Regulatory T Cell Induction and Retention in the Lungs Drives Suppression of Detrimental Type 2 Th Cells During Pulmonary Cryptococcal Infection Free

Cryptococcosis is an emerging infectious disease of humans caused by the fungus, Cryptococcus neoformans (1). Yeasts or spores are inhaled from the environment and enter the lower respiratory tract. Robust CD4⁺ Th cell–mediated immunity controls this initial pulmonary infection, and as a result, immune replete individuals rarely experience overt disease. However, Th cell deficiencies associated with solid-organ transplantation, cancer chemotherapy, and HIV/AIDS dramatically increases susceptibility to invasive cryptococcosis. C. neoformans emigrates from the lung, enters the bloodstream, and traverses the blood–brain barrier to cause cryptococcal meningitis. Despite access to standard antifungal and antiretroviral therapies, patients receiving treatment for cryptococcal meningitis exhibit a wide range of adverse clinical outcomes, including: infection relapse, immune reconstitution inflammatory syndrome due to excessive reaction to persistent Ag, and/or death (2). The reasons why some individuals recover without experiencing complications and others perish remains enigmatic.

Differences in the quality of the impaired Th cell responses in patients with HIV stratify the spectrum of clinical outcomes (3–5). IFN-γ production by type 1 Th1 cells defends against invasive cryptococcal disease and promotes fungal clearance (6–8). Dysregulated reconstitution of protective immunity in patients with recent cryptococcosis can also cause harmful inflammation (9). In addition, C. neoformans (serotype A) subverts protective immunity and exacerbates disease by driving Th2 cell production of IL IL-4, IL-5, and IL-13 (10, 11). Therefore, therapies that dampen detrimental Th cell responses could be used to ameliorate disease.

One mechanism the immune system uses to dampen Th cell responses is by employing regulatory T (Treg) cells. Treg cells are a distinct subset of Th cells that uniquely express the transcription factor Foxp3, which stabilizes the suppressive function of Treg cells. Genetic aberrations in Foxp3 (i.e., IPEX syndrome) cause fatal Th cell–driven autoimmunity in humans, highlighting the importance of Foxp3 in immune homeostasis (12). Treg cells also inhibit effector Th cell responses to microbial infections (13). In particular, conditional depletion of Foxp3⁺ Treg cells in mice infected with C. neoformans increases Th2 cell abundance in the lungs, indicating Treg cells limit the proliferation of Th2 cells primed by cryptococcal infection (14, 15). Beyond these initial observations, little is known about the mechanism of Th2 cell suppression by Treg cells during cryptococcal infection.

Because Treg cell suppression of effector Th cells is contact dependent (16), Treg cells must colocalize with effector cells to function in tissues such as the lung (17). To accomplish this, Treg cells express chemokine receptors and integrins that allow them to home to and be retained at sites of inflammation (18). Separate evidence indicates Treg cells that restrain mucosal Th cell responses exhibit highly specialized control of distinct Th cell subsets by expressing the same lineage-defining transcription factors as their effector Th cell counterpart (19–21). In particular, IFN regulatory factor 4 (IRF4) expression by Treg cells has been implicated in the suppression of Th2-driven autoimmunity (21).

In this study, we used a mouse model of experimental cryptococcosis to investigate Treg cell responses to pulmonary fungal infection. Specifically, we explored the hypothesis that Treg cells use IRF4 and chemokine receptors to colocalize with Th2 cells in the lungs. While in proximity with Th2 cells, Treg cells are able to inhibit the expansion of deleterious Th2 cell responses to cryptococcal infection.

All mice used in this study were derived from a C57BL/6 background. B6.129P2-Ccr5^tm1Kuz/J (22), B6.129P2(C)-Ccr7^tm1Rfor/J (23), B6.Cg-Foxp3^tm2Tch/J (24), B6.129(Cg)-Foxp3^{tm3(DTR/GFP)Ayr}/J (25), Foxp3^{tm9(EGFP/cre/ERT2)Ayr}/J (26), B6.129S1-Irf4^tm1Rdf/J (27), B6.Cg-Gt(ROSA)26Sor^{tm14(CAG-tdTomato)Hze}/J, and B6.PL-Thy1^a/CyJ mice were purchased from The Jackson Laboratory (Bar Harbor, ME). Foxp3-cre/GFP mice were a kind gift from Calvin Williams (28). Foxp3-eGFP mice were crossed with Thy1.1 mice to generate congenic marked mice for transfer experiments. Foxp3-cre estrogen receptor 2 (ERT2) mice were crossed with tdTomato mice for Treg fate-mapping studies. All mice were housed in specific pathogen-free conditions.

C. neoformans var. grubii strain KN99α was streaked on yeast peptone dextrose agar plates and incubated for 2 d at 30°C. Yeast peptone dextrose broth was inoculated with colonies from the aforementioned plate and incubated for 16 h at 30°C with gentle agitation. The inoculum was prepared by pelleting the culture, washing three times with PBS, and resuspending in PBS at a concentration of 2 × 10⁶ cells/ml.

Six- to 8–wk–old, sex-matched mice were anesthetized with pentobarbitol. A total of 5 × 10⁴ serotype A-KN99α (29) cryptococcal cells in 25 μl PBS were placed on the nares of each mouse, and the mice aspirated the inoculum into the lower respiratory tract. Finally, the mice were suspended by their incisors for 5 min and subsequently placed upright in their cage until regaining consciousness. For survival studies, 10 mice per group were infected as described above. Animals were monitored for morbidity and sacrificed when endpoint criteria were reached. Endpoint criteria were defined as 20% total body weight loss, loss of 2 g of weight in 2 d, or symptoms of neurologic disease.

For intravital staining, 3 μg of anti-CD45.2 (104, BV421; BioLegend) was injected into the tail vein of mice or placed on the nares of sedated mice 3 min prior to sacrifice and whole blood/lung harvest (30). Foxp3-cre ERT2 tdTomato mice received 2 mg/d tamoxifen i.p. for 5 consecutive d to induce endogenous fluorescence for Treg cell fate-mapping. For transfer studies, 1 × 10⁶ negatively selected CD4⁺ Th cells from naive mice were injected via tail vein into congenic mice infected 7 d previously, and lungs were harvested at 14 d postinfection for leukocyte isolation and flow cytometric analysis. For CCR5 blockade experiments, mice were treated i.p. with 500 μg/d maraviroc (R&D Systems, Minneapolis, MN) from 9–14 d postinfection. Lastly, wild-type mice were treated 5 and 10 d postinfection with 1 mg IL-10R Ab (1B1-3A; BioXCell) to block IL-10 signaling.

Lung leukocytes were isolated as previously described (31). Briefly, lungs were excised and minced to generate ∼1-mm³ pieces. The lung mince was incubated in HBSS (Invitrogen, Grand Island, NY) + 1.3 mmol EDTA solution for 30 min at 37°C with agitation, then transferred to RPMI 1640 (Invitrogen) medium supplemented with 5% FBS (Invitrogen) and 150 U/ml type I collagenase (Invitrogen), and incubated for 1 h at 37°C with agitation. The cells were passed through a 70-μm filter, pelleted, and resuspended in 44% Percoll RPMI 1640 medium (GE Life Sciences, Pittsburgh, PA). A Percoll density gradient was created (44% top, 67% bottom), and the samples were centrifuged for 20 min at 650 × g. The leukocytes at the interface were removed, washed two times with RPMI medium, and resuspended in PBS plus FBS at a concentration of 10⁷ cells/ml. CD4⁺ T cells were enriched using a Dynabeads CD4⁺ T Cell Negative Isolation Kit (Life Technologies, Grand Island, NY) per the manufacturer’s instructions. For intracellular cytokine analysis, ∼10⁶ CD4⁺ T cells were suspended in 200 μl restimulation buffer (RPMI + 10% FBS plus 1% penicillin/streptomycin plus 5 μg brefeldin A) without (unstimulated) or with (stimulated) 10 ng PMA and 50 ng ionomycin. After 5 h, the cells were washed and immediately prepared for flow cytometry.

Samples were incubated for 5 min with CD16/32 Ab (BioLegend) and LIVE/DEAD Fixable Far Red stain (Invitrogen) to prevent nonspecific Ab binding, as well as mark dead cells. A total of 25 nmol chitin deacetylase 2 (Cda2) tetramer was added to the sample and incubated at 25°C for 1 h in the dark. CCR3 (J07E35, PE; BioLegend), CCR4 (2G12, PE; BioLegend), and CCR5 (HM-CCR5, PE; BioLegend) were added 1:50 during tetramer staining when appropriate. Samples were surface stained at 4°C for 30 min with the following Abs: CD4 (RM4-5, BV605; BioLegend), CD11b (M1/70, PE-Cy5; eBioscience, San Diego, CA), CD11c (N418, PE-Cy5; eBioscience), B220 (RA3-6B2, PE-Cy5; eBioscience), CD25 (3C7, BV650; BioLegend), CD44 (IM7, Alexa Fluor 700; BioLegend), and/or Siglec F (E50-2440, PE; BD Biosciences). When applicable, the cells were then incubated in Foxp3 Transcription Factor Buffer (eBioscience) at 4°C for 30 min. The cells were labeled with antibodies against the following intracellular Ags: Foxp3 (FJK-16s, FITC; eBioscience), IL-5 (TRFK5, allophycocyanin; BioLegend), IL-13 (eBio13A, eFluor 450; eBioscience), GATA3 (L50-823, PE-Cy7; BD Biosciences), and/or IRF4 (3E4, eFluor 450; eBioscience). Ab concentrations of 1:200 were used for most surface staining, and 1:100 Ab concentrations were used for intracellular staining. For data acquisition, events from the entire sample (500,000–1,000,000) were collected on a BD FACSCanto II flow cytometer (BD Biosciences, San Jose, CA), and the data were analyzed with FlowJo X (Tree Star, Ashland, OR).

Analysis of Ag-specific Th cells within the preimmune repertoire was performed, as previously published (32). Briefly, thymi and secondary lymphoid organs were collected from uninfected Foxp3-GFP mice. Cell suspensions were labeled with Cda2-tetramer and enriched using anti-PE MACS cell isolation kits (Miltenyi Biotec, San Diego, CA). Flow cytometry was performed as described above.

Lungs from naive mice or mice 14 d postinfection were excised, snap-frozen in liquid nitrogen, and homogenized in 3 ml T-PER (Thermo Fisher Scientific) with Complete Protease Inhibitor Cocktail (Roche, Indianapolis, IN). The lung homogenate was pelleted, and the supernatant was collected and stored at −80°C until analysis. Samples were diluted 1:4 in assay buffer immediately before processing. Cytokines were quantified using Luminex technology according to the manufacturer’s instructions (Bio-Rad, Hercules, CA).

Lungs were removed from mice 14 d postinfection, perfused via the right ventricle with cold PBS, inflated with 10% formalin (Thermo Fisher Scientific, Rockford, IL), and placed in a container of 10% formalin. Tissues were dried with organic solvent, embedded in paraffin, sectioned, and stained with H&E, before images were captured.

The p values for pairwise comparisons were by Mann–Whitney U with Bonferroni adjustments for multiple comparisons. Global tests were by Kruskal–Wallis ANOVA. Survival curves were compared with log-rank tests. Power calculations were performed to assess appropriate sample size for all experiments. The p values ≤0.05 were considered statistically significant. All statistics and graphs were processed with Prism 6 (GraphPad Software, La Jolla, CA).

All animal experiments were done in concordance with the Animal Welfare Act, U.S. federal law, and National Institutes of Health guidelines. Mice were handled in accordance with guidelines defined by the University of Minnesota Institutional Animal Care and Use Committee protocol numbers 1010A91133 and 1207A17286.

Upon pulmonary infection with the pathogenic fungus C. neoformans (KN99α) mice develop lethal disease that results from a combination of unabated fungal replication and Th2-driven immunopathology. Importantly, these detrimental Th2 cells are primed and accumulate in the lungs (15). Because Treg cell suppression of effector cells requires these cells to be in close proximity (16), we hypothesized that Treg/effector cells colocalize within Cryptococcus-infected lungs.

Th cells are highly heterogeneous with respect to their TCR and cognate functions. Thus, the use of Cryptococcus-specific reagents to track Ag-specific Th cells within a polyclonal repertoire facilitates direct comparisons of Th cell subsets responding to infection. We used a peptide from Cda2, an immunogenic cryptococcal protein (33), to construct a recombinant peptide–MHC class II tetramer to track Cryptococcus-specific Th cells by flow cytometry (15). The Cda2–MHC class II tetramer not only identified Cda2⁺Foxp3⁻ effector Th cells in the lungs of infected mice, but the tetramer also labeled a sizable population of Cda2⁺Foxp3⁺ Treg cells (Fig. 1A).

The lung consists of three physically separate compartments that may contain Th cells: blood vessels, airways, and lung parenchyma. To determine whether the Treg/effector cells were circulating in the blood vasculature or residing in the airways/lung parenchyma, we performed intravital Ab staining. Anti-CD45 Ab administered i.v. was used to label polyclonal Foxp3⁺ Treg cells and Foxp3⁻ effector Th cells present within the peripheral blood. The complete labeling of all Foxp3⁺ and Foxp3⁻ cells retrieved from the blood indicates that the Ab completely penetrated the entirety of the blood vasculature (Fig. 1B). Cells collected from whole lung digests had a mixture of i.v. stained and unstained cells, showing the lung was composed of circulating and lung-resident Treg/effector Th cells (Fig. 1B). In contrast, nearly all Ag-specific cells were unstained, indicating that infection-induced Th cells resided outside of the blood vasculature (Fig. 1B). To distinguish Treg/effector Th cells in the lungs from the airways, we sedated infected mice and instilled fluorescent-coupled anti-CD45 Ab into the nares. Airway-resident alveolar macrophages collected by lavage were fully labeled by this intranasal Ab treatment, showing this method of Ab delivery was effective (Fig 1C). Conversely, Th cells in the peripheral blood remained unlabeled, indicating the Ab did not leak from the airways into blood circulation (Fig 1C). A majority of polyclonal and Ag-specific Treg/effector Th cells obtained from lung digests did not stain with this inhaled Ab (Fig. 1C). Therefore, most of the Treg/effector Th cells were not resident in the airways. By extension, these data collectively demonstrate that Treg and effector Th cells responding to infection coexist in the lung parenchyma.

We next sought to determine the location of Treg cell induction. The mediastinal lymph node (MLN) is the principle origin of Th cell priming to most microbes that breech the airway mucosa. However, we previously showed that effector Th cells gather in the lungs in the absence of dendritic cell trafficking and subsequent T cell activation in the lung-draining lymph node (15). Thus, we compared the presence of Treg cells at the location of traditional Th cell priming in the MLN and the site of C. neoformans infection in the lungs. Cryptococcus-specific Foxp3-expressing Treg cells existed in the MLN and spleen, yet these populations remained relatively small compared with the large population of Cryptococcus-specific Treg cells that accumulated in the lungs (Fig. 2A, Supplemental Fig. 1).

The relatively small Treg cell response in the MLN suggests that either Treg cells immediately migrate after activation in the lymph node to the site of infection or Treg cell induction occurs autonomously in the lungs. We used CCR7^−/− mice (23) to answer this question. CCR7 is required for naive T cell entry into lymph nodes, thus CCR7 deficiency expectedly inhibited naive Th cells (i.e., CD44 low) from accumulating in the MLN (Fig. 2B). Likewise, the MLN of infected CCR7^−/− mice exhibited decreased swelling compared with wild-type mice (Fig. 2C), further indicating dysfunctional Th cell priming in the MLN of CCR7-deficient mice. Despite the aberrant MLN response, bulk and Cryptococcus-specific Treg cells accumulated in the lungs of CCR7^−/− mice similar to levels in wild-type mice after C. neoformans infection (Fig. 2D). Thus, Treg cell induction and accumulation in the lungs does not require mediastinal lymph node priming during pulmonary cryptoccocal infection.

Treg cells develop along two ontologically distinct lineages: peripheral Treg (pTreg) cells and thymic Treg (tTreg) cells. Upon receiving secondary cues of excessive inflammation, naive Th cells can differentiate in the periphery into pTreg cells. Conversely, tTreg cells become regulatory cells during thymic selection based on TCR affinity for self-antigens (34). Of note, tTreg cells emigrate from the thymus with full suppressive potency and do not need to undergo further activation in lymphoid tissues (35). Therefore, we asked whether the Treg cells in cryptococcal infected lungs are tTreg cells that populate the lungs independently of lymph node priming or pTreg cells autonomously induced in the lungs.

Cryptococcus-specific tTreg cells must exist in the preimmune, naive Th cell repertoire, if these cells are the dominant source of Treg cells in the lungs of infected mice. Therefore, we examined the thymus and secondary lymphoid organs of uninfected mice for the presence of Cda2⁺Foxp3⁺ Treg cells. Cda2-specific Treg cells were present in the preimmune repertoire contained in the thymus and secondary lymphoid organs, albeit at lower Treg/effector proportions compared with polyclonal Th cells (Fig. 3A). Therefore, a small number of Cryptococcus-specific tTreg cells can be found in uninfected mice, and these cells could migrate to the lung and proliferate in response to cryptococcal infection.

To further address the question of whether the Treg cells accumulating in lungs of infected mice migrated from the thymus or were induced in the lungs, we developed a genetic fate-mapping system to distinguish where these cells developed. Mice containing a Foxp3-cre ERT2 transgene (26) were crossed with mice that had a Rosa26 stop codon-floxed tdTomato allele to make Foxp3-i-cre tdTomato mice. Effectively, the combination of these transgenes allows for inducible fluorescent marking in vivo of Treg cells and all of the progeny of these cells. Similarly, when tamoxifen is removed, Treg cells produced de novo will not have any detectable fluorescence reporter activity. Ultimately, this allowed us to label tTreg cells (and all cells derived from this progenitor) within the preimmune repertoire, halt new reporter induction by stopping tamoxifen administration, and determine whether the lung-resident Treg cell progenitors existed prior to infection (i.e., tdTomato⁺) or were produced postinfection (i.e., tdTomato-) (Fig. 3B). Less than 1% of Treg cells from Foxp3-i-cre tdTomato mice that did not receive tamoxifen were fluorescent (Fig. 3B), and tamoxifen administered during the peak Th cell response, 9–14 d postinfection, induced fluorescence in >90% of the Treg cells (Fig. 3B). Thus, the genetic fate-mapping system is not leaky and suitably penetrant. When tamoxifen was given 7–12 d prior to infection to label the preimmune Treg cells, a minor fraction of Treg cells retained fluorescence when harvested at 14 d postinfection (Fig. 3B). Therefore, a small proportion of Treg cells in the lungs came from tTreg cells in the preimmune repertoire, and instead, the majority were pTreg cells that acquire a regulatory phenotype as a consequence of fungal infection.

Treg cells generated during cryptococcal infection are poised to uniquely suppress Th2 cells (14, 15). Additionally, our data indicating that Treg cells are induced and reside in the lungs of infected mice led us to investigate features consistent with Treg cells that develop extrathymically, accumulate at mucosal surfaces, and target Th2 cells for suppression (36–38). A prominent feature of pTreg cells that suppress distinct Th cell subsets in mucosal tissues is the expression of transcription factors that mirror the lineage of the effector Th cell populations targeted for suppression (19, 21). Therefore, we examined the expression kinetics of the Th2 cell transcription factor, IRF4, by both Ag-specific Foxp3⁺ Treg cells and cognate Foxp3⁻ effector Th2 cells from mice infected with C. neoformans. As hypothesized, IRF4 expression increased in Treg cells and effector Th cells throughout the course of infection (Fig. 4A). This raised the possibility that IRF4 is used by Treg cells to suppress the Th2 cell response to pulmonary fungal infection.

To test whether IRF4 expression by Tregs was important for Th2 suppression, we bred Foxp3-cre mice (28) with IRF4 floxed mice (27) to generate mice with a conditional IRF4 gene deletion in Treg cells (Foxp3-cre IFR4 fl/fl) (Fig. 4B). Cryptococcus-specific Th2 cells increased >5-fold in the lungs of Foxp3-cre IRF4 fl/fl mice, and this impaired suppression of Th2 cells resembled the situation observed with complete Treg abrogation (25) (Fig. 4C). Consistent with the increase in Th2 cell numbers, Foxp3-cre IRF4 fl/fl mice also had significantly elevated amounts of IL-5 and IL-13 from infected lung homogenates compared with both wild-type animals and Treg cell–depleted mice (Fig. 4D). Importantly, Foxp3-cre IRF4 fl/fl mice did not experience a concomitant increase in Th17 and Th1 cell cytokine production (Supplemental Fig. 2). Altogether, IRF4 is used by Treg cells to suppress Th2 cell responses to pulmonary cryptococcal infection.

The failure to efficiently suppress Th2 cell proliferation and effector function in mice with IRF4-deficient Treg cells also correlated with exacerbation of Th2-mediated disease. IRF4 deficiency in Treg cells aggravated gross lung pathology (Fig 4E), as well as enhanced pulmonary accumulation of macrophages, multinucleate giant cells, and polymorphonuclear cells (Fig. 4E). Foxp3-cre IRF4 fl/fl mice had elevated fungal burden and accelerated cryptococcal disease (Fig. 4F, 4G). Although lethal disease onset appeared to be more rapid in Foxp3–diphtheria toxin receptor (DTR) mice relative to Foxp3-cre IRF4 fl/fl mice, naive Foxp3-DTR mice receiving DT experienced fatal autoimmunity (Fig. 4G). DT treatment also negatively impacted the survival of infected wild-type mice independently of Treg cell ablation by potentially augmenting fungal burden (Fig. 4F, 4G). Thus, DT treatment and autoimmunity additionally contributed to the faster disease experienced by Foxp3-DTR mice. Taken together, IRF4-deficient Treg cells exhibited a profound Th2 suppression defect that was comparable to complete Treg cell deficiency.

How Treg cells use IRF4 to suppress Th2 cells remains incompletely understood. Existing evidence suggests that IRF4 may dictate expression of suppressive factors employed by Treg cells. For example, IRF4 interacts with Blimp-1 to mediate transcription of the suppressive cytokine, IL-10 (39). Chromatin immunoprecipitation of IRF4 confirms that IRF4 binds to the IL-10 locus, and IRF4 has been shown to mediate IL-10 production by Th2 cells (40, 41). However, IL-10 in lung homogenates of cryptococcal-infected mice was unaffected by IRF4 deficiency in Treg cells and actually increased in mice with complete Treg cell abrogation (Supplemental Fig. 3A). Furthermore, blockade with substantial quantities (i.e., 2 mg/mouse over the course of 9 d) of anti–IL-10R Ab did not alter Th2 cell production (Supplemental Fig. 3B) or IL-5 and IL-13 secretion in lungs of infected mice (Supplemental Fig. 3C). Finally, in other systems, IRF4-deficient Treg cells still suppress effector Th cells in an in vitro assay, indicating that IRF4 is dispensable for the direct suppression of effector Th cells by Treg cells (21).

An alternative hypothesis is IRF4 promotes the retention of Treg cells at the site of inflammation. Although IRF4 deficiency in Treg cells did not alter the proportion of Foxp3⁺ Treg cells among total CD4⁺ Th cells in the spleen and MLN (Supplemental Fig. 4A), it significantly decreased Treg cell proportions in the lungs of infected mice (Fig. 5A). Furthermore, Foxp3-cre IRF4 fl/fl mice had substantially fewer Ag-specific Treg cells in the lungs in comparison with wild-type mice (Fig. 5B). However, these studies could not determine whether the decreased Treg cell proportions were due to biased effector Th cell accumulation or defective Treg cell retention in the lungs.

To test the hypothesis that IRF4 promotes Treg cell localization in the lungs, we performed a set of adoptive transfers. Naive CD4⁺ Th cells from Foxp3-cre IRF4 fl/fl mice were transferred into congenic wild-type recipients, as well as the reciprocal transfer of wild-type Th cells into Foxp3-cre IRF4 fl/fl mice. After resting in infected mice for 5 d (9–14 d postinfection of recipient), the donor cells were identified using the congenic markers. Transferred Th cells from Foxp3-cre IRF4 and wild-type mice parked equivalently in the lungs of their respective hosts (Fig. 5C), yet the transferred cells remained vastly outnumbered by the recipient cells. This allowed us to observe the behavior of transferred cells in the context of a recipient dominated inflammatory milieu. Despite a fully competent wild-type suppressive response that should suppress Th2 cell proliferation, we still observe a blunted Treg cell response skewed toward effector Th cells in Foxp3-cre IRF4 fl/fl transferred cells (Fig. 5C), indicating Treg cells lacking IRF4 were inefficiently retained in the lungs. IRF4-deficient Treg cells deposited normally in the MLN and spleens of infected mice (Supplemental Fig. 4B). Thus, a generalized defect in Treg induction of the Foxp3-cre IRF4 cells was not apparent. Finally, Treg cells from Foxp3-cre IRF4 fl/fl mice were overrepresented in the blood vasculature of infected lungs (Fig. 5D), correlating the decrease in pulmonary retention of Treg cells with the diffusion of these cells into the local bloodstream. These data demonstrate that IRF4 intrinsically regulates Treg cell localization in the lungs of cryptococcal-infected mice.

Th cells follow chemokine gradients to traffic to the site of inflammation (18). Therefore, we investigated chemotactic signals that may influence pulmonary localization of Treg cells. CCL3, CCL4, and CCL5 are involved in type 2 immunity (42), and these chemokines increased 5–100-fold in the lungs of infected mice compared with naive controls (Fig. 6A). To determine if the Treg cells could recognize these chemokines, we examined expression of the cognate chemokine receptors by Treg cells in the lungs of infected mice (Fig. 6B). CCR4 and CCR5 were highly expressed by Treg cells, and expression of these receptors decreased in IRF4-deficient Treg cells (Fig. 6B). In contrast, CCR3 was minimally expressed by Treg cells (Fig. 6B). Thus, CCL3, CCL4, and CCL5 were highly abundant in the lungs of infected mice, and the ability to detect these chemokine signals by Treg cells would require IRF4-dependent expression of CCR4 and CCR5.

Due to the elevated expression of CCR5, abundance of cognate chemokine ligands, and the high dependence of CCR5 on IRF4, we tested the causal relationship between CCR5 and pulmonary retention of Treg cells during fungal infection. Maraviroc is a selective inhibitor of CCR5 that is used in patients with HIV to block CCR5-mediated entry of HIV into leukocytes (43). Mice that received 500 μg maraviroc every day from 9–14 d postinfection had significantly reduced accumulation of pulmonary Treg cells compared with similarly infected, vehicle-treated controls (Fig. 6C). To further test the requirement of CCR5 for Treg localization in the lungs. CD45.1/CD90.2 congenic Foxp3-DTR mice were infected, and Treg cells were eliminated at 7 d postinfection by administering DT. One million naive CD4⁺ Th cells from uninfected CD45.2/CD90.1 wild-type and CD45.2/CD90.2 CCR5^−/− mice (22) were transferred into the Foxp3-DTR mice. At 14 d postinfection, the lungs were harvested and analyzed for Treg accumulation in the lungs. Strikingly, although wild-type Tregs readily accumulated in the lungs, the Tregs transferred from CCR5^−/−, mice were absent from the lungs of infected mice (Fig. 6D). Thus, Treg cell induction and retention in the lungs requires CCR5.

Th cells are central to immunity and immunopathology associated with cryptococcal infection. Although Th1 cells correlate with protection, Th2 cells exacerbate cryptococcal disease. Therefore, a deeper understanding of how the diseased host regulates Th cell responses could lead to development of interventions that ameliorate disease in predisposed individuals. One promising target of immune modulation is the Foxp3⁺ Treg cell population. Previously, Treg cells were found to counterbalance pathologic Th2 cell inflammation following pulmonary cryptococcal infection (14, 15). Yet, the mechanism behind this suppression was largely unexplored. In this study, we tracked Cryptococcus-specific Th cell responses with multiparameter flow cytometry and manipulated host immune responses to unravel the mechanism of Treg-mediated suppression of Th2 cells during cryptococcal infection. We showed that Treg cells are induced in the tissues and use CCR5 and IRF4 to colocalize with and suppress Th2 effector cells in the lung parenchyma.

Many fungal pathogens elicit Treg cell responses. In most cases, Treg cells control the axis of Th17 cell responses and fungal clearance (44–47). In contrast, the primary function of Treg cells generated during cryptococcal infection is Th2 cell suppression. This clearly benefits the host, as disease is enhanced when Treg cells fail to adequately control Th2 cell proliferation. The signals the host uses to detect host damage and elicit Treg cell induction are unknown in the case of pulmonary cryptococcal infection. Additional insight into these processes could lead to the identification of potent biomarkers to predict immune dysfunction in patients stricken with cryptococcal disease. Furthermore, therapeutic targeting of these pathways could be used to prompt the host to dampen harmful Th2 cell production.

We report that Treg cells are induced in substantial quantities during cryptococcal infection, even in the absence of CCR7-mediated entry of naive Th cells into the MLN. Th2 cells also accumulate in the lungs independently of dendritic cell trafficking to the MLN and subsequent Th cell priming (15). This begs the question as to the precise location of T cell priming during cryptococcal infection. BALT is comprised of stochastically distributed clusters of lymphocytes in proximity to high endothelial venules and tissue-resident dendritic cells (48). These structures exist under homeostatic conditions and readily increase in size and number (known as “inducible” BALT) in individuals with chronic inflammatory conditions (49, 50). BALT sufficiently supports T cell priming in the absence of canonical lymphoid responses (51, 52). It is plausible that the BALT is responsible for T cell priming in cryptococcal infection. Fungi may not freely diffuse through the lymphatics to reach the lymph nodes due to the relatively large size of individual fungal cells. Consequently, during early cryptococcal infection in mice, an overwhelming majority of Ag is contained in the lungs. The lung is a high blood flow organ, so circulating naive Th cells have consistent access to the depot of cryptococcal Ag. Naive Th cells could be coaxed into the lungs via high endothelial venules in a chemokine/integrin-mediated process, and the BALT could direct naive Th cell activation, proliferation, and differentiation.

Treg cells are required for the suppression of Th2 cells in this model of pulmonary cryptococcal infection. However, the mechanism of Treg cell–mediated suppression was unknown. IL-10 production by Treg cells is a well-known pathway by which Treg cells inhibit pulmonary Th cell responses (53). IL-10 signaling reduces the proliferative potential of Th cells (54), as well as amplifies the suppressive potency of Treg cells (55). However, IL-10 blockade had minimal impact on Th2 cell responses to cryptococcal infection. Previous studies have shown IL-10–independent Treg cell suppression of effector Th cells involves close contact (56). Thus, mechanistically, the colocalization of Treg cells with effector Th2 cells during cryptococcal infection is an important observation, and the ability of Treg cells to inhibit effector Th cell niches affords unique functional opportunities.

Perhaps the most interesting regulatory pathway concerns the potential ability of Treg cells to mediate suppression by starving Th cells of local growth factors. In particular, Treg cells scavenge IL-2 via their high-affinity IL-2R (57). This competition limits IL-2 growth factor availability and restricts Th cell proliferation (58). There is some evidence to suggest this might occur in the context of pulmonary cryptococcal infection. First, IL-2 complexes administered to infected mice massively augment Th2 cell accumulation (15). Thus, IL-2 is not only an important signal for Th2 cell proliferation, but IL-2 is also a limited resource in this setting. Additionally, STAT6 and IRF4 are each individually required for Th2 cell generation during cryptococcal infection (D.L. Wiesner and K. Nielsen, unpublished observations). However, the requirement for these transcription factors can be bypassed by treating infected knockout mice with IL-2 complexes (D.L. Wiesner and K. Nielsen, unpublished observations). Local IL-2 starvation by Treg cells leading to Th2 cell suppression is an intriguing, but still untested hypothesis in this model of pulmonary fungal infection.

Collectively, our data unify several emerging concepts regarding Treg cell suppression of Th2 cells. Peripherally induced Treg cells inhibit Th2 cells at mucosal surfaces (38), and Treg cells use effector cell programs like IRF4 to mediate specific suppression of Th2 cells (21). IRF4 functions as a rheostat for TCR signaling (59), and TCR signaling is required to maintain a portion of the suppressive program of Treg cells (60). Additionally, chemokines promote the migration and retention of Treg cells in inflamed tissues (18), and CCR5 is important for Treg cells to suppress Th cell responses to pulmonary fungal infections (61). In our model, Treg cells were induced in the periphery, and IRF4 expression by Treg cells was required for efficient Th2 suppression. Treg cells in the lungs of cryptococcal-infected mice expressed high levels of CCR5, and the few remaining IRF4-deficient Treg cells in the lungs had significantly decreased expression of CCR5. IRF4 does not directly interact with the promoter region of CCR5 (41) and does not likely influence CCR5 gene transcription. Thus, we favor a model in which diminished TCR signaling in Treg cells due to IRF4 deficiency reduces CCR5 expression. This prevents Treg/effector cell colocalization and hinders Treg suppression of Th2 cells. Thus, our data provide a logical connection between the hitherto disjointed observations of extrathymically induced Treg cells, IRF4-dependent suppression, chemokine-mediated localization, and Th2-specific inhibition.

Skewed type 2 cytokine responses in the peripheral blood and cerebral spinal fluid of patients with cryptococcal meningitis are associated with early mortality and onset of immune reconstitution inflammatory syndrome (5, 62). CCR5⁺ T cells are recruited to the CSF of patients experiencing cryptococcal meningitis and increased presence of CCR5⁺ T cells is associated with poor clinical outcome (63). HIV infects and lyses CCR5⁺ and CXCR4⁺ Th cells equally (64), and the in vivo evolution of CXCR4-tropic virus is assisted by efficient elimination of CCR5⁺ Th cells (65). Moreover, maraviroc treatment selectively eliminates Treg cells in patients with HIV (66). Taken together with our findings that Treg cells require CCR5 to colocalize and suppress detrimental Th2 cell responses, these observations unveil a novel potential etiology of cryptococcal pathogenesis. Perhaps, HIV-directed lysis of CCR5⁺ Treg cells and/or therapeutic targeting of CCR5 in people living with HIV could exacerbate Th2-driven disease experienced by patients with cryptococcal meningitis.

We thank Dr. Calvin Williams (Medical College of Wisconsin) for kindly providing Foxp3-cre mice, as well as Dr. Marc Jenkins for helpful discussions. We also thank the University of Minnesota Flow Cytometry Core Facility for instrumentation and the University of North Carolina Lineberger Comprehensive Cancer Center Animal Histopathology Core Facility.

The authors have no financial conflicts of interest.