Myeloid-derived IL-33 drives γδ T cell–dependent resistance against cutaneous infection by Strongyloides ratti

Abstract

Interleukin 33 (IL-33) is a pleiotropic cytokine released from diverse cell types that regulate both pro- and anti-inflammatory responses during pathogen infection. However, it remains unclear whether IL-33 controls key aspects of cutaneous immunity against skin-penetrating parasites. In this study, mice percutaneously infected with the parasitic helminth Strongyloides ratti were investigated to understand mechanisms of anamnestic immunity at the skin barrier. Surprisingly, mice lacking the Type 2 transcription factor STAT6 (signal transducer and activator of transcription 6) had no defects in secondary resistance to infection, whereas IL-33 gene deficiency or local blockade of IL-33 receptor (ST2) signaling abrogated host resistance. Depletion of CD4⁺ T cells or type 2 innate lymphoid cells had only a moderate impact on protection, but the loss of γδ T cells completely ablated cutaneous immunity against rechallenge. We identified a CD62L^hi IL-33 receptor (ST2)–expressing γδ T cell population that accumulated in the skin of protected mice that was dependent upon IL-33 expression in myeloid lineage antigen-presenting cells. This work suggests a previously unrecognized mechanism wherein noncanonical type 2 immunity operates through myeloid antigen-presenting cells and skin γδ T cells to adaptively repel skin-penetrating helminth larvae.

The Editors have selected this article as a highlight of the issue.

Introduction

Parasitic helminth infection is a significant cause of morbidity worldwide. However, due to their multicellular nature and the general lack of information regarding how they interact with different host cellular systems, preventative measures that block reinfection remain incompletely understood.¹ Diverse parasitic helminth species possess a skin-penetrating infectious larval stage (iL3), but whether immune-mediated host resistance can block the skin entry barrier has not been addressed. It is generally held that host protection against helminths requires type 2 immunity that can be driven by the alarmin cytokine interleukin (IL)-33^2,3 or the transcription factor signal transducer and activator of transcription 6 (STAT6),^4–6 but whether either of these mechanisms protects against skin penetration by helminths is unknown.

The IL-1 cytokine superfamily member IL-33 is constitutively expressed by nonhematopoietic cells of the skin, lung, and intestine.^7,8 Inflammatory tissue damage and/or necrotic cell death can release bioactive IL-33 from structural cells such as epithelia, endothelia, and fibroblasts.⁸ However, IL-33 can also be released from myeloid lineage cells through poorly defined mechanisms involving pore-forming proteins.² Extrinsic activity of IL-33 is initiated by binding to the heterodimeric receptor IL-1 receptor accessory protein/ST2, which is expressed by various subsets of myeloid cells, epithelial cells, and lymphocytes. IL-33 can initiate type 2–dominant inflammation including eosinophilia, activation of innate and adaptive type 2 lymphocytes (type 2 innate lymphoid cells [ILC2s] and T helper 2 [Th2] cells, respectively) and IL-4/IL-13 production by basophils.^3,9–11 Myeloid antigen-presenting cells (APCs) can be both targets of IL-33 signaling, as well as sources of IL-33 that can regulate intestinal Foxp3⁺ T regulatory populations.^2,12 Many aspects of IL-33–induced type 2 inflammation parallel STAT6-driven type 2 responses.^4–6,13 Both IL-33^2,3 and STAT6^4–6 drive lung immunity against reinfection with helminth parasites. However, distinct tissue microenvironments differ in their cellular composition and may use distinct mechanisms for host protection. In mucosal tissues, CD4⁺ T cells and innate lymphoid cells expressing ST2, classified as Th2 cells and ILC2s respectively, are known drivers of protective anti-helminth immunity.¹⁴ Th2 development is dependent on IL-4–driven STAT6 activation¹⁵ that may involve instruction via interaction with specific APC subsets.¹⁶ However, at the cutaneous barrier, it is entirely unknown whether IL-33 and/or STAT6 can promote acquired immunity against invasive helminth larvae and what cell types, or interactions, are critically important to establish immune memory.

Most studies of cutaneous immune responses directed against helminth larvae have employed a hypodermic needle to administer the infectious inoculum into a subcutaneous pocket, which bypasses the epidermal tissue.^5,17,18 The epidermal layer contains an array of immune effector cells including γδ T cells that are known to serve critical roles in response to injury and/or infection with microbial pathogens^19,20 without requirement for antigen specificity.²¹ Obata-Ninomiya et al.⁵ demonstrated in a Nippostrongylus brasiliensis model that STAT6-dependent recruitment of Arg1⁺ M2 macrophages and IL-4⁺ basophils was required to trap and kill migratory larvae during the secondary challenge, which prevented parasite dissemination to the lung. Moreover, iL3 stage larvae of both the hookworm N. brasiliensis and threadworm Strongyloides ratti are killed in the skin by extracellular traps comprised of nucleic acid released by neutrophil/eosinophils^17,18; however, it is unclear whether the subcutaneous inoculation route selectively triggered this mechanism of immunity. Neutrophils can be recruited to infected lung tissue by γδ T cells via IL-17 secretion,²² but there is no literature on whether γδ T cells can mediate host protection against skin-penetrating helminth larvae during percutaneous infection of the skin.^5,17,18 Moreover, it remains unclear whether host immunity involves alterations in proteins controlling extracellular matrix (ECM) or adherens junctions. Importantly, given the global burden (>2 billion people infected) and the recurring nature of human helminth infections that occur via skin penetration, it is imperative that experimental investigations designed to understand immunity are based on natural infection.¹

This study employed the rodent-specific threadworm S. ratti to interrogate mechanisms of cutaneous immunity.²³  S. ratti iL3 penetrates host skin and migrates through the vasculature to enter the head and lung parenchyma within 24 to 48 h post-penetration, molt into L4, and enter the gastrointestinal tract by 72 h postinfection.^24,25 Percutaneous inoculation of iL3 through the footpad was used to mimic the natural infection process,²⁶ and protective immunity was assessed by quantifying the number of iL3 that failed to enter the skin after 30 min. Contrary to expectation, acquired resistance to secondary infection was independent of STAT6 but required the presence of γδ T cells and IL-33 derived from CD11c⁺ APCs. Genetic loss of either IL-33 or γδ T cells impaired the upregulation of α-catenin, which is involved in the tension of adherens junctions between structural cells.²⁷ Interestingly, IL-33 deficiency solely in APCs abrogated expansion of γδ T cell populations that were CD62L^hi28,29 and ST2 positive. Given the importance of γδ T cells in mediating skin barrier homeostasis and wound repair^21,30 this work provides support for a model wherein γδ T cells require communication with skin myeloid cells expressing IL-33 to strengthen the skin barrier against parasite invasion.

Materials and methods

Study design

The objectives of this study were to determine whether mice acquire secondary resistance against skin penetration by the gastrointestinal parasite S. ratti. Mice (n = 4–7, in age- and sex-matched groups) were infected once (primary) or twice (secondary) with S. ratti iL3 or left naïve. Experiments were repeated 3 or more times to assure reproducibility prior to conclusion. Mice were identified using tail markings in red sharpie, or toe tattoo in the case of transgenic strains, with experimental groups randomized across cages to account for any microisolator effects. Initial experiments establishing the model did not find a significant effect of sex on parasite penetration in primary or secondary infection of C57BL/6 mice, so mice of either sex were used in subsequent experiments. Unless indicated otherwise in the figure legend, experiments involving transgenic mouse strains used sex- and age-matched littermate control mice either solely expressed Cre recombinase or lacked Cre expression but possessed the targeted gene flanked by loxP sites (floxed). To address subjectivity during the study, experimental groups were assigned a letter-number code to ensure that experiments were conducted in a blinded manner. Justification for the removal of outliers was solely based on Grubb’s test using GraphPad Prism (version 10; GraphPad Software). Figure legends indicate when outliers were removed based on Grubb’s test. For all flow cytometry data, negative controls were defined by the fluorescence-minus-one strategy.

Mice

Mice were housed under specific pathogen–free conditions in the vivarium at University of Pennsylvania school of veterinary medicine, with food and water provided ad libitum. Mice were housed using a standard 12-h light/dark cycle and were between 6 and 15 wk old at the time that experiments were initiated. Wild-type (WT) C57BL/6 mice were purchased from Taconic Biosciences and bred in-house. The following transgenic strains available from the Jackson Laboratory were obtained for in-house breeding: B6.129S2(C)-Stat6^tm1Gru/J (STAT6 knockout [KO]),³¹ B6.C-Tg (CMV-cre)1Cgn/J (CMV^Cre),³² B6.129P2^Tcrdtm1Mom/J (TCRδ KO),³³ and B6(FVB)-Mgl2^{tm1.1(HBEGF/EGFP)Aiwsk}/J (Mgl2-DTR) mice.³⁴ IL-33^fl/fl IRES-GFP reporter mice were provided by Dr. Paul Bryce and are available at Jax (B6(129S4)-Il33tm1.1Bryc/J).³⁵ LCR1^−/− mice were generated and provided by Dr. Jorge Henao-Mejia.³⁶ CMV^Cre mice were bred with IL-33^fl/fl IRES-GFP mice to generate IL-33–deficient mice. For antibody depletion of CD4⁺ or ST2⁺ cells, separate cohorts of WT mice each received either anti-mouse CD4 antibody (GK1.5 clone, cat. no. BE0003, 700 µg, intraperitoneal; Bio X Cell) or anti-mouse ST2/IL-33R monoclonal antibody (cat. no. MAB10041, 50 µg; Bio-Techne), respectively, by local injection into the hock of the hind leg.³⁷ Diphtheria toxin (DT) from Cayman Chemicals (cat. no. 19657) was used for ablation of Mgl2⁺ cells in Mgl2-DTR mice and similarly used to treat littermate control mice with 50 µg of DT by hock injection between primary and secondary infection. CD11c^Cre mice were crossed to IL-33^fl/fl IRES-GFP mice to generate mice with myeloid-specific IL-33 deficiency. All experimental procedures involving mice and rats were approved by the University of Pennsylvania’s Institutional Animal Care and Use Committee (protocol 805911) and were performed in compliance with the U.S. Department of Health and Human Services Guide for the Care and Use of Laboratory Animals.

S. ratti strain maintenance and mouse percutaneous infection

Parental (ED321) S. ratti was maintained in female rats by subcutaneous injection with 1,500 iL3 and serially passaged through different rat hosts, alternating between a Crl: NIH-Foxn1^rnu (“nude”) immune-suppressed and a Crl: WI (Wistar) immune-competent rat. Specifically, the first Crl: NIH-Foxn1^rnu (“nude”) rat is infected once, then reinfected 6 months later from fecal coprocultures.³⁸ At 6 mo post–secondary infection, parasites derived from the nude rat are used to make coprocultures that generate the iL3 used to infect a second Wistar rat. One month later, the coprocultures from the Wistar rat are used to generate iL3 in a new nude rat to repeat the cycle.

For experimental infections in mice or passage infections in rats, feces were collected from infected animals and coprocultures were generated through mixing with charcoal into a moist paste that was kept humidified in a 22 °C incubator for >5 d. Experimental infections of mice were performed with 1500 iL3 isolated from Crl: NIH-Foxn1^rnu (“nude”) rat. iL3 were isolated from the coprocultures via the Baermann funnel technique. Larvae were washed 3 times in phosphate-buffered saline (PBS) with 1% penicillin-streptomycin (P/S) and enumerated under a light microscope. After isolation, iL3 were resuspended to achieve the indicated inoculum dose of iL3 (1,500 iL3 per 100 µL PBS for each infected animal).

For percutaneous infection of the mouse paw, an additional 100 µL of PBS was combined with the 100 µL iL3 solution in an autoclaved Eppendorf tube with a 600 µL-capacity. Mice were sedated using a 7.1% ketamine/4.6% xylazine solution (80 mg/kg ketamine and 10 mg/kg xylazine, intraperitoneal), and 1 hind paw was submerged into this Eppendorf for 40 min. After exposure, the worms remaining in the tube were counted under a light microscope from a 10 µL aliquot, and the penetration efficiency was calculated based on the initial inoculum in the stock solution (worms per 100 µL stock).

For percutaneous infection in the shaved abdomen or dorsal ear pinnae, a metal ring was placed over the exposure area, the inner barrier was coated with Vaseline to seal the exposure area, and 100 µL of the infection inoculum solution was inside the metal ring. After 40 minutes, 100 µL of fresh PBS was added to the skin to resuspend the remaining worms on the skin that did not penetrate and quantified.

Quantification of S. ratti parasites that migrated to the head (upper airway)

S. ratti larvae primarily migrate to the head of the host en route to the gastrointestinal tract.²⁵ To determine the migration efficiency of iL3 after percutaneous hind paw exposure, methods were adapted from prior studies.¹⁸ Mice were euthanized by CO₂ narcosis at defined time points after primary or secondary infection. The skin was removed from the head and divided into 4 parts by cutting it in half along the sagittal suture of the skull and then cutting just caudal to the eye on either half where the zygomatic arch meets the skull. The 4 pieces of the head were each placed into separate wells of a 6-well Petri dish and submerged in tap water. The bottom of each well contained hand-drawn lines evenly spaced apart to aid in counting. These Petri dishes were incubated at 37 °C with gentle shaking for 3 h to allow worms to migrate from tissue. The tissue was removed from the well and the larva along the lines were counted under a light microscope.

Tissue dissection and digestion of skin and lymph nodes for flow cytometry or culture

To generate single-cell suspensions from skin, the hind paw was severed just above the ankle joint, the skin was degloved from the bone and muscle and placed in RPMI complete media. The skin was mechanically dissociated by mincing with dissection scissors and then incubated in a digest solution containing 2 mg/mL collagenase XI (Sigma-Aldrich), 0.5 mg/mL hyaluronidase (Sigma-Aldrich), and 0.1 mg/mL DNase (Roche) in complete RPMI media for 25 min at 37 °C with agitation. The digested skin was then passed through and 16-gauge needle 3 to 5 times and then incubated in the digestion solution at 37 °C for an additional 20 min. The skin homogenate was then pressed through a 100 μm cell filter by a rubber syringe plunger and supplemented with equal volume RPMI. The popliteal lymph node (pLN), which drains the hind paw, was excised, homogenized by passage through a 70 μm cell strainer (Fisher Scientific), centrifuged at 1,500 rpm for 5 min at 4 °C, and resuspended in complete RPMI solution. Cells in suspensions from skin and pLN were counted using a Muse Cell Analyzer (EMD Millipore), per the manufacturer’s instructions.

Enzyme-linked immunosorbent assay

Popliteal lymph node cells were cultured for 4 d on a plate coated with anti-CD3/28 antibodies or without stimulants (media). Skin biopsies were homogenized in RIPA buffer containing 1× Proteinase cocktail inhibitor (Thermo Fisher Scientific), and their protein concentration was determined using the Pierce BCA Protein Assay Kit (Thermo Fisher Scientific). Serum from mice was collected at indicated experimental time points by retro-orbital bleeds (or cardiac puncture at experimental endpoint) in Fisher Scientific amber serum collection tubes. Serum was centrifuged at 1,600 rpm for 15 min to separate serum from red blood cells and stored at −20 °C until time of analysis. For enzyme-linked immunosorbent assay (ELISA) analysis of cytokine expression, 500 μg of protein, 10 to 50μL of cell-free supernatant, or 50 μL of red blood cell–free serum were used to determine cytokine levels using the following commercially available ELISA kits following the manufacturer’s instructions (Invitrogen): IL-4 (88-7044-22), IL-10 (50-112-5188), tumor necrosis factor α (88-7324-88), interferon γ (88-7314-88), IL-5 (88-7054-88), IL-17A (88-7371-77), IL-22 (88-7422-88), IL-17F homodimer (88-7472-88), IgE (88-50460-88), and IL-33 (50-112-5200).

Histological analysis

To reduce the number of animals used and provide additional internally controlled analyses, paws were bisected longitudinally, and one-half provided skin for flow cytometry or RNA isolation. The other half was used for histological analysis. Paw tissue was fixed in 4% paraformaldehyde for 5 d, then rinsed briefly in distilled water before decalcification by submersion in Cal-Ex II (Fisher Scientific) for 9 to 11 d, with exchange of the solution every 2 to 3 d. The tissues were then washed again in distilled water prior to postfixation in 4% paraformaldehyde for 5 h. At the end of the decalcification, the paws were dehydrated in 70% ethanol in cassettes and submitted to the University of Pennsylvania Center for Molecular Studies in Digestive and Liver Diseases (P30DK050306) and the Molecular Pathology and Imaging Core (RRID: SCR_022420) for paraffin embedding, tissues sectioning, and staining with hematoxylin and eosin.

Flow cytometry

Single-cell suspensions from skin or lymph nodes were stained first for viability using the LIVE/DEAD Fixable Aqua Dead Cell Stain Kit following the manufacturer’s protocol (Invitrogen; cat. no. L34957). After washing the cells in fluorescence-activated cell sorting buffer, Fc block containing anti-CD16/32 was performed for 15 minutes, followed by surface marker staining for 25 min; for both steps, cells were kept on ice and in the dark. The eBioscience Foxp3/Transcription Factor Staining Buffer Set (Invitrogen) was used to permeabilize and fix the cells according to the manufacturer’s protocol. Cells were subject to intracellular staining overnight, also on ice and in the dark. Samples were run on a BD Symphony A3 Lite (maintained by the Penn Cytomics and Cell Sorting Core Facility). Antibodies from BD Biosciences include BUV395 Siglec-F (cat. no. 740280), BUV563 CD45 (cat. no. 752412), BUV737 CD11b (cat. no. 612801), and BUV737 CD4 (cat. no. 612844). Antibodies from BioLegend include Brilliant Violet 605 (BV605) CD11c (cat. no. 117334), BV711 CD45R/B220 (cat. no. 103255), BV711 CD19 (cat. no. 115555), BV711 CD3ε (cat. no. 100349), BV711 CD8α (cat. no. 100747), BV711 NK-1.1 (cat. no. 108745), APC/Cyanine7 Ly-6G (cat. no. 127624), PE FcεRIα (cat. no. 134307), BV605 TCR γ/δ (cat. no. 118129), BV711 CD11b (cat no. 101242), BV711 CD11c (cat. no. 117349), BV785IL-33Rα (IL1RL1, ST2, cat. no. 145321), Alexa Fluor (AF) 700 CD44 (cat. no. 103026), PE Dazzle CD62L (cat. no. 104448), PE Cy7 CD90 (cat. no. 140324), PE Cy5 CD3 (cat. no. 100310), BV711 CD90.2 (Thy1.2, cat. no. 105349), APC CD44 (IM7) (cat. no. 103012), AF700 CD3 (cat. no. 105349), PE Cy5 CD278 (ICOS) (cat. no. 107708), PE Cy7 CD127 (cat. no. 135013), BV711 Ly6G (cat. no. 127643), PE Cy7 CD62L (cat. no. 104418), PE Dazzle CD127a (cat. no. 135032), and PE/Cyanine5 TCR β chain (cat. no. 109210). Antibodies by eBiosciences include BV421 arginase 1 monoclonal antibody (A1exF5; cat no. 404-3697-80 or 82) and AF 488 Gata-3 clone TWAJ (cat. no. 53-9966-42).

Quantitative real-time polymerase chain reaction and RT² profiler

RNA from paw skin was isolated using the Qiagen RNeasy Mini Kit. RNA concentration was determine using Qubit RNA High Sensitivity Assay kit followed by assessment of RNA quality using the High Sensitivity RNA ScreenTape Analysis. Maxima H minus Reverse Transcriptase (Thermo Fisher Scientific) generated complementary DNA (cDNA) from 500 ng/mL of RNA isolated from each skin sample. Quantitative real-time polymerase chain reaction (PCR) was performed using 300 ng/mL of cDNA/sample, gene-specific primers, and the SsoAdvanced Universal SYBR Green Supermix (Bio-Rad) on the CFX96 platform (Bio-Rad), with thermal cycling conditions and melting curves optimized for the Supermix and the primers used. Relative expression of the target gene to GAPDH was determined using the equation Expression = 2^{(GAPDH Cq−Target gene Cq)}. Primers used in the quantitative real-time PCR reactions included GAPDH forward (AGGTCGGTGTGAACGGATTTG), GAPDH reverse (TGTAGACCATGTAGTTGAGGT), and Ctnna2 proprietary primers optimized and sold by Qiagen (PPM03836D-200), which match the primers set used in the RT2 Profiler Array described in the following paragraph.

To perform an exploratory determination of ECM and cell adhesion genes that could be affected by our experimental manipulations we used a 96-well array of primers for 84 genes, 5 internal housekeeping genes, and other internal assay controls lyophilized in the plate, the RT² Profiler PCR Array (Mouse Extracellular Matrix and Adhesion Molecules; cat. no. 330231 PAMM 013ZA) from Qiagen. Skin harvest, RNA isolation, and cDNA generation were performed as described previously. The cDNA from a single biological replicate and SsoAdvanced Universal SYBR Green Supermix was added to each well of the RT² Profiler plate in a ratio of 1:5 so that each well-received 10 μL of the cDNA SYBR Green mixture.³⁹ The RT² Profiler qPCR reaction was run as per Qiagen’s instructions for thermal cycling and fluorescence detection, on a CFX96 platform (Bio-Rad). The Cq values were uploaded to the Gene Globe (Qiagen) RT² Profiler PCR Data Analysis website to normalize and analyze the data. The data was normalized to the arithmetic mean (3 mice per group) of GAPDH, and the WT naïve group was set as the control group. The displayed fold change regulation depicts the normalized gene expression (2^(−ΔCT)) in the test sample divided by the normalized gene expression (2^(−ΔCT)) in the control sample. Of the 84 genes analyzed, we show only the ones with a significant (P < 0.05) fold increase or decrease >2.

Statistical analysis

Error bars depict the SEM and P < 0.05 was considered statistically different for all comparisons. We evaluated all data for outliers using Grubb’s test and none were identified for exclusion. Student’s t test was used to compare 2 groups. For nonrepeating data comparing 3 or more groups, such as cellular content determined by flow cytometry, 1-way analyses of variance (ANOVAs) with Bonferroni post hoc tests were used. For repeated measurements, such as penetration efficiency determined in the same mouse receiving a primary and secondary infection, a 2-way ANOVA with repeated measures and between-group comparisons followed by uncorrected Fisher’s least significant difference post hoc tests was used. All statistical tests were performed in GraphPad Prism 10.

Results

Acquired resistance to S. ratti larval penetration of hind paw skin is accompanied by mixed type 1/2 inflammation

A percutaneous penetration model was developed to evaluate protective immunity against the skin-penetrating nematode S. ratti (Fig. 1A, B). In this system, naïve mice (0 d postinfection [dpi]) are exposed to a percutaneous infection on day 0 (d0) and secondary exposure on d12. Exposure of the hind paw for 40 min to 1 or 2 infections with 2,500 iL3, each time followed by quantification of the nonpenetrating parasites revealed that the proportion of penetrating iL3 was significantly reduced upon secondary exposure, with 60% to 80% penetration during primary infection (d0) versus 40% during secondary infection (d12) (Fig. 1C). Quantifying the number of worms recovered from the cranial upper airway on d2 and d14 revealed that significantly more worms reached this tissue site after primary infection as compared with the second (Fig. 1D).¹⁸ Analysis of inflammatory changes in the footpad skin by hematoxylin and eosin staining at d0, d2, d12, and d14 revealed that dense pockets of inflammatory cells appeared by 12 dpi (Fig. S1A). Resistance to reinfection was not observed when percutaneous infection was performed at other sites such as the abdomen or ear skin (Fig. S1B, C). However, when the secondary challenge involved the contralateral hind paw, there was also reduced larval penetration efficiency, which was not observed to the same degree when the secondary challenge involved the ipsilateral forepaw (Fig. 1E). Collectively, these data indicated that acquired resistance to percutaneous entry by iL3 developed only in hind paw skin.

Figure 1.

Immunity against Strongyloides ratti larval skin penetration is accompanied by mixed type 1/2 inflammation and limited to the hind paw. (A) Schematic of percutaneous infection strategy using the hind paw for exposure to 2,500 iL3. (B) Experimental time points for exposure and assessment of inflammatory pathology in days postinfection. (C) Penetration efficiency in WT C57BL/6 mice exposed at the indicated times. (D) Numbers of iL3 recovered from the head at 2 d (48 h) post–primary or secondary infection. (E) Penetration efficiency at d2 (solid) vs. d14 postexposure (hatched) when the challenge paw was the initial hind paw, the contralateral hind paw, or the ipsilateral forepaw. (F) Cytokines released from popliteal lymph node lymphocyte cell cultures from mice at 2 dpi (white bars) or 14 dpi (red bars) left untreated or stimulated with α-CD3/28 monoclonal antibody as determined by ELISA. (G–K) Flow cytometry performed on single-cell suspensions of hind paw skin at d0, d2, d12, and d14 showing the frequency of eosinophils (FcER1⁻/Siglec F⁺, Ly6G⁺, CD11b⁺)), neutrophils (FcER1⁻/Siglec F⁻, Ly6G⁺, CD11b⁺), M2 macrophages (CD11b⁺, F4/80⁺/CD64⁺, and Agr1⁺), dendritic cells (CD11c⁺/MHCII⁺), and Langerhans cells (CD11b⁻CD11c⁺ MHCII⁺, EpCAM⁺). All data are representative of 3 independent experiments, except panel E, which is representative of 2. Each symbol represents values from an individual mouse, plots represent mean ± SEM. Data were analyzed using Student’s t test in panel C and D and a 1-way ANOVA with Bonferroni’s post hoc tests in all other panels. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. IFNγ, interferon γ; ns, not significant; TNF, tumor necrosis factor.

Evaluation of the immunological changes that accompanied host protection revealed significantly increased secretion of IL-4 and interferon γ, but reduced tumor necrosis factor α secretion and no change in IL-10 from pLN cultures exposed to T cell mitogen stimulation (α-CD3/α-CD28) as determined by ELISA (Fig. 1F). Serum IgE levels also increased over throughout the protocol (Fig. S1D). There was significant paw swelling on d1 and this edema was significantly greater after second exposure (Fig. S1E). Despite this increased edema, total cellularity in the hind paw did not increase (Fig. S1F), but the frequency of CD45⁺ leukocytes and CD11b⁺ myeloid cells in the hind paw skin increased significantly by d2 and remained elevated throughout the period of evaluation (Fig. S1G, H).

Previous reports have demonstrated an important role for myeloid lineage cells in larval trapping and killing in the skin.^5,14,18 There was an increased frequency of CD11b⁺ myeloid cells at 48 h following each infectious exposure on d2 and d14 (Fig 1H). To characterize the myeloid subsets, flow cytometry was used to evaluate affected skin at 0, 2, 12, and 14 dpi. There was a significant increase in the frequency of neutrophils, M2 macrophages, dendritic cells, and Langerhans cells but not eosinophils, mast cells, or basophils (Fig. 1G–K; Fig. S1I, J). Neutrophils expanded following each exposure to iL3, whereas M2 macrophages remained elevated throughout the protocol, whereas dendritic cells were only significantly elevated by d14 and Langerhans cells were elevated on d12 to d14. Taken together, this indicated differential expansion of myeloid subsets in the hind paw skin of S. ratti–exposed mice.

Acquired resistance to secondary percutaneous infection is not dependent on STAT6

The transcription factor STAT6 regulates host immunity against GI nematodes through the anti-parasitic effects it imparts on epithelial, myeloid, and T cell lineages.^5,6,40 Contrary to expectation, STAT6-deficient mice were as resistant to skin penetration by iL3 as their WT matched control mice (Fig. 2A), with no significant impact on the extent of footpad pathology (Fig. 2B), inflammation, or edema (Fig. S2A–C). Evaluation of pLN cytokine responses revealed abrogation of IL-4 and IL-5 secretion as expected in the absence of STAT6 (Fig. 2C). STAT6-deficient mice had significantly fewer eosinophils (Fig. 2D) and M2 macrophages (Fig. S2D) as compared with infected WT mice at d14 despite equivalent secondary resistance. As expected, ILC2 and Th2 cell populations were either significantly reduced or moderately reduced in STAT6-deficient mice (Fig. 2E, F).^13,40 There was no defect in the accumulation of skin GATA3⁺ Foxp3⁺ T regulatory or γδ T cell populations (Fig. 2G, H). Despite the known role for ILC2s and Th2 cells in anamnestic immunity in mucosal tissues,^38,41,42 these data implied they were unlikely to serve a critical role in host resistance against percutaneous infection by S. ratti larvae. Even so, both CD4⁺ T cells and ILC2s increased in WT hind paw skin by d12 (Fig. 3A, B; Fig. S3A shows gating strategy) associated with resistance to reinfection.

Figure 2.

Acquired immunity does not rely on STAT6. (A) Penetration efficiency of the initial inoculum of 1,500 iL3 in WT (gray violin plots) or STAT6 KO (red violin plots) mice. (B) Photomicrographs of individual mouse digits from WT or STAT6 KO mice at d14 following hematoxylin and eosin (H&E) staining at the indicated timepoints. Black arrows point to inflammatory cell infiltrates. Scale bars = 400 µm. (C) Cytokines released from popliteal lymph node lymphocyte cell cultures from mice at d14 in WT (gray box plots) or STAT6 KO (red box plots) mice left untreated or stimulated with α-CD3/28 monoclonal antibody as determined by ELISA. (D–H) Flow cytometry performed on single-cell suspensions of hind paw skin from WT or STAT6 KO at d14 showing the frequency and cells/mg of tissue of eosinophils, ILC2s (lineage (B220, CD11b, CD11c, NK1.1, CD8a)⁻/CD90⁺, CD3⁻/CD4⁻, CD127⁺/GATA3⁺), Th2s (lineage⁻/CD90⁺, CD3⁺/CD4⁺, GATA3⁺/Foxp3⁻, GATA3⁺), GATA3⁺ T regulatory cells (Tregs) (lineage⁻/CD90⁺, CD3⁺/CD4⁺, GATA3⁺/Foxp3⁺), and γδ T cells (as lineage⁻/CD90⁺, CD3⁺/CD4⁻, TCRδ⁺). (A) Representative of 3 or 4 independent experiments. Each symbol represents an individual mouse, and plots represent mean ± SEM. Data were analyzed using Student’s t test in panels A and B, and a 1-way ANOVA with Bonferroni’s post hoc tests in all other panels. *P < 0.05 and **P < 0.01. P > 0.05 was not significant (ns).

Figure 3.

A critical role for IL-33 but not CD4⁺ T cells or ILC2s in resistance to secondary penetration. (A, B) The frequency and number of CD4⁺ T cells and ILC2s at d0 (naïve) and d12 postexposure to S. ratti iL3. (C) Penetration efficiency in WT or LCR1^−/− (ILC2 KO) mice after exposure to 1,500 S. ratti iL3. (D) Penetration efficiency in isotype-treated or α-CD4 monoclonal antibody (mAb)–treated WT mice via intraperitoneal injection. Mice were exposed to 1,500 iL3. (E) IL-33 in the hind paw skin at d0, d1, and d13, as determined by ELISA. (F) Penetration efficiency of iL3 in of WT (gray violin plots) or IL-33 KO (CMV^CreIL-33^fl/fl) (green violin plots) at indicated times. (G) Penetration efficiency of iL3 in isotype-treated (gray violin plots) or α-ST2/IL-33R treated (green violin plots) at indicated times. Data are representative of 1 (C), 2 (A, B), 3 (D, E), or 4 (F, G) independent experiments. Each symbol represents values from an individual mouse, plots represent mean ± SEM. Data were analyzed using Student’s t test in panels A and B, and a 1-way ANOVA with Bonferroni’s post hoc tests in all other panels. *P < 0.05, **P < 0.01, and ****P < 0.0001. P > 0.05 was not significant (ns).

Although CD4⁺ T cells and ILC2s are dispensable, IL-33 is necessary for acquired resistance to percutaneous infection with S. ratti iL3

To directly evaluate the role(s) of ILC2s and Th2 cells, we used genetic KO of locus control region 1 (LCR1^−/− mice)³⁶ and CD4-antibody depletion strategies, respectively, between primary and secondary exposure (Fig. S3B, depletion confirmed in Fig. S3C) to directly test whether ILC2s or CD4⁺ T helper cells were required for acquired resistance in the skin. Two-way ANOVAs revealed only a significant effect of infection (for LCR1^−/−: F_1,8 = 11.06, P = 0.0105; for CD4 depletion: F_1,8 = 11.45, P = 0.0096) but no significant genotype (F_1,8 = 2.398, P = 0.1601) or CD4 depletion effect (F_1,8 = 3.214, P = 0.1108), nor were there significant interactions with infection (for LCR1^−/−: F_1,8 = 2.504, P = 0.1522; for CD4 depletion: F_1,8 = 0.9754, P = 0.3523) with respect to penetration efficiency exhibited by mice lacking ILC2s (LCR1^−/−) or CD4⁺ cells as compared with their respective controls mice. Post hoc comparisons indicated that the control groups recapitulated secondary protection and LCR1^−/− mice and CD4⁺ cell–depleted mice had an intermediate phenotype where the significant difference between primary and secondary is lost, but they also do not differ significantly from control mice, suggesting only a moderate reduction in secondary resistance to larval entry (Fig. 3C, D). These data suggested that Th2 cells and ILC2s were minimally involved in anamnestic immunity against skin-penetrating helminth larvae.

IL-33 is critical for anti-helminth immunity in the lung and intestinal tissues, but its role in cutaneous immunity in helminth infection is unknown.^2,3,10,41 IL-33 protein levels in footpad skin significantly increased by d1 and remained elevated as compared with naïve skin (Fig. 3E). To test the potential requirement for IL-33 in protection against skin penetration, genetic deletion, and antibody-neutralization strategies were used. Mice lacking IL-33 ubiquitously in all cells by were generated by interbreeding the CMV^Cre and IL-33 GFP^flox/flox transgenic strains and compared with Cre-negative control mice.³² Strikingly, IL-33 deficiency in all cells abrogated acquired immunity against secondary infection with S. ratti iL3 (Fig. 3F). Second, In parallel experiments, IL-33 responsiveness was blocked by administration of an ST2/IL-33R neutralizing antibody into the hock of the infected leg between the primary and secondary exposures (Fig. S3D).³⁷ This approach led to a near-complete loss of secondary host resistance (Fig 3G). Thus, intact, local IL-33/ST2 signaling was required for secondary cutaneous immunity against percutaneous larval infection.

γδ T cells are required for secondary resistance to percutaneous infection

Demonstration that among lymphocytes, neither CD4 cells nor ILC2s served an essential role in resistance to secondary infection, prompting a hypothesis that γδ T cells could drive protective immunity, particularly given that γδ T cell numbers were unaffected by STAT6 deficiency (Fig 2H). Indeed, IL-33–deficient mice had significantly fewer γδ T cells in the skin after 2 exposures compared with their IL-33^fl/fl control mice (Fig 4A), and ST2⁺ γδ T cells were also significantly decreased amid IL-33 deficiency (Fig 4B). TCRδ KO mice were subjected to the percutaneous infection protocol to directly test whether γδ T cells were required for host resistance. Strikingly, the absence of γδ T cells in the host significantly abrogated protective immunity at levels equivalent to IL-33 deficiency or ST2 monoclonal antibody neutralization (Fig. 4C). These data pointed to a mechanism involving γδ T cells and IL-33, in acquired resistance to larval skin penetration of hind paw skin.

Figure 4.

γδ T cells are required for secondary resistance to skin penetration and mediated by IL-33 expression. (A, B) Flow cytometry performed on single-cell suspensions of hind paw skin in WT or IL-33 KO mice showing the frequency and number of (A) γδ T cells or (B) ST2⁺ γδ T cells at 14 dpi. (C) Penetration efficiency in WT (gray violin plots) or TCRδ KO (blue violin plots) mice exposed to 1,500 iL3 at indicated times. (D) Representative flow plot gated on CD90⁺ TCR γδ⁺ cells to determine CD44^hi CD62L^lo, CD44^lo CD62L^hi, ST2⁺, and CD62L⁺ ST2⁺ γδ T cell subsets at d14. Frequency and number of (E) CD44^lo CD62L^hi γδ T cells and (F) CD62L⁺ ST2⁺ γδ T cells at d0 and d12. Frequency and number of (G) CD44^lo CD62L^hi γδ T cells and (H) CD62L⁺ ST2⁺ γδ T cells in WT and IL-33 KO hind paw skin at d14. Data are representative of 2 or 3 experiments. Each symbol represents an individual mouse, plots represent mean ± SEM. Data were analyzed using Student’s t test or a 1-way ANOVA with Bonferroni’s post hoc tests. *P < 0.05, **P < 0.01, and ***P < 0.001.

Hind paw skin was evaluated in WT mice at 12 dpi, which corresponded to the timepoint immediately prior to administration of the second infectious inoculum to interrogate the characteristics of skin γδ T cells that were poised for host protection (gating strategy shown in Fig. 4D). The CD44^lo CD62L^hi but not CD44^hi CD62L^lo subset of γδ T cells significantly increased in both frequency and number by 12 dpi (Fig. 4E;  Fig. S4E). There was also an ST2⁺ subset of the CD62L⁺ γδ T population, which increased in frequency by 12 dpi (Fig. 4F). In contrast, in IL-33–deficient mice that lacked protection had significantly fewer the CD44^lo CD62L^hi (Fig. 4G) and CD62L⁺ ST2⁺ γδ T cells (Fig. 4H), while the frequency of the CD44^hi CD62L^lo γδ T cell subset was increased (Fig. S4F). Collectively, this data implicated IL-33 signaling in controlling the accumulation of distinct populations of skin γδ T cells, some of which that had the potential to respond to IL-33.

Epidermal and dermal γδ T cells play important roles in skin wound healing and barrier homeostasis through regulation of keratinocyte survival and proliferation.^21,30 Given the “repulsive-like” nature of anamnestic immunity in this system, we reasoned that γδ T cells and/or IL-33 signaling led to inflammatory cell infiltration, epidermal thickening, or a change in ECM/adherens junction genes that enforced skin barrier integrity. Surprisingly, hematoxylin and eosin–stained tissue sections of the infected hind paw revealed similar levels of hematopoietic infiltrates and overall epidermal thickness in WT and IL-33 KO mice at d14 (Fig. S4C). There was no difference in the extent of tissue edema between strains at this time point (Fig. S4H). To investigate whether IL-33 deficiency dysregulated the expression of ECM and adherens junction molecules an RT Profiler Array of ECM and adhesion molecule genes was used. Data showed an upregulation of 13 ECM and adhesion molecule genes and a downregulation of the metalloprotease gene Mmp13 in the protected skin of WT mice as compared with the naïve skin (Fig. S4I). Critically, Ctnna2 (which encoding αN-catenin) increased 4.5-fold in protected WT skin as compared with naïve skin but did not change in IL-33–deficient skin (Fig. S4J). Ctnna2 expression increased in WT skin following 2 exposures to S. ratti iL3 but not in tissues from infected γδ TCR KO mice (Fig. S4K). These data indicated that acquired resistance to worm penetration of hind paw skin was associated with alterations in skin barrier genes but not with a generalized thickening of the epidermal layer.

APC-derived IL-33 is necessary for the expansion of CD62L⁺ γδ T cells and acquired resistance to percutaneous penetration

That IL-33 was essential for the accumulation of γδ T cells, particularly, a CD62L⁺ ST2⁺ γδ T subset, prompted an inquiry into its cellular source. Although IL-33 is constitutively expressed by cutaneous epithelial and endothelial cells,⁸ myeloid lineages can also express the Il33 transcript and IL-33 protein.^2,43 Therefore, we formulated a hypothesis that skin APCs promoted γδ T accumulation, potentially through IL-33.

Given the importance of Mgl2⁺ (CD301b⁺) dendritic cells in promoting skin type 2 inflammation,^34,44 experiments were designed to test whether a loss of Mgl2⁺ DCs would impair γδ T accumulation. To address this idea, Mgl2-DTR mice were given a primary hind paw exposure to S. ratti iL3, followed by DT administration at d3 and d9 to selectively deplete CD301b⁺ DCs before secondary challenge on d12 (Fig. 5A; see Fig. S4A for flow gating). Data in Fig. 5B revealed that DT treatment of Mgl2-DTR mice abrogated resistance to skin penetration by S. ratti iL3. This approach significantly reduced the skin ST2⁺ γδ T cell population at d14 compared with control mice (Fig. 5C, D). Analysis of depletion specificity following DT treatment showed that Mgl2⁺ DCs, ST2⁺ DCs, and Langerhans cells were all significantly reduced following DT treatment of Mgl2-DTR mice (Fig. S4B). In accordance with previous studies, DT treatment of the Mgl2-DTR strain significantly reduced CD4⁺ T cells,^34,44 but not ILC2s (Fig. S4C). This data supported the notion that certain APC populations were critical for γδ T cell–dependent protective immunity.

Figure 5.

IL-33⁺ APCs are required for γδ T cell–mediated secondary resistance. (A) Schematic of strategy to deplete Mgl2 (CD301b⁺) cells from Mgl2-DTR mice in the percutaneous infection model. (B) Penetration efficiency in WT mice or Mgl2-DTR mice treated with DT via hock injection and infected percutaneously with 1,500 iL3. (C, D) Flow cytometry on single-cell suspensions of hind paw skin showing cell number per milligram of tissue of total γδ T cells and ST2⁺ γδ T cells in WT mice or Mgl2-DTR mice at d14. (E) Penetration efficiency in WT (CD11c^Cre) or myeloid-specific IL-33 KO (CD11c^CreIL-33^fl/fl) mice exposed to 1,500 iL3 at indicated time points. (F) Representative flow plots of the hind paw skin cells gated on CD90⁺ TCRβ⁻ TCR γδ⁺ cells to determine the frequency and cell number of (G) CD44^lo CD62L^hi γδ T cells and (H) CD62L⁺ ST2⁺ γδ T cells. Data are representative of 2 to 3 independent experiments. Each symbol represents an individual mouse, plots represent mean ± SEM. Data were analyzed using Student’s t test in panels A and B, and a 1-way ANOVA with Bonferroni’s post hoc tests in all other panels. *P < 0.05 and ***P < 0.001. P > 0.05 was not significant (ns).

Last, to directly evaluate the role for APC-derived IL-33 in resistance to skin penetration and accumulation of skin γδ T cells, we subjected mice lacking IL-33 specifically in the Itgax CD11c expressing population (CD11c^Cre IL-33^fl/fl) and CD11c^Cre control mice to our percutaneous reinfection model. Strikingly, CD11c^Cre IL-33^fl/fl animals failed to develop secondary resistance to percutaneous infection (Fig. 5E). Although these mice did not have a significant reduction in CD4⁺ T cells or ILC2s (Fig. S4D), they did have a significant reduction in both CD44^lo CD62L^hi (Fig. 5G) and CD62L⁺ ST2⁺ γδ T cell populations (Fig. 5H) as compared with CD11c^Cre control mice. These results mirrored our previous findings in mice deficient in all IL-33 expression (Fig 3F) and indicated a critical role for APC in the accumulation of host protective γδ T cells in the skin by providing IL-33 to repel invasive S. ratti iL3.

Discussion

This study combined a percutaneous infection strategy with precise quantification of penetration efficiency to investigate mechanisms of anamnestic skin immunity against the GI nematode S. ratti. Our experimental approach revealed that host resistance can functionally repel and/or impede the process of skin penetration, which was entirely unexpected, as protection is thought to be mediated by inflammatory cells that attach, trap, and kill larvae once inside host tissues. Acquired immunity to rechallenge did not occur at anatomically distinct skin sites such as the ear or abdomen, although it did manifest at both contralateral and ipsilateral paws (the latter to a lesser extent). Surprisingly, these data support a mechanism of anamnestic immunity in hind paw skin that did not require the canonical type 2 transcription factor STAT6, CD4⁺ T cells, or ILC2s in acquired resistance to larval entry. Instead, the data indicate vital role(s) for IL-33, γδ T cells, and myeloid APCs, the latter serving as an essential source of IL-33. We found an IL-33–dependent accumulation of CD62L⁺ ST2⁺ γδ T cells in footpad skin, which has not been previously described. Interestingly, both IL-33 and γδ T cells were independently required for the infection-induced upregulation of αN-catenin, which is a key skin barrier integrity gene that controls cell-cell adhesion. Whether IL-33 drives γδ T cells to adopt distinct effector functions is unknown, but to our knowledge, this is the first report demonstrating γδ T cell–dependent cutaneous immunity against a parasitic helminth. This is mechanistically distinct from studies of skin immunity against helminths that employed injection-based inoculation methods.^5,17,18

Perhaps the most unexpected finding was that STAT6 was not required for host resistance to percutaneous infection by S. ratti larvae despite the widely consistent evidence for its importance in immunity against diverse helminth species.^4–6,13 Upon binding of IL-4Ra by IL-4 or IL-13, STAT6 binds to the promoter region of various genes associated with the expulsion of helminths from the intestine by myeloid, lymphoid, epithelial, and smooth muscle cell lineages.⁶ Surprisingly, our model shows that in the absence of STAT6, secondary resistance to penetration is maintained. Obata-Ninomiya et al.⁵ demonstrated a critical role for Arg⁺ M2 macrophages and basophils in mediating acquired immunity to secondary infection in N. brasiliensis. Additionally, Ehrens et al. showed an important role for eosinophils in trapping and killing S. ratti larvae in the skin preventing dissemination and systemic infection.¹⁸ However, despite a significant reduction in M2 macrophages and eosinophils in the absence of STAT6, acquired resistance to penetrating larvae after a primary infection was maintained. There were no significant differences in mast cells and basophils at the indicated time points in this system. Instead, the data point to STAT6 independent, but IL-33– and γδ T cell–dependent factors that drive acquired skin immunity against large metazoan pathogens.

Previous studies of cutaneous helminth infection have not investigated any roles for lymphocyte populations in local protective immunity.^5,17,18 Many lymphocytes, including ILCs and CD4⁺ T cells reside in the skin and serve essential roles in immunity against bacterial, fungal, and parasitic infections.^45–47 CD4⁺ T cells can drive acquired immunity in the skin and seed distal skin sites to promote immunity against reinfection.^44–46 In helminth infection, CD4⁺ T cells differentiate into Th2 cells and mediate antigen specific protective immunity against parasitic reinfection in the airway and intestines.³⁸ While we did find an increase in CD4⁺ T cells by 12 dpi, loss of these lymphocyte populations only moderately abrogated acquired resistance to percutaneous penetration. ILC2s also play critical roles in protective immune responses to helminth infection at mucosal barrier sites by inducing and recruiting myeloid cells.^41,42 To evaluate the necessity of ILC2s, we employed previously described LCR1^−/− mice that specifically lack mature ILC2s through interruption of the Id2 promoter interaction with the H3K27ac enhancer.³⁶ Similar to the phenotype of CD4 depletion, ILC2 deficiency only led to a moderate reduction in resistance to secondary penetration. These data indicated that lymphocyte populations other than CD4⁺ T cells and ILC2s contributed to skin immunity in the hind paw.

IL-33 is an alarmin cytokine constitutively expressed by barrier cells (epithelia, endothelia, fibroblasts) and released by these populations during barrier damage/infection.^7,8 In the skin, IL-33 can drive type 2 immunity in the context of allergy and atopy.^7,48  S. ratti iL3 must forcibly penetrate the skin barrier to establish infection,²⁶ which is consistent with our data showing that IL-33 expression increases after both primary and secondary percutaneous infection. Moreover, blocking IL-33 signaling through neutralization of its receptor ST2 only after the initial antigen priming event was enough to abrogate acquired resistance. Therefore, data show that IL-33/ST2 is critical for manifesting recall immunity after the initial priming events of percutaneous infection. That IL-33 deficiency significantly reduced the accumulation of γδ T cells highlight a previously unidentified role for IL-33 in skin γδ T cell biology.

We are unaware of any previous investigations into whether γδ T cells play any role in cutaneous anamnestic immunity against helminth infection. While γδ T cells are well known to serve essential roles in skin homeostasis, wound healing, and cutaneous immunity against microbial and protozoan pathogens,^21,30 how they could function in the context of infection with skin-penetrating metazoans has remained an enigma. In fact, due to the rarity, there are no exhaustive studies investigating the involvement of γδ T cells in human infection with Strongyloides spp. or other skin-penetrating helminths. Many investigations on this front are skin disease and psoriasis-related.^49–52 In both mice and humans, epidermal and dermal γδ T cells express specific repertoires of invariant Vγ chains that are distinct from γδ T cells of other tissues. In the mouse skin, dendritic epidermal γδ T cells express Vγ5Vδ1 and dermal γδ T cells express Vγ4Vδ1 and Vγ6Vδ1.^30,49 γδ T cells that express Vγ4 recirculate and enter inflamed tissue to initiate wound repair while Vγ6 acts as persistent effector cells that terminally differentiate and migrate to distal issue sites.³⁰ Mouse γδ T cells have morphological and phenotypic differences depending on the tissue in which they reside.⁵³ How this affects their functionality in inflammation is unknown, but it has been reported that γδ T cells in footpad skin exhibit higher expression levels of activation markers than ear or abdominal skin.⁵³ Data show that hind paw skin mounted resistance against penetration, while the ear and abdominal skin did not. Whether this is due to unique features of γδ T cells in the hind paw skin remains to be determined. Further investigation will be required to fully understand the Vγ chain identities and cytokine release patterns of the γδ T cell population(s) in the footpad skin of helminth-exposed animals.

γδ T cells are classically considered innate-like lymphocytes due to their ability to activate and carry out functions without antigen presentation.⁵⁴ However, recent studies have shown that γδ T cells expand rapidly in a recall-like response to recurrent infections and migrate to distal noninflamed tissue.^55,56 γδ T cell subsets exhibit distinct phenotypic markers much like αβ T cells.^57–59 In particular, effector/memory markers CD44 and CD62L can identify distinct populations of differentiated γδ and CD4⁺ T cells that play unique roles in acquired immune responses.^57,59 Evidence suggests CD62L⁺ γδ T cells originate from secondary lymphoid organs such as the pLN with the potential to initiate protective adaptive-like immunity in the periphery.^28,29 IL-33–deficient mice lacked the CD44^lo CD62L^hi subset but not the CD44^hi CD62L^lo γδ T cell population, indicating that IL-33 potentially serves a key role in mediating the expansion of the CD62L⁺ γδ T cell subset. Further, in the absence of IL-33 expression the frequency and number of ST2⁺ γδ T cells and more specifically the CD62L⁺ ST2⁺ γδ T cell population was significantly reduced. Thus, IL-33 may be involved in the recruitment or functional behavior of cutaneous γδ T cells after helminth skin penetration, which may help confer acquired resistance.

Because IL-33 had such an important role in acquired resistance and γδ T cell accumulation in the skin, we speculated that IL-33 deficiency would negatively impact skin integrity. Strongyloides spp. secrete extra secretory products such as serine proteases, cysteine proteases, and metalloproteases, which can lead to the degradation of the epithelial barrier and aid in parasite invasion and dissemination.^26,60,61 Two scenarios were envisioned wherein the epidermis changed to impede iL3 penetration: (1) the skin epidermis thickened after a primary infection, making it harder for worms to penetrate the skin; or (2) the skin barrier integrity was altered making the epithelial barrier stronger/tighter, which would make it more difficult for migratory larvae to forcibly penetrate the skin. γδ T cells can mediate hyperkeratinization and epidermal thickening through the release of cytokines and growth factors.³⁰ IL-33 can also promote thickening of the epidermis in certain disease states.^8,62,63 However, results indicated no quantifiable differences in keratinocyte morphology or thickness of the epidermis and failed to reveal significant differences in infection-induced skin inflammation which argued against the first hypothesis. To test the second hypothesis, expression levels of 84 ECM and adhesion gene transcripts were quantified using a RT Profiler Array. IL-33 deficiency did lead to dysregulation of several ECM and adherens junction genes known to promote skin barrier integrity^64–67 and genes that decrease barrier integrity.⁶⁸  Ctnna2 (an α-catenin gene)⁶⁹ was not upregulated in the absence of either IL-33 or γδ T cells. As such, the changes in the epidermis that may have contributed to protective immunity against parasite entry were unlikely a thickened epidermal barrier, but rather a tighter barrier that perhaps slowed burrowing parasites attempting to enter the host. Further investigation into the skin barrier integrity of the protected skin will be required to uncover what specific aspects of IL-33 or γδ T cell biology drive resistance to migratory iL3.

While IL-33 is canonically thought to be released from damaged and dying epithelial and endothelial cells in helminth infection, IL-33 is also expressed by CD11c⁺ APCs.^2,12 Previous reports suggest the function of IL-33 depends on its cellular source.² Data show that CD301⁺ APCs in the skin were required for resistance to penetration and accumulation of IL-33–responsive skin ST2⁺ γδ T cells. Indeed, specific ablation of IL-33 from APCs impaired acquired resistance to larval penetration. Furthermore, APC-derived IL-33 was necessary for the accumulation of CD62L⁺ ST2⁺ γδ T cells in the hind paw skin. These data indicate a previously unknown skin-immune axis in which APC-derived IL-33 drives expansion/accumulation of CD62L⁺ ST2⁺ γδ T cells for impeding large metazoan pathogens from entering the host. Very few investigations have shown a role for IL-33 in γδ T cell biology⁷⁰ and to our knowledge none have investigated how IL-33 may change γδ T cell functions in the skin. We speculate that the CD62L⁺ ST2⁺ γδ T cells could be important mediators of skin alterations that confer resistance to parasite penetration, but this topic requires further investigation.

Herein, results support a previously unappreciated mechanism of resistance to percutaneous infection specifically in the hind paw skin. Resistance is driven by APC-derived IL-33, which drives γδ T cell accumulation and expansion to prevent helminth infection. The demonstration that myeloid APCs serve as a biologically important source of IL-33 further supports the emerging view that hematopoietic cells are an important source of this pleiotropic cytokine and substantiates further investigation in diverse tissue-specific contexts. Moreover, this work emphasizes the importance of employing natural modes of pathogen infection to uncover biological mechanisms pertinent to human disease. As such, further investigation into the importance of IL-33 in γδ T cell biology is warranted.

Supplementary material

Supplementary material is available at The Journal of Immunology online.

Funding

This material is based upon work supported by the National Science Foundation Graduate Research Fellowship Program under Grant No. DGE-2236662 awarded to Erin Evonne Jean. Any opinions, findings, conclusions, or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation. Other funding sources that supported this work include National Institutes of Health grants awarded to D.R.H. (U01AI163062, R01AI164715) and to both De’Broski R. Herbert and Heather Lynn Rossi (R21AI171740).

Conflicts of interest

None declared.

Data availability

Data available upon request.