The Metabolic Cytokine Adiponectin Inhibits Inflammatory Lung Pathology in Invasive Aspergillosis

This article is featured in In This Issue, p.773

Recent advances have identified overlapping pathways that regulate both systemic metabolism and immunity, and as a result, research interest at the nexus of these two fields is rapidly expanding (1). The metabolic cytokine adiponectin is an intercellular signal that regulates insulin sensitivity and inflammatory processes at sites distal from adipose tissue, where it is expressed at high levels, resulting in a plasma concentration of up to 0.01% of total protein in healthy individuals (2–5). Moreover, this high level of expression is significantly decreased in individuals with obesity or asthma, both diseases associated with chronic inflammation (6–8). The anti-inflammatory role of adiponectin in obesity is well documented (9, 10), and a role for adiponectin in attenuation of lung inflammation is becoming increasingly clear (11–16). Mice that lack adiponectin exhibited increased allergic airway inflammation and concomitant eosinophil recruitment when compared with wild-type mice (11), and exogenous administration attenuated allergic inflammation and lung eosinophilia (12). More recently, adiponectin-deficient mice were shown to exhibit impaired bacterial clearance because of excessive inflammation, suggesting an additional role for adiponectin in protection from microbial infection (17). However, the role of adiponectin in protection from infection with pathogens associated with significant lung immune pathology remains unknown.

The incidence of opportunistic fungal infections has risen along with the use of immune-suppressive therapies for transplantation, cancer, and various inflammatory diseases (18). Aspergillus fumigatus is an opportunistic human fungal pathogen that causes lung inflammation and/or a highly fatal invasive infection, depending on the level of host immune competence (19–21). Mortality in patients with invasive aspergillosis (IA) may exceed 50%, and antifungal resistance is increasing (20, 22, 23). Despite the immune suppression that renders many patients susceptible to IA, residual antifungal immunity controls the balance between protection and pathology (18–21, 24), and excessive inflammation may paradoxically result in defective fungal clearance and poor disease outcomes (25, 26). Therefore, complementary therapies for current antimicrobial drugs that limit immune pathology could result in marked improvement in IA patients with dysregulated inflammation (25).

In this study, we investigated the role of adiponectin in inflammatory pathology using a neutropenic mouse model of IA. We report that adiponectin-deficient mice with IA exhibited increased morbidity, mortality, and fungal burden compared with wild-type controls. An extensive increase in pleiotropic inflammatory cytokines and a decrease in anti-inflammatory pathways were observed in adiponectin-deficient neutropenic mice with IA, suggesting a broad anti-inflammatory role for adiponectin in invasive fungal infection in immune-compromised hosts. Consistent with this role, we observed a marked increase in eosinophil recruitment and activation, yet eosinophil deficiency only resulted in delayed mortality in adiponectin-deficient mice with IA. Furthermore, neutrophil depletion was required for increased lung inflammatory cytokine expression in response to swollen A. fumigatus conidia, even as further germination/invasive fungal growth was blocked by fixation. These results identify a therapeutic potential for the adiponectin pathway in the development of treatments for immune-compromised IA patients who do not improve with conventional antifungal therapy.

C57BL/6 and adiponectin-deficient (Adipoq^−/−) mice (B6 background) were received from The Jackson Laboratory. ΔdblGATA1 mice on the B6 background were previously provided by Dr. A. August. Double-deficient Adipoq^−/−ΔdblGATA1 mice were generated by breeding F1 animals heterozygous for both genes with offspring screened by PCR, flow cytometry, and/or ELISA to verify double deficiency. Adiponectin transgenic mice on the B6 background that globally overexpress adiponectin driven by the housekeeping β-actin promoter (adiponectin-overexpressing transgenic [AdipoTg] mice) were generated as previously described (27). Mice with constitutive production of IL-5 via a lung-specific promoter were provided by Dr. J. Lee (28). All animal handling and experimental procedures were performed in accordance with the recommendations found in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The work in this study was approved by the Institutional Animal Care and Use Committee of the host campus of Indiana University School of Medicine–Terre Haute, Indiana State University.

The clinical isolate Af293 was previously obtained from the Fungal Genetics Stock Center. Fungi were grown on Malt Extract Agar plates at 22°C, and conidia suspensions were collected and aseptically prepared, as described (29, 30).

For single aspiration of swollen and fixed conidia, 5 × 10⁷ Af293 conidia were incubated in RPMI 1640 for 4 h at 37°C, fixed with 4% paraformaldehyde, and then washed with ammonium chloride and sterile Dulbecco’s PBS. Isoflurane-anesthetized mice involuntarily aspirated swollen/fixed conidia in 50 μl of suspension, and mouse bronchoalveolar lavage cells and lungs were harvested 24 h later for subsequent analyses. To induce invasive pulmonary aspergillosis in mice, neutrophils were depleted with i.p. injection of 0.5 mg of α-Ly6G (1A8 clone; BioXCell) 24 h pre- and postinfection (p.i.), as described (31). Neutropenic mice were infected with 5 × 10⁶–1 × 10⁷ of A. fumigatus Af293 conidia. Lungs were harvested from between 0 and 72 h p.i. and used for quantification of fungal burden, flow cytometric analysis, lactate dehydrogenase quantification, cytokine ELISA, and quantitative RT-PCR (qRT-PCR) assays or histological staining and analysis. Bronchoalveolar lavage was performed to collect samples to assay immune cell composition by flow cytometry or to determine cell-free concentrations of the eosinophil granule protein major basic protein (MBP) Some mice were further followed for morbidity (with disease severity assessed using a five-point scale) and survival, as previously described (30).

For adiponectin transcription quantification, mouse lungs were harvested, flash frozen, and used for total RNA extraction and analysis, as previously described (30). Primers for qRT-PCR were obtained from SABiosciences. Serum was separated from blood collected by cardiac puncture, and serum adiponectin levels were measured by ELISA, according to manufacturer’s instructions (R&D Systems). MBP ELISA was performed according to manufacturer’s instructions with 10-fold diluted cell-free bronchoalveolar lavage fluid (BALF) (MyBioSource). Cytokine protein concentrations from lung homogenates were determined by ELISA, according to manufacturer’s instructions (Peprotech [IL-1a, TNF] and R&D Systems [IL-6, IL-17A]). Lung injury was assessed by a lactate dehydrogenase assay of lung homogenates, according to manufacturer’s recommendations (Pierce). All flow cytometry reagents were obtained by BD Biosciences or eBioscience. Populations of cells were evaluated by flow cytometric analysis on a Guava EasyCyte 8HT benchtop flow cytometer (EMD Millipore), as previously described (30).

Seventy-two hours p.i., lungs were harvested after perfusion with 5 ml of PBS followed with 5 ml of 10% phosphate buffered formalin, inflated with formalin, and fixed overnight at room temperature. Tissue was embedded in paraffin after dehydration and sectioned at 4 μm. Modified Gomori modified methenamine (GMS) (Sigma-Aldrich) and H&E staining (Richard-Allan Scientific) were performed following deparaffinization.

mRNA sequencing and basic expression comparison analysis of mouse lung tissues from wild-type and adiponectin-deficient mice with IA at 72 h p.i. were performed by Otogenetics (Atlanta, GA) with 20 million reads per sample. Expression of mRNA was compared through paired-end (PE100-125) sequencing using HiSeq2500 (Illumina), alignment, and transcript compilation to produce fragments per kilobase of transcript per million mapped reads values used for differential expression analysis using DNAnexus Platform (DNAnexus). RNA sequencing (RNA-seq) data were deposited at Gene Expression Omnibus (https://www.ncbi.nlm.nih.gov/geo/), with the accession number GSE130456.

The resulting data sets were further analyzed by Ingenuity Pathway Analysis (Qiagen Bioinformatics) with assistance from the Indiana University–Purdue University Indianapolis Center for Computational Biology and Bioinformatics. A false discovery rate threshold of 0.05 and fold-change threshold of 2 (log2 [1]) were used to identify significant differences. The heat map was created using Heatmapper software (Wishart Research Group, University of Alberta).

Analysis of mouse flow cytometric data was performed with FlowJo software (Tree Star). Prism software was used for generation of graphs and figures and for statistical analyses (GraphPad). Unpaired t tests were used to measure statistical significance when two groups were compared, and one- or two-way ANOVA tests were used along with Tukey or Sidak posttests for multiple comparisons, respectively. Survival curves were analyzed with Mantel–Cox log-rank tests. Differences between experimental groups that resulted in a p value <0.05 were considered significant.

Recently, we observed that adiponectin inhibited airway eosinophil recruitment in response to inhalation of chitin, an immune-stimulating fungal cell wall component (32). Furthermore, others have reported enhanced inflammatory pathology in adiponectin-deficient mice in models of lung allergy and systemic bacterial infection (11, 17). Because excessive inflammation is detrimental in IA (25), we wanted to determine if adiponectin-deficient mice with IA exhibited increased morbidity and mortality. To accomplish this, we infected neutropenic mice with the clinical isolate Af293 and compared the survival, disease severity, fungal burden, and histopathology of wild-type B6 with adiponectin-deficient mice or mice that globally expressed high levels of adiponectin controlled by a β-actin promoter (AdipoTg) (33). In AdipoTg mice challenged with chitin or infected with A. fumigatus, no significant differences in chitin-mediated eosinophil recruitment, fungal burden, or severity of disease were observed at day 3 p.i. in comparison with wild-type B6 mice (Supplemental Fig. 1). In contrast, all adiponectin-deficient mice succumbed to invasive infection by day 5, whereas 50% of infected wild-type mice survived until 10 d p.i. (Fig. 1A). Similarly, adiponectin-deficient mice displayed more severe disease symptoms starting at day 2 p.i. (Fig. 1B). H&E sections displayed numerous bronchoalveolar foci with infiltration of leukocytes in both wild-type B6 and adiponectin-deficient mice (Fig. 1C, 1D). GMS staining showed invasive hyphal growth that constituted a higher percentage of the lungs in adiponectin-deficient mice in comparison with wild-type mice (Fig. 1E–G). This increase in fungal burden in adiponectin-deficient mice was confirmed by PCR quantification of fungal DNA (Fig. 1H). These data identify a protective role for adiponectin in IA that is not enhanced by global overexpression.

RNA-seq was performed to compare whole-lung transcription in adiponectin knockout (KO) and wild-type mice with IA. Expression of 252 genes was increased and expression of 544 genes was decreased in adiponectin-deficient mice relative to wild-type mice at day 3 p.i. (data not shown). Increased genes expressed included the inflammatory cytokines IL-6, IL-12, IL-17A, IL-17F, and IL-23 and additional chemokines and adhesion molecules associated with increased inflammation (Fig. 2A, 2B), whereas anti-inflammatory or extracellular matrix genes such as ido1 and mfap4, respectively, were decreased (Fig. 2C). Ingenuity Pathway Analysis identified IL-1β and TNF as upstream regulators and further associated these as causal networks by significant activation of downstream target genes (Fig. 2D). Notably, only Zfp36 was predicted to be inhibited as a causal network regulator, with 106 of 143 genes measured in the direction consistent with inhibition (Fig. 2D and data not shown). Zfp36 encodes the zinc finger protein tristetraprolin, a broadly expressed mRNA-binding protein that mediates poly(A) tail removal and thus promotes mRNA decay (34). Thus, our RNA-seq analysis results are consistent with an inflammatory phenotype in adiponectin KO mice with IA.

We next wanted to confirm transcription of inflammatory cytokines in wild-type and adiponectin-deficient mice at earlier time points p.i. The most significant increases in IL-1a, IL-6, IL-12b, IL-17A, IL-17F, and TNF were in adiponectin-deficient mice at day 2, with a more-modest increase in IL-1a in wild-type mice (Fig. 3A). Similar to neutrophil-depleted mice at day 0, naive undepleted adiponectin-deficient mice also exhibited similar lung cytokine gene expression compared with wild-type animals (Supplemental Fig. 2). In contrast to inflammatory cytokines, lung adiponectin mRNA was significantly decreased in wild-type mice after 1 and 2 d of infection compared with day 0 (Supplemental Fig. 3). Early production of inflammatory cytokines in the lungs of infected adiponectin-deficient mice suggested involvement of other innate cells, such as alveolar macrophages, and thus we compared alveolar macrophage production of TNF by intracellular cytokine staining followed by flow cytometric analysis. The total number of alveolar macrophages was similar between wild-type and adiponectin-deficient mice (Fig. 3B). However, the frequency, not total number, of TNF⁺ alveolar macrophages was significantly increased in adiponectin-deficient mice, as was the intensity of TNF staining (Fig. 3C–F). In contrast, intracellular cytokine staining for TNF and IL-17A in eosinophils, CD3⁺ cells, and CD11c⁺SiglecF⁻ cells did not detect significant differences (data not shown). Increased TNF protein was also observed in lung homogenates of adiponectin-deficient mice at 2 d p.i. (Fig. 3G). These results confirm increased inflammatory cytokine production in adiponectin-deficient mice, specifically in alveolar macrophages.

In a parallel study, we observed that chitin-mediated eosinophil recruitment was decreased by adiponectin (32). We thus wanted to determine if the lungs of adiponectin-deficient mice with IA displayed increased eosinophil recruitment and activation. Airway eosinophils were significantly increased in adiponectin-deficient mice with IA, whereas total leukocytes remained similar to wild-type controls (Fig. 4A, 4B). Although surface expression of the eosinophil activation marker CD86 (35) was only slightly increased on airway eosinophils in adiponectin-deficient infected mice, the number of CD86-expressing cells was increased relative to wild-type airway eosinophils (Fig. 4C, 4D). To determine if eosinophil degranulation was also increased, we compared the cytotoxic eosinophil granule protein MBP in the BALF supernatant of wild-type, adiponectin-deficient, adiponectin/eosinophil-deficient (Adipo^−/−ΔdblGATA1), or naive mice with constitutive expression of IL-5 under a lung promoter (elevated lung eosinophils; IL-5Tg). We observed that cell-free MBP was increased in the absence of adiponectin, with little or no MBP detected in lavage fluid from infected mice also deficient in eosinophils or naive IL5Tg mice (Fig. 4E). These results support an inhibitory role for adiponectin in lung eosinophil recruitment, activation, and degranulation in IA.

Because we previously reported that eosinophils were detrimental in IA (30, 31), we next wanted to determine if the increased disease severity observed in adiponectin-deficient mice was eosinophil dependent. Thus, we bred adiponectin-deficient (Adipoq^−/−) mice with eosinophil-deficient (ΔdblGATA1) mice to generate double-deficient Adipoq^−/−/ΔdblGATA1 mice. Adiponectin-deficient and Adipoq^−/−/ΔdblGATA1 mice with IA were then compared for morbidity, mortality, histopathology, and fungal burden. With this strategy, we observed peribronchoalveolar inflammatory foci in both groups, with less frequency in double-deficient mice at day 3 p.i. (Fig. 5A, top panels). Similarly, A. fumigatus hyphae were observed in inflamed areas in adiponectin-deficient lungs but were less frequent in mice that also lacked eosinophils (Fig. 5A, bottom panels). Mortality was significantly delayed, but not decreased, in double-deficient mice (Fig. 5B). Lung fungal burden assessed by lung fungal DNA or histological quantification was decreased in double-deficient mice at day 3 p.i. (Fig. 5C, 5D). To further determine the effects of adiponectin and adiponectin/eosinophil deficiency on inflammatory cytokine production, we compared lung transcription of IL-1a, IL-6, IL-17A/F, and TNF in wild-type (B6), adiponectin-deficient, and double-deficient mice at day 3 p.i. Of the cytokines examined, IL-6 and IL-17A were decreased in Adipoq^−/−ΔdblGATA1 compared with adiponectin-deficient mice (Fig. 5E). BALF concentrations of lactate dehydrogenase, a marker of lung injury, were increased between days 2 and 3 p.i. in all groups, with the most significant increase at day 3 p.i. in double-deficient mice (Fig. 5F), suggesting a delay in lung injury in the absence of eosinophils. These results suggest that eosinophils contribute to immune pathology but are not required for the increased mortality observed in adiponectin-deficient mice with IA.

Our previous results with adiponectin-deficient mice were found during invasive infection after neutrophil depletion. Thus, it is unknown how adiponectin deficiency affects inflammation and fungal clearance in response to aspiration of A. fumigatus conidia in nonneutropenic mice. Furthermore, although the results of previous studies showed increased inflammation in adiponectin-deficient mice in noninfectious models (11, 36, 37), it is still possible that the inflammatory phenotype in adiponectin-deficient mice is driven by increased fungal burden mediated by another mechanism. To examine these parameters further, we compared leukocyte recruitment and inflammatory cytokine transcription in neutropenic or nonneutropenic wild-type or adiponectin-deficient mice that aspirated preswollen and fixed conidia to prevent invasive fungal growth. Twenty-four hours after aspiration of 5 × 10⁷ swollen/fixed A. fumigatus conidia (isolate Af293), no significant changes in recruitment of total leukocytes, neutrophils, eosinophils, or alveolar macrophages were observed in adiponectin-deficient mice compared with wild-type mice, regardless of neutrophil depletion (Fig. 6A–D). Compared to nonneutropenic mice, total cells and eosinophils were greatly reduced in neutropenic animals (Fig. 6A, 6C). However, lung inflammatory cytokine transcription was increased in neutropenic mice, with the most significant increases in adiponectin-deficient mice (Fig. 6E), yet lung fungal burden remained similar, regardless of neutropenia or adiponectin deficiency (Fig. 6F). We also obtained similar results by aspiration with viable conidia in nonneutrophil-depleted mice, with no significant differences in lung histopathology, leukocyte recruitment, or cytokine expression 48 h postaspiration (Supplemental Fig. 4). Taken together, our results demonstrate that neutropenia results in enhanced inflammatory cytokine expression in response to swollen/fixed A. fumigatus conidia, most significantly in adiponectin-deficient mice.

In this study, we report increased mortality and fungal burden in adiponectin-deficient mice with IA (Fig. 1) further characterized by broad activation of inflammatory pathways (Figs. 2, 3), with a contribution from eosinophils that was essential for the timing of disease but not the outcome (Figs. 4, 5). Similar to previous studies with noninfectious models (11, 36), increased lung inflammation in adiponectin-deficient mice was not dependent on increased lung fungal burden, as the inflammatory phenotype was observed in response to swollen and fixed conidia and the absence of lung fungal growth (Fig. 6). However, inflammation was not increased in nonneutropenic mice, suggesting that a lack of neutrophils resulted in enhanced inflammatory cytokine-mediated pathology. The results of our study identify, to our knowledge, a novel potential for the adiponectin pathway in the development of therapeutics that dampen detrimental inflammation in IA patients.

The array of comorbidities associated with A. fumigatus infection and inflammatory pathology includes cystic fibrosis (38), allogeneic stem cell transplantation (39), chronic granulomatous disease (CGD) (40, 41), hyper IgE syndrome (42), and allergic fungal disease (43). Notably, these pathologies are not all directly associated with immune deficiencies. In some instances, immune pathology may occur in the presence of chronic inflammation, in others, during immune reconstitution following neutropenia/leukopenia in infected patients (25). It will be of interest to further investigate the role of adiponectin in these and other Aspergillus-associated diseases as well as in nonlung tissues and with other fungal diseases, such as candidiasis. Furthermore, the inhibition of systemic Listeria monocytogenes clearance due to enhanced inflammation in the bone marrow in the absence of adiponectin suggests a broad inhibition of inflammation in systemic response to other classes of pathogens (17). In addition to the expansion of disease model, fungal pathogen, and organ/system considerations, it will be important to confirm the inflammatory phenotype beyond the neutrophil depletion used in this study to include cyclophosphamide and cortisone acetate IA models. These strategies will allow a complete picture of the broad effect of adiponectin in dampening immune pathology in fungal infection.

The comparison of whole-lung transcription in mice with IA resulted in a comprehensive activation of inflammatory pathways with inhibition or predicted inhibition of anti-inflammatory pathways in adiponectin-deficient mice (Fig. 2). With pathway analysis, it was predicted that Zfp36 was inhibited in adiponectin-deficient mice (Fig. 2D), based on 106 out of 143 genes modulated in the direction consistent with Zfp36 inhibition (data not shown). Zfp36 encodes the RNA-binding and degrading protein tristetraprolin (34). Tristetraprolin binds to AU-rich elements and enhances the decay of TNF, IL-6, and IL-23 mRNA. Interestingly, inhibition of LPS-stimulated TNF secretion in macrophages by adiponectin was associated with loss of TNF mRNA stability and increased tristetraprolin binding (44), thus providing a direct connection between the adiponectin pathway and tristetraprolin activity in the dampening of inflammation. In addition, ido1, the gene encoding the tryptophan-degrading enzyme IDO, was decreased in the lungs of adiponectin-deficient mice with IA (Fig. 2C). IDO uses superoxide for tryptophan catabolism, resulting in accumulation of the downstream product l-kynurenine that activates downstream regulatory/anti-inflammatory pathways (41, 45). The lack of reactive oxygen species in p47^phox mice with CGD resulted in a hyperinflammatory phenotype in A. fumigatus–infected mice that was characterized by defective IDO activity (41). Notably, the hyperinflammatory phenotype in CGD mice was similar to that observed in adiponectin KO mice, with increased lung IL-17, IL-12, IL-23, and TNF (Figs. 2, 3). Determination of the mechanism of adiponectin-mediated protection in IA will require further investigation of the activation of anti-inflammatory pathways by adiponectin in response to fungal infection.

We observed that adiponectin deficiency resulted in an increase in an array of pleiotropic inflammatory cytokine transcription with multiple potential cellular sources in the lungs of mice with IA. Based on our previous results, we believed that eosinophils could be an important effector cell for detrimental lung inflammation in adiponectin-deficient mice (30–32). Interestingly, eosinophil recruitment and activation were increased (Fig. 4), yet eosinophil deficiency only delayed mortality, and immune pathology did not significantly alter survival in adiponectin-deficient mice (Fig. 5). This suggests that although eosinophils may act to enhance fungal invasion or inhibit fungal clearance, there are additional cells that contribute to immune pathology in the absence of adiponectin. Adiponectin inhibits activation of macrophages (46–49) and innate lymphocytes (4, 50) and inhibits lung epithelial cell apoptosis (51, 52). In our study, intracellular alveolar macrophage TNF was increased in the absence of adiponectin (Fig. 3C, 3D, 3F). Although we did not detect increased intracellular TNF or IL-17A in eosinophils or γδ T cells (data not shown), increased cytokine production by other cells is still likely. Mice with IA that lacked both adiponectin and eosinophils demonstrated decreased IL-6 and IL-17A in comparison with adiponectin deficiency alone (Fig. 5E); these cytokines may be produced directly by eosinophils (53, 54). Furthermore, the adiponectin receptors AdipoR1 and AdipoR2 are expressed by a variety of leukocyte subsets and stromal cells (4), yet it remains unknown if they are required for adiponectin-mediated inhibition of antifungal cytokine production. Additional studies will be required to determine the individual and combined effects of adiponectin on inflammatory cytokine production and effector functions of lung macrophages (alveolar and nonalveolar), innate lymphocytes, eosinophils, and epithelial cells and the associated adiponectin receptor and regulatory pathways in these cells that modulate detrimental pathology in IA.

Our results demonstrate that, despite a heightened inflammatory response in adiponectin-deficient mice with IA that is otherwise necessary for antifungal immunity, mortality and disease severity were increased. Several studies have described excessive inflammation that compromises host immune responses and fungal clearance in aspergillosis patients (25, 26). Interestingly, the inflammatory phenotype we observed after aspiration of A. fumigatus may require the absence of or impaired function of neutrophils, as increased cytokine production and fungal burden were not observed in undepleted mice that lacked adiponectin (Fig. 6). Neutrophils are critical for clearance of A. fumigatus spores and hyphae from infected lungs, and their absence resulted in increased cytokines in neutropenic patients (19, 55). In addition, coculture with neutrophils increased protease degradation of IL-1β and TNF production by human PBMCs in response to Candida albicans (56). Despite a clear protective role for neutrophils in response to A. fumigatus, the absence of neutrophils does not uniformly result in increased susceptibility or severity in response to other fungal pathogens, as increased cytokine production in neutrophil-depleted mice infected with Cryptococcus neoformans resulted in decreased fungal burden and severity of disease compared with undepleted mice (57). Based on these findings, it is possible that the role of adiponectin in protection from detrimental inflammation is more critical for individuals with specific immune deficiencies and associated pathogens that require neutrophil-mediated clearance.

A recent study reported inhibition of clearance of the bacteria L. monocytogenes from adiponectin-deficient or obese mice due to increased bone marrow inflammation and reduced granulopoiesis (17). To our knowledge, our study is the first to discover a role for adiponectin in protection from fungal infection. Based on our results with aspiration of swollen/fixed conidia in neutropenic mice (Fig. 6) and published results with adiponectin-deficient mice in noninfectious models of lung inflammatory pathology (12, 13, 15, 16), we hypothesize that excessive inflammation exacerbates lung immune pathology that inhibits effective fungal clearance in the absence of adiponectin. To date, no alternative explanation is well supported by published data. Although it is difficult to disassociate immune pathology from pathologies directly linked to invasive fungal growth, our identification of potential tristetraprolin and IDO regulation by adiponectin provides an opportunity to directly observe the effect of stimulation of these pathways on survival and fungal clearance in adiponectin-deficient mice with IA. We will manipulate these pathways in future studies to more clearly separate the effects of adiponectin on inflammatory and noninflammatory pathologies in fungal infection. In addition, we are currently using models of obesity to identify common immune effector and regulatory pathways with adiponectin-deficient mice that mediate pathology in IA. We encourage others to initiate studies that examine the role of adiponectin in the programming of immunity to a wide array of microbial pathogens that will determine the potential of the adiponectin pathway as a source of novel immune-modulating molecules for patients that do not respond to conventional pharmacotherapy.

We thank Joe Lewis for animal care.

The authors have no financial conflicts of interest.