An IL-17F.S65L Knock-In Mouse Reveals Similarities and Differences in IL-17F Function in Oral Candidiasis: A New Tool to Understand IL-17F Free

Fungal infections have a serious impact on human health, yet typically receive less attention than other pathogens (1). Candida albicans is a commensal fungus in humans, colonizing mucosal surfaces such as the oral cavity, vaginal tract, and gut as well as skin. Although normally benign in healthy individuals, C. albicans can cause pathogenic infections such as oropharyngeal candidiasis (OPC; oral thrush) under settings of immunodeficiency. Immunity to C. albicans is highly dependent on CD4⁺ T cells, as up to 95% of HIV⁺ patients suffer from recurrent OPC during progression to AIDS, correlating with T cell counts (2, 3).

Human T cell responses to C. albicans are dominated by Th17 cells (4–7). IL-17A (IL-17) is the signature cytokine of Th17 cells and is also expressed by various subsets of innate lymphocytes, such as γδ-T, NKT, and group 3 innate lymphoid cells (ILC3s) (8). IL-17 plays a critical role in antifungal immunity to C. albicans (9, 10). Studies in mice demonstrated the importance of IL-17 signaling against C. albicans oral infections, as mice lacking IL-17RA, IL-17RC, or the signaling adaptor Act1 are all highly sensitive to systemic, oral, and dermal candidiasis (11–15). Clinical data from cohorts of chronic mucocutaneous candidiasis disease (CMCD) patients confirmed that susceptibility to candidiasis is strongly associated with IL-17/Th17 deficiencies. Mutations in IL-17RA, IL-17RC, or Act1 in humans, as in mice, cause susceptibility to mucocutaneous candidiasis (16–19). Clinical treatment with anti–IL-17A Abs for autoimmunity is also linked to OPC (20). Similarly, STAT3, RORC, and STAT1 mutations are linked to reduced Th17 frequencies and CMCD (21–23).

The IL-17 family of cytokines is structurally distinct from other cytokine subclasses, composed of IL-17A, IL-17B, IL-17C, IL-17D, IL-17E (IL-25), and IL-17F (24). Among these, IL-17A shares the most homology with IL-17F at the amino acid level (56%) (25, 26). Both IL-17A and IL-17F form homodimers but can also form a heterodimer (IL-17AF) (27–29). All three isoforms signal through the IL-17RA:IL-17RC receptor complex, although the ligands have different binding affinities for the receptor, with IL-17A > IL-17AF > IL-17F (28, 30, 31). Recent data suggest that IL-17A may also signal through an IL-17RA:IL-17RD receptor (32) and that an IL-17C homodimer may also influence signaling (31, 33). Hence, our understanding of the IL-17 cytokine family is still developing.

IL-17A and, to a lesser extent, IL-17F activate a program of inflammation in target cells, mainly fibroblasts and epithelial cell types (34, 35). IL-17A induces a characteristic gene signature that includes cytokines (IL-6, GM-CSF, G-CSF), chemokines (CXCL1, CXCL2, CXCL8, CCL2, CCL7, CCL20), matrix metalloproteinases (MMP1, 2, 3, 9, 13), transcriptional and posttranscriptional regulators (C/EBPs, IκBξ, Regnase-1, Arid5a), and antimicrobial peptides (β-defensins, S100A proteins, lipocalin 2) (36–40). Emerging studies also implicate IL-17A in metabolic and proliferation gene expression (41, 42). Although less well characterized, IL-17F induces a similar but not entirely overlapping panel of genes (43–46). Studies of the in vitro functions of the IL-17AF heterodimer are comparatively limited, but studies to date indicate a similar gene induction profile induced by IL-17AF compared with IL-17A and IL-17F (27, 28).

Despite their capacity to bind the same receptor, IL-17A and IL-17F exert distinct activities in vivo. Il17a^−/− and Il17f^−/− mice show differential susceptibilities to various diseases, both infectious and autoimmune (47, 48). This dichotomy is illustrated in OPC; whereas Il17a^−/− mice or wild-type mice treated with IL-17A–neutralizing Abs exhibit elevated susceptibility to OPC compared with immunocompetent control mice (49), Il17f^−/− or mice treated with IL-17F–neutralizing Abs are fully resistant to C. albicans infection. However, there is evidence that IL-17F does participate in immunity to OPC, as dual blockade of IL-17A and IL-17F increases susceptibility to OPC over blockade of IL-17A alone (49, 50). In line with this, humans with autoimmune polyendocrinopathy-candidiasis-ectodermal dystrophy caused by AIRE gene deficiencies have circulating autoantibodies against both IL-17A and IL-17F, which is thought to underlie susceptibility to CMCD (51–53).

The physiological role of IL-17AF is still unclear. Although IL-17AF can be reliably detected (e.g., by sandwich ELISA), there are no commercial Abs that block this isoform selectively or efficiently, and hence, determining its specific role in vivo is challenging. Hints regarding IL-17F and IL-17AF function came from an unusual cohort of rare CMCD patients that harbor a heterozygous serine-to-leucine mutation in the IL17F gene at position 65 (IL-17F.S65L) (16). The mutated residue in these individuals lies within the region of interaction of IL-17F with the IL-17R. In vitro, this mutation impaired IL-17F binding to the receptor but had no apparent impact on the ability of IL-17F to dimerize with IL-17F or IL-17A. Cultured human fibroblasts treated with an IL-17F homodimer containing this mutation (IL-17F.S65L/IL-17F or IL-17S65L/IL-17F.S65L) or a mutant IL-17AF heterodimer (IL-17A/IL-17F.S65L) showed strongly impaired signaling in vitro (16). Accordingly, the S65L mutation in human IL-17F appeared to be both a loss-of-function and a dominant-negative mutation, causing functional blockade of IL-17F and IL-17AF leading to CMCD. Importantly, IL-17A homodimers were still found in these patients yet were apparently insufficient to fully protect from candidiasis. These data thus implied that IL-17F and/or IL-17AF are significant contributors to the antifungal immune response in humans.

In this study, we created an IL-17F.S65L mouse strain using CRISPR/Cas9 technology to exploit this “experiment of nature” (16) to better understand the functions of IL-17F and IL-17AF in immune responses. There was no detectable disease susceptibility in mice heterozygous for the mutation, suggesting differences between mouse and human IL-17F function in vivo. However, Il17f^S65L/S65L mice exhibited a similar susceptibility to OPC as Il17a^−/− mice, which contrasted with the previously known resistance of Il17f^−/− mice to OPC (49). Thus, this mutation appears to be more than simply a loss-of-function mutation. The increased susceptibility of these mice to fungal infection was linked to impaired expression of CXC chemokines and concomitantly reduced neutrophil recruitment to the oral mucosa but surprisingly not to expression of key antimicrobial peptides known to control OPC, such as β-defensin-3. As these mice do not phenocopy Il17f^−/− mice, they may provide a valuable tool to interrogate IL-17F and IL-17AF in vivo.

Il17f^S65L/S65L mice were created by CRISPR/Cas9 by the Transgenic and Gene Targeting and Innovative Technologies Development Core facilities in the Department of Immunology, University of Pittsburgh. Briefly, a S.py. Cas9 target sequence overlapping the Ser⁶⁵ codon in the mature Il17f sequence (following signal peptide cleavage) was selected: 5′-GTTCCCCTCAGAGATCGCTG AGG-3′. The protospacer adjacent motif in bold was not included in the single guide RNA (sgRNA). Cas9 mRNA and the sgRNA were produced as described (54, 55). A 127-mer oligonucleotide (Ultramer; Integrated DNA Technologies) was used as template for homology-directed repair: Il17f-S65L-HDRv3: 5′-CATCCTGCTTTACTTTTTATTTTTTTCCTTCAGCATCACTCGAGACCCCCACCGGTTCCCTCTAGAAATCGCTGAGGCCCAGTGCAGACACTCAGGCTGCATCAATGCCCAGGGTCAGGAAGACAGC-3′. The oligonucleotide contains a substitution to convert serine 65 to leucine, which contemporaneously disrupts an HPY188I restriction site (underlined) to facilitate genotyping. The sequence also contains an additional silent substitution, bringing a total of four mismatches (bold) between the sgRNA target sequence and the designed allele, thus limiting further re-editing of the mutant allele by Cas9. C57BL/6J embryos were microinjected with the sgRNA (50 ng/μl), the homology-directed repair oligonucleotide (0.5 μM), and Cas9 mRNA (100 ng/μl). Embryos that developed to the two-cell stage were transferred into pseudopregnant female surrogates. Sixteen founders were identified by PCR amplification of the target region (forward: 5′-ATGGGAGAAACCCCGTTTTA-3′; reverse: 5′-TCCAACCTGAAGGAATTAGAACA-3′), followed by restriction digestion of the PCR product with HPY188I. The correct sequence was validated by Sanger sequencing of the PCR products following TOPO cloning (Invitrogen). Of these, five founders were homozygous for the mutation (Supplemental Fig. 2). Based on the CRISPOR Web site (56), there were only two potential off-target sequences with fewer than four mismatches in the mouse genome (Table I). Among all founders analyzed, there were no mutations introduced at those sites, as determined by PCR and Sanger sequencing. The expanded line was backcrossed twice to C57BL/6J mice prior to colony expansion.

Il17f^Thy1.1 reporter mice were previously described (57). Il17ra^−/− mice were a gift from Amgen. Act1^−/− mice were from U. Siebenlist, National Institutes of Health. Wild-type (WT) mice were from The Jackson Laboratory, Taconic Farms, or generated from breeding colonies. All mice were on the C57BL/6 background. Age-matched mice (6–10 wk) were used for experiments with both sexes. Animal use protocols were approved by the University of Pittsburgh Institutional Animal Care and Use Committee.

OPC was induced by sublingual inoculation with a cotton ball saturated in C. albicans (CAF2-1) for 75 min under anesthesia, as described (58). Tongue homogenates were prepared on a gentleMACS (Miltenyi Biotec), and fungal burdens were determined by plating serial dilutions on yeast extract peptone dextrose agar with ampicillin. Anti–IL-17A or isotype control Abs (BioXCell) (200 μg/μl) were administered i.p. on days −1, 1, and 3 relative to C. albicans infection.

Tongues were digested with Collagenase IV (0.7 mg/ml) in HBSS for 30 min. Cell suspensions were separated by Percoll gradient centrifugation. Abs were from eBioscience, BD Biosciences, and BioLegend, as described (59). Data were acquired on an LSR Fortessa and analyzed with FlowJo (Ashland, OR).

Tongues were homogenized in a Gentle MACS Dissociator, and tongue RNA was extracted using RNeasy kits (Qiagen). RNA from ST2 cells was extracted in RLT buffer (Qiagen). cDNA was generated using a SuperScript III First Strand Synthesis System (Invitrogen). Relative quantification of mRNA expression was determined by real-time PCR with SYBR green (Quanta BioSciences) normalized to Gapdh on the Applied Biosystems 7300 platform. Primers were from QuantiTect (Qiagen).

The ST2 stromal cell line was cultured in α-MEM (Sigma-Aldrich, St. Louis, MO) with l-glutamine, antibiotics, and 10% FBS. A total of 5 × 10⁴ cells were seeded into six-well plates prior to cytokine stimulation. rIL-17F.S65L and IL-17F were synthesized by Bon Opus Biosciences (Millburn, NJ) by expression in Expi293 cells (Thermo Fisher Scientific). Mouse TNF-α was from Peprotech and used at 2 ng/ml (Rocky Hill, NJ).

Naive splenic CD4⁺ T cells were purified by magnetic separation (Miltenyi Biotec). T cells were activated with anti-CD28 (5 μg/ml; BioXCell) and plate-bound anti-CD3 (clone 145-TC11, 5 μg/ml; BioXCell) in RPMI 1640 supplemented with 10% FBS, 2 mM l-glutamine, 100 U/ml penicillin, 100 mg/ml streptomycin, 50 mM 2-ME, and sodium pyruvate for 4 d with IL-1β (50 ng/ml), IL-6 (50 ng/ml), IL-23 (50 ng/ml), and TGF-β (5 ng/ml). Cytokines were from R&D Systems.

Data were analyzed on Prism (GraphPad) using ANOVA or Student t test. Fungal burdens are presented as geometric mean analyzed by ANOVA with Mann-Whitney analysis. The p values are as follows: *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

The S65 residue in IL-17F that is mutated to leucine in humans with CMCD is conserved among many species, including mice (16). The human IL-17F.S65L cytokine has no signaling capacity in vitro (16). To determine whether the mouse ortholog of IL-17F.S65L functions similarly to its human counterpart, a His-tagged murine IL-17F.S65L and a corresponding WT IL-17F control were expressed in Expi293 cells and purified from conditioned media on a nickel-charged affinity resin (Ni-NTA) (Fig. 1A). rIL-17F.S65L and IL-17F migrated according to their predicted dimeric sizes on nonreducing SDS-PAGE (Fig. 1B). The multiple bands in the reduced and nonreduced gels are consistent with the expectation that these recombinant cytokines exist in multiple glycosylated and nonglycosylated forms (Fig. 1B, Supplemental Fig. 1).

To evaluate the signaling capability of murine IL-17F.S65L, murine ST2 stromal cells were treated with increasing concentrations of IL-17F.S65L or WT control IL-17F for 3 h. Two IL-17F–inducible genes, Il6 and Cxcl1, were assessed by quantitative PCR (qPCR) as endpoints of signaling. As expected, control IL-17F induced Il6 and Cxcl1 in a dose-dependent manner (Fig. 1C). In contrast, IL-17F.S65L at most doses (6.25–50 ng/ml) did not detectably upregulate Il6 or Cxcl1. However, both mRNAs were slightly enhanced by IL-17F.S65L at a supraphysiological dose (100 ng/ml) but nonetheless showed significantly reduced activity compared with nonmutated IL-17F (Fig. 1C). Because IL-17F induces gene expression synergistically with other cytokines (43, 46), we evaluated IL-17F.S65L responsiveness in concert with a suboptimal dose of TNF-α (2 ng/ml). Control IL-17F synergized with TNF-α to induce Il6 and Cxcl1, but the IL-17F.S65L mutant did not synergistically enhance expression of these genes (Fig. 1D). Thus, analogous to the IL-17F.S65L mutation–described CMCD patients, the mouse IL-17F.S65L homodimer appears to be a loss-of-function mutation.

To determine the function of the murine IL-17F.S65L mutation in vivo, we created an IL-17F.S65L mutant mouse strain by CRISPR/Cas9 technology in the C57BL/6 background. Sixteen founders were generated, of which five had a homozygous nucleotide replacement (Fig. 2A, Supplemental Fig. 2). None of the founder lines were found to have mutations in the either of the two predicted off-target sites, ascertained by PCR of genomic DNA and sequencing (Table I, see 2Materials and Methods section for details). Similar to IL-17a^−/− and Il17f^−/− mice (47), there were no obvious abnormalities in the health of these mice when maintained in specific pathogen-free conditions, and they bred normally with Mendelian numbers of offspring.

In the one family of humans with this mutation identified, five out of seven affected individuals experienced CMCD as heterozygotes [no homozygote family members were described (16)]. Based on this, we predicted that mice that are homozygous or heterozygous for the IL-17F.S65L mutation would be susceptible to mucosal candidiasis. To test this hypothesis, mice were subjected to OPC by a 75-min sublingual exposure to C. albicans (strain CAF2-1) by standard methods (58, 60). In this model, WT mice typically clear C. albicans from the oral mucosa within 3 d and show no overt signs of illness, such as prolonged weight loss (12, 61). We previously demonstrated that mice lacking IL-17RA or Act1 maintain a high fungal burden in the oral cavity postinfection (p.i.), which we typically measure at day 5 p.i. (12, 14). Mice lacking IL-17A (Il17a^−/− or given anti–IL-17A Abs) have detectable fungal burdens, although consistently lower than Il17ra^−/− or Act1^−/− mice (49). In contrast, mice lacking IL-17F, either by knockout or with neutralizing Abs, are fully resistant to OPC, at least within the detection limits of this system (49). In this study, we observed that Il17f^S65L/S65L mice had a significantly higher oral fungal burden compared with WT controls at 5 d p.i., at levels similar to Il17a^−/− animals (Fig. 2B). Notably, however, the Il17f^+/S65L heterozygous mice did not have elevated fungal loads compared with WT. Approximately 40% of the Il17f^S65L/S65L and Il17a^−/− mice still had a detectable fungal load at this time point, whereas the WT and the Il17^/S65L/+ heterozygous mice almost all fully cleared the infection (Fig. 2C). Il17f^S65L/S65L mice also lost slightly more weight than WT controls, which was most evident at day 2 p.i. (Fig. 2D, 2E). However, oral fungal burdens in Il17f^S65L/S65L mice were not measurably different at day 2 (Fig. 2F). Collectively, these results indicate that IL-17F.S65L mutation contributes detectably, albeit modestly, to susceptibility to OPC but only when the mutation is present on both alleles.

To understand the mechanisms by which the IL-17F.S65L mutation promotes susceptibility to OPC, we evaluated factors known to be critical for antifungal immunity mediated by IL-17R signaling (12, 15, 61). Neutrophils are vital for fungal clearance in OPC (62–64). We have observed that Il17ra^−/− mice show impaired recruitment of neutrophils to the oral mucosa following OPC induction (12, 15, 59). IL-17F upregulates expression of neutrophil-attracting chemokines, such as CXCL1 and CXCL2 (65–67), and both human and murine IL-17F.S65L showed impaired induction of this chemokine in fibroblasts in vitro [Fig. 1 (16)]. Consistent with this, Cxcl1 gene expression in the tongue was downregulated in Il17f^S65L/S65L mice upon C. albicans oral infection, although not completely impaired. However, expression of Cxcl2 was similar between Il17f^S65L/S65L mice and WT controls (Fig. 3A). Flow cytometry revealed that the percentage of neutrophils recruited to the tongue at day 2 p.i. was partly, although by no means fully, decreased in infected Il17f^S65L/S65L mice compared with WT (45 versus 57%) (Fig. 3B). There was a trend to reduced total numbers of neutrophils, although there was high variability in the number of cells recovered from the tongue [a common problem often encountered when isolating cells from this tissue (68, 69)]. Thus, although not proven directly, impaired neutrophil recruitment in Il17f^S65/S65L mice may help explain OPC susceptibility in these mice.

Antimicrobial peptides such as β-defensins and calprotectin (S100A8/S100A9) have antifungal activity toward C. albicans (70–75). IL-17RA knockout mice show impaired antimicrobial peptide expression after C. albicans oral challenge (12, 61). Surprisingly, the induction of Defb3 and S100a9 in the oral mucosa of Il17f^S65L/S65L mice was equivalent to WT upon oral C. albicans infection (Fig. 3C). Therefore, the increased OPC susceptibility in Il17f^S65L/S65L mice is apparently not due to inefficient antimicrobial peptide expression. Moreover, these data suggest that IL-17F and/or IL-17AF signaling may be more critical for the neutrophil response than for induction of antimicrobial peptides.

IL-17A mediates host defense activity in OPC (49, 76). To determine if IL-17A expression was impacted in Il17f^S65L/S65L mice, we analyzed Il17a mRNA from tongue at day 2 p.i. Interestingly, Il17f^S65L/S65L mice showed elevated Il17a as well as Il17f (Fig. 4A), arguing that the higher fungal susceptibility in Il17f^S65L/S65L mice is likely not due to impaired Il17a expression; rather, the mildness of the disease susceptibility in these mice could be due to compensatory IL-17A levels. To determine whether this increased Il17a expression is caused by the IL-17F.S65L mutation or is a result of fungal infection in the oral mucosa, we cultured splenic CD4⁺ T cells for 4 d in Th17 conditions (IL-6, IL-23, TGF-β). IL-17A concentrations in supernatant were measured by ELISA. As shown, there was no significant difference in the amount of IL-17A produced by T cells obtained from Il17f^S65L/S65L, Il17f^−/−, or WT mice (Fig. 4B). Thus, IL-17F.S65L does not appear to directly influence IL-17A production from T cells.

To determine if the elevated IL-17A in the tongue could compensate for IL-17F in Il17f^S65L/S65L mice, we treated mice with neutralizing Abs against IL-17A (76) during OPC induction. In WT mice, blockade of IL-17A caused increased fungal burdens and an increased percentage of mice with fungal loads, measured at day 5 p.i., as previously demonstrated (Fig. 5) (49). However, IL-17A blockade in Il17f^S65L/S65L mice did not further increase the oral fungal load compared with Il17f^S65L/S65L isotype control mice or WT mice treated with anti-IL-17A Abs (Fig. 5). Thus, in contrast to findings in Il17f^−/− mice (49), IL-17A does not compensate for the IL-17F.S65L mutation during OPC.

Unlike humans, mice do not harbor C. albicans as a commensal microbe. Multiple studies have verified that the initial response to oral infection with this organism during this first encounter in mice derives entirely from the innate immune compartment (50, 69, 77–80). The sources of IL-17A during acute OPC were previously shown to be dominantly from an unconventional, innate-acting TCRαβ⁺ cell population and γδ-T cells (69, 77). ILC3s were reported to produce IL-17A as well (50). Using Il17f^Thy1.1 reporter mice (57), we observed an increased level of Il17f-expressing cells 2 d after C. albicans infection (Fig. 6A), the time point at which Il17f mRNA expression peaks (49). γδ-T cells constituted the major Thy1.1⁺ population (64%), and TCRβ⁺ cells also comprised a significant portion of the Thy1.1⁺ cells (21%). A population of TCRγδ-negative TCRβ-negative cells (15.5%) that may be ILC3s was also observed (Fig. 6B).

C. albicans asymptomatically colonizes healthy individuals and usually only causes mucocutaneous infections in immunocompromised individuals (2, 81). Deficits in IL-17 signaling or Th17 cell development are particularly linked to superficial C. albicans infections (21, 82). Although CMCD patients with null mutations in genes encoding IL-17RA, IL-17RC, or the adaptor Act1 have all been described, thus far, no CMCD patients have been reported with a single IL-17A deficiency (22). Even anti–IL-17A biologics used to treat autoimmune disease cause only a modest increase in mucocutaneous Candida infections (20, 83).

In general, observations in mouse studies of OPC have been accurate predictors of the immune correlates in oral candidiasis, especially with regards to the IL-17 axis (12, 15, 84, 85). Mice treated with IL-17A–neutralizing Abs have lower fungal burdens compared with the Il17ra^−/− mice during OPC (49, 76), which parallels the low percentage of patients who only experience mild C. albicans mucocutaneous infections during secukinumab treatment (20, 83, 86). Unlike humans (5, 87), laboratory mice do not harbor C. albicans as a commensal organism, and hence the acute OPC model system reflects events in the innate response (69, 77, 79, 88). Nonetheless, mice do generate potent and protective recall Th17 responses to C. albicans after a secondary encounter, similar to humans (78, 79, 89). Humans with STAT3 mutations experience CMC (21), and although STAT3 is not required in CD4⁺ cells to protect from acute OPC, we recently found that STAT3 is essential in the oral epithelium (59, 77). Mice also express different antimicrobial peptide proteins than humans, especially in saliva (90, 91). Of relevance to C. albicans infections, the antimicrobial peptide β-defensin 3 is essential to prevent OPC in mice (61), although it has no direct ortholog in humans (92, 93).

Although humans with the IL-17F.S65L mutation were described only in a single kindred, the degree to which the phenotype of Il17f^S65L/S65L knock-in mouse is similar or different from humans is a matter of interpretation. Whereas mice with a complete IL-17F deficiency do not recapitulate the CMC phenotype found in humans with the IL-17F.S65L mutation, mice with the analogous IL-17F mutation do exhibit OPC in the homozygous state, pointing to similarities among species. Puel et al. (16) reported that 70% of the individuals carrying the IL-17F.S65L mutation (5 out of 7) had a confirmed diagnosis of mild CMCD. In contrast, almost all the Il17f^S65L/+ mice fully cleared C. albicans from the mouth and were statistically indistinguishable from WT controls. However, susceptibility was seen in mice carrying the mutation on both alleles, contrasting with the lack of susceptibility previously documented in Il17f^−/− mice (49). This observation could be interpreted to mean there is no dominant-negative activity of this mutation in mice. However, the CFU enumeration system used to quantify OPC probably underestimates the actual number of infectious fungi because C. albicans hyphae are not easily separable into single cells, and hence, each colony may represent more than a single organism (94). The finding that Il17f^S65L/S65L mice showed fungal susceptibility that was closer to Il17a^−/− mice than to Il17f^−/− mice may imply there is some level of dominant-negative activity that OPC is not a sufficiently sensitive system to detect (49). Additionally, there is one other report of a human IL-17F mutation, but unfortunately the nature of the mutation was not provided, and so no functional inferences can be made (95).

Blocking IL-17A together with IL-17F increases susceptibility compared with blockade of IL-17A alone (49, 50), which is reminiscent of autoimmune polyendocrinopathy-candidiasis-ectodermal dystrophy (AIRE deficiency) in which patients have neutralizing Abs against multiple Th17 cytokines, including IL-17A and IL-17F (52, 53, 96). However, administration of IL-17A–neutralizing Abs [which also efficiently block IL-17AF (49)] in Il17f^S65L/S65L mice did not further increase susceptibility to OPC, suggesting that residual IL-17A homodimers present in these animals do not provide additional detectable protection. Clearly there is more to learn about IL-17F and antifungal immunity.

Taken together, the above results are consistent with a protective role for the heterodimer IL-17AF in OPC, which has been challenging to determine because of a paucity of reagents to study the heterodimer. Studies in human fibroblasts showed that the IL-17F.S65L mutation impaired signaling of both the IL-17F homodimer and the IL-17AF heterodimer (16). The increased susceptibility to OPC in Il17f^S65L/S65L mice could be due either to a contribution of IL-17F homodimer and/or the IL-17AF heterodimer. Nonetheless, anti–IL-17F neutralization did not cause an increase in fungal loads in WT animals (49), suggesting that a deficiency of the IL-17AF heterodimer rather than (or in addition to) IL-17F could be responsible for increased C. albicans infection in Il17f^S65LS66L mice. A protective role of IL-17AF could also explain why IL-17A blockade failed to further promote a fungal burden in Il17f^S65LS65L mice. If IL-17AF is indeed the primary effector cytokine among the three IL-17RA/IL-17RC receptor ligands, the increased susceptibility caused by anti–IL-17A treatment could be due to its capacity to block IL-17AF rather than the IL-17A homodimer, as generally assumed (49). Because Il17f^S65L/S65L mice have a fully impaired IL-17AF signaling pathway, this could explain why anti–IL-17A Ab treatment did not lead to a higher fungal burden than isotype control Abs.

An unexpected observation made in these studies was that the IL-17F.S65L mutation affected CXCL1 mRNA expression and subsequent neutrophil recruitment, albeit modestly, yet had no detectable impact on expression of key antifungal antimicrobial peptides, β-defensin 3, and S100A8/A9 (calprotectin). IL-17 is a potent regulator of the neutrophil axis, acting on target epithelial cells to induce chemokines that in turn recruit myeloid cells, especially CXCR2-expressing neutrophils (62). It is unclear if the moderate suppression on neutrophils observed in this study accounts for the susceptibility phenotype. In fact, not all studies of OPC have found that IL-17 regulates neutrophil infiltration during OPC (97); the reason for the discrepancy among laboratories is unclear but could relate to effects of different local oral microbiota (98, 99).

Although the IL-17RA/IL-17RC heterodimer is the canonical receptor complex thought to transduce signaling of IL-17A, IL-17F, and IL-17AF (100, 101), recently, several alternative configurations of the receptor have been proposed. IL-17RD was suggested to act in concert with IL-17RA to mediate IL-17A but not IL-17F signaling in keratinocytes to drive psoriasis-like skin inflammation (32). Thus far, the binding capacity of IL-17AF to an IL-17RA/RD receptor complex has not been characterized. In the OPC model, Il17rc^−/− mice phenocopy Il17ra^−/− mice in terms of fungal loads and other signs of disease (15), suggesting that IL-17RC is needed to mediate immunity. Il17rc^−/− mice are also susceptible to dermal candidiasis (102). In agreement with results in mice, humans with IL-17RC null mutations experience CMCD (18). Hence, it is unlikely that an IL-17RC–independent receptor complex mediates host defense during mucocutaneous candidiasis.

A recent structural analysis of the IL-17RC subunit unexpectedly revealed that IL-17F may have the ability to signal through an IL-17RC homodimeric receptor (31), although this would contrast with our prior molecular studies showing that a forced dimer of murine IL-17RC is not signaling competent (46). Interestingly, based on the human IL-17RC structure, the IL-17F.S65L mutation is predicted to cause steric hindrance that decreases its binding affinity to IL-17RC. However, the contacts of IL-17F to IL-17RA are less tight than to IL-17RC, so changes in binding affinity to this subunit would likely be less pronounced (31). Alignment of the IL-17RC residues that are proximal to IL-17F.S65 shows full conservation between the human and mouse receptors, arguing for strong overall structural conservation (J.M. Rondeau, personal communication). Putting this together, we speculate that the increased fungal burden in Il17f^S65LS65L mice is caused by reduced interactions of IL-17F with the IL-17RC homodimeric receptor, rather than with the IL-17RA/IL-17RC heterodimer.

In summary, results from this new IL-17F.S65L mutant mouse strain in the murine OPC model are generally, although not completely, in line with data in CMCD patients with the IL-17F.S65L mutation. These data help reconcile the observations that, whereas a complete IL-17F deficiency does not cause candidiasis in mice, this naturally occurring IL-17F mutation in humans does. The potential utility of this new mouse strain goes beyond studies of candidiasis. Because these mice show a different phenotype from Il17f^−/− mice with respect to OPC, they could be valuable as an additional tool to interrogate the functions of IL-17F as well as the enigmatic IL-17AF heterodimer in other settings in which these cytokines may participate, such as autoimmunity, cancer, or other infectious disease settings.

We are grateful to Casey Weaver (University of Alabama) for Il17f^Thy1.1 reporter mice and Ulrich Siebenlist (National Institutes of Health) for Act1^−/− mice. Il17ra^−/− mice were a kind gift from Amgen. We thank Chunming Bi of University of Pittsburgh Transgenic and Gene Targeting for expert technical help. Drs. Jean Michel Rondeau and Frank Kolbinger (Novartis) and Yufang Shao (Bon Opus) provided valuable discussions. We also thank Drs. Mark Shlomchik, Jean-Laurent Casanova, and Anne Puel for helpful input.

The authors have no financial conflicts of interest.