Innate type 2 immunity controls hair follicle commensalism by Demodex mites

Summary

Demodex mites are commensal parasites of hair follicles (HF). Normally asymptomatic, inflammatory outgrowth of mites can accompany malnutrition, immune dysfunction and aging, but mechanisms restricting Demodex outgrowth are not defined. Here, we show that control of mite HF colonization in mice required innate lymphoid cells (ILC2s), interleukin-13 (IL-13), and its receptor, IL-4Ra-IL-13Ra1. HF-associated ILC2s elaborated IL-13 that attenuated HF and epithelial proliferation at anagen onset; in their absence, Demodex colonization led to increased epithelial proliferation and replacement of gene programs for repair by aberrant inflammation leading to loss of barrier function and HF exhaustion. Humans with rhinophymatous acne rosacea, an inflammatory condition associated with Demodex, had increased HF inflammation with decreased type 2 cytokines, consistent with the inverse relationship seen in mice. Our studies uncover a key role for skin ILC2s and IL-13, which comprise an immune checkpoint that sustains cutaneous integrity and restricts pathologic infestation by colonizing HF mites.

eTOC blurb

Type 2 cytokines are well recognized for their role in mediating allergic pathologies in skin but their role in normal skin physiology is unclear. Ricardo-Gonzalez et al. reveal that type 2 immunity restrains skin inflammation in response to injury that is necessary to control hair follicle commensalism by Demodex mites.

Graphical Abstract

Introduction

Type 2 immunity is orchestrated by the production of canonical cytokines, such as interleukin-(IL)-4, IL-5, and IL-13, that are induced during allergic inflammation or infection by parasitic helminths (Gause et al., 2020). Although host adaptive type 2 immunity engages and typically dominates these chronic afflictions, work over the last decade has revealed innate immune cells, particularly Group 2 innate lymphoid cells (ILC2s), as key sources of initial type 2 cytokines in involved tissues. ILC2s develop first during fetal hematopoiesis and position in tissues where they expand perinatally and express effector cytokines like IL-5 and IL-13 at birth (Schneider et al., 2019). Prevalent at mucosal barriers, ILC2s are also in skin (Kobayashi et al., 2019; Ricardo-Gonzalez et al., 2018), where roles in tissue repair have been demonstrated after injury or bacterial invasion (Sakamoto et al., 2021), although contributions of intermittently or constitutively produced type 2 cytokines to basal physiology remain poorly understood.

While investigating an acquired inflammatory skin condition among type 2 immunodeficient mice in the specific pathogen-free (SPF) animal facility at UCSF, we identified a critical role for innate type 2 immunity involving ILC2s in sustaining skin homeostasis in the presence of otherwise benign commensalism by Demodex. Demodex mites have co-evolved over millions of years as obligate commensal ectoparasites inhabiting the hair follicles (HF) and sebaceous glands of mammals (Palopoli et al., 2015; Thoemmes et al., 2014). Relatively short-lived (2-3 weeks), the minute arthropods eat sebum and keratinocyte debris and produce 1-2 dozen eggs that differentiate through intermediate larvae to generate progeny that spread to nearby follicles unless constrained by undefined host factors (Rather and Hassan, 2014). Among humans, prevalence of asymptomatic Demodex commensalism, particularly on the face, reaches essentially 100% with age, but outgrowths can be provoked by skin damage, immunodeficiency and malnutrition, and have been associated with rosacea (Georgala et al., 2001; Zhao et al., 2010), prompting the addition of anti-mite treatment in patients afflicted with inflammation in the setting of a high Demodex burden. Clinically, inflammation of the eyelids (blepharitis, which may overlap with ocular rosacea) observed in patients treated with anti-IL-4Ra for atopic dermatitis (Barbe et al., 2021; Simpson et al., 2016) is associated with increased Demodex spp. infestation involving the dense HF landscape of the eyelid (Heibel et al., 2021; Quint et al., 2020), suggesting that inhibition of basal IL-4Ra signaling might precipitate Demodex overgrowth and HF inflammation.

Using genetic models of adaptive and innate immunity in combination with co-housing and cell transfer experiments, our studies show that loss of type 2 cytokine signaling underlies failure to control HF colonization by Demodex. Patients with rhinophymatous rosacea, an inflammatory dermopathy of the nose associated with heavy Demodex infestation, had increased HF inflammation but reduced IL-13 as compared with HF incidentally commensalized by Demodex among patients without rosacea. Finally, we reveal an unanticipated role for ILC2s and IL-13 in HF and cutaneous homeostasis after perturbation that is activated and sufficient for control of this ubiquitous commensal of mammalian skin.

Results

Demodex outgrowth and inflammatory dermopathy occurs in the absence of type 2 cytokine signaling

Around 8-12 weeks after birth, homozygously-maintained Il4ra^−/−, Il4^−/−,Il13^−/−, and Stat6^−/− C57BL/6 mice (here designated type 2 immunodeficient) in our SPF facility developed prominent facial folds and blepharitis, with loss of coat pigmentation that progressed with age (Figure 1A). Microscopic examination of the skin from these type 2 immunodeficient mouse strains revealed acanthosis (thickening of the epidermis), aberrant HF morphology, and enlarged sebaceous glands (Figure 1B). We also observed increased proliferation and numbers of HF stem cells (HFSCs), and increased numbers of HFs (Figure S1, A to D).

Figure 1. Loss of type 2 immunity results in dermatitis associated with Demodex infestation.

(A) Representative facial and fur phenotype of wild type (WT), Il4ra^−/−, Il4^−/−,Il13^−/− and Stat6^−/− mice at homeostasis. Mice shown are retired male breeders at 18-20 months of age. (B) Skin sections from WT, Il4^−/−,Il13^−/−, Il4ra^−/− or Stat6^−/− mice stained with H&E. Arrowheads highlight infestation by Demodex mites in hair follicles and sebaceous glands. Brackets emphasize epidermal acanthosis, s.g., sebaceous gland. Scale bar, 100 μm. (C) Brightfield image of a Demodex mite from a fur pluck of Il4ra^−/− mice. Scale bar, 50 μm. (D) Skin sections from WT, Il4^−/−,Il13^−/−, Il4ra^−/− or Stat6^−/− were stained for chitin (green) using an enhanced GFP chitin-binding domain fusion (eGFP-CBP) and DAPI (blue). Arrowheads highlight the Demodex mites. Scale bar, 100 μm. (E) PCR for Demodex chitin synthase (CHS) from fur plucks. Images are representative of n ≥ 5 mice of both sexes for each genotype. (F) SPADE representation of CD45⁺ cells from epidermal fraction of 8-10 weeks old WT, Il4ra^−/−, and Il4^−/−,Il13^−/− mice. One representative plot shown, n = 4 individual mice per genotype. (G) Number of ILC2s in skin from WT, Il4ra⁻⁻, and Il4^−/−,Il13^−/− mice. (H) Total CD4 in skin from WT, Il4ra⁻⁻, and Il4^−/−,Il13^−/− mice. (I) Frequencies of skin Treg cells (CD3⁺CD4⁺FoxP3⁺, as a percentage of total CD4) in skin from WT, Il4ra^−/−, and Il4^−/−,Il13^−/− mice at homeostasis. (J) Total CD8 in skin from WT, Il4ra^−/−, and Il4^−/−,Il13^−/− mice. Data are from one representative experiment (F), or pooled from multiple independent experiments (G–J). * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001 by two-tailed Student’s t test (vs. WT). Please also see Figure S1.

These alterations were accompanied by infestation with HF-dwelling mites with characteristics of Demodex species (Figure 1B). We used morphologic criteria of mites from hair pluck specimens (Figure 1C) and sequencing of the 18S ribosomal RNA gene to identify the mites as Demodex musculi (Table S1), a relative of human-associated Demodex folliculorum. To localize mites by their chitinous exoskeleton, we generated an enhanced-GFP chitin-binding domain protein (eGFP-CBP) and examined involved skin using immunofluorescence (IF). We observed mites, often multiple, in essentially all HF and sebaceous glands of back skin in type 2 immunodeficient mice (Figure 1D). To assess the Demodex burden, we developed a PCR assay for the highly conserved Demodex chitin synthase gene (Zhao et al., 2012). Affected type 2-deficient mice were PCR positive whereas separately maintained, microscopically negative WT mice were PCR negative (Figure 1E). Thus, mice genetically lacking diverse arms of type 2 immunity developed uncontrolled infestation by an otherwise commensal skin ectoparasite, in line with previous reports of infrequent outbreaks of Demodex mites in mouse facilities (Smith et al., 2016).

We next sought to understand how immune cells and cytokines were altered in affected mice to explore immune mechanisms necessary for restraining Demodex overgrowth. Both flow- and mass cytometry (CyTOF) of immune cells in skin revealed significant expansion of ILC2s, regulatory T (Treg) cells, and CD4⁺ and CD8⁺ T cells in type 2 immunodeficient mice as compared to wild-type (WT) C57BL/6 mice (Figures 1, F to J and S1, E to G). Skin ILC2s were expanded 20-30-fold in Demodex-infected type 2 immunodeficient strains. In prior studies, proliferation of resident ILC2s was shown to drive their extrusion from tissues into the blood (Campbell et al., 2019; Huang et al., 2018; Ricardo-Gonzalez et al., 2020). Consistent with marked expansion of skin ILC2s in the presence of high numbers of mites, skin inflammation in Demodex-infested IL4Ra-deficient mice was accompanied by increased circulating ILC2s with the characteristic IL-18R⁺ skin phenotype (Figure S1, H and I). We also quantified elevated serum IL-5 and IL-13 (in IL-13-sufficient Il4ra^−/− mice), prototypical ILC2 cytokines (Figure S1J), but also increased serum IL-22 (Figure S1J), suggesting transition of ILC2s to an inflammatory type 2 and type 3 immunophenotype as reported in other contexts (Bielecki et al., 2021; Reynolds et al., 2021). Indeed, skin ILC2s from Demodex-infected Il4ra^−/− mice expressed Il22 and secreted IL-22 when activated in vitro (Figure S1, K and L). Thus, the absence of type 2 cytokines leads to overgrowth by cutaneous Demodex resulting in lymphocytic infiltrates accompanied by marked ILC2 activation, entry into blood and transition to an inflammatory phenotype.

Human rhinophyma with Demodex is associated with folliculitis and reduced IL13

The morphologic facial changes observed in infected type 2 immunodeficient mice resembled changes observed in patients affected by rhinophymatous acne rosacea associated with Demodex (Forton and Seys, 1993). Rhinophyma is a chronic manifestation of rosacea characterized by acanthosis, inflammation and enlargement of the sebaceous glands affecting the nose of older adults (Figure 2A). To investigate correlates of the type 2 immune axis in humans that promote stable commensalism with Demodex, we analyzed excisions of nasal skin from adults with and without rhinophyma. In normal skin, we identified incidental Demodex mite infestation limited to individual HF accompanied by mild inflammatory infiltrates that included more IL13- than IL4-expressing cells as assessed by RNA in situ hybridization (ISH) (Figure 2, B to E). In contrast, patients with rhinophyma had significantly denser inflammatory infiltrates but fewer IL13- and IL4- expressing cells surrounding HF (Figure 2, B to E). Dual IL13 ISH and CD3ε immunohistochemistry (IHC) showed that approximately 40% of IL13-expressing cells in normal skin were CD3ε negative, consistent with production of this cytokine by both T cells (IL13+CD3ε+) and innate lymphocytes (IL13+CD3ε−) (Figure 2F–G). We also assessed the expression of IFNG and IL22 among these patients for evidence of immune deviation and found that IFNG and IL22 expression was increased in rhinophyma compared to normal skin (Figure 2, H to K). These data show that patients with rhinophyma and Demodex, which phenotypically resembles mite infection in type 2 immunodeficient mice, have inflammation and diminished IL13 production, thus encouraging us to investigate further the mechanisms by which type 2 immunity facilitates stable commensalism by this ubiquitous ectoparasite.

Figure 2. Type 2 cytokines expression is associated with healthy Demodex commensalism in humans.

(A) Low power H&E stained sections of normal skin (top) and rhinophyma (bottom). (B) H&E stained sections (left panels) and corresponding IL13 mRNA in situ hybridization (ISH, right panels) in hair follicles infected with Demodex in representative cases of normal skin and rhinophyma. Arrowheads highlight Demodex infection of the hair follicle. Dashed lines highlight perifollicular inflammation. Scale bars, 250 μm. (C to E) Scoring of hair follicle inflammation (C) and quantification of IL4 (D) and IL13 (E) expression by ISH. (F and G) Representative co-stained section for IL13 by ISH and CD3ε by immunohistochemistry (F) and quantification of proportion of CD3⁺ or CD3⁻ IL13-expressing cells (G). Arrowhead highlights IL13⁺CD3⁻ cell. Scale bar, 10 μm. (H and I). Representative IFNG mRNA ISH (H) and quantification of IFNG expression (I) in hair follicles infected with Demodex. (J and K). Representative IL22 mRNA ISH (J) and quantification of IL22 expression (K) in hair follicles infected with Demodex. Scale bars, 250 μm. Statistical significance shown by * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001 by two-tailed Student’s t test.

Type 2 immunodeficient mice fail to control HF colonization by Demodex

Spontaneous skin inflammation is not typical among type 2 immunodeficient strains in SPF facilities. When affected IL-4Ra-deficient mice were re-derived and bred in separate SPF rooms, none developed dermopathy, skin ILC2 expansion or Demodex infection (Figure 3, A, G, and H). After housing with mite-infested IL-4Ra-deficient mice, however, unaffected IL-4Ra-deficient mice were colonized with Demodex and developed both the dermopathy and the immune phenotypes described above over a period of 6-to-12 weeks (Figures 3, B to G).

Figure 3. Co-housing Demodex-infected with non-infected Il4ra^−/− mice transfers the phenotype.

(A) Facial appearance of Il4ra^−/− mice infected (affected) or uninfected (unaffected) with Demodex musculi. (B) Schematic of the co-housing experiment. 8 week-old (w.o.) Il4ra^−/− mice infected with Demodex (Group 1) were co-housed with unaffected 3 w.o. Il4ra^−/− mice (Group 2). Unaffected littermates of co-housed mice (Group 3) were allowed to age concurrently. WT (Group 4) and Il4ra^−/− from known Demodex infected parents (Group 5) were aged as independent controls. (C) Representative mice at 3 months after the co-housing experiment. G1-G5 indicate experimental groups as outlined in (B). (D and E) Number of skin ILC2 (D) and CD4⁺ (E) cells in back skin. (F) Frequency of Treg cells (as percentage of total CD4) in back skin. (G) Sections from back skin were stained with H&E. Inset highlights Demodex infestation of the hair follicle and sebaceous glands. Scale bar, 100 μm. (H) PCR for Demodex chitin synthase gene (CHS, top) or genomic DNA for the keratin 5 gene (Krt5, bottom) from 2mm punches of back skin. Quantification of relative band intensity (CHS/Krt5) is shown on the right. Data are from one representative experiment of two independent experiments. Statistical significance shown by * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001 by two-tailed Student’s t test. Please also see Figures S2, S3 and S4.

To further examine transmissibility, we bred uninfected WT mice to Demodex-infected Il4ra^−/− mice, which were maintained as homozygotes, and subsequently intercrossed the phenotypically normal-appearing heterozygous F1 littermates to generate WT, heterozygous (Il4ra^+/−), and homozygous Il4ra^−/− F2 mice. Although the F2 generation was derived from previously affected animals, neither WT, heterozygous, nor genetically ablated (Il4ra^−/−) littermates developed the facial or skin inflammation phenotypes when derived from unaffected heterozygous parents descendent from an affected Il4ra^−/− grandparent (Figure S2, A to E). However, F1 Il4ra^−/− mice derived from the intercross of affected Il4ra^−/− and unaffected heterozygous Il4ra^+/− mice developed the inflammatory phenotype (Figure S2F), suggesting that direct parental transfer of mites from and to type 2 immunodeficient offspring is important for development of this phenotype. Appearance of the skin and facial phenotypes was accompanied by increases in skin ILC2s, CD4⁺ T cells and Treg cells, as well as the presence of IL-13 and IL-22 in serum and an increase in follicular mites (Figure S2, G to O). Thus, under these conditions, functional type 2 immunity, even in heterozygous genotypes, suppresses Demodex below what is required for transmission and progression to outgrowth and dermopathic phenotypes in offspring.

Targeted anti-mite but not anti-bacterial therapy prevents the inflammatory dermopathy

Loss of skin ILCs leads to sebaceous gland hypertrophy with altered sebum constituents and effects on the skin bacterial microbiota (Kobayashi et al., 2019). To this end, we performed 16S rDNA amplicon sequencing of skin samples from uninfected WT and Demodex infected and uninfected Il4ra^−/− mice (Figure S3A). Although bacterial diversity was similar between phenotypes, analysis of weighted UniFrac distance between samples showed differences in overall community composition (Figure S3B, p = 0.001, PERMANOVA test). Looking at the broad characteristics of the bacteria, we found that the fraction of obligate anaerobes was higher in uninfected mice (p = 0.028, Figure S3C), consistent with the normal anaerobic environment of the healthy HF (Chen et al., 2018). Despite these microbiota differences, treatment of infected Il4ra^−/− mice with broad-spectrum antibiotics had no effect on the phenotype (Figure S3D). In contrast, treatment of 6-week-old infected Il4ra^−/− littermates (before they show the visual phenotype) with topical anti-mite agents moxidectin and imidacloprid (Nashat et al., 2018), but not vehicle control (ethanol), decreased the mite burden and prevented the skin phenotype, in association with reductions in skin ILC2s and normalization of CD4⁺ T cells and Treg cells (Figure S4, A to G). We obtained comparable results when treating young (4-5 weeks) Demodex-infected Il4^−/−,Il13^−/− mice (Figure S4, H to P). Taken together, these data show that targeted anti-mite therapy reverses the immune disruption and prevents the facial phenotype induced by Demodex in the absence of type 2 immunity.

Colonization by Demodex drives accumulation of activated ILC2s in normal skin

Our data suggested that type 2 cytokine signaling is required to restrict Demodex outgrowth and prevent pathology after colonization. Although neither WT nor heterozygous mice develop the inflammatory dermopathy, we used co-housing to ask how ILC2s and type 2 cytokines respond to Demodex. To this end, we co-housed 3-to-5-week old Il4ra^−/− mice from infected colonies, before they acquired visible inflammation, with uninfected WT mice. While Il4ra^−/− mice developed the cutaneous and facial phenotype with age, as expected, co-housed WT mice remained visually unaffected (Figure 4A). Although Demodex-free at the start of co-housing, however, WT mice were colonized by low numbers of mites as established by histologic exam, eGFP-CBP IF, and PCR for Demodex chitin synthase (Figure 4, B to D). Infected WT mice developed expansion of skin ILC2s and had increased IL-13 and IL-5 expression in skin, but no cytokines in blood; expansion of activated CD4⁺ T cells and Treg cells was modest (Figure 4, E to H). Activated skin ILC2s appeared by 2 weeks and peaked 1-2 months after infection of WT mice at approximately half the frequency seen in infected IL-4Ra-deficient mice (Figure 4, I to N). Alarmin cytokines like thymic stromal lymphopoietin (TSLP), IL-25 and IL-33 are important for homeostatic activation of tissue ILC2s; IL-18 may play a similar role in skin (Ricardo-Gonzalez et al., 2018). To assess whether typical alarmins are required for activation of skin ILC2s in response to Demodex, we generated Tslp^−/−; Il18^−/−; Il1rl1^−/−; Il25^−/− quadruple ablated (Quad-ablated) mice together with Il5^Red5 and Il13^Smart13 reporters, and co-housed the mice with WT and Demodex-infected IL-4Ra-deficient mice for 8 weeks. At baseline, Quad-ablated mice had reduced numbers of IL-5- and IL-13-expressing skin ILC2s as compared to WT mice (Figure 4, O and P). Although reduced, skin ILC2s from alarmin-deficient mice retained the proportionate capacity to expand and increase IL-5 and IL-13 in response to mites in association with an attenuated but intermediate Demodex burden as compared to WT mice and sufficient to block development of the facial phenotype (Figure 4, O to S). Taken together, these data show that Demodex colonization provokes expansion and activation of skin ILC2s by a mechanism partially dependent on endogenous alarmins responsible for setting basal innate type 2 immune tone, but absolutely dependent on elaboration of type 2 cytokines to restrict mite outgrowth and attenuate the inflammatory dermopathy.

Figure 4. Type 2 immunity sustains non-pathologic Demodex colonization.

(A) Representative wild type (WT), and co-housed WT with Demodex infested Il4ra^−/− mice. Three week old WT and infested Il4ra^−/− mice were co-housed for 12 weeks. (B) Skin sections from co-housed WT or Demodex infected Il4ra^−/− were stained for chitin (eGFP-CBP, green) and DAPI (blue). Scale bar, 100 μm. (C) PCR for Demodex chitinase synthase gene (CHS, top) or genomic DNA for the keratin 5 gene (Krt5, bottom) from back skin. Quantification of relative Demodex infestation (CHS/Krt5) is shown on the right. (D) H&E-stained sections of back skin from WT, or co-housed WT and Il4ra^−/− mice. Scale bar, 100 μm. (E–G) Number of skin ILC2s (E), total CD4 (F), and frequency of Treg cells (CD3⁺CD4⁺FoxP3⁺ T cells) as a percentage of total CD4 (G). (H) Quantification of serum IL-13 and IL-22. (I) IL-5 and IL-13 expression by ILC2s in Demodex infection. WT Il5^{tdTomato(Red5)/+}, I13^{hCD4(Smart13)/+} littermates were separated (Control) or co-housed with infested Il4ra^−/− mice (Co-housed) for 2 weeks (wk), 1 month (mo) or 2 mo. Flow cytometry plots show IL-5 and IL-13 expression by skin ILC2 (Red) or CD4 (blue). (J–L) Quantification of ILC2s (J), and their expression of IL-13 (hCD4, Smart13, (K)), and IL-5 (L). (M–N) Quantification of total CD4 (M) and their expression of IL-13 (hCD4, Smart13, (N)). (O) Representative wild type (WT), and Crlf2^−/−,Il1rl1^−/−,Il18^−/−,Il25^−/− (Quadablated) mice co-housed with Demodex infested Il4ra^−/− mice. (P) Number of skin ILC2s and IL-13 expressing (Smart13) ILC2s. (Q) H&E-stained sections of back skin from WT, and Quadablated mice co-housed with Demodex-infected Il4ra^−/− mice. Scale bar, 100 μm. (R) PCR for Demodex chitinase synthase gene (CHS, top) or genomic DNA for the keratin 5 gene (Krt5, bottom). (S) Quantification of percentage of hair follicles infected with Demodex. Data presented as mean ± s.e.m. Statistical significance shown by * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001 by one-way ANOVA. Please also see Figure S5.

Although both IL-4 and IL-13 signal through the IL-4Ra receptor, IL-4-deficient mice (Il4^−/−; (Mohrs et al., 2001) controlled Demodex as well as WT mice when co-housed with infected IL-4Ra-deficient mice (Figure S5, A to G) suggesting that IL-13 is necessary for Demodex control. We generated IL-13Ra1-deficient mice and co-housed these mice with mite-infected IL-4Ra-deficient mice. Il13ra1^−/− mice were highly susceptible to Demodex outgrowth and developed the dermopathic facial phenotype; heterozygous mice showed partial control (Figure S5, H to J). We additionally co-housed Il4^−/− and Il13ra1^−/− mice with infected mice to demonstrate the requirement for IL-13 but not IL-4 for control of hair follicle infestation (Figure S5, K to N). Taken together, these experiments in mice and observations from human patients support a key role for type 2 immunity and IL-13 acting through the IL-4Ra and IL-13Ra1 receptor in regulating stable HF colonization by skin mites. In the absence of these constituents of type 2 immunity, Demodex overgrowth appears to be the key trigger driving inflammation and ILC2 activation, immune deviation, and appearance in blood.

The adaptive immune system is not required to control Demodex

In our initial characterization of Demodex-infected type 2 immunodeficient strains, we observed expansion of ILC2s as well as adaptive immune cells. To ascertain a necessary role for the adaptive immune system in Demodex control, we co-housed infected IL-4Ra-deficient mice with uninfected Rag1-deficient mice, which lack adaptive B and T cells, including Treg cells. Like infected WT mice, Rag1-deficient animals remained visually unaffected and restrained Demodex colonization in contrast to co-housed IL-4Ra-deficient mice (Figure S6, A to D). As in co-housed WT mice, IL-18R⁺ skin ILC2s expanded in cohoused Rag1^−/− mice relative to uninfected controls (Figure S6, E to H). These data suggested a role for innate immune cells in controlling Demodex colonization. To explore this hypothesis, we co-housed infected Il4ra^−/− mice with uninfected Rag2^−/− mice, which lack B and T cells, and Rag2^−/−,Il2rg^−/− mice, which also lack ILCs. Within 4-6 weeks of co-housing, Rag2^−/−,Il2rg^−/− mice, but not Rag2^−/− cage mates, developed the characteristic facial phenotype, blepharitis and increased mite burden observed in Demodex-infected type 2 immunodeficient mice (Figure 5, A to D). Demodex-infected Rag2-deficient mice had minimal epidermal proliferation (Figure 5E), while Rag2^−/−,Il2rg^−/− mice developed hyperplasia, hyperkeratosis and epithelial proliferation. Thus, under these conditions, the adaptive immune system is not required for control of Demodex colonization nor for the expansion of skin ILC2s in response to Demodex, whereas innate lymphoid cells are critical for control of mite burden and prevention of the inflammatory dermopathy.

Figure 5. ILC2s are sufficient to restrain Demodex infection.

(A) Representative Rag2^−/− and Rag2^−/−,Il2rg^−/− co-housed with Demodex infested Il4ra^−/− mice. 5-week old Rag2^−/− and Rag2^−/−,Il2rg^−/− were co-housed with Demodex infested Il4ra^−/− for 6 weeks. n = 3 to 5 mice per group, representative of two independent experiments. (B) PCR for Demodex chitinase synthase gene (CHS, top) or genomic DNA for the keratin 5 gene (Krt5, bottom) from back skin. (C) Quantification of relative Demodex infestation (CHS/Krt5) for the PCR in (B). (D) H&E-stained sections of eyelid (top) and back (bottom) skin from unaffected Rag2^−/− and Rag2^−/−,Il2rg^−/− with or without co-housing with a Demodex infested Il4ra^−/−. (E) Sections from back skin of co-housed Il4ra^−/−, Rag2^−/− and Rag2^−/−, Il2rg^−/− were stained for EdU (green), P-Cadherin (red) and DAPI (blue). (F) Schematic of the adoptive transfer experiment. Il4ra^−/− mice infected with Demodex were co-housed with unaffected 4-8 w.o. Rag1^−/− Arg^YFP/YFP, Il5^{tdTomato(Red5)/tdTomato(Red5)} (Ragl^−/− YYRR) mice for 1 month. ILC2s (CD45⁺Lin⁻ Thy1⁺Red5⁺) were sorted from skin, lung, and fat, pooled, and transferred to Rag2^−/−,Il2rg^−/− mice two weeks prior and at the start of co-housing (see methods). (G) Representative Rag2^−/−,Il2rg^−/− treated with PBS (top) or with transferred ILC2s (bottom) after 6 weeks of co-housing with Demodex infected Il4ra^−/−. (H) PCR for Demodex chitinase synthase gene (CHS, top) or genomic DNA for the keratin 5 gene (Krt5, bottom) from back skin. (I) Quantification of relative Demodex infestation (CHS/Krt5) for the PCR in (H). (J) Flow cytometry plots of ILCs (pre-gated on Live CD45⁺Lin⁻Thy1⁺) showing IL-5 expression (left panels) or Arginase (middle and right panels) on skin and lung as indicated. (K) H&E-stained sections of back skin from Rag2^−/−,Il2rg^−/− treated with PBS or ILC2 adoptive transfer and co-housed with Demodex infested Il4ra^−/−. n = 3 to 5 mice per condition. Arrowheads highlight Demodex mites. Scale bars, 100 μm. For the ILC2 transfer experiments, a representative mouse is shown from 2 independent experiments with n = 2-3 mice per group. Please also see Figure S6.

To test the sufficiency of ILC2s, we sorted ILC2s from Rag1^−/− mice (expressing Arg11 and Il5 reporter alleles as described (Bando et al., 2015; Nussbaum et al., 2013) that were co-housed for one month with Demodex-infected Il4ra^−/− mice in order to activate and expand ILC2s in vivo (Figure 5F). Primed ILC2s were purified and transferred into Rag2^−/−,Il2rg^−/− mice prior to and at the start of co-housing with affected Il4ra^−/− mice. After 6 weeks, we detected transferred ILC2s in the reconstituted Rag2^−/−,Il2rg^−/− mice, which had no facial disfigurement, minimal hyperplasia by skin histology, and reduced Demodex burden as compared to colonized Rag2^−/−,Il2rg^−/− mice that did not receive ILC2s (Figure 5, G to K). Together, these experiments support a fundamental role for skin ILC2s in control of epithelial homeostasis in response to HF colonization by Demodex.

Type 2 immunodeficient mice activate inflammatory rather than repair responses to Demodex colonization

To understand the impact of Demodex HF colonization on skin homeostasis, we analyzed immune and non-immune cells from type 2 immune-sufficient or -deficient mice infected with Demodex using scRNA-seq. In agreement with flow cytometric studies, Demodex HF colonization in WT mice led to expansion of IL-18R⁺ ILC2s (Figure 6, A and B), which were activated as indicated by expression of Il13, Areg, and Il17a (Figure 6, C to E), and mirroring effector programs associated with tissue repair by skin T cells (Harrison et al., 2019). In type 2 immunodeficient mice, there were increases in ILC2s, Treg cells and CD8⁺ T cells, and the appearance of an ILC2-ILC3 transitional population, consistent with our prior findings (Figures 6, F to H; 1, F to J; and S1, J to L). Of note, skin lymphocytes in infected WT mice mounted an enhanced type 2 immune response whereas lymphocytes in type 2 immunodeficient strains deviated towards type 1 (Ifng) and type 3 (Il22) responses (Figure 6, H and I).

Figure 6. scRNA-seq of CD45+ cells from Demodex-infected WT and type 2 immunodeficient mice reveal divergent responses.

(A) UMAP projection of CD45⁺ cells from skin of WT uninfected and Demodex-infected mice. (B) Bar plot showing the percentage of cells in each cluster for all lymphocyte (left) and myeloid and granulocyte (right) populations by experimental condition. (C) UMAP projection of ILC subclustering analyses showing experimental condition. (D) Feature plot of ILC subclusters showing expression of genes associated with ILC2 subset identity in skin. (E) Feature plots showing expression of genes for Il13, Il17a, and Areg. (F) UMAP projection of the skin lymphocyte fraction of WT (uninfected), and Il4ra^−/− or Il4^−/−,Il13^−/− (Demodex-infected) mice showing subset of lymphocytes. Red circle denotes ILC2-ILC3 population present only in Il4ra^−/− or Il4^−/−,Il13^−/− samples. (G) Bar plot showing the percentage of cells in each cluster for the samples in (F). (H) Density plots from the combined object as in F, showing the expression for selected cytokines. (H) Dot plot of a combined analysis for all lymphocytes in all samples. Please also see Figure S7.

Analysis of CD45⁻ skin cells from control and Demodex-infected WT mice yielded similar representation between the samples among major skin cell populations (Figure S7, A and B). Gene ontology pathway analysis in Demodex-infected WT mice showed induction of pathways associated with extracellular matrix conferring tensile strength, including induction of collagen proteins which were not as highly expressed in Demodex-infected Il4ra^−/− or Il4^−/−,Il13^−/− mice (Figure S7, C and D). In contrast, infected Il4ra^−/− and Il4^−/−,Il13^−/− mice revealed induction of pathways associated with host defense and inflammation, response to bacteria (S100a8, S100a9, Defb6, Stfa1, Srpinb1a), and antigen presentation (Tap1, H2-M3, H2-D1, Cd74, H2-Q6) (Figure S7, E and F).

To assess skin cells that respond to IL-13, we examined cell populations that expressed Il4ra, Il13ra1 or both, noting that the epithelium had a high correlation of Il4ra and Il13ral co-expression (Figure S7G). Subclustering and biological processes analyses of the epithelia and Lgr5-expressing HF stem cells (HFSCs) revealed a biological response primed for wound healing in Demodex-infected WT mice (Figure S7H). However, this reparative response was absent among epithelia and HFSCs from infected Il4ra^−/− and Il4^−/−,Il13^−/− mice, and replaced by a response characteristic of replicating cells, including DNA synthesis and ribosomal processing (Figure S7, I and J). Together, these data support a role for the ILC2-IL13 pathway in inducing the restorative and regenerative response necessary to sustain anatomic integrity during colonization by HF mites.

Type 2 cytokines reduce HFSC proliferation and delay HF regrowth

To investigate mechanisms by which IL-13 from ILC2s confers protection of HF after mite colonization, we used well-characterized reporter mice for the type 2 cytokines IL-5 and IL-13 (Liang et al., 2011; Nussbaum et al., 2013) to assess the location and activation of ILC2s in resting skin. ILC2s were the most prevalent cells expressing these cytokines under basal conditions, corroborating prior findings (Mayer et al., 2021; Ricardo-Gonzalez et al., 2018; Schneider et al., 2019), and localized predominantly within the epithelial layer in close association with HF (Figure 7, A and B). We observed cyclic IL-13 expression in skin ILC2s, with high expression in the early postnatal period as previously observed (Schneider et al., 2019), and then oscillating expression with phases of the hair cycle, which remains synchronous in the first months of life, with low IL-13 during telogen (rest) and high IL-13 during anagen (growth) (Figure 7C). To confirm that cells from IL-13 reporter mice accurately reflect Il13 gene expression, we sorted ILC2s from back skin of mice in telogen or anagen phases of the hair cycle and confirmed the increase in Il13 during anagen by quantitative PCR (Figure 7D). These observations suggest that ILC2s and IL-13 may support the cyclic growth and turnover of the HF compartment (Naik et al., 2018), particularly at onset of anagen when this compartment may be susceptible to the 2-3 week reproductive cycle of Demodex.

Figure 7. Type 2 immunity sustains skin barrier and HF regeneration after perturbation.

(A) Localization of IL-5 producing ILC2s (red) within the skin epithelium. E-cadherin (green) is used to highlight hair follicles. DAPI is represented in blue. Scale bar, 20 μm. (B) Quantification (left) or percentage (right) of ILC2s (CD45⁺Lin⁻Thy1⁺Red5⁺) in the epidermis and dermis of adult Il5^Red5/+ mice. (C) Quantification of expression of IL-13 by ILC2s throughout the various stages of the hair follicle cycle. (D) Quantitative PCR for Il13 expression by ILC2s (CD45⁺Lin⁻ Thy1⁺Red5⁺) sorted from mice back skin in telogen or anagen stage of the hair cycle. (E) Effect of IL-4 and IL-13 subcutaneous injection in hair growth. Wild type C57BL/6 mice were treated with PBS, IL-13 or IL-4 complex (IL-4c) on the first 2 days after depilation (Days 0 and 1) and their hair regrowth pattern was tracked over 2 weeks. One representative mouse per treatment arm is shown. (F) Quantification of hair regrowth from PBS, IL-4c, or IL-13 treated mice as in (E). (G) Representative flow cytometry panel of hair follicle stem cells (HFSCs, pre-gated as CD45⁻MHCII⁻CD34⁺Alpha6⁺) incorporation of EdU. Mice were depilated and treated with IL-4 complex (IL-4c) on the first two days after depilation. Mice were harvested on day 4 after depilation. (H) Percent EdU⁺ HFSCs four days after depilation as in (G). (I) Scaffold maps of Ki67 and IdU for the CD45⁻ epidermal cell fraction as analyzed by CyTOF (see methods). Black nodes represent canonical cell populations identified manually as noted in the top left scaffold map. Blue denotes the population has significant lower expression frequency of the marker (Ki67, or IdU, as denoted in the columns) in the experimental arm (IL-4 complex treated WT mice) vs. WT controls; red denotes significantly higher frequency. n = 4 individual mice per group. (J) Scaffold maps for phospho(p)STAT6 in IL-4 complex treated mice compared to WT controls. Blue and red colors denote statistically significant decrease or increase in expression frequency, respectively, in each cluster compared to the control group (q < 0.05 by SAM). (K) Transepidermal water loss at homeostasis in WT, Il4ra^−/− and Demodex-infested Ilra4^−/−, Il4^−/−Il13^−/−, and Stat6^−/− mice. (L) Transepidermal water loss at homeostasis in WT or WT mice co-housed with Demodex-infested Il4ra^−/− mice. (M) Lucifer yellow barrier assay. WT or Il4ra^−/− were shaved and depilated. Two days post-depilation, the epidermal surface was exposed to Lucifer Yellow overnight at 37 °C. Lucifer yellow in green, DAPI in blue. Scale bar, 100 μm. (N) Repeated depilation of wild type (WT) and Il4^−/−Il13^−/− mice. A representative animal from n = 4 individual mice shown. (O) Sections from back skin of WT and Il4^−/−,Il13^−/− mice after 3 cycles of depilation were stained with H&E. Scale bar, 100 μm. (P) Representative images of 2 year-old (yo) WT and Il4^−/−Il13^−/− mice. (Q) qPCR for Cdkn2a(p16) from sorted HFSCs from age-matched 2 yo WT and Il4^−/−Il13^−/− mice. Data presented as mean ± s.e.m and are representative of at least two independent experiments with n ≥ 3 mice per individual group/time point, pooled from two independent cohorts with n=5-10/group (K, L) or representative from an experiment with n ≥ 4 per group (P, Q). *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 as calculated with two-tailed Student’s t test (A to D, Q) or ANOVA (F, H, K, L).

To assess the role of type 2 cytokines on hair growth and HFSCs, we depilated mice to induce anagen and synchronous hair regrowth (Ali et al., 2017; Paus et al., 1999), and then administered IL-13 or IL-4, which signal via the common IL-4Ra receptor. Either IL-13 or IL-4, the latter administered as long-lived antibody complexes, slowed hair regrowth (Figure 7, E and F) and attenuated proliferation of HFSCs (Figure 7, G and H). We used CyTOF to assess the impact of type 2 cytokine signaling on skin epithelia, which express Il4ra and Il13ra1 (Figure S7G), and observed decrease in proliferation of the epithelial fraction as assessed by Ki-67 and uptake of the EdU analog, 5-Iodo-2-deoxyuridine (IdU). IL-4Ra signaling also elicited an increase in phospho-STAT6 (Figure 7, I and J). These findings, in combination with our scRNA-seq data (Figure S7), support a link between type 2 cytokines, epithelia, and HFs in sustaining skin homeostasis that is evident after colonization by mites.

Loss of type 2 immunity compromises cutaneous integrity in the setting of Demodex colonization

The loss of a skin reparative response seen in mite-colonized WT as compared to colonized Il4ra^−/− and Il4^−/−,Il13^−/− mice, together with loss of the normal obligate anaerobes from HF in infected Il4ra^−/− mice (Figure S3C), led us to hypothesize that type 2 immunity may be necessary to preserve HF and skin integrity in the setting of perturbation. To test this hypothesis, we measured transepidermal water loss (TEWL) in WT, IL-4Ra-deficient and Demodex-infected IL-4Ra-deficient mice. We noted barrier dysfunction in Demodex-infected IL-4Ra-deficient mice compared to uninfected IL-4Ra-deficient mice (separately housed) or Demodex-colonized WT mice (housed in the same cage as affected IL-4Ra-deficient mice) (Figure 7, K and L). Further, penetration of lucifer yellow applied to depilated skin was constrained to the epidermis and upper HF in WT mice, but diffused throughout the subcutaneous tissues in IL-4Ra-deficient mice, even in the absence of mite infection (Figure 7M).

The finding that depilation alone was sufficient to reveal loss of skin integrity in the absence of type 2 cytokines led us to hypothesize that the repetitive demands of persistent mite HF colonization on tissue repair might drive the appearance of the inflammatory dermopathy among type 2 immunodeficient mice by exceeding the limits for tissue resilience. To test this, we performed sequential depilations in WT or Il4^−/−,Il13^−/− mice infected with mites. Although WT mice re-grew hair with each cycle as predicted (Paus et al., 1999), Il4^−/−,Il13^−/− mice had increasingly slower kinetics of regrowth with subsequent depilations (Figure 7N). Histologic exam showed that WT mice maintained normal HF morphology and subcutaneous adipose tissue, whereas Il4^−/−,Il13^−/− mice had a more fibrotic dermis and HF structures were further compromised, consistent with alterations accompanying tissue senescence (Figure 7O) (Ge et al., 2020). Supporting this, aged Demodex-infected type 2 immunodeficient mice developed premature coat depigmentation as compared to similarly aged WT mice (Figure 7P). Analyses of HFSCs from 20-24 month-old, age-matched WT and Demodex-infected Il4^−/−,Il13^−/− mice showed an increase of the senescence-associated marker Cdkn2a/p16 in the absence of type 2 cytokines (Figure 7Q). Taken together, these data reveal a key role for skin ILC2s and innate type 2 immunity in enhancing skin resilience following inflammation in response to HF colonization by Demodex.

Discussion

Demodex mites are ubiquitous ectoparasites considered commensals of mammalian HFs (Rosshart et al., 2019; Sastre et al., 2016). As such, Demodex outbreaks remain infrequently reported among SPF mice facilities, but have been noted among strains deficient in type 2 immunity, as occurred here (Smith et al., 2016). As we have shown by co-housing and transfer experiments, activation of ILC2s, IL-13 and signaling through IL-4Ra were necessary and sufficient to limit proliferation and spread of mites once acquired in the pilosebaceous units of the skin. Further study revealed deficits in the ability of type 2 immunodeficient mice to sustain skin homeostasis not only following mite colonization, but also after depilation. Repetitive depilation resulted in evidence for stem cell failure and aged type 2 immunodeficient mice colonized with Demodex developed skin changes consistent with premature HF senescence, uncovering a non-redundant role for type 2 immunity in sustaining cutaneous integrity in the post-birth environment.

Despite their ubiquitous nature and robust reproductive potential, Demodex are maintained at low density on human skin, although mites proliferate in settings of immunodeficiency, malnutrition, and with aging, particularly in sun-exposed areas of the face (Chovatiya and Colegio, 2016; Lacey et al., 2011). Human-associated Demodex possess a reductive genome anticipated by obligate parasites that spend their entire life within the narrow niche of HFs. Demodex share chitinous exoskeletons with other arthropods, and survival in the sebum-rich HF has evolved a rich armamentaria of lipases, proteases and chitinases (Hu et al., 2019), an array of enzymes linked, like chitin itself, with tissue injury, alarmin induction and activation of innate type 2 immunity (Cheng and Locksley, 2014; Gieseck et al., 2018; Van Dyken et al., 2017). As we have shown, loss of canonical alarmins attenuated but did not ablate the ILC2 response, suggesting a prominent role for alarmins in regulating basal ILC2 numbers and tone, but revealing alternative pathways by which ILC2s react to tissue perturbations induced by this evolved HF commensal. Although skin development and basal homeostasis were largely normal in the absence of ILC2s or type 2 cytokines, perturbations like depilation, Demodex and aging resulted in dysregulation and loss of barrier function. In concert with additional signals, including via the epidermal growth fator receptor (EGFR), type 2 cytokines impacted small intestinal epithelial stem cell trajectories, hinting at shared pathways by which type 2 immunity can enhance resilience to perturbations at cutaneous and mucosal barriers (Sanman et al., 2021). Such adaptations in the gut contribute to the phenomenon of concomitant immunity by which helminth-infected hosts become resistant to further infestation by immature forms (Schneider et al., 2018). Whether HF colonization by Demodex imparts similar stem cell and epithelial ‘memory’ and protection against subsequent cutaneous injury will be important areas for further investigation (Gonzales et al., 2021).

Our studies suggest a role for IL-13 in transiently restraining HFSC and epithelial proliferation at anagen onset, revealing barrier abnormalities upon depilation in the absence of this cytokine. Perturbations that impact the skin HF niche by disruptions of cell junctions, migration and adhesion, pathways we show engaged by Demodex colonization, drive precocious anagen entry, resulting in re-directed proliferation of HFSCs towards ‘rescue’ states dedicated to preservation of the HF niche (Gur-Cohen et al., 2019; Harrison et al., 2019; Lay et al., 2018). Precocious anagen entry would be compatible with the increased numbers of proliferating HFSCs and HFs observed in mite-colonized type 2 immunodeficient mice. Of note, the epithelial and HFSC reparative pathways induced by IL-13 in naïve and mite-colonized WT animals were lost in type 2 immunodeficient mice and replaced by epithelial proliferation and an inflammatory cell influx that was not sufficient to control infestation. Given the marked HF anatomic complexities occurring among epithelial and endothelial structures during the transition through anagen (Yang et al., 2017), it is tempting to speculate that effects of IL-13 are critical to keep Demodex compartmentalized to regions of the HF above the isthmus, as compared to the deeper localization of organisms seen in type 2 immunodeficient mice. ILC2s in intestinal mucosa regulate homeostasis with the protozoan, Tritrichomonas (Howitt et al., 2016; Schneider et al., 2018), suggesting a role for innate type 2 immune cells in sustaining tissue function in the presence of barrier-associated eukaryotic pathobionts, perhaps reflecting responsiveness of ILC2s to endogenous damage-associated molecular patterns (DAMPS) rather than to genetically encoded microbial and viral recognition elements. Although we cannot exclude a role for microbiota-mediated inflammation contributing to the inflammatory phenotype, the infection was controlled by topical drugs targeting the mite but not by antibiotics.

Our observations of increased type 2 inflammation in mite-colonized type 2 immunodeficient mice were bolstered by evidence for increased inflammation but diminished IL-4 and IL13-producing innate immune cells among patients with rhinophymatous rosacea associated with Demodex. We cannot exclude contributions by adaptive or unconventional T cells that produce IL-13, particularly with the presence of such cells in human rhinophyma samples under non-SPF conditions with longer life experience as compared to mice. It is possible that the increased prevalence of atopy in the aging population, and perhaps even the early age of onset of type 2 inflammation in atopic skin disease, might represent physiologic efforts to promote tissue integrity and homeostasis during stress induced by senescence and growth, respectively. Efforts to understand the role of traditional immune cells in ‘physiological’ inflammation will be important to anticipate responses to prevalent biologic therapeutics, particularly in aging populations (Medzhitov, 2021). The availability of potent biologics inhibiting type 2 immunity that require continuous therapy may benefit from targeted screening of high-risk patients to curtail symptomatic blepharitis and other cutaneous eruptions while potentially optimizing long-term tissue integrity, particularly in the setting of high Demodex burden.

Limitations of the Study

We characterized an inflammatory dermopathy caused by Demodex among mice at UCSF. Although cases have been noted elsewhere (Smith et al., 2016), the impact of local microbial flora and animal care practices remain unstudied. We analyzed a small human cohort to provide corroborative evidence but further studies are needed across more diverse patient groups to assess the generalizable nature of these findings. Humans are colonized with two types of Demodex and the periodicity of anagen and telogen in humans differs from mice. Although Demodex outgrowths have been noted among patients treated with anti-IL4Ra, controlled trials are needed to assess potential causality linking these drugs with cutaneous outcomes related to Demodex. Finally, increasing information regarding Demodex genomes (Smith et al., 2022) may faciliate identification of moieties or metabolites that sustain quiescent colonization that might represent the target of IL-13-induced host molecules that enable the widespread and successful habitation of the HF niche by these creatures.

STAR Methods

Lead Contact

Further information and requests for reagents may be directed to, and will be fulfilled by, the lead contact Richard Locksley (Richard.Locksley@ucsf.edu).

Materials Availability

All reagents generated or used in this study are available on request from the Lead Contact with a completed Materials Transfer Agreement. Information on reagents used in this study is available in the Key resources table.

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Brilliant Violet 421™ anti-mouse CD19 Antibody, 50 μg	BioLegend	Cat#115549; RRID:AB_2563066
Brilliant Violet 421™ anti-mouse TER-119/Erythroid Cells Antibody, 50 μg	BioLegend	Cat#116234; RRID:AB_2562917
Brilliant Violet 421™ anti-mouse Ly-6G/Ly-6C (Gr-1) Antibody, 50 μg	BioLegend	Cat#108445; RRID:AB_2562903
Brilliant Violet 421™ anti-mouse/human CD11b Antibody, 500ul	BioLegend	Cat#101236; RRID:AB_11203704
Brilliant Violet 650™ anti-mouse/human CD11b Antibody	BioLegend	Cat#101259; RRID:AB_2566568
Pacific Blue™ anti-mouse CD11c Antibody, 100 μg	BioLegend	Cat#117322; RRID:AB_755988
Pacific Blue™ anti-mouse CD49b (pan-NK cells) Antibody, 100 μg	BioLegend	Cat#108918; RRID:AB_2249376
Brilliant Violet 421™ anti-mouse CD335 (NKp46) Antibody, 50 μg	BioLegend	Cat#137612; RRID:AB_10915472
Brilliant Violet 421™ anti-mouse NK-1.1 Antibody, 50 μg	BioLegend	Cat#108741; RRID:AB_2562561
BrilliantViolet 421™ anti-mouse TCR γ/δ Antibody, 50 μg	BioLegend	Cat#118120; RRID:AB_2562566)
Pacific Blue™ anti-mouse FcεRIα Antibody, 100 μg	BioLegend	Cat#134314; RRID:AB_10613298
PE/Cy7 anti-mouse CD3 Antibody, 100 μg	BioLegend	Cat#100220; RRID:AB_1732057
Pacific Blue™ anti-mouse F4/80 Antibody, 100 μg	BioLegend	Cat#123124; RRID:AB_893475
Brilliant Violet 711™ anti-mouse CD45 Antibody, 50 μg	BioLegend	Cat#103147; RRID:AB_2564383
BUV395 Rat Anti-Mouse CD45, 50 μg	BD Biosciences	Cat#564279; RRID:AB_2651134
APC/Cyanine7 anti-mouse CD45 Antibody, 100 μg	BioLegend	Cat#103116; RRID:AB_312981
Brilliant Violet 785™ anti-mouse CD90.2 Antibody, 50 μg	BioLegend	Cat#105331; RRID:AB_2562900
Brilliant Violet 605™ anti-mouse CD90.2 Antibody	BioLegend	Cat#105343; RRID:AB_2632889
Brilliant Violet 711™ anti-mouse CD4 Antibody, 500 μl	BioLegend	Cat#100550; RRID:AB_2562099
Pacific Blue™ anti-mouse CD8a Antibody, 100 μg	BioLegend	Cat#100725; RRID:AB_493425
Brilliant Violet 785™ anti-mouse CD8a Antibody	BioLegend	Cat#100750; RRID:AB_2562610
Pacific Blue™ anti-mouse CD5 Antibody	BioLegend	Cat#100642; RRID:AB_2813916
Brilliant Violet 605™ anti-mouse Ly-6A/E (Sca-1) Antibody, 125 μL	BioLegend	Cat#108133; RRID:AB_2562275
Brilliant Violet 711™ anti-mouse I-A/I-E Antibody	BioLegend	Cat#107643; RRID:AB_2565976
PerCP/Cyanine5.5 anti-mouse CD326 (Ep-CAM) Antibody	BioLegend	Cat#118220; RRID:AB_2246499
FITC anti-human/mouse CD49f Antibody	BioLegend	Cat#313606; RRID:AB_345300
PE/Cyanine7 anti-human/mouse CD49f Antibody	BioLegend	Cat#313622; RRID:AB_2561705
Alexa Fluor® 647 Rat anti-Mouse CD34	BD Biosciences	Cat#560230; RRID:AB_1645200
CD127 (IL-7R) Monoclonal Antibody (A7R34), APC-eFluor 780	Invitrogen	Cat#47-1271-82; RRID:AB_1724012
Gata-3 Monoclonal Antibody (TWAJ), eFluor 660, eBioscience™, 100 tests	Invitrogen	Cat#50-9966-42; RRID:AB_10596663
APC anti-human CD4 Antibody	BioLegend	Cat#300514; RRID:AB_314082
FOXP3 Monoclonal Antibody (FJK-16s), eBioscience™	Invitrogen	Cat#11-5773-82; RRID:AB_465243
BV605 Rat Anti-Mouse IL-33R (ST2), 50 μg	BD Biosciences	Cat#745257; RRID:AB_2742841
CD218a (IL-18Ra) Monoclonal Antibody (P3TUNYA), PerCP-eFluor 710, eBioscience™, 100 μg	Invitrogen	Cat#46-5183-82; RRID:AB_2573764
Living Colors® DsRedPolyclonal Antibody (100μl)	TakaraBio	Cat#632496; RRID:AB_10013483
Ki-67 Monoclonal Antibody (SolA15), FITC, eBioscience™	Invitrogen	Cat#11-5698-82; RRID:AB_11151330
FITC Mouse Anti-E-Cadherin	BD Biosciences	Cat#612131; RRID:AB_2076677
Goat anti-Rabbit IgG (H+L) Cross-Adsorbed Secondary Antibody, Alexa Fluor 555	Invitrogen	Cat#A-21428; RRID:AB_2535849
Anti-human CD3e	Cell Signaling Technology	Cat#85061; RRID:AB_2721019
Purified anti-mouse TER-119/Erythroid Cells (Maxpar® Ready) Antibody	BioLegend	Cat#116241; RRID:AB_2563789
Purified anti-mouse CD45 (Maxpar® Ready) Antibody	BioLegend	Cat#103141; RRID:AB_2562800
BD Transduction Laboratories™ Purified Mouse Anti-Stat3 (pS727)	BD Biosciences	Cat#612542; RRID:AB_399839
Purified anti-mouse CD326 (Ep-CAM) (Maxpar® Ready) Antibody	BioLegend	Cat#118223; RRID:AB_2563743
Purified anti-mouse/human CD11b (Maxpar® Ready) Antibody	BioLegend	Cat#101249; RRID:AB_2562797
Anti-Histone H3 (phospho S28) antibody [HTA28]	Abcam	Cat#ab10543; RRID:AB_2295065
BD Pharmingen™ Purified Hamster Anti-Mouse CD11c	BD Biosciences	Cat#553799; RRID:AB_395058
Phospho-MAPKAPK-2 (Thr334) (27B7)	Cell Signaling Technology	Cat#3007; RRID:AB_490936
Phospho-CREB (Ser133) (87G3) Rabbit mAb	Cell Signaling Technology	Cat#9198; RRID:AB_2561044
BD Transduction Laboratories™ Purified Mouse Anti-Stat1 (pY701)	BD Biosciences	Cat#612232; RRID:AB_399555
BD Transduction Laboratories™ Purified Mouse Anti-Human Stat5 (pY694)	BD Biosciences	Cat#611964; RRID:AB_399385
Phospho-S6 Ribosomal Protein (Ser235/236) (2F9) Rabbit mAb	Cell Signaling Technology	Cat#4856; RRID:AB_2181037
Purified anti-mouse Ly-6C (Maxpar® Ready) Antibody	BioLegend	Cat#128039; RRID:AB_2563783
Phospho-p44/42 MAPK (Erk1/2) (Thr202/Tyr204) (D13.14.4E) XP® Rabbit mAb	Cell Signaling Technology	Cat#4370; RRID:AB_2315112
Anti-Human CyclinB1 (GNS-1)-153Eu	Fluidigm	Cat#3153009A
BD Transduction Laboratories™ Purified Mouse Anti-p38 MAPK (pT180/pY182)	BD Biosciences	Cat#612288; RRID:AB_399605
Purified anti-mouse CD8a (Maxpar® Ready) Antibody	BioLegend	Cat#100755; RRID:AB_2562796
Purified anti-mouse CD4 (Maxpar® Ready) Antibody	BioLegend	Cat#100561; RRID:AB_2562762
BD Pharmingen™ Purified Rat Anti-Mouse CD3 Molecular Complex	BD Biosciences	Cat#555273; RRID:AB_395697
Purified anti-mouse Ly-6G (Maxpar® Ready) Antibody	BioLegend	Cat#127637; RRID:AB_2563784
Phospho-4E-BP1 (Thr37/46) (236B4) Rabbit mAb	Cell Signaling Technology	Cat#2855; RRID:AB_560835
BD Transduction Laboratories™ Purified Mouse Anti-Human ZAP-70 (pY319)/Syk (pY352)	BD Biosciences	Cat#612574; RRID:AB_399863
Phospho-TBK1/NAK (Ser172) (D52C2) XP® Rabbit mAb	Cell Signaling Technology	Cat#5483; RRID:AB_10693472
Purified anti-mouse TCR γ/δ Antibody	BioLegend	Cat#118101; RRID:AB_313826
IκBα (L35A5) Mouse mAb (Amino-terminal Antigen)	Cell Signaling Technology	Cat#4814; RRID:AB_390781
alpha-Fetoprotein/AFP Antibody (SPM334)	Novus	Cat#NBP2-32915
BD Transduction Laboratories™ Purified Mouse Anti-Stat 6 (pY641)	BD Biosciences	Cat#611566; RRID:AB_399012
Anti-pRb [S807/S811] (J112-906)-166Er	Fluidigm	Cat#3166011A
FOXP3 Monoclonal Antibody (NRRF-30)	eBioscience	Cat#14-4771-80; RRID:AB_529583
Purified anti-mouse NK-1.1 (Maxpar® Ready) Antibody	BioLegend	Cat#108743; RRID:AB_2562803
Ki-67 Monoclonal Antibody (SolA15)	eBioscience	Cat#14-5698-82; RRID:AB_10854564
BD Transduction Laboratories™ Purified Mouse Anti-Stat4	BD Biosciences	Cat#610926; RRID:AB_398241
Purified anti-mouse CD62L (Maxpar® Ready) Antibody	BioLegend	Cat#104443; RRID:AB_2562802
BD Pharmingen™ Purified Rat Anti-Mouse Siglec-F	BD Biosciences	Cat#552125; RRID:AB_394340
Purified anti-mouse CD19 (Maxpar® Ready) Antibody	BioLegend	Cat#115547; RRID:AB_2562806
BD Pharmingen™ Purified Rat Anti-Mouse CD34	BD Biosciences	Cat#553731; RRID:AB_395015
BD Pharmingen™ Purified Rat Anti-Mouse CD44	BD Biosciences	Cat#553131; RRID:AB_394646
Purified anti-mouse I-A/I-E (Maxpar® Ready) Antibody	BioLegend	Cat#107637; RRID:AB_2563771
Bacterial and virus strains
Does not apply
Biological samples
Does not apply
Chemicals, peptides, and recombinant proteins
Recombinant mouse IL-4	R&D systems	Cat#404-ML-050/CF
Recombinant mouse IL-13	R&D systems	Cat#413-ML-050/CF
Recombinant mouse IL-13	BioLegend	Cat#575908
Recombinant mouse IL-7	BioLegend	Cat#577804
Recombinant mouse TSLP	R&D systems	Cat#555-TS-010/CF
Recombinant mouse IL-18	BioLegend	Cat#767004
Recombinant mouse IL-33	BioLegend	Cat#580504
Cell Activation Cocktail (without Brefeldin A)	BioLegend	Cat#423301
InVivoMAb anti-mouse CD16/CD32, Clone 2.4G2	BioXCell	Cat#BE0307
InVivoMAb anti-mouse IL-4, clone 11B11	BioXCell	Cat#BE0045
TrueStain FcX (anti-mouse CD16/32 antibody (clone 93)	BioLegend	Cat#101320
Moxidectin	Sigma-Aldrich	Cat#33746-25MG
Imidacloprid	Sigma-Aldrich	Cat#37894-100MG
Ampicillin	Sigma-Aldrich	Cat#A1593-25G
Kanamycin	Sigma-Aldrich	Cat#K1637-25G
Neomycin	Sigma-Aldrich	Cat#N1876-100G
Vancomycin	Sigma-Aldrich	Cat#V2002-5G
Metronidazole	Sigma-Aldrich	Cat#M1547-25G
5-Ethynyl-2’-deoxyuridine (EdU)	Sigma-Aldrich	Cat#900584
Nair depilatory cream	Nivea	N/A
RPMI-1640 Medium	Sigma-Aldrich	Cat#R8758
Liberase TL	Sigma-Aldrich	Cat#5401020001
DNAse I	Sigma-Aldrich	Cat#10104159001
Fetal Bovine Serum	Sigma-Aldrich	Cat#F0926-500ML
4’,6-Diamidine-2′-phenylindole dihydrochloride (DAPI)	Roche	Cat#10236276001
Live/Dead fixable dye	Thermofisher	Cat#L34968
Trypsin-EDTA	ThermoFisher	Cat#15400-054
HEPES	Sigma-Aldrich	Cat#H3537-100ML
Penicillin/Streptomycin	Fisher	Cat#15140122
5-Iodo-2′-deoxyuridine (IdU)	Sigma-Aldrich	Cat#I7125
Cisplatin	Sigma-Aldrich	Cat#P4394-25MG
Paraformaldehyde	Fisher	Cat#15713
Bovine Serum Albumin (BSA)	Sigma-Aldrich	Cat#A7906-500G
Goat Serum	Sigma-Aldrich	Cat#G9023-10ML
Blocking Reagent	Perkin Elmer	Cat#FP1012
Mounting Media	Vector Laboratories	Cat#H-1000
ProLong Gold Antifade	ThermoFisher	Cat#P36930
1x DPBS	Life Technologies	Cat#14190-250
191/193Ir DNA inercalator	Fluidigm	Cat#201192B
Q5 High-Fidelity DNA Polymerase	NEB	M0494L
Yeast cell lysis buffer	Lucigen	Cat#MPY80200
Lucifer yellow	ThermoFisher	Cat#L453
Optimal Cutting Temperature Compound O.C.T.	VWR	Cat#25608-930
Tris-HCl (1M)	Fisher	Cat#50843267
NaCl (5M)	Fisher	Cat#AM9760G
Proteinase K	Thomas Scientific	Cat#C755H28
Citrate buffer (pH 6.0)	ThermoFisher	Cat#00500
3% hydrogen peroxide	JT Baker	Cat#JT-2186-01
ImmPRESS horseradish peroxidase reagent	Vector Laboratories	Cat#MP-7451
Diaminobenzidine substrate	Vector Laboratories	Cat#SK-4100
Commercial Assays
CountBright Absolute Counting Beads	Life Technologies	Cat#C36950
Click-IT Plus EdU imaging kit	ThermoFisher	Cat#C10637
Click-IT Plus EdU flow cytometry kit	ThermoFisher	Cat#C10633
MaxPAR antibody conjugation kit	Fluidigm	Cat#201300
DNEasy PowerSoil Pro Kit	Qiagen	Cat#47014
Chromium Single Cell 3’ GEM, Library & Gel Bead Kit v3.1	10X Genomics	Cat#PN-1000121
Cytokine Bead Array Flex Sets: IL-4	BD Biosciences	Cat#558298
Cytokine Bead Array Flex Sets: IL-5	BD Biosciences	Cat#558302
Cytokine Bead Array Flex Sets: IL-13	BD Biosciences	Cat#558349
Cytokine Bead Array Flex Sets: IL-17A	BD Biosciences	Cat#560283
LEGEND MAXTMMouse IL-22 ELISA Kit	BioLegend	Cat#436307
RNeasy Micro Plus kit	Qiagen	Cat#74034
Superscript VILO cDNA synthesis kit	ThermoFisher	Cat#11754250
Power SYBR Green PCR master mix	ThermoFisher	Cat#4368702
RNAscope® 2.5 HD Reagent Kit-RED	ACD Bio/Bio-Techne	Cat#322350
RNAscope® Probe- Hs-IL4	ACD Bio/Bio-Techne	Cat#315191
RNAscope® Probe- Hs-IL5	ACD Bio/Bio-Techne	Cat#319391
RNAscope® Probe- Hs-IL13	ACD Bio/Bio-Techne	Cat#586241
RNAscope® Probe- Hs-IL22	ACD Bio/Bio-Techne	Cat#560811
RNAscope® Probe- Hs-IFNG	ACD Bio/Bio-Techne	Cat#310501
Experimental models: Cell lines
Does not apply
Experimental models: Organisms/strains
Mouse: Wild type: C57BL/6J	The Jackson Laboratory	Stock# 000664
Mouse: Red5: B6(C)-Il5tm1.1(icre)Lky/J	The Jackson Laboratory	Stock# 030926
Mouse: Yarg: B6.129S4-Arg1tm1Lky/J	The Jackson Laboratory	Stock# 015857
Mouse: Smart13: B6.129S4(C)-Il13tm2.1Lky/J	The Jackson Laboratory	Stock# 031367
Mouse: B6.129S2(C)-Stat6tm1Gru/J	The Jackson Laboratory	Stock# 005977
Mouse: B6.Il4ra−/−: BALB/c-Il4ratm1Sz/J	The Jackson Laboratory	Stock# 003514
Mouse: B6.Il4/Il13−/−	Liang et al., 2012	N/A
Mouse: B6.Il4−/−: Il4tm1(CD2)Mmrs(KN2/KN2)	Mohrs et al., 2001	N/A
Mouse: B6.129S7-Rag1tm1Mom/J	The Jackson Laboratory	Stock# 002216
Mouse: B6.129S6-Rag2tm1Fwa N12: ko/ko	Taconic Biosciences	Cat# RAGN12-F
Mouse: C57BL/6NTac.Cg-Rag2tm1Fwa Il2rgtm1Wjl: ko/ko;ko/ko (Rag2−/−/Il2rg−/−)	Taconic Biosciences	Cat# 4111-F
Mouse: B6.Crlf2−/−	Carpino et al., 2004	N/A
Mouse: B6.Il1rl1−/−	Hoshino et al., 1999	N/A
Mouse: B6.Il25−/−: B6.Il25tm1Anjm	Fallon et al., 2006	N/A
Mouse: B6.Il18−/−:B6.129P2-Il18tm1Aki/J	The Jackson Laboratory	Stock# 004130
Mouse: B6.Il13ra1−/−	This paper	N/A
Oligonucleotides
qPCR Primer: Il13-forward: 5’-CCTGGCTCTTGCTTGCCTT-3’	PrimerBank	PrimerBank ID:6680403a1
qPCR Primer: Il13-reverse: 5’-GGTCTTGTGTGATGTTGCTCA-3’	PrimerBank	PrimerBank ID:6680403a1
qPCR Primer: Il22-forward 5’-ATGAGTTTTTCCCTTATGGGGAC-3’	PrimerBank	PrimerBank ID:21426819a1
qPCR Primer: Il22-reverse 5’-GCTGGAAGTTGGACACCTCAA-3’	PrimerBank	PrimerBank ID:21426819a1
qPCR Primer: Cdkn2a/p16-forward: 5’-CGCAGGTTCTTGGTCACTGT-3’	PrimerBank	PrimerBank ID:6753390a1
qPCR Primer: Cdkn2a/p16-reverse: 5’-TGTTCACGAAAGCCAGAGCG-3’	PrimerBank	PrimerBank ID:6753390a1
qPCR Primer: Rps17-forward: 5’-CCAAGACCGTGAAGAAGGCTG-3’	PrimerBank	PrimerBank ID:6677801a1
qPCR Primer: Rps17-reverse: 5’-GCTGGGGATAATGGCGATCT-3’	PrimerBank	PrimerBank ID:6677801a1
Primer: CHS-forward: 5’-GAAGCGGCGAGTAATGTTCATC-3’	This paper	N/A
Primer: CHS-reverse: 5’-CCTGACTCCATCTTTTACGATGTC-3’	This paper	N/A
Primer: Keratin 5-forward: 5’-TCTGCCATCACCCCATCTGT-3’	PrimerBank	PrimerBank ID:20911031a1
Primer: Keratin 5-reverse: 5’-CCTCCGCCAGAACTGTAGGA-3’	PrimerBank	PrimerBank ID:20911031a1
Primer: 18S-forward: 5’-TCCAAGGAAGGCAGCAGGCA-3’	This paper	N/A
Primer: 18S-reverse: 5’-CGCGGTAGTTCGTCTTGCGACG-3’	This paper	N/A
16s rRNA gene forward: 5’-AGAGTTTGATCCTGGCTCAG-3’	This paper	N/A
16s rRNA gene reverse: 5’-ATTACCGCGGCTGCTGG-3’	This paper	N/A
sgRNA targeting exon 2 of Il13ra1 gene: ACGCTCAAATTCGTCACAGGTGG	This paper	N/A
sgRNA targeting exon 2 of Il13ra1 gene: TCCTGAGCCACAGCATGTACTGG)	This paper	N/A
Genotyping primer: Il13Ra1-Forward: 5’-AAAGTGAAGCCATCAGTAGTCCCTC-3’	This paper	N/A
Genotyping primer: Il13Ra1-Reverse: 5’-AAATGCAAACCCACCCACCACTAAC-3’	This paper	N/A
Recombinant DNA
eGFP-CBP plasmid	This study	https://www.addgene.org/169211/
Software and algorithms
ImageJ Software	https://imagej.net/software/fiji/
Significance Across Microarrays algorithm	Bair and Tibshirani, 2004; Bruggner et al., 2014
Scaffold maps R package	https://github.com/spitzerlab/statisticalScaffold
viSNE and Spade analyses	www.Cytobank.org
Flow Cytometric Analysis Program (FCAP) Array software	https://www.beckman.com/flow-cytometry/software
SoftMax Pro software	Molecular Devices
Flowjo (v10.6.1)	https://www.flowjo.com
BD FACSDiva Software	BD Biosciences
GraphPad Prism	https://www.graphpad.com/scientific-software/prism/
Quantity One (v4.6.1)	Bio-Rad Laboratories
Image Lab (v6.1.0)	Bio-Rad Laboratories
QIIME2	Bolyen et al., 2019
DADA2	Callahan et al., 2016
Greengenes Reference Database	McDonald et al., 2012
MAFFT	Katoh and Standley, 2013
Principal Coordinates Analysis of weighted and unweighted UniFrac distance	Lozupone and Knight, 2005; Lozupone et al., 2007
scikit-bio	Bokulich et al., 2018
Seurat V3 or V4	https://satijalab.org/seurat/
R package Harmony	https://github.com/immunogenomics/harmony
GOrilla	http://cbl-gorilla.cs.technion.ac.il
Nebulosa Package	https://github.com/powellgenomicslab/Nebulosa
Mass cytometry data normalization algorithm	Finck et al., 2013
AxioVision Software	Zeiss
NIS Elements Software	Nikon
Other

Data and Code Availability

All the data supporting the findings of the article are available within the main text or supplementary materials. The published article includes datasets generated during this study. Original single cell RNA-seq data has been deposited in GEO (GSE197983). Original 16S rRNA sequencing datasets analyzed in this study are available at the NCBI Sequence Read Archive (SUB11832976).

Mice

Wild-type (C57BL/6J; Stock 000664) mice were purchased from Jackson Laboratories. Il5^Red5, Arg1^Yarg, and Il13^Smart reporter alleles on C57BL/6J (B6) backgrounds were bred and maintained as described (Bando et al., 2013; Ricardo-Gonzalez et al., 2018). For some experiments, mice encoding the Arg1^Yarg and/or Il13^Smart reporter alleles that do not impact the endogenous gene expression were used as wild-type controls. Il4ra^−/− (BALB/c-Il4ratm1Sz/J; 003514) and Stat6^−/− (B6.129S2(C)-Stat6tm1Gru/J; 005977) mice were initially purchased from The Jackson Laboratory and backcrossed to C57BL/6J for at least eight generations as described (von Moltke et al., 2016). Additional wild-type C57BL/6J and Il4ra^−/− mice on the C57BL/6J background were obtained from the UCSF colony of M.S. Fassett and K.M. Ansel. Il4^−/−,Il13^−/− mice were generated as previously described and backcrossed C57BL/6J for at least eight generations (Liang et al., 2012). Il4^−/− (KN2/KN2 mice) were generated as previously described (Mohrs et al., 2001) and backcrossed to C57BL/6J. Crlf2^−/− (Carpino et al., 2004), Il25^−/− (Fallon et al., 2006), and Il1rl1^−/− (Hoshino et al., 1999) mice on C57BL/6 J backgrounds were intercrossed to generate Crlf2^−/−,Il25^−/−,Il1rl1^−/− triple-deficient mice expressing Il5^Red5, Arg1^Yarg, and Il13^Smart reporter alleles as previously described (Ricardo-Gonzalez et al., 2018; Van Dyken et al., 2016). Triple-deficient mice were crossed to Il18^−/− mice (Stock 004130) purchased from The Jackson Laboratory to generate Crlf2^−/−,Il25^−/−,Il1rl1^−/−,Il18^−/− quadruple-ablated (Quad-ablated) in combination with Il5^Red5, Arg1^Yarg, and Il13^Smart reporter alleles. Rag1^−/− mice (Stock 002216) were purchased from Jackson laboratory, and crossed to Arg1^Yarg and Il5^Red5 reported alleles. Rag2^−/− (Stock RAGN12-F) and Rag2^−/−,Il2rg^−/− (Stock 4111-F) mice were obtained from Taconic Biosciences. For all experiments, sex-matched mice aged 7-14 weeks were used unless otherwise indicated. Mice were maintained under specific pathogen-free conditions. For in vivo cytokine treatments, mice were injected subcutaneously with IL-4 complex (IL-4c) comprised of rmIL-4 (2 μg, R&D systems) combined with anti-IL-4 antibody (10 μg, Clone Mab 11B11, BioXcell), or rmIL-13 (2 μg, R&D systems) in calcium- and magnesium-free phosphate buffered saline (PBS). For co-housing experiments, mice of the appropriate genotypes were co-housed for 6-12 weeks in the ratios noted in the figure legends. For treatment with topical anti-parasitic agents, a stock solution in 70% ethanol of moxidectin (Sigma 33746) and imidacloprid (Sigma 37894) was prepared and diluted to a working solution of (0.83 mg/ml moxidectin and 3.33 mg/ml imidacloprid) and applied topically to intrascapular skin at 3.3 mg/kg moxidectin and 13.3 mg/kg imidacloprid. Control mice were treated with 70% ethanol. For broad-spectrum antibiotics treatment, mice received drinking water supplemented with ampicillin, kanamycin, and neomycin (all 1 mg/ml), vancomycin (0.5 mg/ml), and metronidazole (2.5 mg/ml) in 1% sucrose. For EdU experiments, mice received a peritoneal injection of 1 μg of EdU (Sigma 900584) 14-18 hours prior to harvesting. All animal procedures were approved by the UCSF Institutional Animal Care and Use Committee.

Generation of IL-13Ra1 deficient mice

Two sgRNAs targeting exon 2 of Il13ra1 gene (ACGCTCAAATTCGTCACAGGTGG and TCCTGAGCCACAGCATGTACTGG) and recombinant Cas9 (PNA Bio) were simultaneously injected as RNPs into the zygotes obtained from C57BL/6 J mice carrying Arg1^Yarg/Yarg, Il5^Red5, and Il13^Smart reporter alleles. Founders were screened by PCR using primers: Il13Ra1-Forward: 5’-AAAGTGAAGCCATCAGTAGTCCCTC-3’ and Il13Ra1-Reverse: 5’-AAATGCAAACCCACCCACCACTAAC-3’ to identify a ~144 bp deletion resulted from the end-joining event of two DSBs at the sgRNA-targeted sites. Since the Il13ra1 gene is an X-linked gene, only positive male founders were chosen to ensure the allelic identity. A male founder was first crossed with C57BL/6 J mice (with Arg1^Yarg/Yarg, Il5^Red5, and Il13^Smart reporter alleles) followed by intercrossing to obtain homozygous female I13ra1^−/−.

Depilation Experiment

Mice were anesthetized with isoflurane inhalation. Under anesthesia, the dorsal surface (back) hair of the animals was shaved down to the level of skin with an electric razor. A thin coat (1 fingertip unit) of Nair depilatory cream (Nivea) was applied to the shaved region for a period of 30 s before wiping clean. For monitoring of clinical hair regrowth, standardized pictures were taken with a ruler on the day of depilation (day 0) and then at days 4, 7, 9, 11, 14. Anagen induction was quantified using intensity analysis on ImageJ software (NIH, USA) at each time point or as a percent of pigmented dorsal skin relative to baseline (day 0). For repeat depilation experiments, mice were initially shaved and depilated when they were in second telogen (2 months old). Cycles were repeated once a month once they have completed the anagen and recovery phase. Tissues were analyzed for flow cytometry, stem cell proliferation, and qPCR as described below.

Flow Cytometry and Cell Sorting

Whole skin single cell suspensions were prepared as described previously (Ricardo-Gonzalez et al., 2018). Briefly, back tissue was minced in RPMI-1640 with 5% FBS, then transferred to C tubes (Miltenyi Biotec), containing 5 ml of RPMI-1640 (Sigma R8758) supplemented with Liberase TL (0.25 mg/ml, Sigma 5401020001), and DNAse I (0.1 mg/ml; Sigma 10104159001). Samples were shaken at 250 rpm for 2 hours at 37 °C, then dispersed using an automated tissue dissociator (GentleMACS; Miltenyi Biotec) running program C. Single-cell suspensions were passed through a 70 μm filter and washed twice with RPMI containing 5% FBS. The following antibodies, all from BioLegend (unless otherwise specified, see Key resources table and Supplementary Table S2) were used at 1:300 dilution unless noted: anti-CD3 (17A2, diluted 1:200), anti-CD4 (RM4-5, diluted 1:100), anti-CD5 (53-7.3), anti-CD8α (53-6.7, diluted 1:100), anti-CD11b (M1/70), anti-CD11c (N418), anti-CD19 (6D5), anti-CD25 (PC61, diluted 1:100), anti-CD45 (30F-11, BD Biosciences), anti-CD49b (DX5; eBiosciences), anti-CD127 (A7R34, diluted 1:100), anti-CD218 (P3TUNYA,eBiosciences), anti-F4/80 (BM8), anti-Gr-1 (RB6-8C5), anti-NK1.1 (PK136), anti-NKp46 (29A1.4), anti-Thy1.2 (30-H12; diluted 1:1000); anti-human CD4 (RPA-T4, diluted 1:20; eBiosciences), anti-GATA3 (TWAJ, diluted 1:25), anti-FoxP3 (FJK-16s, Invitrogen, diluted 1:100), anti-T1/ST2 (U29-93; BD Biosciences, diluted 1:200). Live/dead cell exclusion was performed with DAPI (4′,6-diamidine-2′-phenylindole dihydrochloride; Roche) or LIVE/dead fixable dye (Thermofisher). Cell counts were performed using flow cytometry counting beads (CountBright Absolute; Life Technologies) per manufacturer instructions. To assay cell proliferation, the Click-IT EdU flow cytometry or imaging kits (ThermoFisher) were used according to the kit instructions. Sample data were acquired with a 5-laser LSRFortessa X-20 flow cytometer and BD FACSDiva software (BD Biosciences) and analyzed using FlowJo software (Tree Star v10.6.1).

ILC2s were sorted from reporter mice as live (DAPI⁻), Lin(CD3, CD4, CD8, TCRγδ CD11b, CD11c, CD19, NK1.1, NKp46, Gr-1, F4/80, Ter119, DX5)⁻CD45⁺Thy1⁺Red5⁺ using a MoFlo XDP (Beckman Coulter). Epithelial cell preparations for flow cytometry or CyTOF analyses were prepared as previously described (Ali et al., 2017; Nagao et al., 2012). Briefly, mice were shaved and dorsal skin was harvested and trimmed of fat. The tissue was floated in 4 ml of 0.25% Trypsin-EDTA (ThermoFisher) and incubated for 90 min at 37 °C. The epidermis was scrapped from the dermis and mechanically separated by pipetting into complete media (RPMI-1640 supplemented with 10% FBS, HEPES, Penicillin/Streptomycin). The cell suspension was passed through a 100 μm strainer and washed with 10 ml of complete media. The cell suspension was then pelleted, filtered through a 40 μm filter and aliquoted for cell staining. The following antibodies were used for staining epithelial fractions: anti-Sca1 (D7, 1:300); anti-MHCII (M5/114.15.2, 1:400); anti-CD326 (G8.8, 1:300); anti-CD45 (30F-11, 1:400); anti-CD34 (RAM34, 1:200); anti-CD49f (GoH3, 1:300).

Adoptive transfer of ILC2s

For adoptive transfer of ILC2s, Demodex-infected Il4ra^−/− mice were co-housed with Rag1^−/− Arg1^Yarg Il5^Red5 reporter mice for four weeks to establish Demodex infection. ILC2s (LiveCD45⁺Lin⁻Thy1⁺Red5⁺) from the Demodex-infected Rag1^−/− mice were sorted from skin, lung and fat. Recipient Rag2^−/−, Il2rg^−/− mice received 2.5 x 10⁴ pooled ILC2 cells intravenously in 200 μl of PBS at two weeks before, and 7 x 10⁴ pooled ILC2s at the start of co-housing. Control mice received PBS. Mice were co-housed with Demodex-infected Il4ra^−/− for 6 weeks before analysis.

Mass Cytometry

Sample preparation

Mass cytometry was performed as described (Kleppe et al., 2017). Briefly, single-cell suspension from the epidermal fraction of the various mice strains were prepared. To test the responsiveness to type 2 immune signaling, mice were treated with s.c. IL-4 complex for 30-60 min prior to harvest. IdU (200 μl of 5mM stock solution, Sigma 17125) was injected 20-30 min prior to harvesting. Cells were incubated with 25 μM Cisplatin for 1 min and fixed with paraformaldehyde at 1.6% for 10 min at room temperature (RT). Cell pellets were stored at −80 °C until analysis. Prior to antibody staining, cells were thawed and mass tag cellular barcoding of prepared samples was performed by incubating cells with distinct combinations of isotopically-purified palladium ions chelated by isothiocyanobenzyl-EDTA in 0.02% saponin in PBS as previously described (Behbehani et al., 2014; Zunder et al., 2015). After two washes with staining media (PBS + 0.5% BSA + 0.02% NaN₃), sets of 20 barcoded samples were pooled together and washed once more with staining media. Antibody staining was performed using a panel of antibodies and metals that were purchased directly conjugated (Fluidigm) or conjugated using 100 μg of antibody lots combined with the MaxPAR antibody conjugation kit (Fluidigm) according to the manufacturer’s instructions and detailed in Supplementary Table 3. Surface marker antibodies were added to a 500 μL final reaction volumes and stained for 30 min at RT on a shaker. Following staining, cells were washed twice with PBS with 0.5% BSA and 0.02% NaN3. Then cells were permeabilized with 4 °C methanol for 10 min at 4 °C followed by two washes in PBS with 0.5% BSA and 0.02% NaN₃ to remove remaining methanol. The intracellular antibodies were added in 500 μL for 30 min at RT on a shaker. Cells were washed twice in PBS with 0.5% BSA and 0.02% NaN₃ and stained with 1 mL of 1:4000 191/193Ir DNA intercalator (Fluidigm) diluted in PBS with 1.6% PFA overnight. Cells were washed once in PBS with 0.5% BSA and 0.02% NaN₃ and twice with double-deionized (dd)H₂0. We analyzed 1 x 10⁶ cells per animal, per tissue, per condition consistent with generally accepted practices in the field.

Bead Standard Data Normalization

Just before analysis, the stained and intercalated cell pellet was resuspended in ddH₂O containing the bead standard at a concentration ranging between 1 and 2 x 10⁴ beads/ml as described (Finck et al., 2013). The bead standards were prepared immediately before analysis, and the mixture of beads and cells was filtered through filter cap FACS tubes (BD Biosciences) before analysis. Mass cytometry files were normalized together using the mass cytometry data normalization algorithm (Finck et al., 2013), which uses the intensity values of a sliding window of these bead standards to correct for instrument fluctuations over time and between samples.

Scaffold Map Generation

Total live leukocytes (excluding erythrocytes) were used for all analyses. Cells from each tissue for all animals were clustered together (rather than performing CLARA clustering on each file individually as originally implemented in (Spitzer et al., 2015)) Cells were deconvolved into their respective samples. Cluster frequencies or the Boolean expression of Ki67 for each cluster were passed into the Significance Across Microarrays algorithm (Bair and Tibshirani, 2004; Bruggner et al., 2014), and results were tabulated into the Scaffold map files for visualization through the graphical user interface. Cluster frequencies were calculated as a percent of total live nucleated cells (excluding erythrocytes). Scaffold maps were then generated as previously reported (Spitzer et al., 2017; Spitzer et al., 2015). All analyses were performed using the open source Scaffold maps R package available at http://www.github.com/spitzerlab/statisticalScaffold.

Spade and viSNE Generation

viSNE and Spade analyses were performed with Cytobank (www.cytobank.org).

Cytokine quantification

Cytokine in mouse serum or ILC2 culture supernatants were measured according to the manufacturer’s protocol using the following Cytokine Bead Array Flex Sets (BD): IL-4 (#558298), IL-5 (#558302), IL-13 (#558349), and IL-17A (#560283). The data were analyzed using Flow Cytometric Analysis Program (FCAP) Array software (BD). Serum IL-22 was measured using LEGEND MAX™ Mouse IL-22 ELISA Kit (BioLegend, #436307) and analyzed using a SpectraMax M2 microplate reader using SoftMax Pro software (Molecular Devices).

RNA preparation and qRT-PCR

ILC2s were sorted into RLT Plus lysis buffer (Qiagen) and stored at − 80 °C until processing using RNeasy Micro Plus kit (Qiagen) per manufacturer’s protocol. For quantitative reverse transcriptase PCR analyses, RNA was reverse transcribed using Superscript VILO cDNA synthesis kit (ThermoFisher) per the manufacturer’s instructions. The cDNA was used to test the gene expression of selected genes using Power SYBR Green PCR master mix (ThermoFisher) in a StepOnePlus cycler (Applied Biosystems). The following primers sets from PrimerBank were used: Il13: Il13-Forward 5’-CCTGGCTCTTGCTTGCCTT-3’ and Il13-Reverse, 5’-GGTCTTGTGTGATGTTGCTCA-3’; Il22: Il22-Forward 5’-ATGAGTTTTTCCCTTATGGGGAC-3’ and Il22-Reverse 5’-GCTGGAAGTTGGACACCTCAA-3’; Cdkn2a/p16: p16-Forward 5’-CGCAGGTTCTTGGTCACTGT-3’ and p16-Reverse 5’-TGTTCACGAAAGCCAGAGCG-3’; Rps17: Rps17-Forward 5’-CCAAGACCGTGAAGAAGGCTG-3’, Rps17-Reverse 5’-GCTGGGGATAATGGCGATCT-3’. Gene expression was normalized to Rps17 using the ΔCt method.

Chitin Synthase and Demodex 18S rRNA gene PCR

For semi-quantitative analyses of Demodex burden in skin, three random 2 mm punch biopsies of back caudal skin were digested in digest buffer (10 mM Tris pH 8.0, 100 mM NaCl, 10 mM EDTA pH 8.0, 0.5% SDS and 200 μg/ml Proteinase K (Thomas Scientific)) and the genomic DNA was resuspended in 30 μl of TE. The following primers and PCR conditions were used to detect the highly conserved Demodex spp. chitin synthase (CHS) gene (GenBank No. AB080667): CHS-Forward 5’-GAAGCGGCGAGTAATGTTCATC-3’; CHS-Reverse 5’-CCTGACTCCATCTTTTACGATGTC-3’; Cycling conditions 95 °C 3’, (95 °C 30”, 52 °C 30”, 72 °C 30”) x 30 cycles, 72 °C 7’. Amplification of the murine Keratin 5 gene was used as a loading control using the following primers and PCR conditions: Keratin5-Forward: 5’-TCTGCCATCACCCCATCTGT-3’; Keratin5-Reverse: 5’-CCTCCGCCAGAACTGTAGGA-3’, 95 °C 3’, (95 °C 30”, 60 °C 30”, 72 °C 30”) x 30 cycles, 72 °C 7’. PCR products were separated by a 2.5% agarose gel with ethidium bromide and visualized by a GelDoc (Bio-Rad Laboratories). Semi-quantitative analyses of band intensity of CHS and Krt5 was perfomed using Quantity One (version 4.6.1) or Image Lab (version 6.1.0) software (Bio-Rad Laboratories) and reported as the ratio of PCR product of CHS over Krt5. Demodex musculi was confirmed by first amplifying the 18S rRNA gene from DNA samples purified from the hair scrapes alone of affected mice using the universal 18S primers and PCR conditions: 18S-Forward 5’-TCCAAGGAAGGCAGCAGGCA-3’; 18S-Reverse 5’-CGCGGTAGTTCGTCTTGCGACG-3’, 95 °C 3’, (95 °C 30”, 58 °C 30”, 72 °C 30”) x 30 cycles, 72 °C 7’. PCR amplicons (532 bp) were purified and subsequently cloned into pCR4-TOPO plasmid by TOPO-TA cloning (Thermo Fisher) for sequencing. Putative Demodex spp. sequences from individual mouse (>3) were identical to the Demodex musculi 18S rRNA gene (GenBank No. JF834894).

Preparation of eGFP-CBP for immunofluorescence

Enhanced GFP was cloned into pGV358 (gift from Kushol Gupta, Univ. of Pennsylvania), a pETDuet based vector that has an Mxe intein in frame at the C-terminus followed by the non-cleavable hexahistidine tagged chitin-binding domain from Bacillus circulans WL-12 Chitinase A1 (eGFP-CBP). Construct expressing eGFP-CBP (gift from Sy Redding, Univ. of California, San Francisco) was transformed into BL21(DE3) cells and expressed overnight in 50 mL LB media at 37 °C. 10 mL of overnight culture was diluted in 1L LB media and expressed at 37 °C until OD₆₀₀ = 1-1.2. Culture was induced with 1 mL 1M IPTG, and expression continued overnight at 19 °C. We added protease inhibitor (Roche #11836170001) at the temperature change to minimize proteolysis of the expressed protein. Cells were centrifuged at 4,000 x g at 4 °C for 20 mins. The pelleted cells were resuspended in 50 mL of lysis buffer (100 mM Tris pH 8.5, 150 mM NaCl, 30 mg Lysozyme, protease inhibitor, 1 μL universal nuclease (ThermoFisher Scientific #88700)) per 1 L of culture, then lysed using sonication for 6 mins at 50% power. Lysed cells were centrifuged at 10,000 x g at 4 °C for 15-30 mins. The lysate supernatant was filtered with a 0.22 μm filter. The filtered lysate was bound to a 5 mL HisTrap FF column (Cytiva #17525501), washed with 100 mM Tris pH 8.5, 150 mM NaCl, 50 mM imidazole pH 7.0, and eluted with a gradient into 100 mM Tris pH 8.5, 150 mM NaCl, 500 mM imidazole pH 7.0. Fractions were selected based on presence of molecular weight band at 56.8 kDa, corresponding with eGFP-CBP, as assessed via SDS-PAGE. Fractions were pooled and dialysed overnight into 100 mM Tris pH 8.5, 150 mM NaCl, 5% glycerol w/v. The protein solution was concentrated and separated via size-exclusion chromatography on a HiLoad 16/600 Superdex 75 pg (Cytiva #28989333). Fractions were selected based on purity as assessed via SDS-PAGE.

Fixed tissue preparation and immunofluorescence staining

For immunohistochemistry, 3-4 mm wide cephalad-caudal strips of back skin were cut and fixed in 2% paraformaldehyde for 2-6 hrs at room temperature, followed by 2 washes in D-PBS, and incubation in 30% (w/v) sucrose in D-PBS at 4 °C overnight. Portions of the tissue sections were submitted for routine paraffin embedding and hematoxylin and eosin staining using UCSF Histology and Biomarker Core. For frozen sections, tissues were embedded in Optimal Cutting Temperature Compound (Tissue-Tek) and stored at −80 °C until further processing. Tissue sections (8-12 μm) were cut on a Cryostat (Leica). Immunohistochemistry was performed in Tris/NaCl blocking buffer (0.1 M Tris-HCl, 0.15 M NaCl, 5 μg/ml blocking reagent (Perkin Elmer), pH 7.5) as follows at RT: 1 hour in 5% goat serum, 1-2 hours in primary antibody, 60 min secondary antibody (Goat anti-Rabbit AF555, 1:1000), 5 min of a 1:1000 dilution of DAPI (Stock 5 mg/ml in PBS), Roche) and mounted in Mounting Media (Vector Labs). The following primary antibodies were used in this study: Living colors DsRed Polyclonal antibody (Rabbit, 1:300, Takara Bio #632496); FITC Anti-mouse Ki67 (1:100, Invitrogen #11-5698-82), and FITC Mouse Anti-E-Cadherin (1:400, BD Biosciences #612131). For detection of chitin in skin tissue sections, 4-5 μm thick paraffin-embedded sections were baked at 60 °C for 30 min, treated with xylenes, and washed with 100% ethanol. The sections were incubated with an eGFP-CBP fusion protein (5 μg/ml) and DAPI (5 μg/ml) in PBS at RT for 60 min, washed 3 times with PBS and mounted using ProLong Gold antifade (Thermo Fisher, P36930). For quantification of HF infested by Demodex, two 1.5 to 2 cm strips of back skin from H&E stained sections were evaluated for infestation of hair follicles with a minimum of 70 HF inspected per mouse. Data presented as percentage of HF with at least one mite within the follicle or associated sebaceous gland over total HF visualized. Images were acquired using a Zeiss immunofluorescence AxioImager M2 upright microscope with a Plan Apochromat 20X/0.8NA air objective running AxioVision software (Zeiss) or an inverted Nikon A1R scanning confocal microscope with a Plan Apo λ 20X/0.75NA air objective running NIS Elements software (Nikon).

RNAscope of human tissue samples

Eight cases of normal nose skin were identified by searching the Yale Dermatopathology tissue archive (see Supplementary Table S4). These cases consisted of the tips of elliptical tissue excisions removing non-melanoma skin cancer. Areas that were well away from scar were selected for analysis. Eight cases of rhinophyma excision were similarly identified by searching the Yale Dermatopathology archive. RNA in situ hybridization was performed using a chromogenic RNAscope (Bio-techne) assay according to the manufacturer’s instructions and as previously described (Wang et al., 2021). The probes were purchased directly from the manufacturer: IL4 (315191), IL5 (319391), IL13 (586241), IL22 (560811), and IFNG (310501). CD3ε IHC was performed using an anti-CD3ε antibody (D7A6E) acquired from Cell Signaling Technology (Danvers, MA). Double staining for CD3ε (IHC) and IL13 mRNA ISH were performed as previously described (Wang et al., 2021). Briefly, after detection of IL13 with the RNAscope kit, antigen retrieval was performed with citrate buffer (pH 6.0) (00500, Thermo Fisher Scientific, Waltham, MA). Peroxidase activity was quenched using 3% hydrogen peroxide (JT-2186-01, JT Baker, Phillipsburg, NJ). CD3ε was then hybridized and detected using the Imm-PRESS horseradish peroxidase reagent and diaminobenzidine substrate (Vector Laboratories, Burlingame, CA).

The density of inflammation around the isthmus and infundibular portion of each follicular unit was graded using H&E stained slides and the following scale:

Histologic description	Semi-quantitative score	Log2 transformation	Score used for statistical analysis
Negligible inflammation	0	-	0
Mild inflammation	1	12	1
Modest inflammation	2	22	4
Marked inflammation	3	32	8
Dense inflammation	4	42	16
Very dense inflammation (pseudolymphomatous)	5	52	32

The amount of positive staining for each RNAscope probe was quantified on a cell-by-cell basis using the manufacturer’s recommendations, with slight modifications. Only inflammatory cells surrounding or within each follicle were scored (there was negligible staining outside these areas). The scoring rubric is summarized below:

ACD score	Scoring criteria
0	No staining*
1	1-2 dots per cell*
2	2-5 dots per cell and no dot clusters
3	5-10 dots per cell and/or <10% of dots are in clusters
4	>10 dots/cell and/or 10% of dots are in clusters

^*This intensity of staining is considered background staining and was not included in the quantification.

The slides were scored in a blinded fashion by a board-certified dermatopathologist (W.D.). For CD3ε/IL13 doublestains the number of CD3ε positive and CD3ε negative cells was manually quantitated. The values were entered into GraphPad Prism. These studies were approved by the Yale IRB (ID 1501015235).

Single cell RNA-seq

C57BL/6J female mice were obtained from Jackson Laboratory at 3 weeks of age and divided into the following groups 1 weeks later: WT control (n = 3), and WT co-housed with Demodex-infested Il4ra^−/− mice for 1 month (n = 3). In a second experiment, WT (uninfected, n = 3) or Demodex infected Il4ra^−/− (n = 3) and Il4^−/−, Il13^−/− (n = 3) were used for analyses. Single cell suspensions from full thickness skin were stained with antibody to CD45 (to isolate the CD45⁺ fraction) and integrin alpha 6 to generate a 50:50 representation of CD45⁻alpha6⁺ and CD45⁻ alpha6⁻ cells in the CD45⁻ fraction. Cell suspensions were stained with DAPI and live cells (DAPI⁻) were isolated by FACS using a MoFlo XDP (Beckman Coulter). The concentration of sorted cells was determined using CountBright beads on an LSRFortessa. Single-cell libraries from 25,000 cells per sample were prepared with the Chromium Single Cell 3′ GEM, Library & Gel Bead Kit v3.1 (10 x Genomics PN-1000121) following the manufacturer’s protocol. The libraries were sequenced on the NovaSeq 6000 at the UCSF Institute for Human Genetics with the following number of cycles for each read: Read 1, 28 cycles; i7 Index, 8 cycles; i5 Index, 0 cycles; Read 2, 91 cycles.

Data were analyzed in Seurat V3 or V4 in R using SCTransform aggregation, 0.6 resolution, with the following quality parameters: min.cells =5, nFeature_RNA ≥ 500; nFeature_RNA< 5000; nCount_RNA < 25000; percent mitochondrial reads < 15; percent hemoglobin reads 0.001. Percentage mitochondrial and ribosomal content were regressed out of clustering and aggregation. After initial clustering, cell identities in the CD45⁻ fraction (epithelia, fibroblasts, other stromal cells) were assigned using public data sets (Joost et al., 2020). CD45⁺ immune clusters were identified by canonical markers and subclustered. Doublets were removed and the populations were clustered again before differential expression (DEG) analysis for cluster-defining genes and differential expression between treatment conditions. Data generated from different experiments (WT control vs. WT Demodex+; WT uninfected vs. Il4ra^−/− Demodex+ vs. Il4,Il13^−/− Demodex+) were integrated using the R package Harmony prior to clustering cell types. Batch correction was performed by creating a binary batch variable between the two datasets.

For DEG analysis between clusters, genes were detected if they are expressed in at least 10% of cells in a cluster with a log-fold change of at least 0.25. For DEG analysis between comparison groups within each cluster, genes were detected if expressed in at least 10% of cells in the group with a log-fold change of at least 0.25. For DEG analysis comparisons for all cells in aggregate, genes were detected if expressed in at least 10% of cells and with a log fold-change of at least 0.10. Additionally, for within-cluster DEG analysis, analysis was only done for clusters with >30 cells in each comparison group. Gene ontology (GO) term pathway analyses was performed using GOrilla (http://cbl-gorilla.cs.technion.ac.il) by inputting ‘pseudo’-bulk DEGs, from all CD45 negative cells, that were upregulated with a log-fold change of at least 0.2 in Demodex-infected WT, Il4ra^−/− or Il4^−/− Il13^−/− mice vs. the experimental WT control. Additional data visualization was performed using the Nebulosa package in Seurat (https://github.com/powellgenomicslab/Nebulosa).

In Figure S7H–J epidermal stem cells were identified among subclustered epithelial cells by differential expression of Lgr5, Cdh3, Gli1 and/or Mki67. DEGs among all stem/dividing cells between conditions were identified by expression in at least 10% of cells with a log2 fold change of at least 0.10 between condition. GO Term enrichment was performed using EnrichGO in the ClusterProfiler package in R (Yu et al., 2012).

Transepidermal water loss measurements

Transepidermal water loss (TEWL) measurements were performed as described (Mathur et al., 2019). Briefly, mice back were shaved and rested for 18-24 h. Mice were sedated and TEWL was measured on four quadrants of back skin using a Tewameter (TM300, Khazaka Electronic).

Lucifer yellow assay to assess barrier function

Back skin from WT or Il4ra^−/− mice was shaved and depilated. Two days after depilation, skin was harvested, washed twice with PBS and placed on a 7 mm Franz-Cell chamber (PermeGear Inc). The epidermal surface was incubated with Lucifer yellow (L453, ThermoFisher, 1 mg/ml) in PBS for 18 h at 37 °C. Samples were fixed for 1 h in 4 % PFA, washed twice with PBS, embedded in OCT and processed for immunofluorescence imaging.

16S rRNA gene sequence analysis

Sample collection

Mice facial and back skin were swabbed at the end of the experiment using a sterile foam tipped applicator (Puritan) that was moistened with yeast cell lysis buffer (Lucigen) and collected into sterile 2 ml tubes (Eppendorf). All samples were immediately snap frozen in dry ice and stored at −80 °C until DNA extraction.

DNA extraction and V1V3 Library Prep

DNA was extracted from samples using the DNEasy PowerSoil Pro Kit (Qiagen) according to the manufacturer’s protocol. Barcoded PCR primers annealing to the V1-V3 region of the 16S rRNA gene (Fwd, 5’-AGAGTTTGATCCTGGCTCAG-3’; Rev, 5-ATTACCGCGGCTGCTGG-3’) were used for library generation. PCR reactions were carried out in quadruplicate using Q5 High-Fidelity DNA Polymerase (NEB, Ipswich, MA). Each PCR reaction contains 0.5 μM of each primer, 0.34 U Q5 Pol, 1X Buffer, 0.2 mM dNTPs, and 5.0 μl DNA in a total volume of 25 μl. Cycling conditions are as follows: 1 cycle of 98 °C for 1 min; 30 cycles of 98 °C for 10 sec, 56 °C for 20 sec, and 72 °C for 20 sec; 1 cycle of 72 °C for 8 min. After amplification, quadruplicate PCR reactions were pooled and then purified using a 1:1 volume of SPRI beads. DNA in each sample was quantified using PicoGreen and pooled in equal molar amounts. The resulting library was sequenced on the Illumina MiSeq using 2x300 bp chemistry. Extraction blanks and DNA free water were subjected to the same amplification and purification procedure to allow for empirical assessment of environmental and reagent contamination.

Bioinformatics processing

Sequence data were processed using QIIME2 (Bolyen et al., 2019). Read pairs were processed to identify amplicon sequence variants with DADA2 (Callahan et al., 2016). Taxonomic assignments were generated by comparison to the Greengenes reference database (McDonald et al., 2012), using the naive Bayes classifier implemented in scikit-bio (Bokulich et al., 2018). A phylogenetic tree was inferred from the sequence data using MAFFT (Katoh and Standley, 2013).

Statistical analysis for microbiome data

Data files from QIIME were analyzed in the R environment for statistical computing, using the QIIMER library, which we developed (http://cran.r-project.org/web/packages/qiimer). Global differences in bacterial community composition were visualized using Principal Coordinates Analysis of weighted and unweighted UniFrac distance (Lozupone and Knight, 2005; Lozupone et al., 2007). Community-level differences between sample groups were assessed using the PERMANOVA test, which allows sample-sample distances to be applied to an ANOVA-like framework (Anderson, 2001). Linear mixed-effects models were used to ascertain differences in alpha diversity and taxonomic abundances.

Statistical analysis

All experiments were performed using randomly assigned mice without investigator blinding. All data points and n values reflect biological (individual mice) replicates. Experiments were pooled whenever possible, and all data were analyzed using Prism 7 (GraphPad Software) by comparison of means using unpaired two-tailed Student’s t tests or one-way ANOVA for multiple comparison as indicated in the figure legends. Data in all figures displayed as mean ± s.e.m. unless otherwise indicated.

Highlights.

• Type 2 innate immunity is critical for normal commensalism with Demodex

• In the absence of type 2 immunity, Demodex causes aberrant inflammation

• IL-13 from ILC2s is coupled to the hair cycle and restrains stem cell proliferation

• Decreased type 2 cytokine expression is observed in patients with rhinophyma