MAVS expression in alveolar macrophages is essential for host resistance against Aspergillus fumigatus

Abstract

Our recent data demonstrates a critical role of the RIG-I-like receptor (RLR) family in regulating antifungal immunity against Aspergillus fumigatus in a murine model. However, the importance of this pathway in humans and the cell type(s) which utilize this innate immune receptor to detect A. fumigatus remains unresolved. Here using patients who underwent hematopoietic stem cell transplantation (HSCT), we demonstrate that a polymorphism in human MAVS present in the donor genome was associated with the incidence of invasive pulmonary aspergillosis (IPA). Moreover, in a separate cohort of confirmed IPA patients, polymorphisms in the IFIH1 gene alter the inflammatory response, including interferon-responsive chemokines. Returning to our murine model, we now demonstrate that CD11c⁺ SiglecF⁺ alveolar macrophages require Mavs expression to maintain host resistance against A. fumigatus. Our data support the role of MAVS signaling in mediating antifungal immunity in both mice and human at least in part through the role of MAVS-dependent signaling in alveolar macrophages.

INTRODUCTION

Aspergillus fumigatus is a ubiquitous environmental mold that humans inhale on a daily basis. Individuals with normal immune systems readily clear A. fumigatus conidia from their airways without complications. In contrast, immunocompromised individuals are at significantly greater risk of developing invasive pulmonary aspergillosis (IPA), including those receiving chemotherapy treatments for cancer and patients receiving immunosuppressive regimens to prevent GVHD following hematopoietic stem cell or solid organ transplantation (1–5). However, only a small proportion of immunosuppressed patients develop invasive fungal infections indicating that additional risk factors must exist. Numerous human studies have identified genetic polymorphisms in key antifungal pattern recognition receptors and inflammatory cytokines associated with fungal infections (reviewed in (6–8)). Thus, it is important to understand how these responses are coordinated in response to fungal infection of the lungs.

Recently, both type I and type III interferons have been shown to be essential for host resistance against pulmonary A. fumigatus challenge (9). Induction of the type I and type III interferon response following A. fumigatus challenge in the mouse model is dependent on both Dectin 1 (Clec7a) and MDA5 (Ifih1) (10, 11). Engagement of Dectin 1 leads to Syk activation which drive IRF5 activation for the production of IFNβ following Candida albicans challenge (12). MDA5 engagement by dsRNA leads to its interaction with MAVS resulting in the recruitment of IKKε and TBK1 that activate NFκB and IRF3 and IRF7, respectively, for the production of early cytokines and type I interferons (13). Genetic polymorphisms within CLEC7A have been associated with increased risk of developing IPA (14–16), while the role for genetic polymorphisms within IFIH1 or MAVS regarding susceptibility to IPA has not been explored to date.

Herein, our data demonstrates that genetic polymorphisms within IFIH1 and MAVS alter the production of interferon-dependent chemokines and risk for HSCT patients in developing IPA, respectively. Interestingly, increased risk for developing IPA was associated with genetic variation within MAVS only in the donor/hematopoietic compartment. Using our murine model, we identify alveolar macrophages as a key hematopoietic cell in the induction of the MDA5/MAVS-dependent interferon response which is necessary for host resistance against A. fumigatus. Overall, our study reveals a critical role for MDA5/MAVS in host anti-fungal immunity in both mice and humans.

MATERIALS & METHODS

Mice and Aspergillus fumigatus challenge model.

Ifih1^(−/−) (Jackson Laboratory, Stock #015812), Mavs^(fl/fl) (17), and Mavs^(fl/fl) x Itgax-Cre mice (17) were bred in-house at Geisel School of Medicine at Dartmouth. C57BL/6J mice were purchased from Jackson Laboratory. All mice were 8–16 weeks of age at the time of challenge. Both female and male mice were used in these studies with no biological difference between the sexes. Animals were not co-housed to equilibrate their microbiota, which is a limitation of this study.

Preparation of Aspergillus fumigatus conidia and murine challenge model.

A. fumigatus CEA10 strains were used for this study. A. fumigatus was grown on glucose minimal media (GMM) agar plates for 3 days at 37°C. Conidia were harvested by adding 0.01% Tween 80 to plates and gently scraping conidia from the plates using a cell scraper. Conidia were then filtered through sterile Miracloth, were washed, and resuspended in phosphate buffered saline (PBS), and counted on a hemocytometer.

Mice were challenged with A. fumigatus conidia by the intratracheal (i.t.) route. Mice were anesthetized by inhalation of isoflurane; subsequently, mice were challenged i.t. with ~4 × 10⁷ A. fumigatus conidia in a volume of 100 μl PBS. At the indicated time after A. fumigatus challenge, mice were euthanized using carbon dioxide. Bronchoalveolar lavage fluid (BALF) was collected by washing the lungs with 2 ml of PBS containing 0.05M EDTA. BALF was clarified by centrifugation and stored at −20°C until analysis. After centrifugation, the cellular component of the BALF was resuspended in 200 μl of PBS and total BAL cells were determined by hemocytometer count. BALF cells were subsequently spun onto glass slides using a Cytospin4 cytocentrifuge (Thermo Scientific) and stained with the Hema 3™ Stat Pack (Fisher Scientific) stain set for differential counting. For histological analysis lungs were filled with and stored in 10% buffered formalin phosphate for at least 24 hours. Lungs were then embedded in paraffin and sectioned into 5-micron sections. Sections were stained with Grocott-Gomori methenamine silver (GMS) using standard histological techniques to assess lung inflammatory infiltrates and fungal germination, respectively. Representative pictures of lung sections were taken using an Olympus BX50WI microscope with a QImaging Retiga 2000R camera.

Alveolar macrophage isolation and adoptive transfer.

Lungs from naïve Ifih1^(−/−), Mavs^(fl/fl) and Mavs^(fl/fl) x Itgax-Cre were perfused with 20–30 ml of PBS. Lungs were then removed, injected with 700 μl of collagenase buffer [10ml RPMI, 1ml FBS, 25 μl 0.5M CaCl₂, 25 μl 0.5M MgCl₂, 50 μl HEPES, 100 μl L-glutamine, 100 μl Penn-Strep, 2 μl gentamycin, 25 μl DNase (VWR, Catalog # 77001–900), and 2 ml Liberase (Sigma-Aldrich)], and finally placed in 15 ml conical tube containing 2 ml of collagenase buffer. Lungs were digested or 30min at 37°C while shaking at 160–180 rpm. After which, 5 ml of ice cold stop buffer [25 ml RPMI, 1.25 ml FBS, 50 μl 0.5M EDTA] was added to each tube. Lungs were pushed through a 70 μm filter to generate a single cell suspension. Single cell lung suspensions were labeled with anti-Siglec-F MicroBeads (Miltenyi Biotec, Cat. No. 130-118-513) and selected for using LS Columns (Miltenyi Biotec, Cat. No. 130-042-401). Siglec F⁺ were eluted from the column, spun down, and resuspended in DMEM at 5×10⁶ cells per ml. Subsequently 5×10⁵ Siglec F⁺ alveolar macrophages were transferred i.t. to Mavs^(fl/fl) x Itgax-Cre. Twenty-four hours later mice were challenge with A. fumigatus as described above.

Human HSCT cohort for SNP analysis.

A total of 460 hematologic patients undergoing allogeneic hematopoietic stem cell transplantation (HSCT) at the Hospital of Santa Maria, Lisbon and Instituto Português de Oncologia (IPO), Porto, between 2009 and 2015 were included in the study. Cases of IPA were identified and classified as ‘probable’ or ‘proven’ according to the 2008 criteria from the European Organization for Research and Treatment of Cancer/Mycology Study Group (EORTC/MSG) (18). Exclusion criteria included diagnosis of ‘possible’ IPA, infection with invasive molds other than Aspergillus spp. or history of pre-transplant mold infection. In accordance with institutional policies, all patients and donors (or their legal representative) provided informed written consent for data collection, DNA and cell storage, and their use for diagnostic and/or research purposes at the time of referral to transplantation. Study approval was obtained from the institutional review boards (SECVS-125/2014, HSM-632/14 and CES.26/015) and from the National Data Protection Commission (CNPD, 1950/2015) and was in compliance with all local relevant ethical regulations.

Genomic DNA was isolated from whole blood of patients using the QIAcube automated system (Qiagen). SNPs were selected based on their previous association with infection or with putative functional consequences to the gene (19–21). Genotyping was performed using KASPar assays (LGC Genomics) in an Applied Biosystems 7500 Fast Real-Time PCR system (Thermo Fisher Scientific), according to the manufacturer’s instructions.

Cytokine analysis of BALF from human IPA patients.

BALF and blood samples were collected from hospitalized adult patients (≥18 years of age) at the Leuven University Hospitals, Leuven, Belgium, as previously described (22). All patients (or their legal representative) provided informed written consent for data collection, DNA and cell storage, and their use for diagnostic and/or research purposes at hospital admission. This study was approved and carried out in accordance with recommendations of the Ethics Subcommittee for Life and Health Sciences of the University of Minho, Portugal (SECVS-125/2014), and the Ethics Committee of the University Hospitals of Leuven, Belgium. Written informed consent was obtained from all subjects in accordance with the Declaration of Helsinki.

For this cytokine analysis, twenty-three cases of “probable” or “proven” IPA were identified according to the standard criteria from the European Organization for Research and Treatment of Cancer/Mycology Study Group (EORTC/MSG) (18) and included for cytokine analysis. Genomic DNA was isolated from EDTA venous blood and genotyped for the IFIH1 rs1990760 polymorphism. Cytokine levels were determined using the Cytokine & Chemokine 34-Plex Human ProcartaPlex and stratified based on their IFIH1 rs1990760 genotype.

Statistical analysis.

Statistical significance for in vitro and ex vivo data was determined by a Mann-Whitney U test, one-way ANOVA using a Bonferroni post-test, or Kruskal-Wallis one-way ANOVA with Dunn’s post-test through the GraphPad Prism 7 software as outlined in the figure legends. Mouse survival data were analyzed with the Mantel-Cox log rank test using GraphPad Prism. For the human HSCT cohort, the probability of IPA resulting from IFIH1 and MAVS SNPs was analyzed using the cumulative incidence method and compared using Gray’s test (23). The Gray’s test was selected for the analysis because, in contrast with the commonly used Kaplan-Meier and Cox models, it allows accounting for multiple competing risks and provides a better estimation for the risk of the main outcome of interest, in this case the development of IPA (24). Cumulative incidences were computed with the cmprsk package for R version 2.10.1 (25), with censoring of data at the date of last follow-up visit and defining relapse and death as competing events. A period of 24 months after transplant was chosen to include all cases of IPA.

Study approvals.

All animal experiments were approved by the Dartmouth College Institutional Animal Care and Use Committee under protocol number 00002168. For our human studies approval was obtained from the institutional review boards (SECVS-125/2014, HSM-632/14 and CES.26/015) and from the National Data Protection Commission (CNPD, 1950/2015) and was in compliance with all local relevant ethical regulations.

RESULTS

Single Nucleotide Polymorphisms (SNPs) in IFIH1 and MAVS influence the risk for developing invasive pulmonary aspergillosis.

Our recent mouse data suggests that MDA5/MAVS signaling is critical in maintaining host resistance against pulmonary challenge with A. fumigatus (11). Therefore, we wanted to explore the role of these molecules in a human cohort comprised of 460 HSCT patients (Supplemental Table 1). Missense SNPs within the coding region of IFIH1 (rs1990760 and rs3747517) are associated with autoimmune conditions, particularly interferonopathies (19). Moreover, a missense SNP in MAVS (rs17857295) is associated with altered type I interferon regulation (21). While an individual genotype for each SNP was not associated with risk of developing IPA (Supplemental Figure 1), we further examined these SNPs using a recessive genetic model because the autoimmune risk allele in IFIH1 rs1990760 has been shown to work in a dominant fashion (19). Notably, we found a significant association between the CC genotype at rs1990760 in IFIH1 and the risk of developing IPA in HSCT recipients (Figure 1A). This increased risk occurred when the variant was carried by the recipient in a recessive genetic model, but not when it was carried by the donor. In contrast, no association between the IFIH1 rs3747517 SNP and the risk of developing IPA was found (Figure 1B). Additionally, we also found a significant association between the MAVS rs17857295 SNP and the risk of developing IPA in HSCT recipients (Figure 1C). This increased risk occurred when the variant was carried by the donor in a recessive genetic model, but not when it was carried by the recipient. Moreover, the risk from the donor MAVS rs17857295 SNP held up to multivariate analysis accounting for relevant clinical risk factors (Table 1).

Figure 1. IFIH1 and MAVS polymorphisms are associated with invasive pulmonary aspergillosis in human transplant patients using a recessive allele model.

Cumulative incidence analysis of invasive aspergillosis after transplantation according to donor or recipient genotypes at rs1990760 in IFIH1 (A), rs3747517 in IFIH1 (B), or rs17857295 in MAVS (C) over 24 months after HSCT. Data were analyzed by two-sided Gray’s test. (D) Inflammatory cytokine levels in the bronchoalveolar lavage fluid from 23 patients with invasive pulmonary aspergillosis were measured using a 32-plex ProCarta Luminex assay and plotted based on their IFIH1 rs1990760 genotype. Data were analyzed using a Mann-Whitney U-test (* p < 0.05).

Table 1.

Multivariate analysis of the association of IFIH1 and MAVS SNPs with the risk of invasive pulmonary aspergillosis among transplant recipients.

Genetic/clinical variables	Adjusted HR† (95% CI)	P value
Donor rs17857295 in MAVS	2.39 (1.07 – 5.32)	0.033
Recipient rs1990760 in IFIH1	1.46 (0.89 – 2.40)	0.121
Acute GVHD grades III-IV	1.73 (0.94 – 3.19)	0.047

HR, hazard ratio; CI, confidence interval. Multivariate analyses were based on the subdistribution regression model of Fine and Gray.

^†Hazard ratios were adjusted for patient age and gender, and clinical variables with a P<0.15 in the univariate analyses. Only the clinical variables remaining significant after adjustment are shown.

To demonstrate a functional effect of the rs1990760 SNP in IFIH1 and the associated amino acid substitution on the function of MDA5, we analyzed the inflammatory response in the bronchoalveolar lavage fluid (BALF) of 23 patients with IPA after stratification by genotypes at rs1990760 in IFIH1 (Supplemental Table 2). We found that individuals with the CC genotype, which is associated with increased risk of developing IPA (Figure 1A), had a significant reduction in IP-10 (CXCL10), RANTES (CCL5), and MIP-1α (CCL3), but not IL-8 and GROα (CXCL1) (Figure 1D). IP-10 (CXCL10), RANTES (CCL5), and MIP-1α (CCL3) are known to be interferon-responsive genes, while IL-8 and GROα (CXCL1) are not (26). In the BALF samples from this patient cohort, the levels of IFN-α were below the detection limit (data not shown). Given the relatively rare frequency of the rs17857295 genotype in MAVS, it was not possible to analyze cytokine production according to MAVS genotype in the BAL samples (data not shown). Overall, these data support an important role for MDA5/MAVS signaling in regulating the inflammatory response and host resistance against Aspergillus spp. in humans.

Deletion of Mavs in CD11c⁺ cells result in increased susceptibility to Aspergillus fumigatus in mice.

Given the observation that patients in the HSCT cohort with the MAVS rs17857295 polymorphism within the donor population had altered risk for developing IPA, we next wanted to assess which hematopoietic cell population(s) in our invasive aspergillosis murine model relies on Mavs to maintain host resistance against A. fumigatus. To directly address which cells must express Mavs we utilized the recently developed Mavs^(fl/fl) x Itgax-Cre conditional knock-out model, hereafter referred to as Mavs^{(Cd11c/Cd11c)}, that will delete Mavs in all CD11c-expressing cells (17), including both dendritic cells and alveolar macrophages. To test of role of Mavs expression in CD11c-expressing cells for the maintenance of host resistance against A. fumigatus, we challenged C57BL/6J, Mavs^(fl/fl), and Mavs^{(Cd11c/Cd11c)} mice with 4×10⁷ conidia of the CEA10 strain and monitored survival over nine days. Mavs^{(Cd11c/Cd11c)} mice were more susceptible to pulmonary challenge with A. fumigatus than either C57BL/6J or Mavs^(fl/fl) mice (Figure 2A; Mantel-Cox log rank test, p < 0.0001). Mice completely lacking either Ifih1 or Mavs have decreased neutrophil recruitment, which was associated with increased fungal germination after pulmonary challenge with A. fumigatus (11). Thus, we assessed fungal germination in the lung by histological analysis at 48 h after conidial instillation. Strikingly, GMS staining of lung tissue from Mavs^{(Cd11c/Cd11c)} mice revealed high levels of A. fumigatus germination at this time compared with control Mavs^(fl/fl) mice (Figure 2B). When the percentage of germinated A. fumigatus was quantified, Mavs^(fl/fl) mice displayed low levels of fungal germination (23.5% ± 9.1) compared with Mavs^{(Cd11c/Cd11c)} mice (60.1% ± 8.2). Finally, we assessed inflammatory cell accumulation in the airways. The increased susceptibility of Mavs^{(Cd11c/Cd11c)} mice to A. fumigatus challenge was associated with decreased accumulation of neutrophils in the airways compared with Mavs^(fl/fl) mice (Figure 2C). These data match our prior observations with mice completely lacking either Ifih1 or Mavs having increased IPA susceptibility (11), suggesting that Mavs expression within CD11c-expressing cells is critical for host resistance against A. fumigatus challenge.

Figure 2. Mavs-dependent responses are essential in CD11c⁺ cells for host resistance against Aspergillus fumigatus.

Mavs^{<Cd11c/Cd11c>}, Mavs^<fl/fl> and C57BL/6J mice were challenged i.t. with 4×10⁷ resting conidia of the CEA10 isolate of Aspergillus fumigatus. (A) Survival analysis in immune-competent wild-type and knock-out mice were tracked over the first 9 days. ****, P < 0.0001 by Mantel-Cox log rank test. (B) Forty hours after A. fumigatus challenge mice were euthanized and fungal germination was assessed in the lungs by GMS staining. **, P < 0.01 by Mann-Whitney U-test. (C) At the same time point, cell differentials in the lung airways was determined by differential staining of BALF cytospins. Statistical significance was determined using a two-way ANOVA with a Sidak’s post-test.

Alveolar macrophage transfer to Mavs^{(Cd11c/Cd11c)} conditional knock-out mice reduces their susceptibility to Aspergillus fumigatus in mice.

Multiple cells in the lungs can express Igtax (CD11c) including alveolar macrophages, CD103⁺ dendritic cells, CD11b⁺ dendritic cells, monocyte-derived dendritic cells, and plasmacytoid dendritic cells, all of which have be implicated in the antifungal immune response against A. fumigatus (27–33). Utilizing the publicly available data assembled by the ImmGen Consortium (34), in the lungs we see that Mavs is highly expressed in Siglec F⁺ alveolar macrophages, but also moderately expressed in both monocytes and dendritic cells (data not shown). Interestingly, following RSV infection alveolar macrophages have been shown to be the key cell type for initiating the type I interferon through a MAVS-dependent mechanism (35, 36). While alveolar macrophages largely populate the lungs during embryogenesis and are maintained by local proliferation during homeostasis (37, 38) following conditioning for HSC transplantation alveolar macrophage can be repopulated from the donor HSC compartment (39–42). Therefore, to specifically address the role of Mavs in alveolar macrophages, we purified Siglec F⁺ cells from the lungs of naïve Mavs^(fl/fl) or Mavs^{(Cd11c/Cd11c)} mice and transferred 5×10⁵ Siglec F⁺ cells intratracheally into Mavs^{(Cd11c/Cd11c)} mice one day prior to challenging with 4×10⁷ conidia of the CEA10 strain (Figure 3A). Forty-eight hours after challenge, we examined the lungs by GMS staining of lung tissue from Mavs^{(Cd11c/Cd11c)} mice receiving the Mavs^{(Cd11c/Cd11c)} Siglec F⁺ cells and revealed the presence of high levels of germinated A. fumigatus (68.1% ± 12.6), while that was not observed to the same extent in Mavs^{(Cd11c/Cd11c)} mice receiving the Mavs^(fl/fl) Siglec F⁺ cells (31.4% ± 6.7) (Figure 3B). This is in line with control Mavs^{(Cd11c/Cd11c)} mice or Mavs^(fl/fl) mice receiving DMEM (vehicle) administered intratracheally. When we examined the accumulation of leukocytes in the airways by differential cytospin analysis we found that Mavs^{(Cd11c/Cd11c)} mice receiving Mavs^{(Cd11c/Cd11c)} Siglec F⁺ cells had reduced numbers of neutrophils when compared with Mavs^{(Cd11c/Cd11c)} mice receiving Mavs^(fl/fl) Siglec F⁺ cells (Figure 3C). Strikingly, when we examined macrophage numbers in the airways, we found Mavs^{(Cd11c/Cd11c)} mice receiving Mavs^(fl/fl) Siglec F⁺ cells had many macrophages present, but Mavs^{(Cd11c/Cd11c)} mice receiving Mavs^{(Cd11c/Cd11c)} Siglec F⁺ cells had few macrophages in their airways at 48 hours post-inoculation with A. fumigatus (Figure 3D). Importantly, the decrease in macrophages in the airways is not observed in naïve Mavs^{(Cd11c/Cd11c)} mice (Supplemental Figure 2A). Additionally, Mavs^{(Cd11c/Cd11c)} mice receiving an i.t. transfer of Mavs^{(Cd11c/Cd11c)} Siglec F⁺ cells 24h prior only had a moderate decrease in alveolar macrophage numbers (Supplemental Figure 2B).

Figure 3. Alveolar macrophages require a Mavs- and Ifih1-dependent response to maintain host resistance against Aspergillus fumigatus.

(A) Mavs^{<Cd11c/Cd11c>} mice were complemented with 5×10⁵ Mavs^<fl/fl> or Mavs^{<Cd11c/Cd11c>} alveolar macrophages (Siglec F⁺) given intranasally. One day later mice were challenged i.t. with 4×10⁷ resting conidia of the CEA10 isolate of Aspergillus fumigatus. Forty hours after A. fumigatus challenge mice were euthanized. (B Fungal germination was assessed in the lungs by GMS staining. Representative 20x GMS images are shown. Statistically significant different were determined using a one-way ANOVA with a Tukey’s post-test: *p<0.05 (vs. DMEM → Mavs^(fl//fl)), ****p<0.0001 (vs. DMEM → Mavs^(fl//fl)), and ττττ p<0.0005 (vs. Mavs^(fl//fl) → Mavs^(Cd11c)). Neutrophils (C) and macrophages (D) in the lung airways was determined by differential staining of BALF cytospins. (E) Mavs^{<Cd11c/Cd11c>} mice were complemented with 5×10⁵ Mavs^<fl/fl> or Ifih1^(−/−) alveolar macrophages (Siglec F⁺) given intranasally. One day later mice were challenged i.t. with 4×10⁷ resting conidia of the CEA10 isolate of A. fumigatus. Forty hours after A. fumigatus challenge mice were euthanized. Fungal germination was assessed in the lungs by GMS staining. Data are pooled from 2 independent experiments. Each dot represents an individual mouse. Statistically significant different were determined using a one-way ANOVA with a Kruskal-Wallis post-test: ****p<0.0001.

Next, to examine the role of Ifih1 in alveolar macrophages, we purified Siglec F⁺ cells from the lungs of naïve Mavs^(fl/fl) or Ifih1^(−/−) mice and transferred 5×10⁵ Siglec F⁺ cells intratracheally into Mavs^{(Cd11c/Cd11c)} mice one day prior to challenging with 4×10⁷ conidia of the CEA10 strain (Figure 3E). Forty-eight hours after challenge, we examined the lungs by GMS staining of lung tissue from Mavs^{(Cd11c/Cd11c)} mice receiving either DMEM or Ifih1^(−/−) Siglec F⁺ cells, revealing the presence of high levels of germinated A. fumigatus, which was not observed to the same extent in Mavs^{(Cd11c/Cd11c)} mice receiving the Mavs^(fl/fl) Siglec F⁺ cells (Figure 3E). Taken all together, these data demonstrate that Ifih1 and Mavs expression within Siglec F⁺ alveolar macrophages is critical for host resistance against A. fumigatus.

DISCUSSION

Humans inhale Aspergillus fumigatus spores on a daily basis, but individuals with healthy immune systems readily clear A. fumigatus conidia from their airways without problems. One lung sentinel cell that can be critical in the clearance of A. fumigatus is the alveolar macrophage. Alveolar macrophages are known to phagocytose and destroy A. fumigatus (31), but their role in host resistance has remained controversial (33, 43). Patients undergoing HSCT are at increased risk of developing IPA (1–5). In early studies from patients undergoing HSCT, alveolar macrophages are dysfunctional (41) and found at decreased in numbers (40–42) for at least 50 days post-transplantation. However, we currently do not understand the effects of modern HSCT conditioning regiments on alveolar macrophages function and numbers, which is a limitation of our study. Either way it is important to understand the role of alveolar macrophages in the initial inflammatory response induced by A. fumigatus. These lung sentinel alveolar macrophages are also known to participate in the inflammatory response induced by A. fumigatus (44), but their role in driving the interferon response following A. fumigatus challenge was not addressed. During respiratory syncytial virus infection alveolar macrophages have been demonstrated to be key drivers of the type I interferon response through a MAVS-dependent mechanism (35, 36). We recently describe a critical role for MDA5/MAVS-dependent induction of interferons following A. fumigatus challenge (11), but the cellular localization of MAVS-dependent signaling was not elucidated. Here we demonstrated that Mavs expression in alveolar macrophages was at least partially necessary for host resistance against A. fumigatus, through the regulation of neutrophil accumulation in the lung for the prevention of fungal growth. While alveolar macrophage transfer to Mavs^{(Cd11c/Cd11c)} mice largely restores host resistance, it was not absolute suggesting additional CD11c⁺ cells might require Mavs expression for resistance. One potential candidate would be plasmacytoid dendritic cells that are known to produce type I interferons after fungal challenge (27) and are essential for host resistance A. fumigatus (45), but the importance of MAVS signaling in plasmacytoid dendritic cells in driving the interferon response remains controversial. Additionally, in a dendritic cell vaccination model, A. fumigatus conidial RNA was shown to be sufficient for the activation of dendritic cells to preferentially prime a Th1 immune response (46), which was at least partially dependent on TLR3 (47). The role of MAVS signaling in other CD11c⁺ cell populations will be explored in future studies. Alveolar macrophages also participate in the inflammatory response to other fungal pathogens including Cryptococcus neoformans (48), Pneumocystis spp. (49), and Rhizopus spp. (50). Interestingly, heterogeneity within the alveolar macrophage population has been demonstrated particularly in regard to CXCL2 expression (48), which could be important in our findings showing that the mice lacking Ifih1 or Mavs have decreased neutrophil accumulation following A. fumigatus challenge (11). Additionally, Jakubzick and colleagues have identified significant heterogeneity within human alveolar macrophages using single cell RNA-Seq, which include the identification of novel IFN responsive alveolar macrophage populations and a number of chemokine producing populations (51). Understanding the specific role of these alveolar macrophage populations in host resistance is the goal of future studies.

In addition to the role of alveolar macrophages in the interferon-dependent antifungal immune response that we describe herein, it has previously been shown that in vitro human bronchial epithelial cells can secrete IFNβ and CXCL10 in response to A. fumigatus spore challenge (52). Following respiratory virus infections, the lung epithelium can be a critical source of antiviral interferons, which can be driven by both TLRs and RLRs (53, 54). In the human bronchial epithelial cell model, Bals and colleagues demonstrated that fungal dsRNA which was transfected into the human bronchial epithelial cells was sufficient for inducing IFNβ and CXCL10 secretion (52), which is similar to our previous findings using in murine fibroblast (11). However, while Bals and colleagues demonstrated that A. fumigatus dsRNA was sufficient for the activation of TLR3-TRIF signaling, they did not examine the necessity of this pathway in driving IFNβ and CXCL10 secretion by the human bronchial epithelial cells (52). Interestingly, in our preliminary studies, cytosolic delivery of A. fumigatus dsRNA drives both MDA5/MAVS-dependent and TRIF-dependent IFNα and CXCL10 responses in murine fibroblasts (data not shown). Moreover, Romani and colleagues have found both Tlr3- and Ticam1 (TRIF)-deficient mice are highly susceptible to pulmonary A. fumigatus challenge (47, 55), and this was due entirely to TRIF-dependent response in epithelial cells (55). Interestingly, in our HSCT patients, the IFIH1 SNP association was found in the host genotype. Taken together, these data suggest the lung epithelial cells could be another potential source of IFNs after A. fumigatus challenge which needs to be explored in future studies.

While our previous studies have shown the importance of MDA5/MAVS signaling (11) and type I/III interferons (9) in host resistance against A. fumigatus in the murine model, the importance of these pathways in humans has not been described. Herein, we have found that antifungal role of MDA5/MAVS signaling is also critical in humans. Our data from a cohort of HSCT patients demonstrate that SNPs in IFIH (rs1990760) and MAVS (rs17857295) are associated with an altered risk of developing IPA. Interestingly, SNPs within IFIH1 have been associated with numerous autoimmune diseases, particularly those that are highly dependent on type I interferons for their progression (19, 56). Recent evidence suggests that the autoimmune risk genotype of the rs1990760 SNP in IFIH1 (TT) has functional consequences in human PBMCs, specifically it is associated with elevated basal levels of IFNB1 (19). Knock-in of the IFIH1 rs1990760 autoimmune risk genotype (TT) into mice not only recapitulates the susceptibility to autoimmune diseases, but also results in increased host resistance against encephalomyocarditis virus (ECMV), a picornavirus specifically recognized by MDA5 (57), and elevated Ifnb1 expression (19). Moreover, in human pancreatic islets it has been shown that the IFIH1 rs1990760 TC genotype also drove enhanced interferon signaling after Coxsackievirus infection (20). This enhanced interferon response during Coxsackievirus infection was associated with greater signaling from the peroxisomes by the IFIH1 rs1990760 TC genotype (20), which has previously been shown to preferentially induce a type III interferon response (58, 59). Interestingly, our SNP analysis of a human HSCT cohort found that the T allele of IFIH1 rs1990760, which as just mentioned is associated with an increased interferon response (19, 20), was associated with a reduced risk for the development of IPA. In addition to the IFIH1 rs1990760 SNP, we also found a clinical association with the rs17857295 SNP in MAVS with the risk of developing IPA after HSCT. Much less is known about the functional outcome of the rs17857295 SNP in MAVS, but one report suggests overexpression of the GG genotype in 293T cells stably knocked-down for the endogenous MAVS allele resulted in an inability to induce Ifnb1 expression following PolyI:C stimulation (21). Overall, our analysis suggests that HSCT patients harboring these SNPs display a reduced ability to trigger an interferon response and are at greater risk of developing IPA.

In contrast to our clinical data from HSCT patients who develop IPA, the T allele at rs1990760 in IFIH1 is associated with an increased risk for the development of chronic mucocutaneous candidiasis (60). Again, the T allele at rs1990760 is associated with an increased interferon response (19, 20), but the lack of IFNAR-dependent signaling leads to increased susceptibility to invasive candidiasis in mice (12). In their study Netea and colleagues did not observe altered Ifnb expression in the absence of MDA5 after stimulation with Candida albicans hyphae (60), but a role for MDA5 in regulating other interferons was not explored. Our data with A. fumigatus suggests that MDA5/MAVS-signaling are more important in the regulation of type III interferons (IFNλ/IL-28), rather than type I interferons (IFNα/β). Excessive inflammation and tissue pathology can be critical in mediating disease during invasive candidiasis (61). Rivera and colleagues demonstrated that type III interferon (IFNλ/IL-28) signaling enhances the production of reactive oxygen species by neutrophils following A. fumigatus (9), which could enhance disease during invasive candidiasis. Thus, there appears to be an interesting dichotomy in MDA5 signaling that is crucial for tuning host resistance against different fungal pathogens and warrants further research.

Key Points/Findings:

• Mavs expression in alveolar macrophages maintains resistance against A. fumigatus

• A SNP in MAVS is associated with a risk of developing invasive aspergillosis