Influenza A virus directly modulates mouse eosinophil responses

Abstract

Allergic asthma and influenza are common respiratory diseases with a high probability of co-occurrence. During the 2009 influenza pandemic, hospitalized patients with influenza experienced lower morbidity if asthma was an underlying condition. We have previously demonstrated that acute allergic asthma protects mice from severe influenza and have implicated eosinophils in the airways of mice with allergic asthma as participants in the antiviral response. However, very little is known about how eosinophils respond to direct exposure to influenza A virus (IAV) or the microenvironment in which the viral burden is high. We hypothesized that eosinophils would dynamically respond to the presence of IAV through phenotypic, transcriptomic, and physiologic changes. Using our mouse model of acute fungal asthma and influenza, we showed that eosinophils in lymphoid tissues were responsive to IAV infection in the lungs and altered surface expression of various markers necessary for cell activation in a niche-specific manner. Siglec-F expression was altered in a subset of eosinophils after virus exposure, and those expressing high Siglec-F were more active (IL-5Rα^hiCD62L^lo). While eosinophils exposed to IAV decreased their overall transcriptional activity and mitochondrial oxygen consumption, transcription of genes encoding viral recognition proteins, Ddx58 (RIG-I), Tlr3, and Ifih1 (MDA5), were up-regulated. CD8⁺ T cells from IAV-infected mice expanded in response to IAV PB1 peptide-pulsed eosinophils, and CpG methylation in the Tbx21 promoter was reduced in these T cells. These data offer insight into how eosinophils respond to IAV and help elucidate alternative mechanisms by which they regulate antiviral immune responses during IAV infection.

Introduction

Eosinophils are generated in the bone marrow from multipotent and lineage-restricted hematopoietic progenitors. Although a small number of eosinophils enter the bloodstream after differentiation, most remain in the bone marrow in healthy individuals. In contrast, eosinophils are released from the bone marrow en masse following exposure to allergen or infection with helminths.¹ For this reason, eosinophils have had ascribed functions in diseases such as asthma, irritable bowel syndrome, and helminth infection.^2,3 More recently, however, eosinophils have been suggested to have a much broader role in immunity, with reports of antibacterial and antiviral functions as well as contributions to both innate and adaptive immune responses.^4–8

As modulators of adaptive immune responses, activated eosinophils function as auxiliary antigen presenting cells (APCs) through up-regulation of MHC-I and MHC-II in response to stimuli,^9–13 increasing the portfolio of cells that can activate CD8⁺ and CD4⁺ T cells during infection. Eosinophil granules contain a diverse complementation of pre-synthesized cationic proteins (eosinophil peroxidase, major basic protein, eosinophil-associated RNases), as well as a variety of cytokines^14,15 including those with defined roles during viral infection and pathogenesis.^16–18 Virus-induced degranulation has been observed during RSV infection of human eosinophils and pneumovirus infection of mouse eosinophils.¹⁹ Recently, we reported that eosinophils utilize piecemeal degranulation following exposure to influenza A virus (IAV), suggesting virus-induced cytokine release by eosinophils is, at least in part, a deliberate and selective process.¹²

While it is not surprising that eosinophils are recruited during viral infection given the multiple roles they play in host defense,^20–22 knowledge lacunae exist with regard to their direct and indirect functions in response to viruses. Since eosinophils are not generally considered as part of the immune armamentarium against respiratory viruses, and their infiltration is significantly lower compared to that of other leukocytes during infection, most virus-based studies have ignored them. Our own work with IAV shows very low eosinophil presence in the airways during early time points of infection but increasing numbers that correlate with the recruitment of T cells into the lungs²³ conceivably to modulate wound repair processes. However, underlying allergic asthma presents a unique environment where large numbers of eosinophils and T_H2 cytokines are already present in the airways upon virus exposure. Asthma is a common affliction affecting ∼8% of people in the United States²⁴ and is often considered a risk factor for increased susceptibility to respiratory infections²⁵ possibly due to subdued antiviral immune responses in asthmatics.^26,27 However, no differences in infection susceptibility have been noted between bronchial epithelial cells from asthmatics and non-asthmatics,^23,28,29 although viruses are known triggers of asthma exacerbations especially in children.^30,31 Unfortunately, clinical studies seldom report the impact asthma attacks have on respiratory viral diseases. During the 2009 influenza pandemic, asthmatics were at increased risk for hospitalization³² but had less severe influenza morbidity compared to non-asthmatics^33–35 as previously reviewed.³⁶ In order to investigate the mechanistic impact of virus-induced asthma exacerbation on influenza pathogenesis, we built a murine model system in which we showed that while viral replication was not affected, and viral clearance was accelerated in allergic mice infected with the pandemic influenza virus strain A/CA/04/2009²³ (possibly due to more robust CD8⁺ T cell responses that we later identified were elicited through eosinophils¹²).

Immune responses to viruses are dictated both by viral characteristics and the underlying immune status of the host. In this report, using our model of fungal asthma and influenza, we investigated the localization of eosinophils and their phenotypic and functional characteristics when exposed to IAV both in vivo and in vitro. We found eosinophils in various lymphoid compartments in addition to the site of infection during asthma, influenza, and co-morbidity. Eosinophil presence in these niches varied by disease state, as did their activation status. Interestingly, virus exposure caused the transcriptional down-regulation of a large number of genes including those involved in chemotactic sensing and a reduction in mitochondrial respiration. Virus-peptide exposed eosinophils induced epigenetic changes in virus-specific CD8⁺ T cell populations from the lung. Cumulatively, this study suggests that eosinophils respond dynamically to the presence of IAV and may represent an additional avenue to the host’s antiviral defenses that may be beneficial to asthmatics.

Methods

Ethics statement

All experiments with infectious agents and mice were performed in accordance with protocols approved by the Institutional Biosafety Committee and the Institutional Animal Care and Use Committee of the University of Tennessee Health Science Center.

Influenza A viruses

Pandemic clinical isolate A/California/04/2009 (“pH1N1”) was kindly provided by Richard Webby at St Jude Children’s Research Hospital (Memphis, TN), and propagated in MDCK.2 cells (ATCC, Manassas, VA, USA). Mouse adapted strain A/PR/8/1934 (PR8) was cultured in embryonated chicken eggs. Strains were sequenced to confirm genetic integrity of hemagglutinin and neuraminidase genes, titered using the TCID₅₀ method in MDCK.2 cells, and frozen in single-use aliquots at −80°C.

Allergic asthma mouse model

Mice were housed under specific pathogen-free conditions at the University of Tennessee Health Science Center, Memphis TN, and received autoclaved water and food ad libitum. Mice were rendered allergic to Aspergillus fumigatus as previously described.^37,38 Briefly, 6 week-old female C57BL/6J mice from Jackson Laboratories (Bar Harbor, ME, USA) were acclimatized to the in-house specific pathogen-free animal facility for 1 week. Mice were sensitized to A. fumigatus by intraperitoneal and subcutaneous injections with 10 µg A. fumigatus antigen (Greer Labs, Lenoir, NC, USA) adsorbed in Imject Alum (Pierce Biochemicals; Thermo Fisher Scientific, Waltham, MA, USA) and, after 2 weeks rest, 3 weekly intranasal instillations of 20 µg A. fumigatus Ag. One and 3 weeks after the final intranasal sensitization, mice were challenged with live, airborne A. fumigatus fungal conidia (ATCC) for 10 min. Recognizing that asthma is a multifaceted and complicated syndrome that does not spontaneously develop in mice, herein, we use the term “asthma” loosely to reflect the asthma-like changes that occur in the animals after A. fumigatus sensitization and challenge.

IAV infection mouse model

Mice were rested for a week after the final conidia exposure, then lightly anaesthetized with isoflurane and challenged intranasally with 1000 TCID₅₀ A/CA/04/2009 in 50 µl PBS. Mice were weighed before infection and daily thereafter, and sacrificed 7 days post infection by CO₂ asphyxiation with adjunct cervical dislocation. The term “Flu” is used to refer to mice that develop influenza as a consequence of infection with IAV. Bronchoalveolar lavage was performed by 2 washes with 1 ml PBS and blood was removed from the chest cavity. Marrow was extracted from the cleaned right tibia and femur by centrifugation at 1500 × g for 5 min. Spleen and lungs were excised and dissociated in PBS or Miltenyi Biotec Lung Dissociation reagents, respectively, using GentleMACS Octo Dissociator (Miltenyi Biotec, Bergisch Gladbach, Germany). Single cell suspensions were prepared from harvested mediastinal lymph nodes (MLN) and thymus by filtering through a 40 µm cell mesh with gentle pressure applied by a syringe plunger. Samples were enumerated with trypan blue on a Countess Automated Cell Counter (Life Technologies, Carlsbad, CA, USA). A separate group of mice had distal and proximal lobes of the right lungs collected and snap frozen in liquid nitrogen prior to freezing at −80°C for later use.

Determination of select cytokines in the lungs

Collected distal and proximal lobes of the right lungs were mechanically homogenized in PBS containing cOmplete™ protease inhibitor cocktail (Millipore Sigma, Burlington, MA), clarified by centrifugation, and supernatant stored at −80°C until use. Lung homogenates were used to determine the concentration of cytokines using kits from R&D Systems (Minneapolis, MN) and read on a Luminex-MagPix machine with xPONENT 4.2 software. Data are represented as the mean and standard deviation (SD) of the fold change over levels in naïve animals.

Morphological identification of eosinophils

Cytospins were prepared (Thermo Shandon, Waltham, MA, USA), differentially stained and examined at 400× magnifications (40× objective). Eosinophils were identified as cells possessing multi-lobed nuclei and red punctate stained cytoplasm. Numbers of eosinophils in 5 fields were averaged and represented graphically as eosinophils per high-powered field (HPF) with SD.

BMdEos and IAV infection experiments

Eosinophils were derived from female C57BL/6 bone marrow for 14 days in eosinophil growth media.³⁹ Sixteen hours before the commencement of the experiment and thereafter, the media was supplemented with mouse GM-CSF (5 ng/ml, PeproTech, Rocky Hill, NJ, USA) to promote the up-regulation of MHC-II. BMdEos were incubated in 0.2 mg/ml A. fumigatus antigen (Greer Labs), or PBS (pH 7.2), for 1 h. Cells were pulsed with a multiplicity of infection (MOI) of 0.1 IAV for 1 h, washed with PBS, and incubated for 24 h in media supplemented with 1 µg/ml TPCK-trypsin (Worthington Biochemical Corporation, Lakewood, NJ, USA). Cells were then transferred to media without TPCK-trypsin and incubated at 37°C/5% CO₂. After 2 days, cells were harvested for flow cytometric staining and analysis.

Measurement of genes associated with antiviral immunity

Mouse bone marrow (BM) was used to induce eosinophils (Eos) and dendritic cell (DC) differentiation. BMdEos and BMdDC were stimulated with A/PR/8/34 (MOI = 1, or media) for 16 h prior to RNA extraction. IAV infection of BMdEos and BMdDC were conducted as described,^40–42 and infected cells were incubated in complete DMEM (cDMEM)/7.5% BSA in 6-well plates for the remainder of the experiment.

Extraction of total RNA by TRIzol (Invitrogen, Carlsbad, CA, USA), cDNA synthesis by reverse transcription (RT) and quantitative PCR (qPCR) were performed as described elsewhere.^43–46 Briefly, total RNA was extracted from cells using TRIzol after the indicated treatment or infections, followed by RT with random hexamer primers and GoTaq qPCR Master Mix (Promega, Madison, WI, USA). qPCR was performed on an iCycler IQ5 real-time PCR system (Bio-Rad, Hercules, CA, USA). The gene-specific oligonucleotide primers used for qPCR are shown in Table 1. Primers used to amplify the 28S rRNA housekeeping gene were described previously.⁴⁵ The relative abundance of each target was normalized to mouse 28S rRNA.

Flow cytometry (FCM)

Fc receptors were blocked by incubation with human gamma globulin (Sigma–Aldrich, St. Louis, MO, USA) and extracellular Ags of interest were stained by incubation with fluorochrome-labeled Abs (Table 2) for 30 min on ice. All antibodies were used at an optimized working concentration of 4 µg/ml. PIR-A/B was conjugated to PE/Cy5 using Abcam PE/Cy5 Conjugation Kit according to the manufacturer’s instructions (Abcam, Cambridge, MA, USA). Cells stained with the matched concentrations of relevant isotype Abs were used to determine background binding. Stained cells were acquired on a BD LSR Fortessa and subsequently analyzed using FlowJo V10 (Treestar, Ashland, OR, USA). Due to niche variation in background levels determined by unstained and isotype controls, gates for each Ag analyzed were adjusted on a niche-to-niche basis. This must be taken into consideration when comparing Ag levels on eosinophils recovered from different niches. Geometric mean fluorescence intensity was used to compare expression of Ags in different groups within the same niche, with isotype control fluorescence subtracted mathematically. Samples where negative values resulted from subtracting the isotype control fluorescence were designated a value of 0. Up to 1 outlier per group were eliminated using Grubb’s test if deemed an outlier.

Microarray analysis

BMdEos derived from individual mice were used as biologically distinct replicates for +/− virus conditions with 5 mice each. On day 14 of differentiation, BMdEos were pulsed for 1 h with PR8 at MOI 0.1. Cells were washed in PBS and cultured for a further 23 h at a density of 1 million live cells/ml in media supplemented with 1 µg/ml TPCK-Trypsin (Worthington) and 5 ng/ml mouse GM-CSF. At 6 h prior to harvest, 1 µg/ml GolgiPlug (BD Biosciences) was added to the cells to reduce the release of endogenous RNases. Eosinophils infected with an MOI of 1 or 3 underwent piecemeal degranulation (as shown previously¹²), resulting in poor RNA integrity and yield.

RNA was extracted from mock or PR8-pulsed BMdEos using RNeasy Mini (Qiagen Inc., Germantown, MD, USA) with β-mercaptoethanol added to the RLT buffer as directed by the manufacturer and resuspended in nuclease-free water. RNA quality was determined using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). Microarray analysis was performed using Affymetrix Mouse Gene 2.0 ST arrays according to the manufacturer’s recommended procedures (Affymetrix, Santa Clara, CA). Agilent Bioanalyzer testing and microarray hybridization were performed at the Hartwell Center, St. Jude Children’s Research Hospital (Memphis, TN). Microarray expression data underwent background correction, quantile normalization, and probe set signal summarization by the RMA⁸⁴ and PLIER⁸⁵ methods, and were then subjected to independent t-tests with and without the Benjamini-Hockberg false discovery rate (B-H FDR) correction method after removal of probe sets with group average log₂ signal values < 1 for both sample groups. Differentially expressed (DE) probe sets were identified from RMA and PLIER data sets by filtering consecutively for all 4 of the following criteria: (1) for B-H FDR corrected P-values less than or equal to 0.1, (2) independent t-test P-values less than or equal to 0.05, (3) absolute log₂-fold change values greater than 0.585 (1.5-fold change), and (4) at least 1 group average probe set signal value higher than the maximal group average probe set signal value for Ddx3y (which should be absent in cells from female mice). The PLIER method yielded a greater number of DE probe sets than the RMA method (PLIER = 610 probe sets vs. RMA = 176 probe sets), but 83% of the RMA DE probe sets were concordant with the PLIER DE results and 91% of RMA DE probe sets that detect known protein-coding gene transcripts (41/45) were concordant with those in the PLIER DE data set (173 probe sets for known protein-coding gene transcripts). The larger PLIER DE data set was used for subsequent pathway analysis. Microarray data has been deposited into the Gene Expression Omnibus (GEO) database under GEO submission GSE140580. Supplementary Tables provide raw and analyzed data.

Unsupervised hierarchical clustering and heat map generation were performed in TM4 MeV⁸⁶ using row mean centered log₂-transformed signal values. Probe set clustering was performed by average linkage based on Pearson correlation as the similarity metric. Pathway enrichment analyses using KEGG, Panther, Reactome, and Wikipathway databases and transcription factor target enrichment using the mSigDB database were performed in Webgestalt⁴⁷ based on the list of differentially expressed genes encoding identified proteins (Supplementary Table S2). We noted that the non-normal distribution of genes listed in Supplementary Table S2 due to the large representation of vomeronasal receptor and olfactory receptor genes skews the calculation of hypergeometric test P-values and lowers the apparent significance of enrichment of pathways containing other genes. However, the arbitrary removal of these genes from the query and reference lists did not change the pathways identified by these analyses, and thus they are provided as is (Supplementary Table S3). Protein-protein association analysis was performed in STRINGdb⁴⁸ based on a list of differentially expressed genes encoding identified proteins that was filtered to remove G-protein coupled receptors and murine urinary proteins (Supplementary Table S4); unlike Webgestalt enrichment analysis, STRING analysis does not depend upon a user-selected reference gene list. The pathway map shown in Fig. 4B was a compilation of the results obtained from Webgestalt and STRING analyses, the pathway maps and supporting documentation from the KEGG, Panther, Reactome, and Wikipathway databases, and additional references on the interaction of HSF1 and MAP kinase pathways^49,50 and the interaction of NF-Y with AP-1.^51,52

Glycolytic and mitochondrial respiration tests

OCR and ECAR were measured by using the Seahorse XFe-96 Analyzer (Agilent Santa Clara). Agilent (Santa Clara) Seahorse XF Cell Mito Stress Test and XF Glycolysis Stress Tests were used to assess eosinophil bioenergetics following the manufacturer’s protocols. Cell viability was determined using the Nexcelom cellometer (Nexcelom Bioscience Lawrence, MA) with aridine orange and propidium iodide staining and the assay was optimized per titration (Supplementary Fig. 1). Requisite numbers of live cells were suspended in freshly prepared seahorse assay medium supplemented with 2% FBS. The appropriate supplements were added to the Seahorse assay medium for glyco or mito stress test (2 mM final l-glutamine or 10 mM glucose, 2 mM l-glutamine, and 1 mM sodium pyruvate, respectively) followed by pH adjustment to 7.4. Cells were incubated for 1 h at 37°C. Manufacturer instrument settings were used for rate determination (28 reads per well). Oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) were measured at 6.5 min intervals for 73 min resulting in 3 measures per response as recommended by manufacturer. Optimal cell density for infected and uninfected eosinophils (300,000 cells per well), and optimal concentration of oligomycin (1 µM) and FCCP (2 µM) were determined prior to experiments using titration assays.

Kinetic profiles were generated in GraphPad Prism from normalized aggregate data with technical variability enumerated with error bars reflecting the standard error of the mean (SEM). Data were normalized and assay parameters calculated with the Seahorse report generators then experimental data were graphed in GraphPad. Assay parameters from 3 Mito Stress assays and 2 Glyco Stress assays were then grouped into control (uninfected) and IAV (PR8 virus infected) groups and mean experimental summary data graphed with experimental variability enumerated with error bars reflecting the SD.

Epigenetic experiments

BMdEos were differentiated from C57BL/6 bone marrow and pulsed with 2 µM PB1_703–711 peptide SSYRRPVGI (or the same volume of media for controls) in eosinophil growth media supplemented with 5 ng/ml GM-CSF for 1 h with periodic gentle rocking. Cells were washed twice with media. CD8⁺ T cells were enriched from lungs and spleens of pH1N1-infected C57BL6/J mice (Day 10) by negative selection using the EasySep™ Mouse CD8⁺ T Cell Isolation Kit (StemCell Technologies, Vancouver, BC, Canada). Purity was confirmed by flow cytometry to be ∼90%. Purified CD8⁺ T cells and PB1-pulsed BMdEos (or controls) were co-incubated at a 1:1 ratio for 2 days with 5 ng/ml GM-CSF in 96-well round bottom cell culture plates.

Cells were stained with PB1-Tetramer-PE (a kind gift from Paul Thomas, St. Jude Children’s Research Center) and anti-CD44-APC/Cy7, anti-CD62L-FITC and anti-CD8-PerCP Abs. PB1-specific CD8⁺ T cells were sorted using a FACS Aria (St Jude Children’s Research Center), and genomic DNA was extracted using a Qiagen DNeasy kit. Genomic DNA was bisulfite treated using the EZ DNA methylation kit (Zymo Research, Irvine, CA, USA), PCR amplified with Tbx21 locus-specific primers.⁵³ The amplicon was cloned into the pGEM-T TA cloning vector (Promega, Madison, WI, USA) overnight at 4°C, then the cloning vector was transformed into XL10-Gold ultracompetent E. coli bacteria (Stratagene, San Diego, CA, USA). Individual bacterial colonies were grown overnight over Luria-Bertani (LB) agar containing ampicillin (100 mg/l), X-gal (80 mg/l), and IPTG (20 mM). White colonies were selected and sub-cultured into LB broth with ampicillin (100 mg/l) overnight at 37°C; the cloning vector was purified; and the genomic insert was sequenced. The DNA methylation frequency was calculated as described before.⁵³

Statistical analyses

Each treatment group contained 5 animals in the in vivo studies and replicate numbers for in vitro studies are indicated in the figure legends. All studies were independently repeated for scientific rigor and reproducibility. Data were analyzed using GraphPad (San Diego, CA) Prism v6.07 and tests used to determine statistical significance are provided in the figure legends.

Results

Eosinophil presence in various niches is generally not affected by IAV infection

Eosinophils migrate from the bone marrow to the bloodstream during acute atopic asthma, and many traffic to the lungs and airways, which are the sites of allergen exposure. To investigate how influenza affects eosinophil migration during fungal asthma, C57BL/6J mice were rendered allergic to A. fumigatus, a common fungal allergen. During the acute phase of pulmonary inflammation, allergic mice or naïve counterparts were challenged intranasally with pH1N1 (Fig. 1A). Uninfected allergic mice served as asthma controls. Our previous characterization of this model with a complete time course of IAV kinetics showed that eosinophils were minimal in the airways after IAV infection alone.²³ Because the peak inflammation in the Flu Ctr mice occurred at Day 7 post infection that also was when the most eosinophils were observed in the BAL of these mice, the number of eosinophils at this timepoint were sufficient to measure in detail by FCM. Therefore, we elected day 7 post infection to compare eosinophils between experimental conditions. We have found that the viral titer and overall airway inflammation (with neutrophils) are higher in the Flu Ctr animals compared to co-morbid mice at this timepoint.²³

Cytokines are key regulators of cell functions within the lungs after stimulation. Since our model employs A. fumigatus that triggers a mixed T_H1/T_H2/T_H17 cytokine profile while IAV is largely T_H1 biased, we measured several cytokines deemed important to immune responses during infection with each agent and compared them as a fold change over cytokine levels in the naïve animals (Fig 1B). While canonical T_H1 cytokines, TNFα, IL-1α, and IFN-γ were elevated in the Flu Ctr animals, they also had significant levels of GM-CSF, CCL11, IL-33, and IL-5 (Fig. 1B) suggesting that at the cytokine milieu in these animals may be changing toward a T_H2 profile at this timepoint that correlates with the infiltration of T cells into the lungs.²³ In general, mice with fungal asthma alone and those with asthma and influenza had a similar cytokine profile except that co-morbid mice had significantly more CC11 and IL-33 in the lungs (Fig. 1B). Fungus or IAV exposure did not affect IL-17 levels in the lungs at this timepoint (Fig. 1B).

Since both differential staining and FCM are used to identify eosinophils, data may not always correlate between the 2 techniques.⁵⁴ Therefore, we determined the number of eosinophils in the bone marrow, spleen, lungs, BAL, MLN, and thymus by enumerating xanthene-positive lobed leukocytes on cytospin preparations (Fig. 1C) as well as FCM staining normalized to the total cell number in each niche (Fig. 1D). At the selected timepoint, eosinophils were identified by staining pattern in the cytospin preparations as shown in the micrograph and found in all tissues investigated (Fig 1C). The majority of eosin-positive eosinophils in allergen-exposed mice with/without IAV were found in the BAL and lungs, while most eosinophils in the Flu Ctr mice remained in the BM (Fig. 1C). Mature eosinophils can be identified by high scatter properties and co-expression of CCR3 and Siglec-F by FCM. Interestingly, this enrichment of eosinophils in the BM of Flu Ctr mice was not observed when eosinophil numbers were quantified by FCM (Fig. 1D), but the predominance of eosinophils in the BAL (not at statistical significance) and lungs in mice with asthma was reproducible between methods.

Eosinophil Ags are differentially regulated in mice with fungal asthma and influenza

We examined the surface expression of several Ags important in eosinophil development, survival, or trafficking. Organs harvested from mice with asthma, influenza, or both morbidities were examined using FCM. Eosinophil numbers in each niche are shown in Fig. 1 as Siglec-F⁺CCR3⁺ singlets; herein, we measured Ag levels on the eosinophil surface as the geometric mean fluorescent intensity to determine how expression patterns were altered by each disease condition (Fig. 2A). The human paralog of Siglec-F, Siglec-8, mediates eosinophil apoptosis,^58,59 and Siglec-F null mice have increased eosinophilia in the lungs, blood, and BM after allergen suggesting that it regulates cell survival in mouse eosinophils.⁶⁰ Eosinophils in the MLN, BAL, and lungs of allergic mice expressed more Siglec-F compared to eosinophils in Flu Ctr mice. Except in the MLN, there were no differences in Siglec-F expression in allergic mice with and without IAV infection (Fig. 2B). IL-5Rα (CD125) expression is associated with eosinophil-committed progenitors and mature eosinophils, IL-5Rα expression alone is insufficient to identify eosinophils.⁵⁴ Our analysis showed that eosinophils in the spleen had the highest IL-5Rα expression compared to the other niches (Fig. 2B). Eosinophils in the thymus, MLN, and BAL of Flu Ctr mice had more IL-5Rα compared to allergic mice +/− IAV (Fig. 2B).

In the bone marrow, PIR-A and -B expression regulate eosinophil apoptosis during maturation, with PIR-B counteracting the pro-apoptotic signal provided by PIR-A.⁶¹ Peripheral eosinophils retain both PIR-A and PIR-B expression in the allergic lung, and are a prominent source of PIR-B.⁶² While BM and thymic eosinophils had minimal PIR-A/B expression, those in the spleen and MLN had increased expression in all 3 treatment groups with some indication that IAV presence was a regulator of expression as Flu Ctr and A+Flu had similar trends in these niches (Fig. 2B). Interestingly, mice with asthma had significantly lower PIR-A/B expression compared to Flu Ctr mouse eosinophils in the BAL and lungs (Fig. 2B). Surface expression of CD62L (l-selectin), which is involved in eosinophil rolling along the vascular endothelium and is negatively regulated by allergy-associated factors such as IL-5 and TSLP,^63–65 was equivalent between groups in all niches except in the BAL wherein eosinophils in A+Flu group had significantly higher expression (Fig. 2B). Although CD62L has been shown to be downregulated on eosinophils in the BAL of patients with allergic asthma,^63,66 we did not observe a similar phenotype in our mouse model.

Recently, we showed that IAV-pulsed eosinophils can activate CD8⁺ T cells by presenting viral Ags on MHC-I.¹² Interestingly, eosinophils from the Flu Ctr mice had higher expression of MHC-I in the MLN, BAL and lungs compared to the allergen-exposed mice (Fig. 2B). Eosinophils can also activate CD4⁺ T cells by MHC-II-restricted Ag presentation.⁹ In humans, eosinophils express MHC-II in the BM but lose expression by maturation.⁶⁷ However, both mature human and mouse eosinophils can upregulate MHC-II upon activation by cytokines or exposure to pathogens.^9–11,68 In our model, we observed that MHC-II was mostly expressed on lymphoid organ eosinophils (except BM) and that eosinophils from the Flu Ctr mice had higher expression of MHC-II in the BAL compared to the allergic mice +/− IAV (Fig. 2B).

Siglec-F^hi BMdEos up-regulate IL-5Rα following infection with IAV

A shift in Siglec-F expression on eosinophils after IAV exposure¹² and altered expression between niches (Fig. 2) may be suggestive of subpopulations of eosinophils that may have differing functions. We found that the Siglec-F high expressing eosinophils were reduced after IAV (Eos+Flu) exposure (Fig. 3A). While A. fumigatus Ag stimulated eosinophils (AgEos) had a slight reduction in the Siglec-F^hi population compared to cells that were not exposed to the Ag, they also down-regulated Siglec-F after IAV exposure (Fig. 3B). We then investigated how other prominent surface markers necessary for eosinophil survival and migration were regulated on the cell surface based on high and low Siglec-F expression (Fig. 3C).

As IL-5 is a crucial growth factor for eosinophils, up-regulation of IL-5Rα on the cell surface is important for survival. We found that eosinophils that did not up-regulate Siglec-F had higher expression of IL-5Rα compared to Siglec-F^hi eosinophils (Fig. 3D). While the Siglec-F low expressors reduced the expression of IL-5Rα after IAV exposure, the Siglec-F^hi eosinophils up-regulated the expression of this cytokine receptor, and pre-exposure to A. fumigatus Ag did not affect this pattern of IL-5Rα expression (Fig. 3D). Since PIR-A/B can also support eosinophil survival and responsiveness to IL-5, we measured its expression after IAV exposure and found the opposite trend where the Siglec-F^hi eosinophils had higher expression of this marker compared to the Siglec-F low expressors (Fig. 3E). As before, allergen stimulation did not have an impact on this dichotomous expression patterns (Fig. 3E). Expression of CD62L is down-regulated on activated eosinophils,^69,70 and we found that the Siglec-F^lo population had lower CD62L expression after IAV irrespective of allergen exposure (Fig. 3F). In contrast, the Siglec-F^hi eosinophils up-regulated CD62L after IAV in cells that were not pre-exposed to A. fumigatus Ag (Fig. 3F).

Virus exposure alters the eosinophil transcriptome

Since eosinophils function as APCs during IAV infection,¹² we examined levels of eosinophil mRNA that encode proteins implicated in IAV recognition, including dsRNA binding proteins retinoic acid-inducible gene I (RIG-I) and MDA-5,^71,72 and TLR3, the latter of which is positively regulated by IAV infection⁷³ in comparison to a professional APC (Fig. 4A). Transcripts for Ddx58 and Ifih1, genes encoding RIG-I and MDA-5 respectively, were significantly upregulated in the IAV-exposed BMdDC controls, while Tlr3 mRNA increased but was not significant (Fig. 4A). Differential expression of these genes was similar in IAV-exposed eosinophils, albeit at lower expression levels (Fig. 4A).

These data, together with our previous finding that eosinophils are susceptible to IAV infection¹² suggest that IAV may impact eosinophils at the transcriptional level. Mature BMdEos were virus/mock exposed for 24 h and global transcriptional profiles, determined using Affymetrix Mouse Gene 2.0 ST arrays, were compared for gross changes of transcript levels in the infected cells (Fig. 4B). The vast majority of transcripts that were significantly altered in IAV-infected eosinophils (604 of 610) were down-regulated compared to mock-infected eosinophils (Fig. 4B). When considering only known protein-coding gene transcripts (173), only 1 (Dusp6) was upregulated. The most dramatically reduced transcripts encode heat shock proteins (Hsp1a, Hsp1b, Hsph1) and heme oxygenase-1 (Hmox1), while many of the remaining down-regulated transcripts encode vomeronasal type 1 and 2 receptors (66 genes), olfactory receptors (38 genes), and major urinary proteins (11 genes). Transcripts for 2 components of the AP-1 transcription factors (Jun, Fosl1) were also down-regulated. Compilation of the results from pathway enrichment, transcription factor target enrichment, and protein–protein interaction analyses (Supplementary Fig. S2 and Table S3) suggest that these transcript changes may alter the activity of MAP kinase signaling pathways and its regulation of heat shock protein and Hmox1 gene expression (Fig. 4C).

Eosinophils reduce mitochondrial respiration in response to IAV

Eosinophil bioenergetics is poorly characterized with a single recent report on human peripheral blood eosinophils,⁷⁴ and no information in the context of viral infection. Eosinophils and neutrophils contain few mitochondria. In neutrophils, energy is derived almost exclusively from glycolysis, with mitochondrial involvement limited to the initiation of apoptosis.^75–78 Similarly, eosinophil mitochondria were initially considered to not respire and primarily function in the initiation of apoptosis.⁷⁹ However, with the use of more advanced equipment, Porter and colleagues demonstrated that human peripheral blood eosinophils do utilize mitochondrial respiration, exhibiting higher basal mitochondrial respiration, ATP-linked respiration, maximum respiratory capacity, and spare respiratory capacity compared to human neutrophils.⁷⁴ Since we observed that virus exposure caused a general transcriptional down-regulation in eosinophils, we hypothesized that IAV-exposure will also reduce the bioenergetic processes in eosinophils.

To quantify eosinophil mitochondrial respiration efficiency, oxygen consumption, and glycolysis following pulsing with IAV, we performed mitochondrial and glycolytic stress assays using the Xfe96 bioanalyzer. While the cell viability was affected by the Seahorse assay media, there was no difference in the viability between the uninfected (84.00 ± 7.70%) and virus-infected (88.20 ± 5.89%) groups in the assays performed. IAV exposure lowered the basal respiration rate in eosinophils and overall, led to a reduction in mitochondrial respiration (Fig. 5). Similarly, maximal respiration measured in the presence of FCCP, and the spare capacity calculated after the addition of antimycin and rotenone, were both lower in eosinophils that were exposed to IAV (Fig. 5). Non-mitochondrial respiration was also reduced. Virus exposure affected eosinophil glycolysis marked by significantly reduced (P < 0.001) ECAR in IAV-exposed cells (137.75 ± 18.88) compared to mock-exposed eosinophils (183.57 ± 26.07).

PB1-pulsed eosinophils enhance DNA demethylation of the Tbx21 locus in CD8⁺ T cells

Following exposure to IAV, eosinophils can enhance cellular immunity by priming and activating naive CD8⁺ T cells, initiating their differentiation into IFN-γ-producing effector cytotoxic T lymphocytes.¹² Recent studies have shown that CD8⁺ T cells undergo extensive epigenetic programming in many effector genes, including Tbx21, during their Ag-driven differentiation in IAV or chronic virus-infected hosts.^53,80 The Tbx21 locus encodes T-bet, which is a key transcription factor that regulates IFN-γ production and plays an important role in the differentiation and function of effector and memory CD8⁺ T cells.^81,82 To determine whether eosinophils promote epigenetic changes in effector CD8⁺ T-cells in an Ag-dependent manner, we investigated the epigenetic DNA methylation changes in the Tbx21 locus in IAV-specific CD8⁺ T-cells after co-culture with IAV peptide-pulsed eosinophils (Fig. 6). CD8⁺ T-cells were purified by negative selection from the lungs and spleen of IAV-infected mice and co-cultured with the PB1_703–711 or mock-pulsed BMdEos. After 2 days, we FACS-purified PB1-tetramer⁺/CD44⁺ CD8⁺ T-cells (Fig. 6A) through PB1 tetramer staining (Fig. 6B), and measured DNA methylation levels in the Tbx21 locus in IAV-specific CD8⁺ T-cells and found that CD8⁺ T-cells co-cultured with peptide-pulsed BMdEos had reduced DNA methylation levels in the Tbx21 locus (Fig. 6C). These data suggest that eosinophils may enforce epigenetic changes in effector IAV-specific CD8⁺ T cells through Ag-presentation.

Discussion

IAV is a leading cause of morbidity and mortality worldwide, with over 1 billion infections resulting in 3–5 million severe lower respiratory tract illnesses and 290,000–650,000 deaths per annum.⁸³ Asthma affects over 26 million Americans,²⁴ where allergic asthma is the most common subtype. Despite the morbidity associated with each individual disease, there is increasing clinical and basic science evidence that morbidity and mortality associated with influenza are reduced in hosts with underlying acute allergic asthma.^{12,23,33,36,84} Using a mouse model that recapitulated the epidemiologic findings associated with Swine Flu in asthmatics,²³ we reported that eosinophils enhanced antiviral defenses in mice with fungal asthma during influenza by releasing granule contents via piecemeal degranulation and activating CD8⁺ T cells.¹² Herein, we further examined how eosinophil physiology and function are impacted by IAV.

Mice with underlying fungal asthma that were protected from severe influenza morbidity²³ had reduced levels of a number of cytokines in the lungs compared to mice with influenza alone. Since excessive pro-inflammatory cytokines in response to IAV infection can contribute to host pathology,⁸⁵ this reduction observed in comorbid mice may be beneficial. The increase in CCL11 in response to IAV likely synergized with elevated IL-5 levels to enable eosinophil recruitment into the lungs of Flu Ctr mice as they had concurrent upregulation of IL-5Rα. It is not surprising that IL-17A levels were not significantly elevated at this time point, as this cytokine has been previously been shown to be increased early after fungal challenge⁸⁶ and IL-33 has been shown to hinder IL-17A production in response to A. fumigatus.⁸⁷ Virus infection alone caused an increase in IL-33 as previously shown,⁸⁸ and since IL-33 can directly stimulate eosinophils⁸⁹ and enhance their survival,⁹⁰ it is possible that the cytokine milieu in the lungs was optimal for eosinophil activation. Despite the prominent T_H1 bias in the Flu Ctr mice, the availability of factors associated with recruitment and activation of eosinophils likely contributed to their survival and function within the tissue. Similarly, the mixed cytokine profile in the comorbid mice is likely to have supported eosinophil sustenance and activation within the lungs. Therefore, delineation of the impact of canonical T_H1, T_H2, and T_H17 cytokines on eosinophil function after IAV stimulation will be of importance.

The notion that eosinophils are terminally differentiated without plasticity in activation/function is now obsolescent with the emergence of new evidence of their role in tissue development, repair, and host defense.^4,91–93 Since eosinophils can alter their expression of surface markers based on cues received from the environment, the fluctuations we noted in the trends between the morphologic and FCM-based quantifications were expected especially because CCR3 and Siglec-F have previously been shown to be differentially expressed in certain niches.⁵⁴ Despite also observing similar numbers of eosinophils in the bone marrow of all 3 groups of mice based on hallmark surface Ags by FCM, the mature eosinophil population (eosin-positive polymorphonuclear cells) in the bone marrow was lower in Asthma Ctr and A+Flu mice compared to Flu Ctr at this time point. After allergen exposure in an individual with allergic asthma, cytokines including IL-5 and prostaglandin D₂ incite eosinophil egress to the bloodstream,^95,96 using CD62L and VLA-4⁹⁷ to migrate to the site of allergen insult. Since eosinophils shed CD62L upon reaching the lungs,^64,98 recruited eosinophils are typically CD62L low or negative while resident eosinophils are identified by high levels of CD62L.⁷⁰ Eosinophils in the lungs had the lowest expression of CD62L irrespective of disease when compared to eosinophils in other niches examined, suggesting that these cells may have been recruited. Based on the expression of surface markers, it is also likely that in response to allergen and/or virus, eosinophils were generated in the bone marrow and recruited into the lymphoid organs and lungs whereby their activation was most likely regulated by the cytokines and other leukocytes within each environment.

Eosinophil presence in the blood and airways of asthmatics, and the correlation between their reduction and alleviation of asthma symptoms, implies a role in asthma pathogenesis. Although corticosteroids commonly prescribed to asthmatics do reduce eosinophilia and symptoms,^99,100 long-lasting detrimental side effects¹⁰¹ and the steroid-resistance in some patients¹⁰² have led to the development of alternative strategies to treat asthma symptoms. The importance of IL-5 to eosinophil maturation, survival, and activation^95,103–105 and the reduction of asthma pathogenesis in IL-5 deficient mice¹⁰⁶ led to the development of anti-IL-5/IL-5R biologics for treatment of asthma symptoms each with specific efficacies in patients.¹⁰⁷ The discovery that Siglec-8 is expressed on human eosinophils and that its ligation triggers apoptosis,⁵⁸ suggests that this lectin may provide a more specific approach to remove eosinophils with minimal off-target effects.¹⁰⁸ Herein, we confirmed our previous observation that eosinophils down-regulate Siglec-F in response to IAV,¹² and since viability was maintained in eosinophils after IAV exposure, we explored eosinophil surface markers associated with Siglec-F expression to determine if perhaps two subsets of eosinophils arose after IAV exposure. IL-5R provides a survival signal while PIR-A/B provide pro- and anti-apoptotic signals to eosinophils.⁶¹ Since the expression of PIR-B is higher than PIR-A on mature eosinophils, and PIR-B inhibits the function of PIR-A,⁶¹ it is possible that the increased expression of PIR-A/B on Siglec-F^lo eosinophils may have aided in their survival as the levels of IL-5Rα expression were reduced. Eosinophils in airways (BAL) of Flu Ctr mice reduced Siglec-F expression and increased PIR-A/B and its ligand MHC-I, suggesting that mechanisms to enhance eosinophil survival were active in vivo. Reduced levels of CD62L on the Siglec-F^lo eosinophils also suggest these cells were activated after IAV exposure. The available Ab cannot differentiate between PIR-A and PIR-B, hence, we would need to measure the level of RNA for each to delineate which PIR molecule would have dominated in eosinophils after IAV exposure. Although we did not sub-categorize eosinophils by their expression levels of Siglec-F in vivo, doing so may provide additional important information regarding their activation state. In humans, eosinophil IL-5R is negatively regulated by IL-5.¹¹⁰ We found that eosinophils in Flu Ctr mice had higher expression of IL-5Rα in the airways (BAL) than those in mice with asthma, and it is possible that local release of IL-5 by T cells and eosinophils in the allergic lung is responsible for the downregulation of its cognate receptor on eosinophils in these animals. Together, our data suggest that eosinophils enhance their survival after IAV exposure by increasing IL-5Rα and reducing pro-apoptotic Siglec-F and possibly increasing expression of PIR-B on their surface.

Mature eosinophils are considered largely transcriptionally inactive, synthesizing granule proteins, mediators, enzymes, and cytokines during their development and storing them intracellularly in granules.^111,112 Granule (or content) release is an elegant mechanism to generate an immediate response to pathogens and allergens, as they can be rapidly released by exocytosis or piecemeal degranulation. However, mature eosinophils are also capable of de novo protein synthesis in some situations.^111,113 Using qPCR analysis, we observed that transcripts for Ddx58 (RIG-I), Tlr3, and Ifih1 (MDA5) were up-regulated in IAV-exposed eosinophils, indicating that eosinophils protect themselves from infection by reducing genes involved in endocytosis and increasing expression of the pattern recognition receptors for sensing RNA viruses and viral products. This hypothesis is consistent with our previous report that viral production is low in eosinophils compared to MDCK cells after infection.¹² Additionally, while glycolysis is increased in IAV-infected human epithelial cells, which is a proposed mechanism to replenish cellular components utilized in the production of progeny virions,¹¹⁴ we found it to be significantly reduced in BMdEos following IAV infection. Up-regulation of NADH dehydrogenase subunit 4 and mitochondrially encoded cytochrome c oxidase II result in increased ATP production in HeLa cells infected with vaccinia virus and is postulated to be a strategy to increase progeny virus production.¹¹⁵ Similarly, avian influenza viruses, H6N2 and H7N1, cause increased ATP production in infected chicken lung epithelial cells by regulating mitochondrially encoded genes involved in oxidative phosphorylation.¹¹⁶ Conversely, in this study IAV-infected BMdEos reduced basal mitochondrial respiration, maximal respiration, non-mitochondrial respiration, and ATP production. Spare respiratory capacity was also dramatically reduced, indicating these cells have low energetic reserves and are operating very close to their maximal respiratory and bioenergetic capacity. If eosinophil apoptosis were triggered through their mitochondria,⁷⁹ reducing mitochondrial function may be another mechanism by which eosinophils protect themselves from viral replication (and IAV-induced damage), thereby allowing them a survival advantage to moonlight as APCs. A thorough proteomics analysis will help decipher the level of inducement in eosinophils following IAV exposure.

Our analysis of global transcriptomic changes in IAV-exposed eosinophils suggests a general down-regulation of many genes that control or respond to MAP kinase signaling. Specifically, we observed reduced levels of a large number of ectopic chemosensory G-protein coupled receptors (V1Rs, V2Rs, OLFRs, P2rx7) and ligand chaperones (MUPs),^117–119 which would result in reduced MAP kinase activity. The increased expression of Dusp6 would also reduce Erk1/2 kinase activity similar to PIR-B^61,62 thereby inhibiting Erk1/2-mediated apoptosis. Furthermore, reduced levels of Erk1/2 will lead to reduced expression and/or activity of the MAP kinase-regulated AP1 transcription factors (Jun, Fosl1), and thereby reduce AP1-stimulated gene expression, including that of heat shock proteins (Hsp1a, Hsp1b, Hsph1) and heme oxygenase-1 (Hmox1), as observed. Such an involvement of MAP kinase pathways in IAV-exposed eosinophils is consistent with studies demonstrating that inhibition of MAP kinase pathways can reduce IAV propagation¹²⁰ and the pathogenic hyper-induction of pro-inflammatory cytokines during infection,¹²¹ and may alleviate host mortality.^121,122 A potential caveat to these findings is our use of eosinophils exposed to IAV at a relatively low MOI (0.1) in order to maximize the extraction of good quality RNA for the microarray analysis. We previously showed that piecemeal degranulation and surface expression of activation and viral markers occurred in response to IAV at MOI of 0.1,¹² indicating functional activity was triggered in the cells at this level of infection. However, we also found that eosinophils were resistant to IAV-induced cytopathology at higher MOIs that were used in confocal studies.¹² It is, therefore, possible that the number of eosinophil genes differentially expressed in response to high level IAV infection is greater than what we observed here.

CD8⁺ T cells are crucial in the control and clearance of IAV infection, clonally expanding upon recognition of MHC-I-restricted virus peptide presented by professional APCs. Accordingly, professional APCs such as DCs up-regulate MHC-I in response to IAV infection.¹²³ We have previously reported MHC-I levels are gradually elevated on mouse eosinophils after IAV exposure in vitro.¹² Here, we report that peripheral eosinophils in Flu Ctr mice also up-regulate MHC-I, most notably in niches where they may come into direct contact with virus: lungs and airways. Although we did not observe elevated expression of MHC-I on eosinophils in the airways or lungs of A+Flu mice, it may be possible that we missed the window in which MHC-I is up-regulated in eosinophils that would have already been in these niches during IAV infection since the data were collected 7 days after virus infection. We also observed that eosinophils in the thymus of A+Flu mice expressed both MHC-I and MHC-II. Thymic eosinophils have been reported previously, and were predominantly localized to the medulla where thymocytes undergo negative selection.^124,125 In neonates, an increase in the apoptotic thymocytes 6 h after administration of a cognate peptide to H-Y TCR transgenic mice correlated with increased numbers of eosinophils in the thymus, suggesting they may play a role in negative selection.¹²⁵ Given that we have previously shown that eosinophils may also present MHC-I-restricted viral Ags to CD8⁺ T cells,¹² it is tempting to speculate that they participate in T cell education during co-morbidity.

Activation of cellular immune responses dominated by CD8⁺ T cells is crucial for anti-influenza host defenses. We previously showed that eosinophils were capable of presenting viral Ags to CD8⁺ T cells and participated in priming and boosting their antiviral functions.¹² The differentiation of naïve CD4⁺ and CD8⁺ T cells to their effector and memory counterparts is influenced by epigenetically controlled transcription factors. T-bet (encoded by Tbx21) is one of the crucial transcription factors governing the generation of effector CD8⁺ T cells, which play a major role in viral clearance. T-bet deficiency diminished IAV-specific CD8⁺ T cell responses¹²⁶ and early expression of T-bet in CD8⁺ T cells optimizes their effector functions.¹²⁷  Tbx21 in CD8⁺ T cells is regulated by bivalent methylation of histone 3 during influenza. Its transcription in naïve CD8⁺ T cells is blocked by the repressive trimethylation of histone 3 lysine 27 (H3K27me3) that overrides the permissive trimethylation of histone 3 lysine 4 at the Tbx21 locus.⁸⁰ The loss of bivalency due to demethylation of H3K27me3 during influenza allows the generation of cytotoxic T cells.⁸⁰ Interestingly, our results indicate when CD8⁺ T cells are activated by PB1 peptide-pulsed eosinophils, CpG dinucleotides at the Tbx21 locus undergo substantial demethylation. Changes in DNA methylation programs were previously reported to enforce transcriptional changes in the effector genes, including Tbx21, during effector or exhausted CD8⁺ T cells differentiation.^53,80 Our observation that peptide-pulsed eosinophils reduce the DNA methylation levels within the Tbx21 locus in IAV-specific CD8⁺ T cells indicate that eosinophils may contribute to further epigenetic remodeling in effector flu-specific CD8⁺ T cells during influenza infections in asthmatic hosts.

Eosinophils are often portrayed as adversaries during allergic asthma, with current treatment strategies including the use of anti-IL-5 mAbs to reduce eosinophil generation in the bone marrow, mobilization, and survival in the periphery. However, increasing evidence from us and others suggests that eosinophils may play a protective role against a co-current viral infection. In this study, we suggest that exposure to IAV initiates self-preservation responses in eosinophils that enable them to survive the influenza virus infection and participate in the antiviral host response including the induction of epigenetic changes promoting the generation of cytotoxic CD8⁺ T cells. Elucidating how eosinophils sense IAV and contribute to antiviral host defense mechanisms may be important to select effective therapeutics for asthmatics that may benefit from their presence in the airways during influenza.

Abbreviations

 
• BAL

  bronchoalveolar lavage

 
• BMdEos

  bone marrow-derived eosinophils

 
• IAV

  influenza A virus

 
• Siglec

  sialic acid-binding immunoglobulin-type lectin

Acknowledgments

The authors wish to thank Lavanya Bezavada (UTHSC, Department of Pediatrics) for running mitochondrial respiration and glycolysis assays on the Seahorse XF analyzer, Kyle Cavender (formerly at CFRI) for assistance in cytospin differential cell counts, and Ghana Gurung at the Animal Resource Center at UTHSC for animal husbandry and care. This work was supported by grants from the Le Bonheur Young Investigator Award (A.E.S.) and the National Institutes of Health (grant AI-125481) (A.E.S.).

Authorship

K.S.L., H.G., K.L., H.S.S., and A.E.S. designed the experiments. K.S.L., H.G., and B.L. performed experimental work. R.R. and H.S.S. ran transcriptome and bioenergentics assays, respectively. K.S.L., H.G., R.R., H.S.S., B.L., and A.E.S. contributed to data analyses. K.S.L. and A.E.S. wrote the manuscript. All authors edited and approved the final version.

Disclosure

The authors declare no conflict of interest.

References