PINK1 attenuates mtDNA release in alveolar epithelial cells and TLR9 mediated profibrotic responses

We have previously shown that endoplasmic reticulum stress (ER stress) represses the PTEN inducible kinase 1 (PINK1) in lung type II alveolar epithelial cells (AECII) reducing mitophagy and increasing the susceptibility to lung fibrosis. Although increased circulating mitochondrial DNA (mtDNA) has been reported in chronic lung diseases, the contribution of mitophagy in the modulation of mitochondrial DAMP release and activation of profibrotic responses is unknown. In this study, we show that ER stress and PINK1 deficiency in AECII led to mitochondrial stress with significant oxidation and damage of mtDNA and subsequent extracellular release. Extracellular mtDNA was recognized by TLR9 in AECII by an endocytic-dependent pathway. PINK1 deficiency-dependent mtDNA release promoted activation of TLR9 and triggered secretion of the profibrotic factor TGF-β which was rescued by PINK1 overexpression. Enhanced mtDNA oxidation and damage were found in aging and IPF human lungs and, in concordance, levels of circulating mtDNA were significantly elevated in plasma and bronchoalveolar lavage (BAL) from patients with IPF. Free mtDNA was found elevated in other ILDs with low expression of PINK1 including hypersensitivity pneumonitis and autoimmune interstitial lung diseases. These results support a role for PINK1 mediated mitophagy in the attenuation of mitochondrial damage associated molecular patterns (DAMP) release and control of TGF-β mediated profibrotic responses.


Introduction
Idiopathic pulmonary fibrosis (IPF) is a progressive and incurable interstitial lung disease of unknown etiology with median survival of only 2 to 3 years from the time of diagnosis [1]. In the pathogenesis of IPF, chronic microinjury and aberrant activation of type II (AECII) epithelial cells drive accumulation of fibroblasts and myofibroblasts with exuberant deposition of extracellular matrix and abnormal remodeling [2,3]. Our previous work has shown that AECII in IPF lungs contain swollen and dysfunctional mitochondria secondary to deficiency of a key mediator of mitochondrial homeostasis and mitophagy, the PTEN-induced putative kinase 1 (PINK1) [4,5].
Owing to its prokaryotic origins, mitochondrial DNA (mtDNA) contains unmethylated CpG islands akin to bacterial DNA [6], which are recognized by several innate immune pattern recognition receptors including the NLRP3 inflammasome system, Toll-like receptor 9 (TLR9), and the cyclic GMP-AMP synthase-stimulator of interferon genes (cGAS-STING) pathway [7-10]. In addition, activation of TLR9 has been shown to be integral to the development of vascular dysfunction in hypertension, heart failure secondary to myocarditis and liver fibrosis in animal models [11][12][13] suggesting that TLR9 signaling plays a role in mediating pathologic responses to released mtDNA as a danger-associated molecular patterns (DAMPs). Increased levels of circulating mtDNA are associated with tissue injury and disease [14][15][16][17], suggesting that mitochondrial DAMPs play a role in the development or severity of human disease [14]. Recently, circulating mtDNA has been proposed as a predictor of mortality in IPF [18].
We hypothesize that mitochondrial dysfunction in AECIIs increase the release of mtDNA that can function as a DAMP and autonomously activate TLR9 to initiate downstream inflammatory and pro-fibrotic signaling.

Study approval
Human lung tissue used for mitochondria isolation was collected from excess pathologic tissue after lung transplantation and organ donation, under University of Pittsburgh IRB-and CORID-approved protocols (970946, PRO14010265, and CORID No. 300). Samples from Mexican patients were acquired at the Interstitial Lung Disease Clinic of the National Institute of Respiratory Diseases in Mexico City. Plasma and BAL samples were obtained under protocols approved by the local ethics committees, and all participants gave written informed consent.
All animal protocols were approved by the IACUC of the University of Pittsburgh (protocol number 16119337) and adhered to NIH guidelines for the use of experimental animals.
(0.5 ml) at room temperature and was centrifuged at 1500 rpm at 4˚C for 5 min; supernatants were collected and stored at −80˚C for mtDNA measurement. C57BL6 mice were acquired from Jackson Labs (Bar Harbor, ME) and TLR9 KO mice were obtained from the Shlomchik Lab (University of Pittsburgh) [19]. Six-month-old mice (C56BL6 and TLR9 KO) were used to procure the lung tissue for precision-cut lung slices. Mice were euthanized by isofluorane overdose followed by cervical dislocation according to the guidelines from the Institutional Animal Care and Use Committee from the University of Pittsburgh.
For mtDNA or gDNA stimulation, intact mitochondria and nuclei were isolated from fresh lung tissue or cell culture by differential centrifugation in STE buffer (250 mM sucrose, 10 mM Tris, 1 mM EGTA, pH 7.4) as previously described [4]. DNA was isolated from mitochondria (mtDNA) or nuclei (gDNA) using QiAmp DNA extraction kit (Qiagen, 51306), quantify spectroscopically and stored at -80C in "one-use only" aliquots. To assess the purity of the mtDNA isolation for stimulation, human or mouse derived mtDNA was cut at a single site with BglII or PvuII (respectively) and run in a 0.5% agarose gel to generate a sharp band at 16.5kb.

Mitochondrial DNA oxidation and lesion quantification
Mitochondrial DNA oxidation was evaluated by ELISA detection of 8-hydroxy-2'-deoxyguanosine (Cayman #589320) according to the manufacturer's protocol in 3 μg of isolated mtDNA from freshly isolated total lung tissue mitochondria. Mitochondria DNA lesions were quantified using long-run qPCR technique for DNA-damage quantification (LORD-Q) method calculating 10kb lesions using amplification efficiency from standard curve based on Ct values [21,22].

Experimental cell culture, gene silencing and treatments
To knock down protein expression, A549 cells were transfected with siRNA for PINK1 (s35168, Thermo Fisher), TLR9 (s288772, Thermo Fisher) or scramble siRNA controls (Select Negative Control No. 2 siRNA, Thermo Fisher 4390847). Treatments were performed beginning 24 hr after transfection and for an additional 24 hr.
Cells were transfected with PINK1-3xFlag in pReceiver M14 (Genecopoeia) or GFP pReceiver M14 (control) for PINK1 overexpression using a transfecting agent (Lipofectamine 3000, Thermo Fisher). All the transfections were performed in no antibiotic and low FBS (1%) media for 6h, then the media was change to DMEM (Gibco) with 10% FBS (Gibco) and 50 U/ml penicillin with 50 μg/ml streptomycin (Gibco) for the rest of the experimental set-up. Treatments with TM were performed 24h after transfection, and supernatant was harvested 24h post stilulation.
Supernatant and cell pellets (from human primary fibroblast derived from IPF patients or aged-matched controls) here harvested after 48h of culture in the presence and absence of stimulation with 1μg/ml of exogenous mtDNA or TGF-β (5ng/ml).
For the co-culture experiments, A549 cells were grown in transwell inserts overnight and then stimulated with PBS 1μg/ml mtDNA or tunicamycin 10μg/ml for 2h. Supernatant was collected to confirm increase in TGF-β release and then epithelial cells were washed extensively with cell media. Those A549 cells in transwell inserts were place over overnight cultures of human primary fibroblast derived from IPF patients or aged-matched controls. After 48h of co-culture, cell pellets from epithelial cells and fibroblast were collected for RNA extraction.

RNA isolation and qRT-PCR
Purification of RNA was performed in cells and lung tissue using RNA isolation kits (Qiagen 74104), according to the manufacturer's recommendations. To quantify relative gene expression, 1-step qRT-PCR (Affymetrix 75700) was performed for the genes of interest and normalized using 18S RNA as housekeeping gene. Premixed primers and probes were used for TGFB1, TLR9, PINK1, COL1A1, FN1, ACTA2, NOX4, IL6 and IL1B (Integrated DNA Technologies, assay numbers in S6 Table).

Transforming Growth Factor beta 1 quantification
Levels of total TGF-β1 in BAL or cell culture supernatant were quantified using ELISA kits (ThermoFisher Scientific 88-850-22) following the manufacturer's protocol.

Precision-cut lung slices (PCLS)
For human PCLS, single lung segments were dissected and warmed in a 37˚C water bath for 30 min. Low-melting-point agarose (2%, Invitrogen Ultrapure) in sterile medium (DMEM; Gibco, supplemented with 50 U/ml penicillin with 50 μg/ml streptomycin and 2.5 μg/ml amphotericin B) was maintained at 37˚C. The lung segments were filled with agarose (by instillation into airways with an 18-gauge cannula) and inspected for appropriate expansion, followed by airway clamping and packing in ice for 30 min (or until the agarose had set). Tissue was cut to the required block size (2 cm x 1cm x 1 cm) and sliced in ice-cold saline with a vibratome (28mm/sec; Leica VT 1200) at a slice thickness of 300 mm. Uniform slices (1 cm x 1 cm) were cultured in 0.5ml of DMEM containing penicillin-streptomycin and amphotericin B without serum in 24-well dishes at 37˚C in a tissue incubator with 5% CO 2 for 2h. Then, the slices were cultured in DMEM (Gibco) with 10% FBS (Gibco) and 50 U/ml penicillin with 50 μg/ml streptomycin (Gibco) overnight. Finally, slices were stimulated with different doses of exogenous human mtDNA. Supernatant and slice were harvested 24h after stimulation for analysis.
For murine derived PCLS, wild type or knock-out mice were anaesthetised with a mixture of ketamine and xylazine hydrochloride. After intubation and dissection of the diaphragm, lungs were loaded with low-melting-point agarose (2%) in sterile medium (DMEM; Gibco, supplemented with 50 U/ml penicillin with 50 μg/ml streptomycin and 2.5 μg/ml amphotericin B) at 37˚C. The trachea was ligated with thread to retain the agarose inside the lung. The lung was excised, transferred into a tube with medium and cooled on ice for 10 min to allow the agarose hard set. The lobes were separated and cut with a vibratome (28mm/sec; Leica VT 1200) at a slice thickness of 300 mm. Uniform slices were sectioned into 1 cm x 1 cm sections and cultured in 0.5ml of DMEM containing penicillin-streptomycin and amphotericin B without serum in 24-well dishes at 37˚C in a tissue incubator with 5% CO 2 . The medium was changed after 2h to DMEM (Gibco) with 10% FBS (Gibco) and 50 U/ml penicillin with 50 μg/ ml streptomycin (Gibco) and kept in culture overnight. Slices were exposed to 1μg/ml mtDNA from PINK1+/+ mice (WT) or PINK1 -/-(KO) mice and then supernatant and slice were harvested for analysis after 24h.
For viability assessment, (live/dead) extra slices from the same agarose block were taken (at time of preparation) and stained with AOPI Staining Solution (Fisher, NC0285242).

Statistical analysis
Differences between groups were calculated by 2-tailed unpaired Student's t test with Welch's correction or by non-parametric one-or two-way ANOVA followed by post-hoc tests (Sidak's multiple comparisons test). Correlations were calculated by the Spearman method. Statistical significance was considered if the p < 0.05. Statistical analyses were carried out using Prism 7 (GraphPad) and Stata 14.2 (StataCorp) software.

PINK1 deficiency lead to selective cellular release of mitochondrial DNA
We have previously shown that IPF lung epithelial cells present low levels of PINK1 [4], driven by persistent ER stress [23] and chronic transcriptional repression [24], that dysregulate mitochondria homeostasis and induce accumulation of damaged mitochondria. PINK1 deficiency is linked to increase opening of the mitochondrial permeability transition pore (mPTP), which allows nonselective traffic to the cytosol liberating mitochondrial content, including mtDNA [25]. To investigate the relationship between ER stress, PINK1 deficiency and mtDNA release, A549 cells were treated for 24h with increasing concentrations of tunicamycin (TM) to induce ER stress (S1A Fig). PINK1 transcript was reduced after ER stress stimulation. Concomitantly, mtDNA was detected in the cell culture supernatant in a dose dependent manner (Fig 1A). In high contrast, levels of extracellular nuclear DNA in the cell media were very low and were not different in the presence or absence of ER-stress induced PINK1 repression (Fig 1B). Primary human alveolar epithelial cells show higher number of mtDNA copies released ( Fig 1C) when subjected to ER stress (S1B Fig), even at low concentrations of tunicamycin (0.1 μg/ml) with no differential detection of nuclear DNA in the supernatant ( Fig 1C). Similarly, knockdown of PINK1 expression in A549 cells (S1C Fig) resulted in increased release of mtDNA ( Fig 1D). To determine if selective release of mtDNA linked to PINK1 deficiency occurs in vivo, we analyzed levels of mtDNA in BAL fluid of wild type and PINK1 knockout mice (S1D Fig) and found a significant increase in mtDNA in the PINK1 -/mice ( Fig 1E). This ER stress induced mtDNA release is driven by the loss of PINK1 (Fig 1F), since the overexpression of PINK1 (S1E Fig) in TM-treated cells inhibits mtDNA leak to the cell media ( Fig 1F). These data show that, in vitro and in vivo, PINK1 deficient lung alveolar epithelial cells (induced by ER stress or by silencing/knockout) release extracellular mtDNA.

PINK1 deficiency and isolated mtDNA increase pro-inflammatory and pro-fibrotic cytokines release
To confirm that mtDNA specifically triggered pro-inflammatory and pro-fibrotic responses in lung epithelial cells, we analyzed A549 cells stimulated with exogenous mtDNA (1μg/ml) in the media. TGF-β release was increased by mtDNA stimulation (Fig 2A). In addition, TGF-β production was mtDNA concentration-dependent also in mouse lung epithelial cell lines as MLE12 (S2A Fig). As pro-inflammatory marker, we evaluated IL-6 expression that was upregulated after mtDNA stimulation (S2B Fig). IL-6 mRNA transcript level was significantly higher in whole lung lysates of PINK1 knockout mice, while total TGF-β was 10-fold higher in the BAL fluid of the deficient mice (S2C Fig). Similarly, IL-6 and TGF-β increased in A549 cells with PINK1 siRNA knockdown for 48 hours (S2D Fig). This TGF-β response was specific to mtDNA external stimulation since pre-treatment of media with DNAse (1U/ml) abrogated this response, and the stimulation with nuclear DNA (1μg/ml) had a much lesser effect than mtDNA (Fig 2A). Additionally, we tested transfection of mtDNA or nuclear DNA into A549 cells, but this method did not significantly change the expression of TGF-β (Fig 2A). To test if the trigger of this TGF-β activation resides in the cell surface, cells were concurrently stimulated with mtDNA and an endosome active agent ( Fig 2B). Blocking early endocytosis with Dynasore or later endosomal acidification with chloroquine decrease the release of TGF-β to the media after mtDNA stimulation ( Fig 2B). Altogether, this data suggest that mtDNA can externally trigger pro-inflammatory and pro-fibrotic responses in lung epithelial cells. It also suggests that the stimulus driving the activation of TGF-β occurs on the cell surface and on the inner surface of the endosome after endocytic uptake of mtDNA rather than in the cytosol.
Toll-like receptor 9 mediates TGF-β release in response to mitochondrial DNA Non-mitochondrial located free mtDNA can engage multiple innate immunity pattern recognition receptors. Since the TGF-β response to mtDNA stimulation is mediated by endocytosis ( Fig 2B) and TLR9 is known to translocate to the endosome, we examined the role of TLR9 in upregulating pro-inflammatory and pro-fibrotic cytokines in the setting of PINK1 deficiency and mtDNA stimulation. Cells stimulated with a TLR9 agonist (ODN M362) upregulated the mtDNA drives profibrotic responses in AECII secretion of TGF-β to the culture media, however in the presence of dynasore or chloroquine, this upregulation was reduced ( Fig 3A). As well, we examined the effect of PINK1 deficiency on TLR9 expression. We found that TLR9 expression was upregulated in total lung lysate of PINK1 knockout mice when compared their wild type littermates (Fig 3B), at 48 hours after PINK1 siRNA knockdown in A549 cells (Fig 3C) as well as after treatment with tunicamycin ( Fig 3D). In primary human lung epithelial cells, TRL9 expression is upregulated not only when PINK1 is reduced by tunicamycin treatment but also in the presence of mtDNA external stimulation (Fig 3E). Furthermore, when primary human lung epithelial cells are pretreated with DNAse and then stimulated with mtDNA, TLR9 expression upregulation is eliminated (Fig 3E).
In addition, TLR9 signaling can induce pro-inflammatory and pro-fibrotic signaling through the NFκB pathway [26]. To demonstrate the involvement of the TLR9-NFκB signaling pathway, A549 cells were treated with PINK1 siRNA to induce mtDNA release (Fig 3F) or were directly stimulated with 1 μg/mL extracellular mtDNA (Fig 3G). Both cases resulted in a statistically significant increase in TGF-β release, and, concurrent pre-treatment with a TLR9 antagonist (ODN-TAG), the NFκB antagonist BAY11-7082, or transfection with TLR9 siRNA returned the level of TGF-β to baseline (Fig 3F and 3G). Finally, this TLR9-mediated response in primary human lung epithelial is more sensitive to mtDNA stimulation, since addition of nuclear DNA only induced a very modest increase in TGF-β release and treatment with mtDNA had a similar effect when compared to treatment with the TLR9 agonist, ODN M362 (Fig 3H). Concurrent treatment of primary human lung epithelial cells with mtDNA and either DNAse or a TLR9 antagonist resulted in similar mitigation of induced TGF-β (Fig 3H). Similarly, IL-6 mRNA transcripts increased to comparable levels when primary human lung epithelial cells were treated with mtDNA or with ODN M362. Pretreating with DNAse or ODN M362, once again, returned TGF-β and IL-6 expression to baseline (S3B Fig). In sum, these data suggest that, in lung epithelial cells, extracellular mtDNA increases TGF-β via the TRL9-NFκB signaling pathway and indicates the presence of a positive feedback loop with activation of TLR9 expression upregulation.

IPF and PINK1-deficient lungs have increased oxidation and instability of mitochondrial DNA
Since PINK1 has a role in mtDNA metabolism [4,27], we analyzed mtDNA damage in lungs of PINK1 deficient mice (Fig 4A). We quantified DNA lesions in lung mtDNA using a long extension qPCR technique for DNA-damage quantification (LORD-Q) method [21,22] and show more lesion in the PINK1 -/-mice (Fig 4B). To determine whether oxidation of mtDNA was increased and/or accumulative, we analyzed the content of 8-hydroxy-2'-deoxyguanosine (8-OH-dG), the most frequent oxidative DNA modification, on purified mtDNA samples from PINK1 WT and KO mice at different ages. At 3 months of age, mtDNA oxidation was greater in PINK1 KO mice ( Fig 4C) and increases with age, in both WT and KO mice (S4A Fig). To determine if changes in PINK1 expression can influence the extent of mtDNA lesions caused by oxidative stress, we exposed A549 lung epithelial cells (with and without silencing PINK1 expression) to 400 μM H 2 O 2 for 15 minutes. The mtDNA of PINK1 depleted cells (S4B Fig) were more susceptible to genotoxic stimulation relative to a scrambled control. In addition, after washing the cells for 1h, siPINK1 cells were unable to repair its damaged mtDNA as shown by the persistence of lesions (S4C Fig). In humans, as the lung age, the levels of PINK1 drop and further decrease in the IPF lung [4] (Fig 4D and S1 Table). Mitochondrial DNA lesions were higher in old and IPF lungs compared to the young group (Fig 4E and S2 Table). These data show that the lung accumulate unrepaired mtDNA damage with aging and disease.

Release of oxidized mitochondrial DNA increases downstream production of TGF-β
To confirm that mtDNA exposure induced a pro-fibrotic response in human lung, precisioncut lung slices (PCLS) of control donors (60±3) were treated with increasing doses of mtDNA for 24 hours (S5A Fig). We observed a dose-dependent increase in the release of total TGF-β upon stimulation (Fig 5A). In addition, higher TGF-β release happened when the mtDNA was  mtDNA drives profibrotic responses in AECII isolated from lung of older patients (87yo) vs younger counterparts (with higher expression of PINK1); suggesting a role for the mtDNA oxidation state in the induction of the pro-fibrotic response (Fig 5B). To evaluate further the role of mtDNA oxidation state upon PINK1 deficiency, we cultured murine PCLS derived from C57BL6 mice and treated them with mtDNA isolated from PINK1 WT or KO mice (S5B Fig). A more oxidized mtDNA from PINK1 KO mice induced a higher upregulation of TGF-β secretion when compared to PINK1 WT derived mtDNA (Fig 5C). To confirm the role of the TLR9 signaling pathway in whole lung tissue, we cultured murine PCLS derived from TLR9 KO mice. The addition of mtDNA, from either PINK1 WT or KO mice, had no effect on total TGF-β levels indicating that TLR9 plays a key role in profibrotic signaling in response to mtDNA (Fig 5C). In summary, these data suggest that mtDNA oxidation state plays a role in its reactivity and mtDNA derived from PINK1 deficient cells has a greater potential to stimulate pro-fibrotic and pro-inflammatory responses.

Circulating mitochondrial DNA levels correlate with severity of disease in IPF and is linked to PINK1 deficiency
Cell-free mtDNA is shed into circulation in many medical conditions that involve significant tissue injury [15,16,28] and in IPF [18]. To investigate the link between PINK1 levels, cellfree mtDNA and ILD, we obtained plasma and bronchoalveolar lavage (BAL) samples from a cohort of patients from the Interstitial Lung Disease Clinic of the National Institute of Respiratory Diseases in Mexico City. This cohort included healthy controls and patients diagnosed with IPF, hypersensitivity pneumonitis (HP) or autoimmune-related ILD (S5 Table). First, we confirmed that lungs from IPF patients (which are characterized by high levels of TGFβ [29][30][31]) show low PINK1 expression, compared with age-matched donor controls; but also do other ILD, in a smaller magnitude (Fig 6A and S4 Table). In the patients, plasma and BAL fluids were obtained at the moment of diagnosis and before any treatment was indicated. In BAL, mtDNA was statistically increased in all the ILD disease groups compared with controls. Of note, BAL mtDNA levels were statistically higher in IPF when compared with HP and autoimmune-related ILD (Fig 6B and S7 Table). In plasma, while the level of circulating mtDNA was increased in all three groups compared to controls, no difference were found among them ( Fig  6C and S8 Table). In IPF patients, increased levels of circulating mtDNA in plasma correlated with lower diffusing capacity for carbon monoxide (DLCO), lower percentage of forced expiratory volume for 1 second (FEV1%), and poorer performance on a 6-minute walk (6MWD) test supporting a direct correlation between levels of circulating mtDNA and IPF severity (S6A Fig). Although the release of mtDNA demonstrates potential clinical utility in IPF [18], these data suggest that it could be possible to expand those studies to other ILDs since all share a lower expression of PINK1 linked to mtDNA release.

Mitochondrial DNA crosstalk signal between epithelial cells and fibroblasts
To test if the role of fibroblast as the source of mtDNA release, we isolated human lung primary fibroblast from young donors (33±4), old age-matched controls (67±3) or IPF patient (68±4). As we previously reported [4], there is no change in PINK1 expression between IPF and non-IPF lung isolated fibroblasts. After 24h of culture in FBS free media there was no differential release of mtDNA in to the supernatant from age-matched control or IPF fibroblast, only small contribution from the normal young fibroblast (Fig 7A). When the fibroblasts were cultured in 10% FBS, young donor derived fibroblast differentially release mtDNA but neither the old or IPF fibroblast since the mtDNA detected in their media correlate with the leak of nuclear DNA (probably due to cell death) (Fig 7A). Non-stimulated primary human lung fibroblasts isolated from IPF lungs had increased expression of pro-fibrotic markers (collagen and fibronectin), markers of activation (α-SMA and Nox4), and a slight higher expression of TLR9 (1.5-fold) compared to fibroblasts isolated from aged-matched non-IPF controls (IPF = 67±3 y; Control = 68±4 y) (Fig 7B). To demonstrate that both IPF and non-IPF fibroblasts were capable of increasing the expression of these markers of fibrosis they were treated with 5ng/mL TGF-β for 48 hours. Both groups of fibroblast upregulated the expression of collagen, Nox4, fibronectin and α-SMA; while TLR9 and PINK1 expression remained unchanged. However, the relative increase in fibronectin (IPF, 8-fold vs non-IPF, 4-fold) and α-SMA (IPF, 13-fold vs non-IPF, 3-fold) was significantly greater in IPF fibroblasts (Fig 7C). To test the reactivity of fibroblast to exogenous DNA, we treated IPF and non-IPF lung fibroblasts with 1 μg/mL of mtDNA or nDNA for 48 hours. Upon stimulation with mtDNA, only the non-IPF derived fibroblast show statistically significant increase in α-SMA, Nox4, and TLR9 ( Fig 7D). Primary human lung fibroblasts exposed to nDNA did not show changes in the expression of the reposted marker (Fig 7E) regardless of their origin. These findings suggest that non-IPF fibroblasts have a pro-fibrotic response to mtDNA DAMPS, while IPF fibroblasts that already show a myofibroblast-like phenotype are less sensitive to this stimulus (Fig 8).

Discussion
Our previous studies have shown that AECII of IPF patients have altered mitochondrial morphology and function associated with deficiency of PINK1 [4]. In this study, we report that low levels of PINK1 in AECII resulted in mitochondrial DNA damage and release of mtDNA. Released mtDNA was recognized by TLR9 inducing downstream TGF-β-mediated profibrotic responses, as well as the well-known established inflammatory mediators for this pathway. This pro-fibrotic response was originated in AECII by a NFκB dependent mechanism and led to lung fibroblast activation. We have also shown that increased levels of circulating mtDNA are found in IPF as well as HP and ILD associated with autoimmune diseases patients. However, in bronchoalveolar fluids mtDNA was significantly increased in IPF compared to the other two ILDs, and moreover, a significant correlation between levels of plasma mtDNA and disease severity was found only in IPF patients.
Damaged mitochondria have been implicated in the induction of inflammation through the production of reactive oxygen species and the release of DAMPS, including mtDNA. For instance, deficiency of Parkin and PINK1 in mice under the stress of exhaustive exercise, provoke cytosolic release of mtDNA with activation of the STING-IRF3-dependent signaling pathway inducing interferon-mediated inflammatory responses [32]. Interestingly, our results revealed that leakage of mtDNA in the lung of PINK1 deficient mice can also induce inflammatory and fibrotic responses through the TLR9 endosomal pathway. A potential explanation is that PINK1 and Parkin not only regulate mitophagy but also participate in the modulation of mitochondrial endosomal degradation, a pathway where mitochondria can be sequestered inside the early endosomes and subsequently be delivered to lysosomes for degradation and clearance [33]. TLR9 recognition of mtDNA DAMPs leading to pro-fibrotic signaling through upregulation, activation, and release of TGF-β has been previously described in other fibrotic disorders such as systemic sclerosis and in a model of prostate cancer [34,35]. The primary function of TLR9 is to recognize viral or bacterial DNA pathogen-associated molecular patterns, specifically hypomethylated CpG motifs, and initiate an innate immune response through upregulation of inflammatory mediators such as IL-6, IL-12, and TNF in response to the presence of these pathogens. TLR9 recognition of mtDNA is likely an unfortunate consequence of the theorized origins of mitochondria as incorporated prokaryotic symbionts. In lung tissue, TLR9 is generally expressed on professional immune cells such as resident B cells and dendritic cells, but is upregulated in both, lung epithelial cells and fibroblasts in various disease states [36,37]. .001 and �� p<0.01 vs Young, ### p<0.001; two-way ANOVA with multiple comparisons). (B) Baseline expression (non-stimulated) of different pro-fibrotic genes in age-matched human lung fibroblast from control and IPF patients (n = 3; ��� p<0.001, �� p<0.005, � p< 0.05 vs control; two-way ANOVA with multiple comparisons). (C) Expression of different pro-fibrotic genes in age-matched human lung fibroblast from control and IPF patients after 48h of TGF-β stimulation (5ng/ml) (n = 3; ��� p<0.001, vs unstimulated; two-way ANOVA with multiple comparisons). (D) Fold change (over non-stimulated baseline) of pro-fibrotic markers expression after 48h of stimulation with 1μg/ml of exogenous mtDNA in age-matched human lung fibroblast from control and IPF patients (n = 3; ��� p<0.001 vs unstimulated; two-way ANOVA with multiple comparisons). (E) Fold change (over non-stimulated baseline) of profibrotic markers expression after 48h of stimulation with 1μg/ml of nuclear DNA in age-matched human lung fibroblast from control and IPF patients (n = 3; two-way ANOVA with multiple comparisons).
https://doi.org/10.1371/journal.pone.0218003.g007 mtDNA drives profibrotic responses in AECII Interestingly, increased TLR9 expression has been reported in IPF lung from patients who rapidly deteriorated, and in this context, activation of TLR9 signaling has been implicated in accelerated progression of IPF [38]. In epithelial cells, transcription of TLR9 can be upregulated by activation of TLR9 signaling but in the presence of severe cell stress NF-kB signal can directly modify TLR9 transcription [39,40]. In addition, stimulating TLR9 signaling with CpG-containing oligonucleotides in normal fibroblast is capable of inducing a more invasive phenotype on those cells [41,42]. While those experiments were conducted with CPG ODN rather than mtDNA as a TLR9 agonist, our findings that show that mtDNA derived from PINK1-depleted AECII can be a more effective driver of fibroblast activation in the progression of IPF. Finally, previous studies using Balb/c TLR9 deficient mice have failed to demonstrate a role of TLR9 in the pathogenesis of lung fibrosis because this background is resistant to the classical model of bleomycin induced lung fibrosis [43]. Moreover, TLR9 deficient mice have attenuated antiviral responses (by a higher production of IFN-β) limiting its use in the MHV68-induced lung fibrosis model [43].
Previous studies have shown increase of circulating mtDNA in plasma and BAL of IPF patients and strong association with disease progression [18]. We found that high levels of circulating mtDNA are not exclusive of IPF (in the context of ILDs) but only in this disease, the levels correlate with the severity of the disease at the time of diagnosis. Whether the mechanisms of mtDNA release in other ILD is different from the one in IPF or if there is a threshold in the microenvironment were the effect of mitochondrial DAMPs is even more deleterious, will require further studies. In IPF, our data suggest that defective mitochondrial homeostasis by PINK1 deficiency in AECII leads to oxidative damage and release of mtDNA with the Chronically injured, ER-stress sensitive, PINK1-deficient AECII in IPF lungs are the driver to this pro-fibrotic signaling through the release of mtDNA and upregulation of TGFb (via a TLR9-NFκB axis) and it will affect the effector cell (the fibroblast) in different ways. TFGb release could perpetuate the migration of wound healing responding fibroblast to this microenvironment, while mtDNA DAMPs could be key to the progression of the disease by transforming the pool of these fibroblasts to a more pro-fibrotic myofibroblast-like phenotype.
https://doi.org/10.1371/journal.pone.0218003.g008 mtDNA drives profibrotic responses in AECII subsequent activation of TLR9 and expression of TGF-β. The AECII-derived mtDNA DAMPs could play a key role in the progression of the disease by transforming the pool of normal fibroblast to a more pro-fibrotic myofibroblast-like phenotype. Thus, potential therapeutic approaches that target the clearance of damage mitochondria might be and effective therapy for this complex disease.
Supporting information S1 Table. Demographic characteristics of lung's patient cohort in Fig 4D. (DOCX) S2 Table , diffusing capacity for carbon monoxide DLCO% (n = 32) and distance on a 6-minute walk test (6MWD n = 25). (B) Levels of nuclear DNA (gDNA) detected in plasma of IPF patients does not correlate with the mtDNA copy numbers found (n = 60). Also, copy numbers of mtDNA in BAL does not correlate with mtDNA copy numbers in plasma (n = 12). (C) IPF circulating levels of nuclear DNA in plasma does not correlate with functional characteristics and markers of disease severity FEV1% (n = 43), DLCO% (n = 32) or 6MWD (n = 25). (D) Hypersensitivity pneumonitis (HP) circulating levels of mtDNA in plasma does not significantly correlate with functional characteristics and markers of disease severity FEV1% (n = 37), DLCO% (n = 22) or 6MWD (n = 4). Spearman correlation with graphical representation of the 95% confidence range. (TIF)