TFAM-Dependent Mitochondrial Metabolism Is Required for Alveolar Macrophage Maintenance and Homeostasis

This article is featured in Top Reads, p.

Alveolar macrophages (AMs) are the main tissue-resident macrophages in the respiratory tract. AMs mainly derive from yoke sac precursors of fetal monocytes, which populate the alveolar space in the first week after birth (1, 2). During their differentiation and maturation, AM undergo profound phenotypical changes, which are characterized by increased expression of CD11c, Siglec-F, F4/80, and CD64 and concomitant downregulation of CD11b (3, 4). After their development, AMs can persist independently of adult bone marrow (BM) and monocyte contribution via self-renew in situ. AMs are critical in the maintenance of lung homeostasis, surfactant clearance, antimicrobial immunity, and tissue repair in the respiratory tract (5). The absence of AMs or impaired AM function can cause pulmonary alveolar proteinosis, which is a rare disease due to alveolar accumulation of surfactant and respiratory insufficiency (6, 7). AMs are also important for the protection against a variety of respiratory bacterial, fungal, and viral infections, including influenza virus infection (5). AMs can phagocytize free virus and virus-infected cells, produce antiviral cytokines, protect alveolar type I cells, clear cellular debris, and repair damaged lungs following influenza virus infection (8–13).

GM-CSF and TGF-β are the two critical cytokines required for the differentiation and maintenance of AMs. The transcription factor peroxisome proliferator–activated receptor gamma (PPARγ), induced by GM-CSF and TGF-β, is vital for AM differentiation and maintenance (1, 11, 14). Bach2, Bhlhe40, and Bhlhe41, hematopoietic protein-1, mammalian target of rapamycin complex 1 (mTORC1), phosphoinositide kinase, FYVE-type zinc finger (PIKfyve), Lkb1, and l-plastin, etc., were also shown to be important in AM development, maintenance, and/or function (15–20). Interestingly, both PPARγ and Bach2 deficiency profoundly altered AM lipid metabolic responses (11, 15), suggesting that functional cellular metabolic state is likely required for AM homeostasis and/or function. In support of this idea, mTORC1-mediated glucose metabolism has been linked to AM maintenance and proliferation in vivo (20). Nevertheless, the exact function of individual metabolic pathways on AM formation and/or function remains to be fully elucidated in vivo.

The mitochondria are the central hub of cellular metabolic responses, which are required for ATP generation, macromolecule biosynthesis, calcium homeostasis, redox balance, and cellular epigenetic programming (21–24). The mitochondrial transcription factor A (TFAM) is a nuclear-encoded transcription factor that plays a critical role in mitochondrial DNA (mtDNA) replication, metabolism, and stability (25). As such, TFAM deficiency causes severe mtDNA depletion, mitochondria damage, and nonfunctional oxidative phosphorylation (OXPHOS) (26). Furthermore, TFAM has been shown to play a crucial role in mtDNA stress-mediated inflammatory responses (27, 28). The roles of mitochondrial metabolism and TFAM function in regulating AM development, maintenance, and/or proliferation are unknown currently. We have recently shown that AM self-renewal and inflammatory responses were uncoupled by mitochondrial metabolism (29). As such, the inhibition of mitochondrial respiratory complex function resulted in diminished AM proliferation in vitro and in vivo (29). Given the roles of TFAM in maintaining cellular mitochondrial metabolism and suppressing inflammatory responses (27, 28), we hypothesized that TFAM-dependent mitochondrial oxidative metabolism is likely required for AM self-renewal and/or function.

In this report, we showed that TFAM deficiency caused impaired AM mitochondrial fitness, altered metabolism, and defective lipid catabolic responses, leading to diminished AM quantity and altered AM phenotypes. We demonstrated that TFAM deficiency did not impair AM differentiation from their precursors, but caused reduced AM maintenance. RNA-sequencing (RNA-seq) experiments revealed that TFAM deficiency caused impaired AM self-renewal capability. We showed that TFAM deficiency in AMs caused impaired clearance of respiratory debris and surfactant and enhanced susceptibility to severe influenza virus infection. Finally, we found that influenza infection resulted in disruption of TFAM-mediated mitochondrial fitness and metabolism in AMs.

C57BL/6, CD45.1⁺, CD11c-cre, ubiquitin C (Ubc)–creERT2, Tfam^fl/fl mice were purchased from The Jackson Laboratory. CD11c^creTfam^fl/fl or CD11c^creTfam^fl/+ mice were generated by crossing Tfam^fl/fl mice with CD11c-cre mice (30). Ubc^creERTfam^fl/fl mice were generated by crossing Tfam^fl/fl mice to mice expressing a cre-ERT2 fusion gene under the control of the human Ubc promoter (31). All mice were maintained in a specific pathogen-free environment at the Mayo Clinic animal facility. For influenza virus infection, influenza A/PR8/34 strain was diluted in FBS-free DMEM media (Life Technologies) on ice and inoculated in anesthetized mice by the intranasal route as described before (32).

After mice were sacrificed, bronchoalveolar lavage (BAL) fluid was obtained by flushing the airway three times with a single inoculum of 600 µl sterile cold PBS via a trachea incision. Cells in the 600 µl BAL fluid were centrifuged at 1600 rpm at 4°C for 5 min, and supernatants were collected for the determination of turbidity, protein concentration, and cytokines/chemokines levels, as indicated in the text. Cell pellets were resuspended in 1 ml PBS and used for further flow cytometry analysis.

Lungs were collected from sacrificed mice. Lung tissues were cut into small pieces by scissors and digested with Collagenase Type 2 (183 units/ml; Worthington Biochemical) for 30 min at 37 ̊°C. The single-cell suspension was passed through a 70-μm cell strainer (Falcon) within a 50-ml Falcon tube. Cells were pelleted at 1600 rpm at 4°C for 5 min, and the pellet was suspended after lysing RBCs with ACK lysis buffer (0.15 M NH₄Cl, 1 mM KHCO₃, and 0.1 mM Na₂EDTA, pH 7.2). After centrifugation at 1600 rpm at 4°C for 5 min, the cell pellet was suspended in cold flow cytometry buffer for flow cytometry analysis.

AMs were obtained from BAL as described previously (29). Briefly, mice were euthanized, the skin and muscles were removed on the trachea, and a small incision was made below the larynx. This incision was used to insert a 1-ml pipette toward the lungs. Alveolar lavages were performed with multiple 1-ml flushes of the lungs with 2 mM EDTA in 2% FBS/PBS (PBS). Pooled BAL washes were centrifuged at 1600 rpm at 4°C for 5 min to pellet cells. AMs were purified by adherence to culture plate for 2 h in complete medium (RPMI 1640, 10% FBS, and 1% penicillin/streptomycin/glutamate; Thermo Fisher Scientific) at 37°C and 5% CO₂. The nonadherent cells were washed off with warm PBS.

For AM infection in vitro, seeded cells were initially incubated with 10 multiplicities of infection of influenza PR8 virus in PBS for 15 min on ice. Then, the cells were cultured at 37°C and 5% CO₂ for 1 h. Virus was subsequently washed out with warm PBS twice. The cells were further cultured in warm complete medium containing 10 ng/ml GM-CSF for an additional 24 h at 37°C and 5% CO₂. Cells were analyzed by quantitative RT-PCR, Western blot, and transmission electron microscopy (TEM).

To assess AM proliferation, AMs were first cultured overnight in complete medium without GM-CSF. Then, AMs were cultured in complete medium supplemented with 10 ng/ml murine rGM-CSF (BioLegend) for 24 h. The cells were digested by 0.25% trypsin-EDTA solution and collected for flow cytometric staining, including staining cell surface markers and intracellular staining of Ki-67.

BM cells were collected from tibias and femurs by flushing with cold PBS through a 25-gauge needle. The single-cell suspension was passed through a 70-μm cell strainer in a 50-ml Falcon tube. Cells were pelleted at 1600 rpm for 5 min, and the pellet was suspended in ACK lysis buffer (0.15 M NH4Cl, 1 mM KHCO3, and 0.1 mM Na₂ EDTA, pH 7.2) at room temperature for 5 min. After centrifugation at 1600 rpm for 5 min, the cell pellet was suspended in 90 µl cold MACS buffer per 10⁷ total cells. According to the protocol for anti-Ly6G Microbeads (Miltenyi Biotec), 10 µl anti-Ly6G beads was added, incubated for 10 min at 4°C, and then washed by MACS buffer at 1600 rpm for 5 min. Cells were resuspended and applied onto the column, the column was washed, and then Ly6G⁺ cells were collected (BM-neutrophils). Anti-CD11b Microbeads (Miltenyi Biotec) were added to Ly6G⁻ cells, and similarly, Ly6G⁻CD11b⁺ cells were collected (BM-monocytes).

In vitro cultured AMs or sorted AMs, as indicated in the text, were fixed in 2% paraformaldehyde and 2.5% glutaraldehyde in 0.1 M sodium cacodylate. Following fixation, cells were embedded and sliced for TEM. The grids were imaged with a JEM 1400 TEM (JEOL). Mitochondrial morphology was scored as normal or abnormal. Structurally abnormal mitochondria were defined operationally as those with loss of cristae, decreased electron density of the matrix, loss of integrity of mitochondrial membrane, and/or the formation of autophagosomes structures, as reported before (33). The numbers of damaged mitochondria and total mitochondria per cell were quantified.

The turbidity of BAL fluid was assessed with the OD at a wavelength of 590 nm (OD590) by spectrophotometry. Total protein of BAL fluid was determined by BCA protein assay kit (Thermo Fisher Scientific) according to the manufacturer’s manual. A total of 50 µl each BAL sample was used. A VersaMax microplate reader (Molecular Devices) was used for colorimetric quantification and analysis at OD 562-nm wavelength.

Cytokine and chemokine levels in BAL fluid were shipped in dry ice to Eve Technologies and measured with the Mouse Cytokine Array/Chemokine Assay 31-Plex (MD31).

Real-time oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) of AMs were measured with a Seahorse XFp Analyzer (Seahorse Bioscience) as described previously (23). AMs (10⁵) were seeded into each well of Seahorse XFp Cell Culture Miniplates overnight at 37°C and 5% CO₂. On the following day, the cells were washed twice and incubated at 37°C for 1 h in the absence of CO₂ in unbuffered assay medium (pH 7.4; Agilent Technologies, Santa Clara, CA) with 10 mM glucose for mitochondria stress test. OCR was measured under basal conditions and after the addition of the following compounds: 1 μM oligomycin (ATP synthase inhibitor), 1.5 μM carbonyl cyanide-4-(trifluoromethoxy) phenylhydrazone (FCCP), and 0.5 μM rotenone plus 0.5 μM antimycin (complex I and III inhibitor). ECAR was measured with the addition of following compounds: 10 mM glucose, 1 μM oligomycin, and 50 mM 2-deoxy-d-glucose. All compounds obtained from Sigma-Aldrich. Data were analyzed with Wave Desktop software version 2.6 (Agilent Technologies).

BAL cells were harvested in flow cytometry staining buffer (FACS buffer; 2 mM EDTA, 0.05% sodium azide, and 2% FBS in PBS) and stained with surface markers (CD11c, CD11b, and Siglec F). The cells were washed twice with PBS and then resuspended in 100 μl PBS containing 500 nM MitoTracker Green FM (Thermo Fisher Scientific). Cells were stained for 45 min at 37°C according to the manufacturer’s protocol. Then, cells were washed twice in PBS and analyzed by flow cytometry.

BAL cells were harvested in FACS buffer and stained with surface markers. According to the protocol of fluorescent 4, 4-difluoro-1,3,5,7,8-pentamethyl-4-bora-3a, 4a-diaza-s-indacene (BODIPY 493/503; Thermo Fisher Scientific), the cells were washed twice in PBS and then resuspended in media containing fluorescent BODIPY 493/503 for 10 min at room temperature. Cells were washed three times in FACS buffer and analyzed by flow cytometry.

Cell staining was performed with the appropriate Ab mixture in FACS buffer. The cell subsets were identified based on following cell surface markers: AM (CD11c⁺Siglec F⁺CD11b^lowCD64⁺MerTK⁺), immature AM (CD11c⁺Siglec F^lowCD11b^lowCD64⁺MerTK⁺), AM precursor (pre-AM) (CD45⁺CD11c⁺Siglec F⁻CD11b^intF4/80⁺Ly6C⁻), neutrophils (CD11b⁺Ly6G⁺), Ly6C^hi monocytes (Ly6G⁻Siglec F⁻CD11b⁺Ly6C^hi), and dendritic cells (DCs; CD11c⁺MHC class II⁺). Fluorescence-conjugated Abs CD11b (M1/70), CD11c (N418), CD64 (X54-5/7.1), MerTK (2B10C42), F4/80 (BM8), Ly6G (1A8), Ly6C (HK1.4), MHC class II (M5/114.15.2), CD45 (30-F11), and Ki67 (SolA15) were purchased from BioLegend; Siglec F (E50-2440) was from BD Biosciences; and NP366 tetramer⁺ cells (CD8⁺NP366-tet⁺) and PA224 tetramer⁺ cells (CD8⁺PA224-tet⁺) were supplied by the National Institutes of Health Tetramer Facility. For Ki67 staining, cell suspensions were stained for surface marker at 4°C for 30 min. Cells were washed twice with FACS buffer prior to fixation and permeabilization with the Foxp3 Transcription Factor Staining Buffer Set (eBioscience) for 1 h at room temperature in the dark. Cells were washed twice with Perm Wash buffer (eBioscience) and stained with Abs against Ki67 (eBioscience) and control Ig (BioLegend) in Perm Wash buffer for at least 30 min at room temperature in the dark. Cells were subsequently washed twice with Perm Wash buffer before samples were processed with flow cytometer. Samples were collected on FACS Attune NxT flow cytometer (Life Technologies) and analyzed using FlowJo software (Tree Star).

Mice were perfused with PBS (10 ml) via the right ventricle following euthanasia. Paraformaldehyde (10%) was then gently instilled into the lung and left inflated for 1 min before excising and moving the lobe to 10% paraformaldehyde. Samples were shipped to the Mayo Clinic Histology Core Laboratory (Scottsdale, AZ), where they were embedded in paraffin, and 5-μm sections were cut for H&E stain. To quantify percent of inflamed or disrupted alveolar area, slides were scanned through the Aperio whole-slide scanning system (Leica) and exported to image files. Computer-based image analysis was performed using the ImageJ software (National Institutes of Health, Bethesda, MD). We first determined the total lung area by converting the image into grayscale, followed by red highlighting through the adjustment of the threshold. For determination of the inflamed and disrupted area, color images were split into single channels. We then used the green channel, highlighted the inflamed areas in red by adjusting the threshold, and measured the areas based on pixel. The percentages of disrupted and inflamed lung areas were calculated based on the ratio of highlighted disrupted areas to the total lung area in each lung section.

To induce gene recombination in Ubc^creERTfam^fl/fl mice, tamoxifen [(E/Z)-4-hydroxy Tamoxifen; Cayman Chemical] was dissolved in 0.5 ml ethanol and 9.5 ml warm sunflower oil (Sigma-Aldrich). Then, the suspension was administered at 2 mg/mouse at a concentration of 20 mg/ml via i.p. injection for 5 consecutive d.

To generate mixed BM chimeric mice, wild-type (WT) recipient mice were lethally irradiated (1000 rads for females or 1100 rads for males) and then i.v. injected an equal mix of ∼6 million BM cells from WT CD45.1 BM and Tfam^fl/fl or Ubc^creERTfam^fl/fl CD45.2 donors (mixed at 1:1 ratio). Following 7 wk of reconstitution, recipient mice were daily administered 2 mg/mouse of tamoxifen via i.p. route for 5 consecutive d. BAL and lung samples were obtained at indicated time in the text postinjection.

Total RNA of AMs was extracted using RNeasy Plus Mini Kit (Qiagen) following the manufacturer’s protocol. Two pools per genotype were used for RNA-seq. After quality control, high-quality (Agilent Bioanalyzer RNA integrity number >7.0) total RNA was used to generate the RNA-seq library. cDNA synthesis, end-repair, A-base addition, and ligation of the Illumina indexed adapters were performed according to the TruSeq RNA Sample Prep Kit v2 (Illumina, San Diego, CA). The concentration and size distribution of the completed libraries was determined using an Agilent Bioanalyzer DNA 1000 chip (Agilent Technologies) and Qubit fluorometry (Invitrogen, Carlsbad, CA). Paired-end libraries were sequenced on an Illumina HiSeq 4000 following Illumina’s standard protocol using the Illumina cBot and HiSeq 3000/4000 PE Cluster Kit. Base-calling was performed using Illumina’s RTA software (version 2.5.2). Paired-end RNA-seq reads were aligned to the mouse reference genome (GRCm38/mm10) using Bowtie (v2.3.4). Pre- and postalignment quality controls, gene level raw read count, and normalized read count (i.e., fragments per kilobase per million) were performed using RSeQC package (v2.3.6) with the National Center for Biotechnology Information mouse RefSeq gene model. For functional analysis, gene set enrichment analysis (GSEA) was performed to identify enriched gene sets using the hallmark collection of the Molecular Signatures Database (MSigDB), having up- and downregulated genes, and using a weighted enrichment statistic and a log₂ ratio metric for ranking genes. Data were submitted to the Gene Expression Omnibus repository (accession number GSE188279, https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE188279).

Unpaired two-tailed Student t test (two group comparison), one-way ANOVA (multiple-group comparison), multiple t tests (weight loss and Multiplex studies), or log-rank (Mantel-Cox) tests (survival data) were used to determine statistical significance by GraphPad Prism software. A p value <0.05 was considered significant (*p < 0.05, **p < 0.01, ***p < 0.001).

To investigate the role of TFAM in AM development and/or function, we crossed Tfam floxed mice to transgenic mice that express Cre recombinase under CD11c (Itgax) promoter (34). CD11c is highly expressed in AM precursors, mature AMs, and DCs (1, 30). Previously, we have shown that CD11c-Cre mediates high efficiency of gene recombination of loxP-flanked DNA in DCs and AMs, but very moderately in other immune cell types in the respiratory tract (30). Indeed, we confirmed that CD11c^creTfam^fl/fl mice had impaired TFAM expression in AM compartment, moderately diminished expression in whole lung cells (potentially due to diminished TFAM expression in AMs and lung DCs), but not in neutrophils or monocytes isolated from the BM compartment (Supplemental Fig. 1A).

At the adult age (7 to 8 wk old), percentages and cell numbers of mature AM (MERTK⁺CD64⁺CD11c^hiSiglec F^hi) (35) were significantly decreased in the BAL and lungs of CD11c^creTfam^fl/fl mice compared with those of Tfam^fl/fl or CD11c-cre Tfam^fl/+ (Het) mice (Fig. 1A, 1B, gating strategy shown in Supplemental Fig. 1B) (29). In contrast, immature AMs (MERTK⁺CD64⁺CD11c^hiSiglec F^low) (36, 37) were increased following TFAM deficiency. The levels of lung interstitial macrophages (IMs; MERTK⁺CD64⁺Siglec F⁻) (35) were moderately elevated in CD11c^creTfam^fl/fl mice comparable those of WT or Het mice (Fig. 1B). There were increased levels of DCs and neutrophils in the BAL of CD11c^creTfam^fl/fl mice, but the levels of these cell types in the lungs were comparable across WT, Het, and CD11c^creTfam^fl/fl mice (Supplemental Fig. 1C, 1D). After birth, mature AMs highly expressed CD11c and Siglec F and but low CD11b (1, 38). Consistent with the increased levels of immature AM presence, BAL AMs in CD11c^creTfam^fl/fl mice expressed higher levels of CD11b than those of WT or Het mice (Fig. 1C).

TFAM expression is crucial for mitochondrial fitness and function (26). We therefore examined whether TFAM deficiency affected AM metabolism. To this end, we measured mitochondrial OXPHOS and glycolysis of BAL AMs from Tfam^fl/fl or CD11c^creTfam^fl/fl mice. We found that basal oxygen consumption rate (OCR) and maximal mitochondrial respiratory capacity (maximum respiration) were decreased in AMs following TFAM deficiency (Fig. 1D, 1E). Furthermore, ECAR of AMs were also decreased (Fig. 1F). These data suggest that TFAM deficiency impairs AM metabolic fitness. Consistent with decreased OCR, TFAM-deficient AMs had decreased mitochondrial mass as measured by MitoTracker staining (Fig. 1G). We then evaluated mitochondrial morphology of BAL AMs from Tfam^fl/fl or CD11c^creTfam^fl/fl mice by TEM (Fig. 1H). TFAM-deficient AMs had decreased overall mitochondrial numbers and increased percentages of abnormal mitochondria (mitochondrial swelling and loss of cristae) (Fig. 1H). Together, these data suggest that TFAM deficiency impairs mitochondrial fitness and metabolism. Mitochondria are essential for lipid catabolism in cells. Consistently, we found that TFAM-deficient AMs had increased lipid content in AMs and enlarged lipid droplets within the cytoplasm (Fig. 1I, 1J).

We next sought to track the kinetics of AM responses following TFAM deficiency. To this end, we first examined AM numbers and phenotypes at the juvenile stage (weeks 3 to 4 postnatally). We found that the percentages and numbers of mature BAL AMs were significantly diminished in CD11c^creTfam^fl/fl mice compared with those of Tfam^fl/fl mice, albeit to a lower extent as in the adult age (7 to 8 wk) (Fig. 2A, Supplemental Fig. 2A). Accompanied with this, immature AMs were significantly elevated in the lung of CD11c^creTfam^fl/fl mice (Supplemental Fig. 2B). Lung neutrophils, DCs, and Ly6C^hi monocytes were at similar levels between CD11c^creTfam^fl/fl mice and Tfam^fl/fl mice at the 3- to 4-wk age (Supplemental Fig. 2C, 2D). Similarly, TFAM-deficient AMs had decreased mitochondrial OXPHOS as measured by OCR (Fig. 2B, 2C).

AMs are mainly developed from fetal monocytes through a pre-AM stage in the first week of birth (1). To explore the function TFAM in early AM differentiation, we examined the monocyte and macrophage populations in the lung at day 3 postbirth (PND3). We found that overall lung myeloid cells (CD11b⁺F4/80⁺) were comparable between Tfam^fl/fl and CD11c^creTfam^fl/fl mice at PND3 (Fig. 2D). Using the gating strategy reported before (1), we found that lung AMs (CD11c⁺Siglec F⁺Ly6C⁻F4/80⁺), pre-AMs (CD11c^intLy6C⁻F4-80⁺Siglec F⁻), monocytes (CD11c⁻Ly6C⁺Siglec F⁻F4/80⁻), and fetal macrophages (CD11c⁻Ly6C⁺Siglec F⁻F4/80⁺) were comparable between Tfam^fl/fl and CD11c^creTfam^fl/fl mice at PND3 (Fig. 2D, 2E). Similar results were observed at PND7 (Fig. 2F). These data suggest that TFAM deficiency likely does not affect the formation of pre-AM population and the differentiation of pre-AM into AM lineage.

Our kinetic analysis suggests that TFAM may not be needed for the initial formation of AM compartment, but is likely required for the maintenance of AM population. However, CD11c-Cre transgenic mice constitutively express Cre recombinase under the CD11c promoter, which likely causes TFAM deficiency in AM precursors (37). We therefore cannot determine whether TFAM is required for AM maintenance using CD11c^creTfam^fl/fl mice. To address the function of TFAM in AMs following their development, we crossed Tfam^fl/fl mice to Ubc-Cre ERT2 transgenic mice (31), in which all cells express a CreERT2 recombinase under human Ubc promoter, to generate Ubc^creERTfam^fl/fl mice. The presence of tamoxifen in Ubc^creERTfam^fl/fl mice is expected to result in the translocation of the recombinase into the nucleus, where it can recombine loxP-flanked exons 6 to 7 of Tfam to acutely deplete TFAM expression. We injected adult Tfam^fl/fl or Ubc^creERTfam^fl/fl mice with tamoxifen. We first confirmed the diminished TFAM levels in AMs and other cell types following tamoxifen injection (Supplemental Fig. 3A).

Three to 4 wk following injection, we examined BAL AM responses. Acute depletion of TFAM resulted in decreased mature AM numbers in the BAL (Fig. 3A), suggesting that TFAM is likely needed for AM maintenance. Tamoxifen injection in Ubc^creERTfam^fl/fl mice will deplete TFAM protein ubiquitously. To determine whether intrinsic TFAM expression in AMs is required for their maintenance, we generated mixed BM chimera mice, in which we transplant 1:1 mixed WT (CD45.1⁺) and Ubc^creERTfam^fl/fl (CD45.2⁺) BM cells into lethally irradiated CD45.1⁺/CD45.2⁺ WT mice. After a 7-wk reconstitution, we injected tamoxifen into the BM chimeric mice and followed AM maintenance (Fig. 3B). At 2 d following the last tamoxifen injection, CD45.2⁺ TFAM-deficient AMs moderately outnumbered WT CD45.1⁺ AMs in both BAL and the lung compartment, suggesting that TFAM deficiency did not result in the acute loss of AMs at this early time point (Fig. 3C). However, at 7 wk after tamoxifen injection, we found that WT CD45.1⁺ AMs significantly outnumbered TFAM-deficient CD45.2⁺ AMs in both BAL and the lung compartment, indicating that intrinsic TFAM function is required for AM maintenance (Fig. 3C). Consistent with this, we found that TFAM deficiency diminished Siglec F but increased CD11b expression in AMs (Supplemental Fig. 3B, 3C), suggesting that TFAM expression may be required for maintaining the mature AM phenotype. In contrast, TFAM deficiency did not dramatically impair the maintenance of neutrophil, DC, IM, and Ly6C^hi monocyte populations (Fig. 3D).

To begin to understand the underlying mechanisms by which TFAM modulates AM maintenance, we preformed bulk RNA-seq of BAL AMs from Tfam^fl/fl or CD11c^creTfam^fl/fl (knockout [KO]) mice at the age of 3 to 4 and 7 to 8 wk. The transcriptional profiles of WT AMs and TFAM-deficient AMs were marked different at both time points (Fig. 4A, 4B). GSEA of hallmark gene sets found that TFAM deficiency increased the expression of hypoxia, adipogenesis, and IFN-γ– and -α–responsive genes, but decreased expression of NF-κB, cholesterol, G2M checkpoint, and mTORC1 signaling (Fig. 4C). Additionally, stem cell–associated genes, which were indicated in AM proliferation and self-renewal (Fig. 4D) (29), were significantly downregulated in TFAM-deficient AMs. These data suggest that TFAM deficiency may impair AM proliferation and self-renewal capability. To test this idea, we measured Ki67 expression in AMs ex vivo from Tfam^fl/fl, CD11c-cre Tfam^fl/+ (Het), or CD11c^creTfam^fl/fl (KO) mice at 3 to 4 wk of age. TFAM deficiency significantly was downregulated Ki67 expression in vivo (Fig. 4E). Furthermore, TFAM deficiency diminished Ki67 expression in AMs after in vitro culture with GM-CSF (Fig. 4F), a major AM growth factor (1, 39). Together, these data suggest that TFAM deficiency impairs AM proliferation and self-renewal, thereby diminishing AM maintenance in vivo. GSEA analysis further revealed that TFAM deficiency caused significant upregulation of inflammation-associated genes at 7 to 8 wk age (Fig. 4G, 4H). Together, these transcriptional analyses revealed that TFAM is important in regulating AM proliferation and immune homeostasis.

AMs are critical in clearing debris in the alveolar space and in surfactant homeostasis (40, 41). Consistent with the greatly diminished mature AM numbers in CD11c^creTfam^fl/fl mice, BAL from CD11c^creTfam^fl/fl mice contained significantly more elevated levels of dead cells or cellular debris than those of Tfam^fl/fl mice (Fig. 5A). Furthermore, supernatant of BAL from CD11c^creTfam^fl/fl mice exhibited a “milky” appearance and had elevated absorbance at OD590 (Fig. 5B, 5C), suggesting that CD11c^creTfam^fl/fl mice had impaired clearance of surfactant proteins.

AMs are critical for host antimicrobial responses, and AM deficiency or functional impairment can lead to increased susceptibility to severe influenza virus infection (8–13). To this end, we infected adult Tfam^fl/fl or CD11c^creTfam^fl/fl mice with influenza A PR8 strain and followed host morbidity and mortality postinfection. We found that CD11c^creTfam^fl/fl lost significant more weight and were succumbed to infection by 10 d postinfection compared with WT mice (Fig. 5D, 5E), which were consistent with the critical function of AMs in antiviral immunity and the maintenance of lung homeostasis during influenza infection (42, 43). BAL total cells, neutrophils, Ly6C^hi monocytes, and DCs were elevated in CD11c^creTfam^fl/fl mice at day 4 postinfection (Fig. 5F), Importantly, compared with Tfam^fl/fl mice, CD11c^creTfam^fl/fl mice showed similar levels of DCs, total and Ag-specific D^b NP_366-374, and D^b PA_224-233 CD8⁺ T cells in the draining lymph node and the lung at day 5 postinfection (44) (Supplemental Fig. 4A, 4B), suggesting that Tfam deficiency in CD11c⁺ cells may not change the quantity and function of DCs. Furthermore, CD11c^creTfam^fl/fl mice exhibited increased tissue inflammation and pathology compared with Tfam^fl/fl mice (Fig. 5G, Supplemental Fig. 4C). Additionally, there was greatly increased pulmonary inflammation in CD11c^creTfam^fl/fl mice as evidenced by their elevated airway proinflammatory cytokine and chemokine levels compared with those of Tfam^fl/fl mice. Taken together, these data suggest that TFAM-mediated AM maintenance is critical for preventing severe host inflammation and diseases following influenza virus infection.

Proper AM homeostasis is critical for host protective responses following influenza virus infection. Notably, influenza virus can directly infect AMs to impair their function and/or prime inflammatory responses (13, 45). Given the important roles of TFAM-mediated mitochondrial metabolism in regulating AM homeostasis and function, we examined whether influenza infection could decrease TFAM expression in WT AMs. Influenza virus infection caused diminished TFAM protein levels in mouse AMs (Fig. 6A). Consistently, influenza virus infection resulted in increased accumulation of abnormal mitochondria (Fig. 6B). We then analyzed a publicly available microarray data set (accession number GSE30723, https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE30723) of human AMs that were infected with influenza PR8 strain for 4 or 24 h (46). TFAM expression in infected AMs was decreased compared with uninfected AMs (Fig. 6C). In consistency, GSEA revealed that there was a decreased enrichment of mitochondrial OXPHOS-related genes in influenza-infected AMs (24 h) compared with those of uninfected AMs (Fig. 6D). Together, these data suggested that influenza infection may impair TFAM expression, thereby causing defective mitochondrial fitness and function in AMs.

Tissue-resident macrophages are important in regulating tissue homeostasis and inflammation and are essential for tissue-specific function and the host protection from infection. Better understanding of the molecular and metabolic pathways regulating tissue-resident macrophage development, maintenance, and function will aid to identify macrophage-targeting therapies for various acute and chronic diseases (41, 47, 48). In this study, we have identified a critical role of TFAM in regulating AM proliferation and phenotypes, thereby providing (to our knowledge) the first genetic evidence on the roles of mitochondrial metabolism in regulating AM maintenance in vivo.

A number of factors have been identified to be important in AM development, maintenance, and/or function, including GM-CSF, TGF-β, Bach2, PPARγ, and mTORC1, etc. Particularly, PPARγ is considered a “master” regulator of AM development and function, although the underlying mechanisms by which PPARγ regulates AM development and function remain to be fully elucidated. Interestingly, PPARγ-deficient AMs had altered mitochondrial and lipid metabolism (1, 14), suggesting that mitochondrial metabolism may be important for AM development and/or maintenance. Furthermore, Bach2-deficient AMs had altered lipid metabolism, leading to enhanced lipid accumulation inside the cells (15). Consistent with these findings, we recently have shown that the inhibition of mitochondrial electron transport chain, but not the inhibition of glycolysis, impaired AM proliferation and self-renewal in vivo (49). Using genetic deficiency of TFAM, in this study, we have provided strong evidence that mitochondrial fitness and metabolism are critical for AM maintenance and function. Interestingly, CD11c-cre–mediated TFAM deficiency did not diminish pulmonary DC numbers. Furthermore, Ubc-creERT2 mediated acute depletion of TFAM-impaired AM maintenance but not DCs, IMs, and other immune cell types, arguing that TFAM is selectively required for AM homeostasis in the respiratory tract. The exact reason underlying this phenomenon requires further investigation. However, AMs are the only immune cell type directly exposed to the ambient air in adult mouse lung during homeostasis. Furthermore, alveolar space is full of lipids and lipid-rich surfactants, which are phagocytized by AMs (but are almost deficient with glucose) (36, 37). Therefore, AMs may be selectively adapted to use mitochondrial metabolism and fatty acid oxidation as their energy source due to the oxygen-rich and lipid-rich environment in the alveolar space. Consistent with this idea, TFAM deficiency leads to increased accumulation of lipid inside the cells, most likely due to the impaired lipid oxidation by mitochondria. Of note, TFAM deficiency did not impair AM precursor development and their differentiation into AMs, suggesting that AMs may gradually adapt to the airway environment during lung development and the maturation process, which normally end at 4 wk after birth (50, 51).

AMs are primarily maintained through proliferation and self-renewal independent of monocyte infiltration in the adulthood during homeostasis. Although still controversial, a recent report suggests that monocytes may contribute to a significant portion of AMs in adult mice, and their contributions to the AM compartment gradually increase during aging (52). To this end, a recent report has found that TFAM deficiency did not decrease AM numbers in aged mice (53). We did not explore the roles of TFAM in regulating the AM compartment in our mouse colony due to the time limitation. Nevertheless, it is possible that monocytes eventually infiltrate into the empty AM niche, differentiate into monocyte-derived AMs, and compensate the loss of embryonic-derived AMs in CD11c-Cre Tfam^fl/fl mice. Additionally, it is possible that AMs may eventually adapt to impaired mitochondrial metabolism and use other metabolic pathways for their energy source and homeostatic proliferation. These possibilities require further studies.

Macrophage metabolism is critical in regulating pro- and anti-inflammatory responses of macrophages (54–57). For example, glycolysis is critical for macrophage inflammatory cytokine production, whereas mitochondrial metabolism and fatty acid oxidation promote M2 macrophage differentiation and the production of anti-inflammatory cytokines, such as IL-10 (58, 59). As the resident macrophages in lung alveoli, AMs exhibit M2 macrophage features and are considered as immunosuppressive during homeostasis. We believe that TFAM-mediated mitochondrial metabolism is vital to maintain the immunoquiescent state of AMs, as TFAM deficiency resulted in increased expression of inflammatory gene sets. In contrast, AMs can rapidly upregulate inflammatory responses following viral exposure (60). To this end, we have shown before that AMs from influenza-infected mice exhibited decreased mitochondrial fitness (29), although whether influenza can directly impair AM fitness is not known. To this end, influenza virus is known to cause impaired mitochondrial fitness, diminished lipid oxidation, and mtDNA release in epithelial and/or structural cells (61, 62). We have confirmed in AMs that direct influenza infection caused mitochondrial abnormality, which is associated with decreased TFAM expression. Because TFAM deficiency could result in increased inflammatory responses, it is possible that the impaired TFAM expression in influenza-exposed AMs may help to prime inflammatory responses in AMs following infection in vivo. Furthermore, AM inflammation could ignite and contribute exuberant pulmonary inflammation following severe influenza infection (63, 64). It is thus possible that agents that can increase mitochondrial fitness may be used therapeutically to dampen exaggerated pulmonary inflammatory responses mediated by the exuberant macrophage responses in the future.

In summary, our data have shown the first genetic evidence, to our knowledge, on the roles of mitochondrial metabolism in regulating resident macrophage self-renewal in the respiratory tract. Furthermore, our data have provided insights on the potential beneficial roles of the restoration of mitochondrial function for therapeutic interventions of respiratory viral infection.

CD11c is also expressed by respiratory DCs. We found that TFAM deficiency did not impair DC numbers or subsequent Ag-specific T cell responses following influenza infection, indicating that TFAM deficiency in CD11c⁺ cells may not impair DC quantity and functions following influenza infection. However, we cannot definitively rule out the possibility that TFAM deficiency in DCs or other uncharacterized lung CD11c⁺ cells, other than AMs, may contribute to the observed phenotypes (excessive pulmonary inflammation and increased mortality) in the CD11c^creTfam^fl/fl mice following influenza infection. Further studies are required to address this limitation.

We thank the Mayo Clinic Flow Cytometry Core for assistance.

The authors have no financial conflicts of interest.