Th17 cells confer long term adaptive immunity to oral mucosal Candida albicans infections

Pre-exposure to Candida albicans promotes accelerated fungal clearance

To address the role of the adaptive response in OPC, we first evaluated susceptibility in Rag1−/− mice. WT (C57BL/6) or Rag1−/− mice were subjected to a 4-day sublingual infection with 2×10⁷ CFU C. albicans (6, 29). As controls, IL-23−/− or WT mice immunosuppressed with cortisone acetate were subjected to OPC. As reported previously, unmanipulated WT mice generally cleared the fungus by the end of the infection period, whereas IL-23−/− and cortisone-treated WT mice had high oral loads of Candida and concomitant weight loss (Fig 1A). Moreover, IL-23−/− mice failed to clear the infection for at least 3 weeks (Fig 1B). Rag1−/− mice exhibited fungal loads ~3 logs higher than WT, which were statistically indistinguishable from IL-23−/− mice, though the Rag1−/− mice did regain some weight towards the end of the experiment (Fig 1A). These results indicate that a rearranged antigen receptor is necessary for immunity to OPC, and further indicate that innate lymphoid-like cells (ILCs) such as those reported in certain forms of colitis do not appear to mediate immunity in this setting (30, 31).

Taking advantage of the fact that C. albicans is not a commensal microbe in mice, we developed a system that permits experimental separation of the effects of innate and adaptive responses (Fig 2A). WT mice were inoculated orally by means of a Candida-soaked cotton ball placed sublingually for 75 minutes. 6–8 weeks after primary exposure, mice were re-challenged orally with Candida. As controls, mice were sham-inoculated with PBS at one or both time points. On days 1, 2 and 3 following re-challenge, tongue was harvested to analyze fungal burden and cervical lymph nodes (cLN) were analyzed for cytokine production. Mice pre-exposed to C. albicans (1° + 2° infection) had significantly lower fungal burdens (~1 log) after secondary challenge than those receiving only a 1° acute infection (Fig 2B). By day 3, most mice had cleared the infection, though residual fungal CFUs were still higher in mice receiving only a primary compared to secondary challenge. Therefore, prior exposure to Candida induces more efficient pathogen clearance.

To evaluate the nature of the T cell compartment in Candida recall responses, draining cervical LN (cLN) cells were stained for CD4, IL-17 and IFNγ (Fig 2C–D). There were no CD4+IL-17+ cells in Sham-treated cLN. However, there was an increased frequency of CD4+IL-17+ cells in cLN of infected mice, comprising approximately 4% by day 3 post re-challenge. There was no detectable induction of CD4+IFNγ+ cells upon Candida infection, either during a first exposure or re-challenge, nor were there IFNγ+IL-17+ double-expressors evident in either setting. There was a small increase in CD4+IL-17+ cells in mice undergoing only a primary response. To evaluate the surface phenotype of the CD4+IL-17+ cells, total cLN cells isolated 3 days after re-challenge were gated on CD4, and surface expression of CD44 and CD27 were examined relative to intracellular IL-17. CD4+IL-17+ cells were primarily CD27-CD44+ (Fig 2E–F), consistent with an activated effector phenotype.

To determine whether Th17 cells have the capacity to protect from OPC, CD4+ or CD4- cells were isolated from cLNs of mice subjected to a recall response and transferred into susceptible Rag1−/− hosts (Supplementary Fig S1). After 24 h, mice were subjected to OPC. Mice receiving CD4+ but not CD4- cells showed complete protection from OPC (Fig 2G). Notably, il17a mRNA expression in tongue correlated with protection, indicative of a local Th17 response (Fig 2H). In contrast, there was no detectable ifng mRNA expression in the tongue following transfer of CD4+ cells (Fig 2H). These data suggest that Th17 cells comprise the dominant functional CD4 response upon Candida re-challenge.

The IL-17 recall response to OPC comes from the CD4+ not the CD8+ compartment

To evaluate the nature of the T cell response in more detail, cells isolated from cLN 3 days post re-challenge were stained for CD3, CD4 and CD8, followed by intracellular staining for IL-17 and IFNγ (Fig 3A–B). As shown, only mice subjected to re-challenge generated significant numbers of IL-17+ cells in the CD4+ compartment (Fig 3B, left). There was no increase in the frequency of CD8+ IL-17+ cells (Fig 3B, left). Additionally, there was no change in the number of CD4+IFNγ+ cells (Fig 3B, right). There was a small increase in IFNγ+ cells in the CD8+ compartment, (Fig 3B, right), but this was transient and not sustained at 5 days post-infection (Supplementary Fig S2A). There was a low baseline level of IL-17+ cells in the CD3-negative population, which did not change after infection (Supplementary Fig S2B).

The CD4+ IL-17 recall response to OPC is antigen-specific

To confirm that induction of cytokines is Candida-specific, cLN cells were harvested 3 days after re-challenge and stimulated in vitro with heat killed (HK) C. albicans for 5 days in vitro. Supernatants were evaluated for IL-17 and IFNγ by ELISA. Consistent with the intracellular cytokine staining (ICS) data (Fig 3A–B), high levels of IL-17 (~500 pg/ml) were produced in mice exposed to Candida during a secondary response, but not in Sham-infected mice and only modestly in 1° infected mice (Fig 3C, left). Production of IL-17 was blocked by anti-CD4 (32) and anti-MHC Class II Abs, indicating that antigen presentation is required. To further verify that IL-17 production is antigen-specific, LN cells from infected mice were cultured in vitro with a peptide derived from the Candida adhesins Als1 and Als3, recently shown to be an epitope recognized by CD4+T cells (33). Treatment with the ALs1/3 peptide (pALS) triggered IL-17 production, though not to the same magnitude as HK Candida, and was blocked by anti-MHC Class II Abs (Fig 3C). Therefore, the Th17 response is antigen-specific.

Unexpectedly. there was a modest but reproducible CD4-dependent induction of IFNγ following treatment with HK Candida in re-challenged mice, ranging from ~15–100 pg/ml, depending on experiment, Fig 3D. The induction of IFNγ-producing cells was slightly increased in re-challenged mice compared to mice receiving only a primary infection, and was blocked with anti-CD4 or anti-MHC II Abs. However, there was no detectable IFNγ production after in vitro challenge with pALS. Although inducible IFNγ following in vitro challenge contrasts with the apparent lack of a Th1 response in cells taken immediately ex vivo measured by flow cytometry, it is most likely that Th1 cells are induced during infection, but at levels below the detection limit of FACS. Together, these data indicate that the response to oral Candida infection is primarily IL-17-driven, though we cannot entirely rule out a role for IFNγ. Moreover, generation of a strong CD4+Th17 recall response promotes accelerated oral clearance of C. albicans.

Compensatory sources of IL-17 in CD4-deficient mice

To determine whether generation of CD4+Th17 cells is necessary for clearance of Candida from the oral cavity, WT or CD4−/− mice were subjected to re-challenge and evaluated for symptoms of OPC, including weight loss and oral fungal burden. CD4−/− mice were resistant to OPC, as they regained weight equivalently to WT mice after infection (Fig 4A) and had no detectable oral fungal burden (Fig 4A, inset). To evaluate the possibility that IL-17 might be produced from a compensatory source in CD4-deficient mice, cLN cells from WT and CD4−/− mice were analyzed for cytokines 5 days after re-challenge. In WT mice, CD4+ but not CD8+ T cells expressed IL-17 during re-challenge (Fig 4B). However, in re-challenged CD4−/− mice, CD8+ T cells produced abundant IL-17 (Fig 4B–C). The frequency of IFNγ-expressing CD8+ cells was similar in WT and CD4−/− mice (Fig 4C, right), further confirming that IFNγ responses do not seem to play a role in protection. Although somewhat variable, levels of IL-17 from the CD8-compartment in CD4−/− mice were usually comparable to that of the CD4 compartment in WT mice (Fig 4D left, and data not shown). IL-17 secretion from in vitro-stimulated cLN cells was observed only during 2° challenge (Fig 4D, left). There was a weak but significant increase in the IFNγ response as well (Fig 4D, right). There was a low frequency of IL-17-producing cells in the CD3-negative population in both WT and CD4−/− mice, which did not change following infection (Supplementary Fig S2).

Since the resistance of CD4−/− mice to C. albicans re-challenge correlates with compensatory IL-17 production by CD8+ T cells, we tested the susceptibility of CD8−/− mice to OPC. Like CD4−/− mice, CD8−/− mice were resistant to OPC upon re-challenge, based on weight gain after infection and the complete absence of an oral fungal load (Fig 5A). Consistently, equivalent numbers of CD4+IL-17+ cells were present in cLN of CD8−/− and WT mice subjected to re-challenge but not in cLN of mice that only received a 1° infection, whereas CD4+IFNγ+ cell frequencies were unchanged (Fig 5B). High levels of IL-17 were induced after in vitro culture with HK Candida (Fig 5C), which was blocked with anti-CD4 Abs (Fig 5C, left). Although the difference between IL-17 levels in the 1° and 1°+2° in vitro cultures did not reach statistical significance, this was likely due to a single outlier in the 1° sample, and the overall trend strongly resembled what was observed in the CD4 compartment in WT mice. There was induction of a weak IFNγ response in CD8−/− mice during both 1° and 2° responses (Fig 5C, right).

Functional compensation by Tc17 and DN cells in CD4−/− mice

We next determined whether the CD8+ “Tc17” cells from re-challenged CD4-deficient hosts were capable of mediating immunity to OPC. To that end, Rag1−/− mice were adoptively transferred with CD4+ cells, CD8+ cells or CD4-CD8- double negative (DN) cells derived from WT or CD4−/− donors (Supplementary Fig S1B). Cells from Sham-infected WT donors did not provide any protection from OPC. As expected, CD4+ cells but not CD8+ or DN cells from infected WT donors were protective. Also as predicted, CD8+ cells from infected CD4−/− mice provided detectable (although variable) protection from OPC. Unexpectedly, the DN population from CD4−/− mice were also protective, even more consistently than the CD8+ population (Fig 6A). Based on these observations, we evaluated IL-17 expression in the DN population from WT and CD4−/− mice. IL-17 was indeed expressed in the DN cells, but only in the context of CD4 deficiency (Fig 6B).

To verify that the transferred cell populations were mediating a local response in the oral mucosa, we performed qPCR on tongue mRNA samples. As shown, CD3ε expression in tongue correlated with protection from OPC, suggesting that the relevant effector populations were being recruited and/or expanding in situ (i.e., CD3+CD4+ cells from WT donors and CD3+CD8 and CD3+CD4-CD8- [DN] cells from CD4−/− donors) (Fig 6C). A similar observation was made for CD4 and CD8 mRNA expression, which correlated with the protective IL-17-producing populations (Fig 6C). Collectively, these results demonstrate that under normal circumstances, CD4+ Th17 cells and not CD8+ or DN cells mediate IL-17-dependent immunity during adaptive responses to OPC. However, in conditions of CD4-deficiency, CD8 “Tc17” cells and DN cells can serve as alternate sources of IL-17 that are able to promote effective immunity to Candida albicans. This finding illustrates the highly flexible nature of the anti-fungal immune response.

Mice

Mice were on the C57BL/6 background and were age- and sex-matched. IL-23p19−/− mice were provided by Genentech (S. San Francisco, CA) and bred in-house. All other mice were from The Jackson Laboratory (Bar Habor, ME). Protocols were approved by the University of Pittsburgh IACUC, and adhered to guidelines in the Guide for the Care and Use of Laboratory Animals of the NIH.

OPC recall infection model

Mice aged 5–6 weeks (minimum n=3 per cohort) were inoculated sublingually with a 0.0025g cotton ball saturated in a C. albicans suspension of 2 × 10⁷ CFU/ml for 75 mins under anesthesia (6, 29). Six weeks after the primary infection, mice were re-challenged by the same procedure. The oral cavity was swabbed immediately prior to each infection and streaked on YPD to verify the absence of Candida. 1–5 d post re-challenge, the tongue was homogenized on a Gentle MACS Dissociator (Miltenyi Biotec Auburn, CA) and plated on YPD-AMP for 48 h at 30°C followed by colony enumeration in triplicate. For cell transfer experiments, cLN cells were harvested from WT mice 3 days after re-challenge and pooled. CD4+ cells were isolated by positive selection using Miltenyi Biotec CD4 (L3T4) MicroBeads and LS columns. The resulting CD4+, CD4-, CD8+ or CD4-CD8- (DN) fractions were transferred into Rag1−/− recipients (stock number 002216) i.p., and 24 h later mice were infected orally as described above. A total of 10⁷ CD4+ cells and 2 × 10⁷ CD4- cells were transferred in 200µl PBS. For sham transfers, transferred cells were from total body LN. Data were analyzed on Graphpad Prism (v. 5) using t-tests with Mann-Whitney correction (p<0.05 considered significant).

Culture of C. albicans and stimulation in vitro

C. albicans (strain CAF2-1) was cultured in YPD at 30°C overnight with agitation. A heat-killed (HK) extract was prepared with 1 ml (~ 4×10⁸ cells) culture boiled for 45 mins. ALS1/3 peptide (KGLNDWNYPVSSESFSYT) (33) was obtained from Biosynthesis (Lewisville, TX). Cervical LN stimulation was performed by plating 10⁶ cells per well in the presence or absence of 2×10⁶ HK C. albicans or 5–10 µg/ml ALS1/3 peptide (pals) diluted in serum-free AIM V media (GIBCO, Invitrogen) with 20 U/ml IL-2 ± anti-CD4 Abs (BD Biosciences, clone RM4-5, 5µg/ml) or anti-MHC II (eBiosciences, clone M5/114.15.2, 10µg/ml). Supernatants were analyzed in duplicate for IL-17 and IFNγ by ELISA (eBiosciences).

Intracellular cytokine staining and flow cytometry

10⁶ cLN cells/well were rested overnight at 37°C in serum-free AIM V media + 20 U/ml IL-2. Cells were stimulated with 50 ng/ml PMA and 500 ng/ml Ionomycin (Sigma, St Louis MO) for 4 h in the presence of Golgi Plug (BD Biosciences, San Diego, CA). Following stimulation, cells were stained for CD3 APC-Cy7 (clone 145-2C11), CD4 FITC (clone GK1.5) and CD8 PerCP (clone 53–6.7). IL-17 and IFNγ were detected with anti-IL-17-PE and anti-IFNγ-APC. Activation markers were assessed with CD44 eFluor 450 (clone IM7) and CD27 APC (clone LG.7F9). Dead cell exclusion was performed using Live/Dead Violet Viability/Vitality Dye Kit (Molecular Probes/Invitrogen). ICS was performed with the Cytofix Cytoperm kit (BD Biosciences). Data was acquired on an LSR II and analyzed with Flowjo.

RNA extraction and real time RT-PCR

Frozen tongue was homogenized in RLT lysis buffer (RNAeasy Kit, Qiagen) with a GentleMACS Dissociator (M-tubes, RNA02 program, Miltenyi). RNA was extracted with RNAeasy Kit. 1µg total RNA was used to synthesize cDNA with a SuperScript III First-Strand Synthesis System (Invitrogen Carlsbad CA). Relative quantification of indicated genes was determined by real-time PCR with SYBR Green (Quanta BioSciences, Inc.) normalized to GAPDH. Primers were from SA Biosciences (Qiagen Valencia, CA). Results were analyzed on a 7300 Real Time PCR System (Applied Biosystems Carlsbad, CA).