**1. Introduction**

Hepatic fibrosis is a histological consequence of the wound-healing process resulting from chronic liver injuries induced by various causes. Advanced fibrosis progresses to liver cirrhosis leading to various life-threatening complications and hepatocellular carcinoma [1]. During long-standing liver injuries, the activation of hepatic stellate cells (HSCs) following hepatocyte damage and the recruitment

of inflammatory mediators lead to the accumulation of extracellular matrix (ECM) [2]. At this time, reactive oxygen species (ROS) are primarily generated in the mitochondria and endoplasmic reticulum of hepatocytes, leading to further hepatocyte damage that results in HSC activation and enhanced ECM production [3]. These vicious pathogenic events of involving hepatocyte damage, inflammation, ROS production, and excessive ECM accumulation can accelerate hepatic fibrosis.

Mitochondria in hepatocytes serve as the primary source of energy; however, their dysfunction is commonly associated with increased ROS production. Moreover, along with being the source of ROS, mitochondria and mitochondrial DNA (mtDNA) can su ffer damage by ROS. Thus, mitochondrial ROS homeostasis is critical for preventing oxidative injury in hepatocytes [4,5]. Once mtDNA is damaged by ROS produced in mitochondria, a cascade of events culminating in apoptosis or cell death proceeds. Studies have suggested that mitochondrial dysfunction in injured hepatocytes can initiate the apoptotic pathway, leading to increased collagen production via HSC stimulation [6–8]. Growing evidence supports a link between mitochondrial dysfunction and liver fibrogenesis, and mitochondrial quality control-based therapy has emerged as a new therapeutic strategy. However, it remains unknown whether mitochondrial dysfunction, specifically in hepatocytes, plays a role in the fibrogenesis, or whether mediators from hepatocyte mitochondrial damage promote liver fibrosis.

High-temperature requirement protein A2 (HtrA2, also known as Omi) is a nuclear encoded serine protease that localizes in the intermembrane space of mitochondria under normal conditions, and it is released into the cytosol upon apoptosis in response to various cellular stresses [9]. The pro-apoptotic function of HtrA2/Omi protease is at least partially mediated via the binding and proteolytic removal of inhibitor of apoptosis proteins. Recent studies illustrated that HtrA2/Omi inactivation does not cause early lethality in non-neuronal tissue, unlike its e ffects in neuronal tissue, but it leads to increased accumulation of mtDNA deletions and premature aging in mammals [10,11]. It has also been demonstrated that HtrA2/Omi deficiency causes mtDNA damage through ROS generation and DNA mutation, which can lead to the accumulation of unfolded proteins in the mitochondria, oxidative stress, and defective mitochondrial respiration, suggesting that HtrA2/Omi is important for mitochondrial homeostasis. Furthermore, our previous studies indicated that HtrA2/Omi deficiency or point mutations in its protease domain cause mtDNA conformational changes through ROS production in cultured cells [12]. The Ser276Cys (S276C) missense mutation in HtrA2/Omi was found to be the cause of symptoms such as muscle wasting, neurodegeneration, involution of the spleen and thymus, and death by 40 days of age in mnd2 (motor neuron degeneration 2) mutant mice. In these mice, the protease activity of HtrA2/Omi is greatly reduced.

Given that both ROS and mitochondrial dysfunction contribute to liver fibrogenesis and that hepatocyte mtDNA can exacerbate HSC activation, we hypothesized that HtrA2/Omi plays a pivotal role in liver fibrosis by modulating mitochondrial homeostasis.

In the present study, we demonstrated that the progression of fibrosis in both animal models and patients is associated with decreased expression of HtrA2/Omi, which modulates mitochondrial function and ROS generation. The modulation of HtrA2/Omi through mitochondrial homeostasis might be a promising anti-fibrotic therapeutic approach. These findings sugges<sup>t</sup> the therapeutic value of HtrA2/Omi in the treatment of liver fibrosis.

#### **2. Materials and Methods**

#### *2.1. Clinical Samples and Animal Studies*

Five liver fibrosis tissues were obtained from patients with diagnosed chronic liver diseases who underwent liver transplantation (Seoul St. Mary's Hospital, Seoul, South Korea) prior to 2010 and stored in liquid nitrogen. None of them had history of any treatment. In addition, three liver tissues (as controls) were also obtained from patients without viral hepatitis during surgical procedures, and they were described in a previous report [13]. All patients provided written informed consent for the storage of liver tissue samples according to the ethical guidelines of Seoul St. Mary's Hospital in the Catholic University of Korea. Their personal information was restricted to analytical purposes. Such information is not available to the public.

All animal care and experimental protocols were conducted in accordance with the guidelines for the Care and Use of Laboratory Animals provided by the Research Supporting Center for Medical Science of the Catholic University of Korea (2016-0005-03). BALB/C and heterozygous *mnd2* (mnd2+/−) mice of the B6(Cg)-Htra2mnd2/J strain were purchased from Orient Bio (Seongnam, Republic of Korea) and Jackson Laboratory (stock no. 004608). Mnd2/mnd2, mnd2/+, and WT mice were obtained by crossing mnd2 heterozygous (mnd2/+) mice. The genotypes of the mice were identified via PCR-AluI-RFLP genotype analysis as previously described [14]. Mice were used when 6–8 weeks old, excluding mnd2/mnd2 mice, which were used at 3 weeks of age, and housed in a standard laboratory animal facility.

To establish an animal model of liver fibrosis, male BALB/C or mnd2/+ mice (from five to seven mice per group) were treated via intraperitoneal injections of CCl4 (Sigma, St. Louis, MO) as previously described [15]. Briefly, mice received CCl4 dissolved in mineral oil (1/4 ratio) or mineral oil alone at a dose of 0.5 mL/kg body weight twice a week for 8 weeks to induce liver fibrosis. The control group received mineral oil alone at the same time. For the preventive study, liver-targeted hydrodynamic gene delivery to the mice was performed as previously described [16,17]. In brief, saline containing 30 μg of pFLAG-HtrA2/Omi plasmid, an expression vector containing the murine HtrA2/Omi open reading frame [18], or its control plasmid was hydrodynamically injected into the liver via a catheter with temporal blood flow occlusions. The injection volume and flow rate were fixed at 5% body weight and 1 mL/s, respectively. CCl4 and the pFLAG-HtrA2/Omi plasmid were administered from five to seven mice in each group every 3 days for 8 weeks. The mice were sacrificed, and their livers were harvested.

#### *2.2. Histological Analysis and Immunohistochemistry*

Liver tissues were fixed in 3.7% bu ffered formalin, and then embedded in para ffin wax. The samples were cut into 3-μm sections and stained with hematoxylin & eosin (H&E) and Sirius Red (Direct Red 80, Aldrich, Milwaukee, WI) to detect collagen deposition. For immunohistochemistry, serial sections were depara ffinized and hydrated through a graded alcohol series. Antigen retrieval was performed by heating the sample in 0.01 M citrate bu ffer (pH 6.0) using a microwave vacuum histoprocessor (RHS-1, Milestone, Bergamo, Italy) at a controlled final temperature of 121 ◦C for 15 min. To block endogenous peroxide activity, the sections were quenched in 3% hydrogen peroxide in methanol and then blocked with 1% bovine serum albumin in PBS. Sections were incubated with primary antibodies against α-SMA and HtrA2/Omi diluted 1:500 in Antibody Diluent (Golden Bridge, Mukilteo, WA) at 4 ◦C. After washing, the peroxidase EnVision System (HRP rabbit/Mouse Envision System TM, Dakocytomation, Denmark) was applied at room temperature for 5–10 min. Peroxidase activity was detected with 3,3-diaminobenzidine tetrachloride (DakoCytomation) and hematoxylin counterstain (DakoCytomation). The percent staining was calculated by the software of the Optimas 6.5 system.

## *2.3. TUNEL Assay*

The TUNEL assay was performed using an in-situ cell death detection kit (Roche Diagnostics GmbH, Mannheim, Germany) following the manufacturer's protocol. After staining, the sections were mounted with mounting medium with 4, 6-diamidino-2-phenylindole (DAPI; Sigma). Apoptotic cells were quantified by counting TUNEL-positive nuclei. For each sample, the number of TUNEL-positive cells was observed under a fluorescent or confocal microscope (Zeiss, Jena, Germany) and counted under ×400 magnification. Six representative fields were evaluated for each mouse in all the experimental groups.

#### *2.4. Isolation of Mouse Primary Hepatocytes and Cell Culture*

Mice were intraperitoneally anesthetized with Rompun (10 mg/kg) and Zoletil (40 mg/kg). These mice were then exsanguinated. Livers were perfused in situ through portal vein with calciumand magnesium-free Hanks' balanced salt solution (HBSS, Welgene, Daegu, Republic of Korea) until the firm texture was lost. After perfusion, soft liver tissue was removed and placed in a 1:1 mixture of Dulbecco's modified Eagle's medium and Ham's F-12 medium (DMEM/F12, Invitrogen, Carlsbad, CA). Subsequently, the liver suspension was poured through sterile 70-μm nylon mesh (BD Sciences, San Jose, CA) and then the homogenate was centrifuged at 50 × g for 2 min. The pellet containing parenchymal cells was washed twice with DMEM/F12 containing 10% fetal bovine serum (FBS, Invitrogen). Isolated primary hepatocytes were plated onto collagen coated plates and cultured in DMEM/F12 supplemented with 10% FBS. The non-tumorigenic mouse hepatocyte cell line FL83B cells was cultured in Ham's F-12K medium containing 10% FBS (Invitrogen), 100 μg/mL penicillin, and 0.25 μg/mL streptomycin. The LX-2 human hepatic stellate cell line (Merck Millipore, Billerica, MA; SCC064) with key features of hepatic stellate cytokine signaling and fibrogenesis was used as described previously [19]. LX-2 cells were cultured in DMEM supplemented with 10% FBS (Invitrogen), 100 μg/mL penicillin, and 0.25 μg/mL streptomycin. The cells were maintained in a humidified incubator at 37 ◦C with 5% CO2.

#### *2.5. Western Blot Analysis*

Protein was extracted from cell lysates using RIPA lysis bu ffer (10 mM Tris-HCl, pH 7.5; 10 mM EDTA; 1% NP-40; 0.1% SDS; 150 nM NaCl; 0.5% sodium deoxychloride; protease inhibitors) for western blotting. Protein extracts were heated at 100 ◦C for 5 min before loading followed by separation on 10% or 12% SDS-polyacrylamide gels, transfer onto nitrocellulose membranes (Schleicher & Schuell, Dassel, Germany), and blocking for 1 h at room temperature in 5% skim milk. The membranes were incubated with primary antibodies overnight at 4 ◦C, followed by incubation (2 h at room temperature) with HRP-conjugated secondary antibodies (Amersham Biosciences, Cardi ff, UK). Target proteins were detected using an enhanced chemiluminescence system (Amersham Pharmacia Biotech, Uppsala, Sweden) according to the manufacturer's instructions. The density of each band was analyzed using the Multi Gauge V3.0 program (Fujifilm, Tokyo, Japan).

#### *2.6. Transmission Electron Microscopy (TEM)*

Cells were collected and fixed with 4% paraformaldehyde and 2.5% glutaraldehyde in 0.1 M phosphate bu ffer, pH 7.2, at 4 ◦C overnight. After rinsing with 0.1 M phosphate bu ffer three times for 30 min each, the cells were treated with 1% osmium tetroxide in 0.1 M phosphate bu ffer for 1 h, dehydrated through a graded series of ethanol and acetone, embedded in Epon 812, and polymerized at 60 ◦C for 3 days. Ultrathin sections (60–70 nm) were prepared using an ultramicrotome (Leica Ultracut UCT; Leica Microsystems GmbH, Wetzlar, Germany). The sections on Formvar-coated slot grids were examined under a transmission electron microscope (JEM 1010; JEOL Ltd., Tokyo, Japan) operated at 60 kV. Images were recorded using a CCD digital camera (Orius SC1000; Gatan, Pleasanton, CA). All experiments were repeated three to five times to ensure reproducibility.

## *2.7. Immunofluorescence Staining*

Cells were fixed with 4% paraformaldehyde for 20 min and permeabilized with 0.5% Triton X (Sigma). After washing three times each with PBS, the cells were blocked with 1% bovine serum albumin in PBS. Subsequently, the cells were incubated overnight at 4 ◦C with the primary antibodies. The cells were washed three times with PBS and incubated with Alexa Flour 488-labeled anti-rabbit IgG (Life Technologies, Carlsbad, CA). Nuclei were visualized by staining for 5 min with 1 μg/mL DAPI. After washing, the preparations were mounted using Kaiser's Glycerol gelatin (Merck, Darmstadt, Germany). The fluorescence intensity of the preparations was detected using a

confocal microscopy (Zeiss). Fluorescence intensity in cells per sections was measured with microscope image-analysis software (ZEN blue software) by a single investigator who was blind to sample identity. In each of four replicate experiments, 20 images were recorded, and the fluorescent cells in each image were counted (every image contained approximately 20 cells). Finally, the cell concentration was calculated from the average number of cells per image.

#### *2.8. Serum Aminotransferase Activity and Hydroxyproline Determination*

Hepatotoxicity was assessed by quantifying the activities of serum alanine aminotransferase (ALT) using an ALT assay kit according to the manufacturer's protocol (Vettest 8008 Chemistry Analyzer; IDEXX Lab., UK). The accumulation of collagen in the liver tissue was determined by estimating the hydroxyproline content, an amino acid characteristic of collagen. Hydroxyproline levels in mouse livers were measured using a hydroxyproline assay kit (BioVision, Milpitas, CA) according to the manufacturer's instructions. The results are reported as milligrams of hydroxyproline per gram of wet liver tissue.

#### *2.9. Lentiviral Vector Transduction*

To establish a stable HtrA2/Omi-depleted cell line, FL83B cells were infected with a mouse HtrA2/Omi specific shRNA-encoded lentivirus (Sigma; SHCLNV-NM\_019752). An shRNA negative control lentiviral particle (LV-Control) was used as a negative control. To generate a stable cell line, FL83B cells were plated at a density of 1 × 10<sup>5</sup> cells per 60-mm culture dish and infected overnight with five multiplicities of infection (MOI) lentiviral particles in the presence of 8 μg/mL hexadimethrine bromide (Sigma). After infection, the transduced cells were selected using 10 mg/mL puromycin (Sigma) for 2 weeks and incubated at 37 ◦C in a humidified incubator with 5% CO2. Suppression of HtrA2/Omi expression in selected cells was confirmed by western blot analysis.

#### *2.10. Quantitative Real Time-PCR-Based Gene Expression*

Total RNA was extracted with TRIzol reagen<sup>t</sup> (Invitrogen) and treated with DNase I (Invitrogen). For first-strand cDNA synthesis, 1.5 μg of total RNA were reverse-transcribed at 42 ◦C for 1 h using a random hexamer primer (Applied Biosystems) and Superscript II reverse transcriptase (Invitrogen). mRNA levels were measured using SYBR Premix Ex Taq (Takara, Japan). The relative mRNA levels were quantified using the comparative ΔCT method, normalized to β-actin. Primer sequences are listed in Supplementary Table S1.

#### *2.11. Mitochondrial Fractionation and mtDNA Extraction*

Following cell lysis, mitochondria were prepared using a mitochondria isolation kit (Pierce Biotechnology, Inc., Rockford, IL), according to the manufacturer's protocol. Isolation and DNase treatment of mitochondrial pellets were performed as described previously [20]. The DNase and RNase-treated mitochondrial pellet was resuspended in lysis bu ffer via gentle pipetting and the suspension was incubated at 37 ◦C for 1 h. A measure of 2 mg of proteinase K (Roche Diagnostics) was added and the lysate was incubated for 1 h at 37 ◦C. mtDNA was purified according to the genomic DNA extraction protocol using a DNeasy Blood & Tissue Kit (Qiagen, Santa Clarita, CA).

#### *2.12. Genomic DNA Extraction and Quantitative PCR (qPCR)*

Preparation of total genomic DNA from cell or liver tissue was performed using a DNeasy Blood & Tissue Kit (Qiagen, Santa Clarita, CA). Kits were used according to the manufacturer's instructions with the inclusion of RNAse A treatment to generate RNA-free genomic DNA, and genomic DNA was eluted using sterile deionized water. Quantitative PCR (qPCR) was conducted on genomic DNA using SYBR Premix Ex Taq in triplicate for each sample.

mtDNA damage was determined as a ratio of the copy number of short mtDNA-79 bp fragments (indicative of damaged mtDNA) to the copy number of long mtDNA-230 bp fragments (indicative of undamaged mtDNA) of the mitochondrial 16S-RNA gene as previously reported (Supplementary Table S1) [21]. In addition, the mtDNA copy number was compared to determine the relative mtDNA:nDNA ratio. Primers were designed within the mitochondria NADH dehydrogenase 1 (*mt-ND1*), and cytochrome oxidase 1 (*mt-COX1*) region of the mitochondrial genome (Supplementary Table S1). The nuclear NADH dehydrogenase flavoprotein 1 (*Ndufv1*) gene was used to standardize the mtDNA copy number to the diploid chromosomal DNA content [22]. Relative gene expression was normalized to that of the single-copy nuclear *Ndufv1* gene ( ΔCT) in each sample.

#### *2.13. Mitochondrial Membrane Potential and ROS Production*

Cells were incubated with 2 uM CM-H2DCFDA (Molecular Probes/Invitrogen) resuspended in warm HBSS or HBSS alone for unstained controls for exactly 15 min. The cells were analyzed on a FACSCalibur flow cytometer (BD). The cellular subset was identified according to size and granularity. We used a mitochondria-specific dye (MitoTracker Green FM) that binds the mitochondrial membrane independently of the membrane potential, and thus, the staining intensity is considered an index of mitochondrial mass. For MitoSOX Red-based flow cytometric detection of mitochondrial superoxide, cells were then incubated with MitoSOX Red superoxide indicator (Invitrogen) for 30 min, washed, and then analyzed on a FACSCalibur. The mean channel fluorescence was converted to absolute fluorescence using an inverse log transformation and normalized to that of untreated cells or WT hepatocytes.

#### *2.14. Measurement of Mitochondrial Respiration*

The OCR and extracellular acidification rate of cells were measured using a Seahorse XF24 extracellular flux analyzer (Seahorse Bioscience, Billerica, MA). In brief, hepatocytes were plated on Seahorse XF 24well plates at a density of 5 × 10<sup>4</sup> per well to achieve 80–90% confluency at the time of assay. Following the overnight attachment of cells, the medium was replaced with Seahorse XF medium, and the manufacturer's protocol for the Mitostress kit was followed (Seahorse Bioscience). In this analysis, sequential injections of 1 μM oligomycin, 1 μM FCCP, and 0.5 μM rotenone/antimycin A were added to the cells to define the basal OCR, ATP-linked OCR, proton leak, maximal respiratory capacity, reserve respiratory capacity, and non-mitochondrial oxygen consumption. Results for mitochondrial respiration were normalized to the total protein content.

#### *2.15. Serine Protease Activity Assy*

The protease activity of HtrA2/Omi in liver sections from WT and mnd2 heterozygous (mnd2/+) mice was assayed with the substrates β-casein. Liver lysates were immunoprecipitated (IP) with HtrA2/Omi-specific polyclonal antibody. The IP complexes were incubated for the indicated times at 37 ◦C with β-casein as a substrate. The reaction samples were resolved by 15% SDS-PAGE, and the processing pattern of β-casein was visualized by staining with Coomassie Brilliant Blue dye (CBB). The level of the HtrA2 was analyzed by western bolt with HtrA2/Omi Ab.
