggested that ILC1lc are not mere bystanders attracted only secondarily to the HFs by CD8 T-cells, similar to, but more pronounced than we have recently observed regarding perifollicular Vδ1+T cells in non-lesional AA skin (Uchida et al., 2020). This invited the hypothesis that ILC1lc is actively involved in transforming healthy human scalp HFs into lesional AA HFs. T-betlo/Eomeshi ILC1lc can be expanded from human peripheral blood mononuclear cells (PBMCs) in vitro To functionally probe this hypothesis, we isolated, purified, and characterized human peripheral blood-derived ILC1lc as the most suitable cell source for the planned HF-immunocyte co-culture studies. The scarcity of ILC1lc in healthy human skin, compared to their relative abundance in peripheral blood (Colonna, 2018; Artis and Spits, 2015) necessitated to isolate autologous ILC1lc from the latter source rather than from skin (Teunissen et al., 2014). To facilitate ILC1lc isolation, PBMCs of healthy volunteers were first cultured with high-dose IL-2 (100 U/mL) in the presence of IL-18 (1 µg/1 ml), IL-33 (1.5 µg/5 ml), and IL-12 (1.5 µg/5 ml), since these cytokines induce ILC1lc expansion (Salimi and Ogg, 2014; Silver et al., 2016; Orimo et al., 2020; Ohne et al., 2016). When ILC1lc were sorted by FACS Aria and characterized by FACS analysis on day seven of culture, low T-bet, and high Eomes expression were observed (Figure 2A), in contrast to classical T-bethi and Eomeslo ILC1 cells (Jiao et al., 2016; Vivier et al., 2018; Zhang et al., 2018). In addition, the ILC1lc expressed and shared the following markers with classical ILC1 cells: LIN- CD3/CD1a/D14/CD19/CD34/CD123/CD11c /BDCH2/FcεR1α/TCRαβ/TCRγδ/CD56, CD127+, CD161+, c-KIT-, and CRTH2- (Zook and Kee, 2016; Bernink et al., 2017; Simoni and Newell, 2017; Figure 2A). Figure 2 with 2 supplements see all Download asset Open asset Circulating ILC1lc expanded and characterized by FACS analysis. (A) PBMCs activated by IL-18, IL-33 and IL-12 were sorted by FACS Aria and characterized by FACS analysis. ILC1lc markers were identified by the expression of CD127+, CD161+, c-KIT-, and CRTH2-, high levels of integrin α1 (CD49a) expression, combined with the absence of integrin α2 (CD49b) and transcription factors Eomeshi and T-betlo (B) unstimulated PBMCs (C) isotype controls. N=10 blood donors, 1.5 × 106 cells/blood donor, analysis was performed in triplicates from each of the blood donors. Following Shapiro-Wilk test, Student’s t-test, p<0.05. Figure 2—source data 1 Quantitative data for circulating ILC1lc expanded and characterized by FACS analysis. https://cdn.elifesciences.org/articles/80768/elife-80768-fig2-data1-v1.xlsx Download elife-80768-fig2-data1-v1.xlsx This immune phenotype suggests that the immune cells used in our study are best classified as ILC1lc (Nabekura and Shibuya, 2021a), and documents that all experiments reported below were indeed performed with autologous ILC1lc rather than with NK cell subpopulations. Indeed, the FACS analysis (Figure 2—figure supplement 1A,B and C) revealed that ILC1lc demonstrates the ILC1 phenotype (CD200R, CD127, CXCR6) (Lopes et al., 2023; Curio and Belz, 2022) but not of the classical NK cell lineage (IRF8, Perforin, NKp80, CD16) (Sagebiel et al., 2019; Brownlie et al., 2021; Krämer et al., 2023), thus further serving as an evidence that EOMES +ILCs represent distinct ILC1 lineage-defining markers. In contrast to NK cells, ILC1lc also expressed the expected high levels of integrin α1 (CD49a), combined with the absence of integrin α2 (CD49b) (Jiao et al., 2016; Figure 2A). All these characteristic markers of ILClc were absent in the control unstimulated PBMCs (Figure 2B and C). This immune phenotype suggests that the immune cells used in our study are best classified as ILC1lc (Nabekura and Shibuya, 2021a), and documents that all experiments reported below were indeed performed with autologous ILC1lc rather than with NK cell subpopulations. Note that we had previously shown that NKG2D+/CD56 +NK cells suffice to induce AA lesions in human skin in vivo (Gilhar et al., 2013a; Laufer Britva et al., 2020) while iNKT cells are AA-protective in the humanized AA mouse model (Ghraieb et al., 2018). Subsequently, these ILC1lc were either used for HF co-culture assays or injected into healthy human scalp skin xenotransplants on SCID/beige mice (Gilhar et al., 2013a; Ito et al., 2005b). As controls, we also isolated ILC2 and ILC3 cells, which failed to induce AA phenotype in a sharp contrast to the ILC1lc (see Materials and methods). In order to exclude the possibility that contamination from ILC3s during sorting the ILC1lc and thus contributing to the observed results, we generated a new set of FACS data on sorted ILC1lc. Given that ILC1 cells are RORγt negative while ILC3 cells are RORγt positive (Peng et al., 2022; Fiancette et al., 2021), the data clearly demonstrate that the contamination hypothesis is highly unlikely (Figure 2—figure supplement 1D). ILC1lc induces HF cytotoxicity ex vivo Next, we functionally probed the interaction of ILC1lc with HFs that were investigated here as a model human (mini-) organ in which the interactions of a healthy human tissue system with defined, autologous immunocyte populations can be interrogated ex vivo in the absence of any confounding systemic immune or neural inputs (Uchida et al., 2021). For this, microdissected, organ-cultured human scalp HFs (Langan et al., 2015) were co-cultured for six days with autologous, peripheral blood-derived, purified, IL-12/IL-18/IL-33-prestimulated ILC1lc, or with autologous human CD8 +NKG2D+ cells (=positive control), ILC2, ILC3 cells, or PBMCs non-specifically activated with PHA (PBMCs/PHA) (=negative controls). Importantly, only scalp HFs in the anagen VI stage of the hair cycle were used (identified as described) (Kloepper et al., 2010) that had been freshly placed into HF organ culture for 24 hr, since these HFs are maximally ‘stressed,’ in contrast to non-cultured HFs, that is immediately after isolation, or that had already undergone several days of adjusting to the harsh conditions of serum-free organ culture (Uchida et al., 2021; Langan et al., 2015). These ‘stressed’ day 1 HFs temporarily up-regulate MHC class Ia and ß2-microglobulin while the expression of IP guardians, that is αMSH and TGFβ2 remain unchanged (Uchida et al., 2021), indicating a transiently weakened, but partially maintained HF immune privilege (Bertolini et al., 2020; Ito et al., 2004). The expression of molecules associated with tissue stress, that is the intrafollicular produced neurohormone, CRH (Ito et al., 2005a), and the NKG2D ligand MICA/B is also higher in day 1 organ-cultured HFs compared to freshly microdissected HFs or after day 3 of organ culture. Day 1 HFs also show signs of mild HF dystrophy (as evidenced by increased lactate dehydrogenase [LDH] release into the medium), and express chemokines recognized for their relevance in AA pathobiology, that is CXCL10 and CXCL12 (Uchida et al., 2021; Ito et al., 2020). Thus, day 1 HFs are ideally suited for interrogating human immunocyte interactions with a transiently ‘stressed,’ but otherwise healthy human (mini-) organ that overexpresses the NKG2D-activating ‘danger’ signal, MICA/B, under physiologically relevant ex vivo conditions (Uchida et al., 2021; Langan et al., 2015). First, we studied the cytotoxic effects of ILC1lc on healthy human scalp HF ex vivo by measuring the HF release of LDH into the culture medium. This not only showed significantly higher LDH release induced by ILC1lc than by co-culture with all three negative control cell populations (ILC2s, ILC3s, or PBMCs/PHA) but also even higher HF cytotoxicity levels than those induced by CD8+/NKG2D+ cells (p<0.01), namely after three days of co-culture (Figure 3). These HF cytotoxicity results were fully corroborated by characteristic morphological signs of HF dystrophy following co-culture with ILC1lc; while CD8+/NKG2D+ cells induced similar dystrophy phenomena, these were not seen after co-culture with PBMC/PHA (Figure 4A, B and C). The induction of significant HF dystrophy by ILC1lc ex vivo was further documented by the presence of pathological melanin clumping and ectopically located intrafollicular melanin granules (Bodó et al., 2007; Hendrix et al., 2005; Figure 4D, E, F and G) and by decreased proliferation and increased apoptosis of hair matrix keratinocytes (Figure 4H, I, J and K). Both were also seen in the CD8+/NKG2D+group (positive control), but not in HFs co-cultured with PBMCs/PHA (negative control) (p<0.001, p<0.01, respectively). Thus, autologous ILC1lc alone suffice to induce substantial HF cytotoxicity ex vivo if co-cultured with transiently ‘stressed,’ but otherwise healthy human scalp HFs. Figure 3 Download asset Open asset Cytotoxic effects of CD8+/NKG2D+and ILC1lc on normal human scalp HF ex vivo. These cell populations were placed separately into wells with (+Hair) dissected HFs and without (-Hair). Cytotoxic effects of these cell populations on normal human scalp HF ex vivo were studied by measuring the spontaneous release of lactate dehydrogenase (LDH) from the microdissected HFs. Increased cytotoxicity of ILC1lc co-cultured with HFs compared to CD8+/NKG2D+, as well as to ILC2s and ILC3s, and PBMCs/PHA cells. N=20–24 HFs/group derived from three independent donors analyzed in three independent HF organ culture experiments. Following Shapiro-Wilk test and Dunn’s test §p<0.05, §§p<0.01, §§§p<0.001. Figure 3—source data 1 Quantitative data for cytotoxic effects of CD8+/NKG2D+ and ILC1lc on normal human scalp HF ex vivo. https://cdn.elifesciences.org/articles/80768/elife-80768-fig3-data1-v1.xlsx Download elife-80768-fig3-data1-v1.xlsx Figure 4 Download asset Open asset Hair follicles (HFs) dystrophy, melanin clumping, and apoptosis in normal human scalp HF ex vivo co-cultured with ILC1lc and CD8+/NKG2D+ cells. (A–C) H&E staining revealed undifferentiated and prominent matrix cells, condensed dermal papilla, and the appearance of apoptotic cells, N=15–19 HFs/group from three independent donors. (D–G) Masson-Fontana histochemistry revealed melanin clumping and ectopic location of melanin granules only in HFs co-cultured with CD8+/NKG2D+and ILC1lc, but not in HFs cultured with PBMCs/PHA. N=7–11, HFs/group from three independent donors. Following Shapiro-Wilk test,Student’s t-test: *p<0.05, **p<0.01, ***p<0.001. (H–K) HFs co-cultured with ILC1lc or CD8+/NKG2D+ cells showed a significantly decreased proliferation (pink, arrowhead) and increased apoptosis (green, wide arrows). N=6 HFs/group from two independent donors, three areas were evaluated per section. Following Shapiro-Wilk test, Student’s t-test: *p<0.05, **p<0.01, ***p<0.001 in the anagen hair bulb compared to HFs cultured with PBMCs/PHA. Scale bars, 50 µm. DP - dermal papilla, HM - hair matrix. Figure 4—source data 1 Quantitative data for HFs dystrophy, melanin clumping and apoptosis in normal human scalp HF ex vivo co-cultured with ILC1lc and CD8+/NKG2D+ cells. https://cdn.elifesciences.org/articles/80768/elife-80768-fig4-data1-v1.xlsx Download elife-80768-fig4-data1-v1.xlsx ILC1lc induces HF immune privilege collapse ex vivo via NKG2D stimulation Given that AA cannot occur without the prior collapse of HF immune privilege [HF-IP] (Gilhar et al., 2012; Bertolini et al., 2020), we also investigated the impact of ILC1lc on key HF-IP markers. Indeed, the co-culture of HFs with ILC1lc triggered IP collapse, as evidenced by ectopic and overexpressed HLA-A,B,C, ß2-microglobulin (ß2 MG), and HLA-DR, along with overexpression of the ‘danger’/tissue distress signals, MICA and CD1d, which interact with and stimulate NKG2D (Uchida et al., 2021; Fan et al., 2022) as compared to HFs interacting with PBMC/PHA or with ILC3 cells (Figure 5A, B, C, D, E and F). Figure 5 Download asset Open asset Characteristic immunopathology features of alopecia areata (AA) hair follicles (HFs). (A) MICA, (B) HLA-A,B,C, (C) β2 MG, (D) HLA-DR, and (E) CD1d, expression by HFs epithelium, which had been co-cultured with either ILC1lc or CD8+/NKG2D+ cells but not in the control HFs, which had been co-cultured with either ILC3s, PBMCs/PHA, ILC1lc /NKG2D neutralization or in the untreated HFs. (F) quantitation. (G) The immune inhibitory HF immune privilege guardians, α-MSH and (H) TGF-β1 almost disappeared in HFs/ ILC1lc and HFs/NKG2D but were prominently present in ILC1lc /NKG2D neutralization and control HFs, N=9–12 HFs/group from three independent donors, three areas were evaluated per section. Following Shapiro-Wilk test,Student’s t-test, *p<0.05. Scale bar, 100 µm. ORS - outer root sheet. Figure 5—source data 1 Quantitative data for characteristic immunopathology features of AA HFs. https://cdn.elifesciences.org/articles/80768/elife-80768-fig5-data1-v1.xlsx Download elife-80768-fig5-data1-v1.xlsx Notably, quantitative immunohistomorphoemtry (qIHM) also showed that protein expression of the immunoinhibitory HF-IP guardians, TGF-β1 and α-MSH (Gilhar et al., 2012; Bertolini et al., 2020; Paus et al., 2018; Ito et al., 2004), almost disappeared in the epithelium of HFs co-cultured with autologous ILC1lc or CD8+/NKG2D+ cells (=positive control) (Figure 5G, H and I), while these critical HF-IP guardians were still prominently expressed in negative control HFs (Figure 5G, H and I). Importantly, adding anti-NKG2D blocking antibodies prevented HFs IP collapse and preserved the IP in the ILC1lc /NKG2D treated group (Figure 5G, H and I). This demonstrates that autologous ILC1lc induces human HF-IP collapse ex vivo – incidentally, the first time that the induction of IP collapse by ILC1lc has been documented in an intact human tissue/organ. ILC1lc are activated by ‘stressed’ HFs and induce premature catagen development via IFN-γ secretion Next, we examined how autologous ILC1lc impacted on human HF cycling, given that premature induction of apoptosis-driven HF regression (catagen) is one of the hallmarks of AA (Gilhar et al., 2012; Bertolini et al., 2020; Messenger et al., 1986). This showed that ILC1lc significantly accelerated the transformation of anagen into catagen HFs ex vivo (Paus et al., 2005) compared to all three negative controls (ILC2, ILC3, or PBMCs/PHA) – thus eliciting the third hallmark of the AA phenotype besides HF-IP collapse and dystrophy ex vivo (Gilhar et al., 2012; Bertolini et al., 2020; Messenger et al., 1986; Figure 6A), just as we had previously shown for Vδ1+ γδT cells (Uchida et al., 2021). As expected (Gilhar et al., 2012; Pratt et al., 2017; Ber