HDAC3 deficiency protects against acute lung injury by maintaining epithelial barrier integrity through preserving mitochondrial quality control

Abstract

Sepsis is one common cause of acute lung injury (ALI) and acute respiratory distress syndrome (ARDS), which is closely associated with high mortality in intensive care units (ICU). Histone deacetylase 3 (HDAC3) serves as an important epigenetic modifying enzyme which could affect chromatin structure and transcriptional regulation. Here, we explored the effects of HDAC3 in type II alveolar epithelial cells (AT2) on lipopolysaccharide (LPS)-induced ALI and shed light on potential molecular mechanisms. We generated ALI mouse model with HDAC3 conditional knockout mice (Sftpc-cre; Hdac3^f/f) in AT2 and the roles of HDAC3 in ALI and epithelial barrier integrity were investigated in LPS-treated AT2. The levels of HDAC3 were significantly upregulated in lung tissues from mice with sepsis and in LPS-treated AT2. HDAC3 deficiency in AT2 not only decreased inflammation, apoptosis, and oxidative stress, but also maintained epithelial barrier integrity. Meanwhile, HDAC3 deficiency in LPS-treated AT2 preserved mitochondrial quality control (MQC), evidenced by the shift of mitochondria from fission into fusion, decreased mitophagy, and improved fatty acid oxidation (FAO). Mechanically, HDAC3 promoted the transcription of Rho-associated protein kinase 1 (ROCK1) in AT2. In the context of LPS stimulation, the upregulated ROCK1 elicited by HDAC3 could be phosphorylated by Rho-associated (RhoA), thus disturbing MQC and triggering ALI. Furthermore, we found that forkhead box O1 (FOXO1) was one of transcription factors of ROCK1. HDAC3 directly decreased the acetylation of FOXO1 and promoted its nuclear translocation in LPS-treated AT2. Finally, HDAC3 inhibitor RGFP966 alleviated epithelial damage and improved MQC in LPS-treated AT2. Altogether, HDAC3 deficiency in AT2 alleviated sepsis-induced ALI by preserving mitochondrial quality control via FOXO1-ROCK1 axis, which provided a potential strategy for the treatment of sepsis and ALI.

Graphical abstract

HDAC3 inspired the disturbance of mitochondrial quality control (MQC) mediated by ROCK1 in AT2 by deacetylating FOXO1 and activating the transcription of RCOK1, thus triggering epithelial barrier damage during ALI. HDAC3 suppression by chemicals with RGFP966 may have therapeutic potential for alleviating sepsis-induced ALI.

1. Introduction

Sepsis is a life-threatening organ dysfunction syndrome caused by uncontrolled host response to infection, giving rise to serious health burden each year [1]. According to the statistics released by Global Burden of Disease study, there were over 48.0 million sepsis patients in 2017 globally and sepsis-related death accounted for nearly 20% of all-cause deaths [2]. Organ injury is one of the major complications during sepsis, and sepsis-induced acute lung injury (ALI) is highly prevalent, which may deteriorate into acute respiratory distress syndrome (ARDS) and respiratory failure [3,4]. Despite the poor understanding of accurate cellular and molecular basis contributing to ALI, acute inflammation and disturbance in epithelial barrier integrity have been proved to participate in the pathogenesis of sepsis-induced ALI [5,6]. Alveolar epithelium possesses critical biological functions. On the one hand, the alveolar epithelium builds a natural physical barrier against exogenous microbes and fine particulate matters [7]. On the other hand, alveolar epithelium also initiates the innate immunity by establishing an antiviral state and activating various immune cells [8]. During ALI, inflammatory cells, proteinaceous fluid, as well as hyaline membranes enter the alveolar space via impaired epithelial barrier integrity, resulting in hypoxemia, decreased lung compliance, and respiratory distress [9]. Therefore, it is reasonable to treat sepsis-induced ALI by maintaining an intact alveolar epithelial barrier.

Mitochondria are elongated double-membrane-bound organelles in the cytoplasm of almost all eukaryotic cells, possessing independent self-replicating genome [10]. Apart from providing adenosine triphosphate (ATP) via oxidative phosphorylation (OXPHOS), mitochondria are also implicated with a great many biological events, including oxygen consumption, calcium homeostasis, intracellular reactive oxygen species (ROS) generation, and cell signaling transduction [11,12]. Persistent mitochondrial damage and dysfunction can result in organ failure and poor outcome in patients with sepsis, thus cleaning unhealthy mitochondria and generating healthy mitochondria by mitochondrial quality control (MQC) are critical for maintaining structural and functional integrity of the mitochondria. MQC could be modulated through various processes involving mitochondrial dynamics, mitophagy, mitochondrial biogenesis, and mitochondrial redox regulation [13]. Wang XR et al. reported that mitoquinone (MitoQ) relieved alveolar epithelial cell apoptosis and barrier breakdown by regulating mitochondrial fission during sepsis-induced ALI [14]. In lipopolysaccharide (LPS)-induced ALI, mitophagy could be activated in type II alveolar epithelial cells (AT2) [15,16]. Upon LPS stimulation, Parkin could be recruited to the mitochondria, followed by the phosphorylation by PINK1 in the outer mitochondrial membrane. And inhibiting mitophagy in AT2 contributed to the mitigation of LPS-induced ALI [17,18]. Mitochondrial fatty acid oxidation (FAO) makes up a main proportion to the energy needs during infection and metabolic stress [19]. In lung tissues from mice with sepsis, PGC-1α-mediated FAO was significantly impaired in AT2. Ablation of AT2 PGC-1α further aggravated ALI while activating PGC-1α-mediated FAO using fenofibrate significantly relieved LPS-induced ALI [20]. These findings suggest that MQC in alveolar epithelial cells may act as a promising therapeutic candidate for sepsis-induced ALI and ARDS.

Histone deacetylase 3 (HDAC3) is one of the members in HDAC family, possessing four splicing variants including HD3α, -β, -γ, and -δ. HDAC3 could not only regulate gene transcription by removing acetyl from histone but also modify non-histone proteins including certain mitochondrial and cytoplasmic proteins posttranslationally [21]. One recent study reported that HDAC3 could translocate to mitochondria to deacetylate mitochondrial trifunctional enzyme subunit α, one FAO enzyme, thus blocking FAO in macrophages [22]. Additionally, inhibition of HDAC3 via RGFP966 improved mitochondrial membrane potential and decreased the mitochondria-mediated apoptosis [23]. Similarly, pharmacologic inhibition of class I HDACs via SAHA or MS275 in the context of diabetes and obesity enhanced mitochondrial function as well as oxidative capacity in adipose tissue and skeletal muscle [24]. Under hypoxic conditions, HDAC3 accelerated the progression of pulmonary fibrosis by promoting epithelial-mesenchymal transition in alveolar epithelial cells [25]. However, it remains unknown whether HDAC3 deletion could attenuate LPS-induced ALI and epithelial barrier dysfunction by modulating MQC in alveolar epithelial cells.

In the present study, we explored the effects of HDAC3 on LPS-induced ALI and epithelial barrier dysfunction, uncovering that HDAC3 inspired ROCK1-mediated MQC in AT2 in the context of LPS stimulation. In addition, HDAC3 directly decreased the acetylation of forkhead box O1 (FOXO1) and increased its nuclear translocation, thus promoting the transcription ROCK1. Our findings revealed that targeting HDAC3 might be a potential strategy for sepsis-induced ALI.

2. Materials and methods

2.1. Animals and treatments

All experimental animal-used care as well as procedures were carried out in line with the National Institutes of Health (NIH) guidelines, which were approved by the Animal Care and Use Committee of Renmin Hospital of Wuhan University. All mice were fed in an environment-controlled specific pathogen-free (SPF) barrier system with free access to water and food. Wild type (WT) male C57BL/6 mice were provided by the Institute of Laboratory Animal Science, Chinese Academy of Medical Sciences. HDAC3floxed (HDAC3^flox/flox) mice were generated via the CRISPR-Cas9 system in the Shanghai Model Organisms Center, Inc. HDAC3^flox/flox mice were crossed with the tamoxifen-inducible Sftpc-CreERT2 mice kindly provided by GemPharmatechCo., Ltd (Jiangsu, China). To generate AT2 HDAC3-deficient mice, the Sftpc-CreERT2 mice were crossed with HDAC3^flox/flox mice to obtain Sftpc-CreERT2⁺-HDAC3^flox/flox (HDAC3 CKO). And the Sftpc-CreERT2^--HDAC3^flox/flox (HDAC3-C) mice from littermates were as controls. Male mice (8–10 weeks old) were treated with tamoxifen (10 mg/kg) dissolved in corn oil intraperitoneally for five consecutive days before experiments. Mice treated with RGFP966 were given subcutaneous injection of RGFP966 (15 mg/kg) twice a day for 7 days before modeling. The ALI mice model was established on the basis our previous description [26]. In detail, the mice were subjected with LPS (10 mg/kg) purified by phenol extraction (0111:B4) Sigma-Aldrich Co., Ltd., Shanghai, China) dissolved in 50 μL sterile saline intratracheally for 12 h. At end point after the LPS challenge, the murine lungs were collected and frozen in liquid nitrogen, which were then transferred to −80 °C for further biochemical measurement or preserved in formalin (10%) for subsequent histological analysis.

2.2. Construction of adenovirus carrying overexpression plasmid

Firstly, an appropriate restriction endonuclease is selected to digest the vector, and the purified linearized vector is recovered by agarose gel electrophoresis. The target fragment is amplified by PCR using designed primers, and the correctly sized fragment is obtained by agarose gel electrophoresis recovery. The linearized vector and the target fragment are ligated using homologous recombination or T4 ligation methods. The resulting product is transformed and cultured for 12–16 h. Single colony picking is performed to validate the clones, and positive clones with correct validation are selected for sequencing. Finally, plasmid extraction is performed on the correctly sequenced clone samples.

After performing high-purity endotoxin-free plasmid extraction, the plasmids were transfected into 293A cells using transfection reagent. After virus production, high-titer adenovirus stocks were obtained either by virus amplification or by ultracentrifugation-based purification, depending on the customer's experimental requirements. Finally, the adenovirus was purified and stored for later use.

2.3. Isolation of primary AT2 in mice and treatment

Primary AT2 were isolated and purified from neonatal wild type mice by membrane filtration and immune-adhesion according to previous description [27]. To be more specific, after the mice were disinfected with alcohol (75%), their lung tissues were excised and digested with trypsin and collagenase, followed by the elimination of red blood cell using lysis buffer. Subsequently, the cell suspension was preliminary filtrated to obtain AT2 on the basis of the difference of cell size. AT1 and macrophages were larger than AT2 so that large tissue debris could be removed by primary filtration via a 74 μm membrane first. Next, all AT1 cells as well as some residual macrophages were thoroughly removed by second filtration via a 38 μm membrane. At last, the AT2 were filtrated a filter membrane with a pore size of 19 μm, which were further was purified by immune-adhesion using plastic plate coated with IgG. The primary AT2 were cultured at 37 °C with CO₂ (5%) in medium containing 10%FBS and Ham's F–12K for 72 h. Then the growth status and cell morphology were observed and identified. The AT2 cells obtained were identified by flow cytometry with SP-C positive as the standard. The results showed that AT2 cells could still basically meet the experimental requirements after 72 h of culture in vitro, but a large part of cells could not be used after 144 h of differentiation (Fig. S1 A-B, C-D). Therefore, we shortened the culture time of AT2 cells and processed them in time for subsequent experiments. To overexpress HDAC3 or ROCK1 in AT2, cells were transduced with adenovirus encoding Hdac3 (multiplicity of infection = 100) or Rock1 (multiplicity of infection = 150) (DesignGene Biotechnology, Shanghai, China) particles for 12 h, and Ad-NC was used as a control. To knockdown HDAC3 or ROCK1, AT2 were transfected with siHDAC3 or siROCK1 (50 nmol/L) for 4 h using Lipo 6000TM, which were then cultured for an extra 24 h. HDAC3 selective inhibitor RGFP966 (15 μM) dissolved in dimethylsulfoxide (DMSO) was used to prove the effects of HDAC3 inhibition on LPS-induced AT2 in vitro. To mimic the in vitro ALI model, LPS (100 μg/mL) was added to stimulate AT2 for 6 h [26].

2.4. Collection and analysis of bronchoalveolar lavage fluid (BALF)

After the end of LPS stimulation, the chest cavity of the mice was opened layer by layer, followed by the insertion of cannula into the lower end of the right main bronchus. Then the surgical sutures were used to ligate the trachea and lavage needle. To obtain BALF, 1 ml of saline was slowly injected into trachea to flush lung lobes three times. Subsequently, the BALF was collected and centrifuged at 4 °C for 10 min at 2500 rpm, and the supernatants were stored for subsequent measurement. The total protein in BALD was analyzed using bicinchoninic acid (BCA) method via a commercial BCA kit (#ST2222, Beyotime Biotechnology, Shanghai, China). To analyze total cells, macrophages, and neutrophil count in BALF, cell specimens were dropwise added on glass slides and stained with Giemsa dye (#C0133, Beyotime Biotechnology, Shanghai, China). The number of total cells, macrophages, and neutrophils and macrophage was counted manually under a microscope (Olympus, Tokyo, Japan).

2.5. Histological analysis

Lung tissues were inflated with 4% paraformaldehyde and embedded in paraffin. Then these paraffin blocks were sectioned into 4 μm thick slices. And the paraffin-embedded lung sections were dewaxed and rehydrated using dimethylbenzene as well as ethanol solution. To access the degree of lung pathological injury, the lung sections were subjected to hematoxylin and eosin (H&E). Then the degree of neutrophil infiltration and pulmonary edema were blindly investigated and scored by two independent physicians using the Image-Pro Plus 6.0 software. The scoring principle for lung injury was based on our previous description [26].

For immunofluorescence staining, the frozen lung sections were labeled with primary antibodies ZO-1 (#ab221547), Occludin (#ab216327) and Claudin3 (#ab214487), then incubated with Alexa Fluor 488- or 568-conjugated secondary antibody (#ab175471 & #ab150077) (Cambridge, United Kingdom). Immunofluorescence images were captured with a microscope (Olympus, Tokyo, Japan) and the positive fluorescence area was semi-quantified with the Image-Pro Plus 6.0 software.

2.6. Western blot and coimmunoprecipitation

Lung tissues and AT2 samples were lysed via radioimmunoprecipitation lysis buffer (#G2002) (Sevicebio, Wuhan, China). Next, the protein concentration of each sample was detected via a commercial BCA kit (#ST2222, Beyotime Biotechnology, Shanghai, China). Subsequently, 30 μg

of protein from each sample was loaded and separated on an sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gel (10%), which was then transferred onto the polyvinylidene fluoride (PVDF) membrane (Merck Millipore, Germany). Afterwards, the membranes were blocked with skimmed milk (5%), followed by the incubation with the primary antibody at 4 °C. The next day, the membranes were incubated with a horseradish peroxidase-conjugated goat antirabbit secondary antibody at room temperature for 1 h. The chemiluminescence of each protein band were detected via ECL Western blot detection kits, which were then visualized by a Chemidoc Touching Imaging System (BIO-RAD, USA) and semi-quantitated with the Image Lab software (BioRad Laboratories, Inc.). The information of primary antibodies used in this study were listed in Supplementary Table S1.

As for coimmunoprecipitation, immunoprecipitation lysate buffer octylD-glucoside (2%) was used to lyse AT2, followed by centrifugation at 4 °C at 13,000 rpm for 10 min before the supernatant was taken. Subsequently, the supernatant protein and HDAC3 antibody were spun at 4 °C overnight. Then the mixture was spun with Protein A/G Plus MaqPoly beads for 4 h at 4 °C. After that, the magnetic bead–antibody–antigen complex was washed by Tris-buffe and eluted with Laemmli sample buffer. Then the obtained solution was mixed evenly and heated at 100 °C for 10 min, followed by the western blot analysis. The interaction was expressed as the ratio of the expression intensity of HDAC3 to the expression intensity of FOXO1.

2.7. Relative quantitative RT-PCR

Relative quantitative RT-PCR was performed according to our previous studies [6,28]. Total RNA was extracted using TRIzol reagent (#5596-026, Invitrogen) and subsequently reversely transcribed to cDNA with the Maxima First Strand cDNA Synthesis Kit (#04897030001, Roche, Basel, Switzerland).

Next, qPCR was perfomed using iTaq Universal SYBR Green Supermix in Step-one Plus Real-time PCR System. Gapdh was used as the endogenous reference gene and the primer sequences used in the present study are listed in Supplementary Table S2.

2.8. Evan's blue staining

The permeability of epithelial barrier in lung tissues was indicated by Evan's Blue staining. In detail, mice were injected with 2% Evans blue dye (2 mL/kg) via tail vein at the end of LPS stimulation. Five minutes later, the abdominal aorta of mice was cut and bled. After that, one catheter was quickly inserted into the right ventricle to the pulmonary artery. Next, 10 mL cold normal saline was injected into the catheter continuously to wash the blood and the entire mouse lungs were isolated. Then lung tissues were immediately weighed, homogenized in 1 mL of trichloroacetic acid solution (50%), followed by the centrifugation at 13,000 rpm for 0.5 h. Then, the supernatant was diluted with ethanol (1:3), the absorbance of which was detected at 630 nm via a microplate reader.

2.9. TUNEL staining

Cell apoptosis was detected using terminal deoxyribonucleotidyl transferse (TdT)-mediated biotin-16-dUTP nick-end labelling (TUNEL) (#G1501) and 4′,6-diamidino-2-phenylindole (DAPI) (#G1012) staining solution (Servicebio, Wuhan). After fixation with paraformaldehyde (4%) at room temperature for 30 min, the lung sections were incubated with Triton X-100 (0.5%) for 10 min. Subsequently, the TUNEL working mixture was added onto the section, followed by the incubation for 1 h at 37 °C. And cell nucleus was stained with DAPI as blue, which was then observed under a fluorescence microscope.

2.10. The determination of oxidative stress

Total superoxide dismutase (SOD) activity was determined using a commercial assay kit (#A001-3-2, Nanjing) obtained from Nanjing Jiancheng Bioengineering Research Institute Co., LTD (Nanjing, China) by WST-1 method. Thiobarbituric acid reactive substance (TBARS) activity (#ab118970) was detected by colorimetric method according to previous study [26]. The nicotinamide adenine dinucleotide phosphate (NADPH) oxidase activity was detected using the

commercially available kit (#A127-1-1) (Nanjing, China) by colorimetric method on the basis of instruction. In addition, the levels of reactive oxygen species (ROS) in AT2 and lung tissues were reflected using a DCFH-DA probe and a ROS probe, respectively via referring our previous report [26,29].

2.11. Transmission electron microscopy

Once the lung tissues were excised from mice, they were cut into small cubes (1x1x1 cm) and fixed in phosphate buffer solution (pH = 7.4) which contained glutaraldehyde (2.5%) for 4 h. Next, on the basis of previous description, the fresh lung tissues were permeated, dehydrated, and sectioned at 70 nm [30]. Subsequently, the ultrathin sections were stained with uranyl acetate (3%) as well as lead citrate, followed by the observation under a transmission electron microscope at 80 kV. And only the AT2 containing cytoplasm and nucleus with mitochondria was photographed. Mitochondrial number and size were further analyzed by Image-Pro Plus 6.0 software. Then the number of mitochondria together with their size were determined according to our previous description [30].

2.12. Real-time cell metabolism assay

The real-time status of cell metabolism including oxygen consumption rate (OCR) as well as extracellular acidification rate (ECAR) was recorded via XF24 extracellular Flux Analyzer (Seahorse Bioscience, USA). Oligomycin was added to block ATP synthase/production of AT2, the residual respiration is triggered by proton leak. Mitochondrial un-coupler (carbonyl cyanide 4-[triflfluoromethoxy] phenylhydrazone) (FCCP) was added to detect maximum respiratory function (maximal OCR) of AT2. And the electron transport chain inhibitors rotenone together with antimycin A were added to investigate non-mitochondrial oxygen consumption. The detailed operation procedure was on the basis of previous studies [20,31].

2.13. ROCK1 activity assay

Samples were dissolved using RIPA lysate and slightly sonicated, and the supernatant was obtained after centrifugation at 4 °C for later use. ROCK1 activity detection kit (ab211175, Abcam, England) was used and operated according to the instructions. The absorbance values of samples in each group after treatment were detected by microplate reader and the data were recorded.

2.14. Chromatin immunoprecipitation (CHIP) and dual-luciferase reporter gene assay

AT2 were transduced with Ad-Hdac3 or Ad-NC, which were then fixed with formaldehyde (4%) for 10 min to form DNA-protein crosslinks. 2 mg/mL of glycine was utilized to neutralize superfluous paraformaldehyde. Then AT2 were disrupted using an ultrasonic disruptor followed by centrifugation at 12000g for 3 min. Anti-HDAC3 antibody and negative control normal mouse IgG antibody was added into the supernatant. After that, Protein Agarose/Sepharose was used to precipitate the DNA-protein complexes by centrifuging at 12000g, followed by the extraction and purification of the DNA fragments. At last, the binding of HDAC3 to ROCK1 promoter was detected via PCR. In addition, in the context of LPS, the binding of HDAC3 to ROCK1 promoter in AT2 was also investigated by the same method.

As dual-luciferase reporter gene assay, the promoter of ROCK1 was amplified via PCR, followed by identification and separation by agarose gel (3%) electrophoresis. Then the luciferase reporter plasmid ROCK1-LUC was generated in E. coli DH5 competent cells. And AT2 were transfected with the ROCK1 plasmid for 48 h in combination with Ad-HDAC3 or LPS stimulation. After that, the luciferase activity was measured and quantified with the Dual Luciferase Reporter Assay Kit based on the instruction.

2.15. Survival analysis

The extra mice in HDAC3-C and HDAC3-CKO group (n = 10 per group) that had free access to water and food were used to investigate the effect of HDAC3 deficiency on 7-day percent survival. The death number was recorded every day and the percent survival was calculated within 7 days after LPS instillation (10 mg/kg).

2.16. Data analysis

All data in this study are presented as mean ± standard deviation (SD, n≥3). One-way or two-way analysis of variance was applied for statistical comparisons among 3 or more groups, followed by Tukey's post hoc test while student's t-test was applied for statistical comparisons between 2 groups.

Statistical analysis was carried out using the software SPSS 23.3. P＜0.05 represents a statistical difference.

3. Results

3.1. HDAC3 is upregulated in AT2 during LPS-induced ALI

To investigate whether HDAC3 was associated with sepsis and LPS-induced ALI, we assessed the protein and mRNA expression of HDAC3 in lung tissues from LPS (10 mg/kg)-treated mice at different time points. We found that LPS stimulation could significantly upregulate the protein and mRNA levels of HDAC3 in murine lung tissues. And the expression of HDAC3 peaked in murine lung tissues at the 12th hour after LPS stimulation (Fig. 1A and B). Subsequently, we investigated the alterations of HDAC3 in vitro ALI model with primary murine AT2 induced by LPS (100 μg/mL). In line with the in vivo data, we found that LPS stimulation could promoted the mRNA and protein expression of HDAC3 in AT2. And the protein expression of HDAC3 peaked at the 8th hour after LPS stimulation while the mRNA expression peaked at the 6th hour after LPS stimulation (Fig. 1C and D). Hence, the intervention time of LPS was set as 12 h in in vivo model and 6 h in in vitro model in subsequent experiments. Furthermore, immunofluorescent staining also showed that LPS stimulation for 6 h could significantly upregulate the protein level of HDAC3, meanwhile LPS stimulation also promoted the expression of HDAC3 in both cytoplasm and cell nucleus in AT2 (Fig. 1E). These results indicated that the aberrant expression of HDAC3 in AT2 may participate in the development of LPS-induced ALI.

Fig. 1.

HDAC3 is upregulated in AT2 during LPS-induced ALI. (A). Western blot images of HDAC3 protein at different time points in murine lung tissues undergoing LPS (10 mg/kg)-induced ALI, and the statistical results (n = 6). (B). Relative mRNA level of Hdac3 in murine lung tissues (n = 6). (C). Western blot images of HDAC3 protein at different time points in primary AT2 under LPS (100 μg/mL) insult (n = 6). (D). Relative mRNA level of Hdac3 in primary AT2 (n = 6). (E). Representative images of immunofluorescence staining of HDAC3 (Green) along with cell nucleus (Blue) in primary AT2 treated with vehicle or LPS (100 μg/mL) for 6 h (n = 3). Differences were accessed by One-way ANOVA followed by a post hoc Tukey test. Values represent the mean ± standard deviation (SD). *P < 0.05 versus the indicated group. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

3.2. HDAC3 deficiency in AT2 relieved LPS-induced apoptosis, oxidative stress, and inflammation during LPS-induced ALI

To examine the effects of HDAC3 deficiency in AT2 in LPS-induced ALI, HDAC3^flox/flox mice were generated and crossed with Sftpc-CreERT2⁺ mice to generate the tamoxifen-inducible, AT2-specific Hdac3 knockout mice (HDAC3 CKO) (Fig. 2A). And the protein expression of HDAC3 in lung tissue and primary AT2 from HDAC3 CKO mice were identified via western blot (Fig. 2B). H&E staining and lung injury score showed that mice in ALI/HDAC3-C group showed obvious lung pathological injury and had higher lung injury score compared with Control/HDAC3-C group, evidenced by thickened alveolar wall, inflammatory cells infiltration, and edema (Fig. 2D and E). However, after HDAC3 was knocked out in AT2, lung pathological injury was significantly alleviated (Fig. 2D and E). HDAC3 deficiency in AT2 also decreased the ratio of lung wet weight to dry weight in mice with ALI (Fig. 2F). LPS stimulation significantly increased the protein ratio of Bax to Bcl-2 as well as the percentage of TUNEL positive cells in lung tissues from mice, which were attenuated by HDAC3 deficiency in AT2 (Fig. 2G and H). In addition, the markers of oxidative stress in lung tissues from mice were also detected. The results showed that HDAC3 deficiency in AT2 significantly enhanced SOD activity, and decreased TBARS activity as well as NADPH oxidase activity in lung tissues from mice with ALI (Fig. 2I-K). Also, we found that the ROS level in ALI/HDAC3-CKO group was significantly lower than that in ALI/HDAC3-C group (Fig. 2L). Furthermore, the mRNA levels of proinflammatory genes involving Il-6, Tnf-α, Il-1β, and Mcp-1 were detected. The results showed that HDAC3 deficiency significantly inhibited inflammatory response in lung tissues from LPS-treated mice (Fig. 2M − P). It is worthy note that HDAC3 deficiency at baseline showed effects on apoptosis, inflammation, and oxidative stress. Kaplan-Meier curve also demonstrated that HDAC3 deficiency could improve 7-day survival rate in mice treated with lethal dose of LPS (40 mg/kg) as mentioned before [32] (Fig. 2Q). Collectively, these data showed that HDAC3 deficiency in AT2 gave rise to the remission of apoptosis, oxidative stress, and inflammation during LPS-induced ALI.

Fig. 2.

HDAC3 deficiency in AT2 relieved LPS-induced apoptosis, oxidative stress, and inflammation during LPS-induced ALI. (A). HDAC3^flox/flox mice were generated by inserting two loxP sequences in the same direction into the introns flanked with the exon 4, 5, 6, and 7 of HDAC3 using CRISPR-Cas9 system, producing a nonfunctional HDAC3 protein. Sftpc-CreERT2 transgenic mice was then crossed with HDAC3^flox/flox mice to generate the AT2-specific HDAC3-knockout mice, named as Sftpc-CreERT2⁺-HDAC3^flox/flox (HDAC3 CKO). Subsequently, the HDAC3 CKO mice were treated with tamoxifen (10 mg/kg) dissolved in corn oil intraperitoneally for five consecutive days before LPS stimulation. (B). Western blot images of HDAC3 protein in murine lung tissues from HDAC3-C and HDAC3-CKO mice, and the statistical results (n = 6). (C). Western blot images of HDAC3 protein in AT2 from HDAC3-C and HDAC3-CKO mice, and the statistical results (n = 6). (D-E). Representative images of H&E staining and lung injury score (n = 6). (F). The ratio of lung wet weight to dry weight (n = 6). (G). Western blot images of Bax and Bcl-2 protein in murine lung tissues, and the statistical results (n = 6). (H). Representative images of TUNEL staining in lung tissues and relative quantification (n = 6). (I–K). The markers of oxidative stress in lung tissues including SOD activity, TBARS activity, and NADPH oxidase activity (n = 6). (L). Representative images of ROS fluorescence staining in lung tissues and relative quantification (n = 6). (M-P). Relative mRNA levels of Il-6, Tnf-α, Il-1β, and Mcp-1 in murine lung tissues (n = 6). (Q). Kaplan-Meier curve of 7-day survival in mice treated with lethal dose of LPS (40 mg/kg) (n = 10). Differences in (B) and (C) were assessed by Student's test. Differences in other panels were accessed by two-way ANOVA followed by a post hoc Tukey test. Survival rate was accessed by Mantel-Cox test. Values represent the mean ± standard deviation (SD). *P < 0.05 versus the indicated group, NS represents no significance.

3.3. HDAC3 deficiency in AT2 maintained alveolar epithelial barrier integrity in mice with ALI

Epithelial barrier integrity is closely associated with the pulmonary function as well as the severity of ALI [33,34]. We next observed whether HDAC3 knockout in AT2 affect alveolar epithelial barrier integrity in mice with ALI. Western blot showed that the protein levels of ZO-1, Occludin, Claudin 3, and Claudin 18 were significantly higher in ALI/HDAC3-C group than those in Control/HDAC3-C group, indicating that LPS disrupted alveolar epithelial barrier. As expected, HDAC3 knockout in AT2 significantly promoted the protein expression of ZO-1, Occludin, Claudin 3, and Claudin 18 in lung tissues in mice with ALI (Fig. 3A). Immunofluorescent staining further confirmed the inhibitory effect on HDAC3 on ZO-1, Occludin and Claudin 3 in AT2 (Fig. 3B). Evan's Blue staining also disclosed that HDAC3 knockout in AT2 decreased permeability of epithelial barrier in lung tissues from LPS-treated mice (Fig. 3C). Meanwhile, mild bleeding in BALF was found in HDAC3 CKO mice with ALI (Fig. 3D). Cell counts and total protein level in BALF were also analyzed. HDAC3 knockout in AT2 could significantly decreased the levels of total cells, macrophages, neutrophil, as well as total protein in BALF collected from HDAC3 CKO mice with ALI (Fig. 3E-H), which was suggestive of ameliorative epithelial barrier integrity. Collectively, HDAC3 deficiency in AT2 could preserve alveolar epithelial barrier integrity in mice with ALI.

Fig. 3.

HDAC3 deficiency in AT2 maintained alveolar epithelial barrier integrity in mice with ALI. (A). Western blot images of ZO-1, Occludin, Claudin 3, and Claudin 18 in murine lung tissues, and the statistical results (n = 6). (B). Representative images of immunofluorescent staining of ZO-1 (red), Occludin (green), Claudin 3 (red) in lung tissues. Blue represents cell nucleus stained with DAPI (n = 3). (C). Representative images of lung after Evans blue dye (2 ml/kg) via tail vein (n = 6). (D). Bleeding in BALF from HDAC3 C and HDAC3 CKO mice 4 h after LPS stimulation for 12 h (n = 3). (E-H). Total cells, macrophages, neutrophil, and total protein in BALF were detected (n = 6). Differences were accessed by two-way ANOVA followed by a post hoc Tukey test. Values represent the mean ± standard deviation (SD). *P < 0.05 versus the indicated group, NS represents no significance. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

3.4. HDAC3 deficiency in AT2 preserved MQC in lung tissues from mice with ALI

Our previous study has demonstrated that MQC involving mitochondrial dynamics, mitophagy, as well as mitochondrial redox regulation in AT2 played very important roles in alleviating LPS-induced ALI [26]. And HDAC3 is also implicated with mitochondrial homeostasis by affecting FAO and mitophagy [22,23,35,36]. We detected the protein level of Drp1 in mitochondria and cytoplasm, respectively. The results showed that LPS stimulation could decrease the protein levels of Drp1 and Fis1 in cytoplasm but upregulate the level Drp1 in mitochondria and Opa1 in cytoplasm, indicating that LPS shifted the dynamic course of mitochondria from fusion to fission. By contrast, HDAC3 knockout could remarkedly promoted mitochondria fusion in lung tissues during ALI (Fig. 4A). Meanwhile, we observed mitochondria number and size in AT2 using transmission electron microscopy. The ultrastructure in AT2 showed that HDAC3 deficiency alleviated mitochondria damage, evidenced by mitigatory vacuolization and increased mitochondrial cristae number (Fig. 4B). The quantifying mitochondria number and size, we also found HDAC3 in AT2 may contribute to mitochondria fission during LPS-induced ALI (Fig. 4C and D). What is more, the level of mitophagy in lung tissues from LPS-treated mice was also investigated. As shown in Fig. 4E, LPS stimulation significantly induced mitophagy in lung tissues, evidenced by increased protein levels of LC3II/LC3I, Parkin, and PINK1, and decreased protein level of P62. And HDAC3 deficiency could restore the level of mitophagy in lung tissues. At last, the FAO level in lung tissues were also detected via RT-PCR. We found that LPS stimulation could decrease FAO level in lung tissues. However, HDAC3 deficiency in AT2 obviously enhanced FAO capacity in LPS-treated lung tissues, evidenced by increased mRNA levels of Pgc-1α, Cpt1a, Mcad, Acsm2, and Acat1 (Fig. 4F-J). These data suggested that HDAC3 in AT2 contributed to the disturbance of MQC during LPS-induced ALI.

Fig. 4.

HDAC3 deficiency in AT2 preserved MQC in lung tissues from mice with ALI. (A). Western blot images of Drp1 in mitochondria and cytoplasm, Fis1 and Opa1 in murine lung tissues, and the statistical results (n = 6). (B-D). The morphology of mitochondria by transmission electron microscopy as well as mitochondria number and size in AT2 of lung tissues (n = 6). (E). Western blot images of LC3II/I, P62, Parkin, and PINK1 in murine lung tissues, and the statistical results (n = 6). (F-J). Relative mRNA levels of FAO markers including Pgc-1α, Cpt1a, Mcad, Acsm2, and Acat1 in murine lung tissues (n = 6). Differences were accessed by two-way ANOVA followed by a post hoc Tukey test. Values represent the mean ± standard deviation (SD). *P < 0.05 versus the indicated group, NS represents no significance. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

3.5. HDAC3 inhibition protected against LPS-induced apoptosis and oxidative stress and maintained cell junctions in vitro

To further clarify the role of HDAC3 in LPS-induced ALI, we used small interfering ribonucleic acid to downregulate the protein expression in murine primary AT2 (Fig. 5A). We found that LPS stimulation significantly increased the protein ratio of Bax to Bcl-2 in AT2, which were restored after HDAC3 was knocked out (Fig. 5B). TUNEL staining further proved the anti-apoptosis effect of HDAC3 inhibition in LPS-treated AT2 (Fig. 5C). Cell viability was also detected after HDAC3 was knocked out. The data showed that HDAC3 inhibition also improved cell viability in LPS-treated AT2 (Fig. 5D). Meantime, the markers of oxidative including SOD activity, CAT activity, NADPH oxidase activity, as well as ROS level in AT2 were investigated. Similarly, we found that HDAC3 knockout could increase SOD and catalase activity, and decrease NADPH oxidase activity as well as ROS level, suggesting that HDAC3 could trigger oxidative stress in the context of LPS stimulation (Fig. 5E-H). Also, we found that siHADC3 significantly upregulated the markers of cell junction including ZO-1, Occludin, Claudin-3, and Claudin-18 in LPS-treated AT2 (Fig. 5I). These data further demonstrated the role of HDAC3 in LPS-induced ALI in vitro.

Fig. 5.

HDAC3 inhibition protected against LPS-induced apoptosis and oxidative stress and maintained cell junctions in vitro. (A). Western blot images of HDAC3 protein in AT2 transfected with siHDAC3 from wild type mice (n = 6). (B). Western blot images of Bax and Bcl-2 protein in AT2, and the statistical results (n = 6). (C). Representative images of TUNEL staining in AT2 and relative quantification (n = 6). (D). Cell viability detected via CCK-8 assay kit (n = 6). (E-G). The markers of oxidative stress in AT2 including SOD activity, CAT activity, and NADPH oxidase activity (n = 6). (H). Representative images of ROS fluorescence staining in AT2 and relative quantification (n = 6). (I). Western blot images of ZO-1, Occludin, Claudin 3, and Claudin 18 in AT2, and the statistical results (n = 6). Differences in (A) were assessed by Student's test. Differences in other panels were accessed by two-way ANOVA followed by a post hoc Tukey test. Values represent the mean ± standard deviation (SD). *P < 0.05 versus the indicated group, NS represents no significance.

3.6. HDAC3 inhibition maintained MQC in LPS-induced AT2 in vitro

Western blot showed that LPS stimulation remarkedly increased the protein level of Drp-1 in mitochondria, accompanied by a concomitant increase in Fis-1 protein levels and a decrease in Opa1 protein levels across the entire cell population. As expected, HDAC3 inhibition by siRNA could shift the dynamic course of mitochondria from fission to fusion (Fig. 6A). In line with protein markers of mitochondria dynamics, we also found HDAC3 inhibition reduced mitochondrial number and increased mean mitochondrial size in LPS-induced AT2 using transmission electron microscopy (Fig. 6B-D). In addition, we also found that HDAC3 knockdown could suppress mitophagy in LPS-induced AT2, evidenced by decreased LC3II/I, Parkin, and Pink1, as well as increase P62 (Fig. 6E). We also found that HDAC3 knockdown downregulates the mRNA levels of Pgc-1α, Cpt1a, Mcad, and Acsm2, promoting FAO in LPS-treated AT2 (Fig. 6F-I). LPS stimulation impaired OCR compared in AT2 with control group. Consistent with the upregulation of FAO markers, HDAC3 siRNA significantly improved OCR in LPS- treated AT2 (Fig. 6J). Taken together, these data further supported that HDAC3 inhibition maintained MQC in LPS-induced AT2 in vitro.

Fig. 6.

HDAC3 inhibition maintained MQC in LPS-induced AT2 in vitro. (A). Western blot images of Drp1 in mitochondria and cytoplasm, Fis1 and Opa1 in AT2, and the statistical results (n = 6). (B-D). The morphology of mitochondria by transmission electron microscopy as well as mitochondria number and size in AT2 (n = 6). (E). Western blot images of LC3II/I, P62, Parkin, and PINK1 in AT2, and the statistical results (n = 6). (F–I). Relative mRNA levels of FAO markers including Pgc-1α, Cpt1a, Mcad, and Acsm2 in AT2 (n = 6). (J). Primary AT2 were transfected with siHDAC3 and were treated with PBS or LPS. The AT2 were plated in XF-24 microplates overnight. The media were then replaced with FAO assay media supplemented with sequential incubation with oligomycin (3 mg/ml), FCCP (6 mM), and Rot (1 mM) and Ant (0.5 mM). Real-time oxygen consumption rate (OCR) was detected (n = 6). Differences were accessed by two-way ANOVA followed by a post hoc Tukey test. Values represent the mean ± standard deviation (SD). *P < 0.05 versus the indicated group, NS represents no significance. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

3.7. HDAC3 promoted ROCK1 expression and enhanced ROCK1 activity in LPS-treated AT2

Previous studies have unveiled that SIRT1 [37], SIRT3 [26], AMPK [38,39], ROCK1 [[40], [41], [42]], and NRF2 [43] were involved in MQC by regulating mitochondrial dynamics, FAO, and mitophagy. Hence, we next explored whether HDAC3 knockout could affect the above-mentioned proteins in AT2. Western blot showed that LPS stimulation could significantly decrease the protein levels of SIRT1, SIRT3, and NRF2, and increased the protein level of P-AMPK in AT2. However, HDAC3 inhibition showed no effects on the levels of these proteins (Figs. S2A–E). And HDAC3 inhibition also showed no effects on SIRT3 activity and SIRT1 activity in LPS-treated AT2 although LPS stimulation gave rise to the reduction of their activity (Figs. S2F–G). Intriguingly, we found that LPS stimulation could significantly activate RohA/ROCK1 pathway in AT2. In detail, LPS stimulation not only upregulated the protein levels of RohA, ROCK1, and P-ROCK1 in AT2, but also enhanced transcriptional levels of RhoA and ROCK1. And HDAC3 knockout did not affect the mRNA and protein level of RhoA but it could decrease the protein levels of ROCK1 and P-ROCK1, and the mRNA ROCK1. At baseline, HDAC3 knockout also inhibited the mRNA and protein level of ROCK1 in AT2, but showed no effects no the protein level of P-ROCK1 and RhoA, as well as the mRNA level of RhoA (Fig. 7A-C). We also detected ROCK1 activity in AT2 and the result showed that HDAC3 knockout decreased ROCK1 activity only in the context of LPS stimulation (Fig. 7D). These data showed that HDAC3 deficiency could decrease the mRNA and protein expression of ROCK1 regardless of LPS stimulation, which gave rise to the reduction of total HDAC3 protein content in AT2. Once LPS stimulation, only a smaller amount of ROCK1 could be phosphorylated and activated in AT2, which maintained the homeostasis of MQC. To further confirm our hypothesis, we also used Ad-HDAC3 to overexpress HDAC3 protein in AT2 (Fig. S3). Similarly, we found that HDAC3 overexpression in AT2 significantly upregulated the protein and mRNA levels of ROCK1 but showed no effects on RhoA level regardless of LPS stimulation. As expected, HDAC3 overexpression further enhanced the phosphorylation and activity of ROCK1 in LPS-treated AT2 (Fig. 7E-H). Taken together, these data suggested that HDAC3 contributed to the activation of ROCK1 during ALI.

Fig. 7.

HDAC3 promoted ROCK1 expression and enhanced ROCK1 activity in LPS-treated AT2. (A). Western blot images of ROCK1, P-ROCK1, and RhoA in AT2 transfected with siHDAC3, and the statistical results (n = 6). (B–C). Relative mRNA levels of RhoA and Rock1 in AT2 transfected with siHDAC3 (n = 6). (D). ROCK1 activity in AT2 transfected with siHDAC3 (n = 6). (E). Western blot images of ROCK1, P-ROCK1, and RhoA in AT2 transduced with Ad-HDAC3, and the statistical results (n = 6). (F-G). Relative mRNA levels of RhoA and Rock1 in AT2 transduced with Ad-HDAC3 (n = 6). (H). ROCK1 activity in AT2 transduced with Ad-HDAC3 (n = 6). Differences were accessed by two-way ANOVA followed by a post hoc Tukey test. Values represent the mean ± standard deviation (SD). *P < 0.05 versus the indicated group, NS represents no significance.

3.8. HDAC3 contributed to epithelial barrier damage and disturbance of MQC in a ROCK1-dependent manner in LPS-treated AT2

Next, we confirmed the role of ROCK1 in HDAC3-mediated epithelial barrier damage and disturbance of MQC. To begin with, we used siROCK1 to downregulate the expression of ROCK1 in primary AT2 (Fig. 8A). Then we observed the effects of HDAC3 overexpression on apoptosis, oxidative stress, cell junction, as well as MQC in LPS-treated AT2 transfected with siROCK1. We found that in LPS-treated AT2 transfected with siROCK1, HDAC3 overexpression did not affect the Bac/Bcl-2 mRNA level, SOD2 activity, CAT activity, or NADPH oxidase activity, indicating that ROCK1 was essential for HDAC3-mediated epithelial apoptosis and oxidative stress (Fig. 8B-E). Meanwhile, in LPS-treated AT2 transduced with Ad-HDAC3, ROCK1 knockout significantly decreased Bac/Bcl-2 mRNA level, and increased SOD2 activity, CAT activity, as well as NADPH oxidase activity, which further suggested the critical role of ROCK1 in HDAC3-meidated epithelial barrier damage during ALI (Fig. 8B-E). In addition, we found that HDAC3-mediated downregulation of Zo-1, Occludin, and Claudin 3 in LPS-treated AT2 was also dependent on ROCK1 (Fig. 8F-H). In line with the above results, HDAC3 also affected mitochondrial dynamics and FAO in a ROCK1-dependent manner in LPS-treated AT2, evidenced by the mRNA levels of Fis1, Opa1, Pgc-1α, and Cpt1a (Fig. 8I-L). We also observed the effects of HDAC3 knockout on apoptosis, oxidative stress, cell junction, and MQC in LPS-treated AT2 transduced with Ad-ROCK1 (Fig. S4A). Similarly, in LPS-treated AT2 transduced with Ad-ROCK1, ROCK1 overexpression abolished the effects of HDAC3 deficiency on apoptosis, oxidative stress, cell junction and MQC (Fig. S4B-L). In brief, these data demonstrated that HDAC3 contributed to epithelial barrier damage and disturbance of MQC in a ROCK1-dependent manner in LPS-treated AT2.

Fig. 8.

HDAC3 contributed to epithelial barrier damage and disturbance of MQC in a ROCK1-dependent manner in LPS-treated AT2. (A). Western blot images of ROCK1 in primary AT2 transfected with siROCK1 (n = 6). (B). Relative mRNA levels of Bax/Bcl-2 in AT2 transfected with siROCK1 and transduced with Ad-HDAC3 (n = 6). (C-E). The markers of oxidative stress including SOD activity, CAT activity, and NADPH oxidase activity in AT2 transfected with siROCK1 and transduced with Ad-HDAC3 (n = 6). (F-L). Relative mRNA levels of Zo-1, Occludin, Claudin 3, Fis-1, Opa1, Pgc-1α, and Cpt1a in AT2 transfected with siROCK1 and transduced with Ad-HDAC3 (n = 6). Differences were accessed by two-way ANOVA followed by a post hoc Tukey test. Values represent the mean ± standard deviation (SD). *P < 0.05 versus the indicated group, NS represents no significance.

3.9. HDAC3 promoted the mRNA transcription of ROCK1 by deacetylating FOXO1 in LPS-treated AT2

It is well known that histone acetylation is usually associated with chromatin decondensation as well as gene activation [44]. Hence, HDAC3 together with other members in HDACs are supposed to repress gene expression by directly regulating histone deacetylation of certain genes. In view of the fact that the level of HDAC3 showed positive correlation with the expression of ROCK1 in this study, we thus excluded the possibility that HDAC3 could deacetylate the histones of ROCK1 gene promoter. Transcription factor FOXO1 serves as a critical regulator, which plays important roles in MQC by regulating transcription of ROCK1 directly [45]. Meanwhile, HDAC3 could decrease the acetylation of FOXO1 and promoted its nuclear translocation, thus transcriptionally activating downstream genes by binding to their promoters [46,47]. Next, we explored whether FOXO1 was involved in HDAC3-mediated ROCK1 activation. First, we found LPS stimulation significantly increased the protein level of total FOXO1 and decreased the protein level of Ac-FOXO1 in AT2 (Fig. 9A). After that, the protein level of FOXO1 in cell nucleus was also detected. The result showed that LPS remarkedly promoted the accumulation of FOXO1 in cell nucleus (Fig. 9B). We also observed that HDAC3 overexpression in AT2 promoted the nuclear translocation of FOXO1 by decreasing its acetylation (Fig. 9C and D). Co-immunoprecipitation also unveiled that LPS stimulation promoted the interaction between HDAC3 and FOXO1 in AT2 (Fig. 9E). In addition, we found that HDAC3 knockout significantly decreased the protein level of FOXO1 in cell nucleus (Fig. 9F). And in the context of FOXO1 knockout, LPS failed to upregulate the mRNA level of Rock1 (Fig. 9G and H). Finally, we found that both HDAC3 overexpression and LPS stimulation could promote FOXO1 to bind to ROCK1 promoter and enhance ROCK1 promoter activity in AT2 (Fig. 9I-L). Briefly, these data further demonstrated that HDAC3 promoted the mRNA transcription of ROCK1 by deacetylating FOXO1 in LPS-treated AT2.

Fig. 9.

HDAC3 promoted the mRNA transcription of ROCK1 by deacetylating FOXO1 in LPS-treated AT2. (A). Western blot images of FOXO1 and Ac-FOXO1 in LPS-treated AT2, and the statistical results (n = 6). (B). Western blot images of FOXO1 in cell nucleus in LPS-treated AT2, and the statistical results (n = 6). (C). Representative images of immunofluorescent staining of FOXO1 (green) in AT2 transduced with Ad-HDAC3. Blue represents cell nucleus stained with DAPI (n = 3). (D). Western blot images of FOXO1 and Ac-FOXO1 in AT2 transduced with Ad-HDAC3, and the statistical results (n = 6). (E). Representative images of co-immunoprecipitation of HDAC3 and FOXO1 in LPS-treated AT2 (n = 6). (F). Western blot images of FOXO1 in cell nucleus in LPS-treated AT2 transfected with siHDAC3, and the statistical results (n = 6). (G). Western blot images of FOXO1 AT2 transfected with siFOXO1 (n = 6). (H). Relative mRNA level of Rock1 in LPS-treated AT2 transfected with siFOXO1 (n = 6). (I-J). Binding of FOXO1 to ROCK1 verified via CHIP and luciferase reporter gene assay in AT2 transduced with Ad-HDAC3 (n = 3). (K-L). Binding of FOXO1 to ROCK1 verified via CHIP and luciferase reporter gene assay in LPS-treated AT2 (n = 3). Differences in (A–E), (G), and (I–L) were assessed by Student's test. Differences in other panels were accessed by two-way ANOVA followed by a post hoc Tukey test. Values represent the mean ± standard deviation (SD). *P < 0.05 versus the indicated group, NS represents no significance. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

3.10. Pharmacological inhibition of HDAC3 by RGFP966 prevented epithelial barrier damage and preserved MQC in ALI

To further identify the role of HDAC3 inhibition in LPA-induced epithelial barrier damage, we

treated primary AT2 with HDAC3 selective inhibitor RGFP966 (Fig. 10A). Primary AT2 were isolated from male C57BL/6 mice. In LPS-treated AT2, RGFP966 treatment significantly decreased oxidative stress and apoptosis, evidenced by decreased ROS level and lower percentage of TUNEL positive cells in AT2 (Fig. 10B and C). Also, we found that RGFP966 could significantly upregulate the protein levels of Occludin and ZO-1 in LPS-treated AT2, hinting that HDAC3 inhibition by RGFP966 also alleviate LPS-induced epithelial barrier damage (Fig. 10D). At last, we detected the status of MQC in AT2. Western blot showed that RGFP966 downregulated the protein levels of Parkin and PINK1 in LPS-treated AT2, suggestive of a lower level of mitophagy (Fig. 10E). In addition, RGFP966 also downregulated the mRNA level of Fis-1 and upregulated the mRNA levels of Opa1, Pgc-1α, as well as Cpt1a in LPS-treated AT2 (Fig. 10F-I). Similarly, pretreatment with RGFP966 and intraperitoneal injection of LPS to induce ALI showed that RGFP966 significantly ameliorated the pathological damage and lung edema in mice（Fig. S5 A-C）. In addition, the level of apoptosis in lung tissue was significantly inhibited, and the damage of epithelial barrier was effectively improved (Fig S5 D-G). FOXO1 and ROCK1 expression levels were consistent with what we observed in vitro (Fig S5 H-K). At the same time, mitochondrial function was significantly restored (Fig S5 L-O). These data suggest that targeting HDAC3 by RGFP966 in AT2 following LPS stimulation alleviated ALI.

Fig. 10.

Pharmacological inhibition of HDAC3 by RGFP966 prevented epithelial barrier damage and preserved MQC in LPS-treated AT2. (A). The molecular structure of RGFP966. (B). Representative images of ROS staining. Blue represents cell nucleus stained with DAPI (n = 6). (C). Representative images of TUNEL staining. Blue represents cell nucleus stained with DAPI (n = 6). (D). Western blot images of ZO-1 and Occludin in AT2, and the statistical results (n = 6). (E). Western blot images of Parkin and PINK1 in AT2, and the statistical results (n = 6). (F–I). Relative mRNA levels of Fis-1, Opa1, Pgc-1α, and Cpt1a in AT2 (n = 6). Differences were accessed by two-way ANOVA followed by a post hoc Tukey test. Values represent the mean ± standard deviation (SD). *P < 0.05 versus the indicated group, NS represents no significance. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

4. Discussion

Sepsis is one of the critical diseases threatening global health, which is mainly featured by high morbidity and mortality. Sepsis-elicited hyperactivated inflammatory response could cause multi-organ functional disturbance and even failure [48]. Lung is one of the organs vulnerable to infection and sepsis [49]. During sepsis, alveolar–capillary membrane barrier, and edema, immune/inflammatory reaction could be observed in lung tissues [50]. Previous data suggested that ALI was partially mediated by dysregulation of key tight junction proteins in alveolar epithelial cells [51,52]. Herein, maintenance of an intact and functional alveolar epithelial barrier is of great significance to combat sepsis-induced ALI. However, there was no effective treatment for ALI apart from supportive care till now because of its complicated mechanisms. Based on these research evidence, our study regarded HDAC3 as the point of penetration to deeply investigate the association between HDAC3 in AT2 and ALI following sepsis. It was clear that HDAC3 served as one of the undesirable genes contributing to the development of LPS-induced ALI. In our study, we revealed that HDAC3 promoted transcription and activation of ROCK1 by deacetylating FOXO1 and facilitated disturbance of MQC in AT2, ultimately leading to epithelial barrier damage and ALI. The present study systematically explained the role of HDAC3 in ALI during sepsis and uncovered its possible mechanism, providing a promising strategy for treating patients with sepsis and ALI in clinical practice (Fig. 11).

Fig. 11.

HDAC3 inspired the disturbance of mitochondrial quality control (MQC) mediated by ROCK1 in AT2 by deacetylating FOXO1 and activating the transcription of ROCK1, thus triggering epithelial barrier damage during ALI. HDAC3 suppression by chemicals with RGFP966 may have therapeutic potential for alleviating sepsis-induced ALI.

Histone acetyltransferases and deacetylases synergistically determine chromatin remodeling and gene expression. One the one hand, HDACs could give rise to a net decrease of histone acetylation, chromatin condensation, as well as transcriptional repression. On the other hand, HDACs could also regulate the deacetylation of non-histone protein involving certain mitochondrial and cytoplasmic proteins via post translation by HDACs [53]. These acetylated proteins participate in various physiological and pathological cellular processes including gene transcription, signal transduction, cell division, cell proliferation, metabolism, DNA damage repair, inflammation, autophagy and cell death [[54], [55], [56]]. Therefore, it determines the breadth and diversity of HDACs. Recently, HDAC3 has attracted particular attention due to its ability to regulate inflammation and oxidative stress in immune cells and tumor cells. For instance, during cerebral ischemia reperfusion injury, HDAC3 aggravated neuroinflammation by deacetylating p65 at K122 in the cytosol and interacting with p65 in the nucleus, thus activating cGAS-STING pathway [57]. In chronic constriction injury, HDAC3 in hippocampus triggered memory impairment by deacetylating H3 and H4 [58]. However, the function of HDAC3 in parenchyma cells has not been largely studied yet. To the best of our knowledge, our study firstly underlines that the upregulated HDAC3 in AT2 could determine susceptibility to LPS-induced sepsis and ALI. To be more specific, LPS stimulation could significantly upregulate HDAC3 in both cytoplasm and cell nucleus in AT2. Both genetic and pharmacological inhibition of HDAC3 could impede epithelial barrier damage and lung pathological injury during sepsis, indicating that the effects of HDAC3 were mediated by its enzymatic activity.

Mitochondria serve as intracellular organelles possessing important roles in producing ATP, regulating a great many catabolic and anabolic processes, as well as maintaining cellular redox homeostasis in eukaryotic cells [59]. In addition, mitochondria are also core hubs coordinating signaling cascades associated with cell survival and death. As a major intracellular source and a primary target of ROS, mitochondria are extremely vulnerable to damage in response to certain stressful stimulation [60]. Multiple mechanisms contributing to MQC have been evolved to combat stress and preserve its integrity and function, involving protein quality control, mitochondrial fusion and fission, mitochondrial biogenesis, mitophagy, and mitochondrial redox regulation. These mechanisms work coordinately to maintain mitochondrial integrity and function, loss of which may cause cell death, tissue injury and even organ failure [13,61]. Previous study has reported that mitochondrial damage and dysfunction could promote epithelial barrier damage through calcium dysregulation, energy failure, apoptosis, and loss of heme homeostasis [62]. Hence, preserving MQC in alveolar epithelial cells can promote cell survival to support alveolar function during sepsis-induced ALI [63]. Our previous study also showed that melatonin could prevent LPS-induced ALI by maintaining MQC including mitochondrial dynamics, FAO, as well as mitophagy in AT2 [26]. In fact, HDAC3 was closely associate with energy metabolism and mitochondrial homeostasis. In brown adipose tissue lacking HDAC3, mitochondrial oxidative phosphorylation genes that gave rise to diminished mitochondrial respiration were significantly downregulated, indicating that HDAC3 was essential for thermogenesis in brown adipose tissue [64]. In macrophages, HDAC3 triggered acute and chronic inflammation by promoting NLRP3-dependent caspase-1 activation by deacetylating and inactivating an FAO enzyme in non-histone-dependent manner [22]. In brown adipocytes and myotubes treated with HDAC3 inhibitor, oxygen consumption, mitochondrial biogenesis, as well as the expression of PGC-1α were significantly enhanced [24]. In line with these studies, our data also showed that HDAC3 deletion or inhibition could improve mitochondrial function in LPS-treated AT2. To be more specific, HDAC3 partially contributed to the imbalance of mitochondrial fission/fusion, activated mitophagy, as well as decreased FAO during sepsis-induced ALI.

Rho-associated, coiled-coil containing protein kinase 1 (ROCK1) is one of the effectors in the Rho family of small GTPases, belonging to a family of serine/threonine kinases [65]. As a downstream effector of the RhoA small GTPase, ROCK1 plays an important role in mediating the effects of RhoA on energy metabolism, mitochondrial homeostasis, motility, regulated cell death, proliferation, as well as viability. Once activated by RhoA, ROCK1 could promote mitochondrial fission via the activation of Drp1 [40]. Also, under high-glucose conditions, activated ROCK1 also triggered mitophagy in endothelial cells [41]. In A549 cells, caveolin‐1 could induce ROCK1-meidated mitophagy in a Parkin-dependent manner. However, the mechanism by which ROCK1 regulate the expression of Parkin was still not clear [66]. Apart from mitophagy and mitochondrial dynamics, ROCK1 also participated in the impairment of FAO. To our knowledge, disorder of energy metabolism in mitochondria could result in redox imbalance directly. ROCK1 deficiency protected against abnormal fatty acid metabolism as well as mitochondrial fragmentation in mice with diabetic kidney disease [42]. Huachun Cui et al. also reported that alveolar epithelial cells conducted robust FAO in physiologic condition. However, during LPS-induced ALI, FAO was strikingly impaired in murine lung tissues and re-activation of FAO significantly alleviated lung pathological injury [20]. In the present study, we found that HDAC3 deficiency could maintain MQC in AT2 by decreasing the activation ROCK1 but showed no effects on the expression of RhoA. We also found that the suppressive effects from HDAC3 deficiency was associated with the reduction of total content of ROCK1. In addition, we excluded other possible proteins or enzymes including SIRT1, SIRT3, AMPK, as well as NRF2 which possess the potential to modulate MQC.

Given that the direct effects of HDAC3 were to repress gene expression by directly regulating histone deacetylation of certain genes, we thus excluded the possibility that HDAC3 could deacetylate the histones of ROCK1 gene promoter. FOXO1 serves as a transcription factor which has been reported to activate the transcription of ROCK1, and the acetylation level of FOXO1 was negatively correlated with transcriptional activity [45,67]. In the presence of LPS stimulation, the deacetylation level and nuclear translocation of FOXO1 was obviously enhanced, resulting in its increased transcriptional activity [46]. In line with others [68], we also found that LPS stimulated nuclear translocation of FOXO1 during ALI. Furthermore, our study also demonstrated that LPS could promote deacetylation and nuclear translocation of FOXO1 in an HDAC3-dependent manner, which subsequently bound to the promoter of ROCK1 and activated its transcription, indicating that HDAC3/FOXO1/ROCK1 axis was essential for LPS-indued disturbance of MQC in AT2 during ALI.

In conclusion, we explored the effects of HDAC3 on ALI during sepsis and uncovered that HDAC3 impaired MQC in AT2 under LPS stimulation. Moreover, HDAC3 contributed to the disturbance of MQC in AT2 through a ROCK1-dependent manner by deacetylating FOXO1 and promoting its nuclear translocation. The current study filled the gaps and unveiled aspects of HDAC3-mediated epithelial barrier damage during sepsis. Meanwhile, selective suppression of HDAC3 by chemicals such as RGFP966 displays therapeutic potential for treating LPS-induced lung epithelial damage. Our findings disclose that targeting HDAC3 might be a promising strategy for sepsis-induced ALI.

Data availability

All data that support the findings in this study are available from the corresponding author upon reasonable request.

Authors’ contribution

Li Ning, Xiong Rui, Liu Bohao, Wang Bo, Geng Qing contributed to conception, designed experiments and were responsible for the whole work; Liu Bohao, Xiong Rui, and Li Guorui performed experiments; Li Ning and Liu Bohao analyzed experimental results and wrote the manuscript.

Sources of funding

This work was supported by grants from the National Natural Science Foundation of China (No. 8210082163, 81800343), the Fundamental Research Funds for the Central Universities (No. 2042021kf0081) and Science Fund for Creative Research Groups of the Natural Science Foundation of Hubei Province (No. 2020CFA027).

Declarations

Ethics approval Animal experiments were approved by the Animal Ethics Committee of Renmin Hospital, Wuhan University (Approval Number: WDRM20210305) and strictly performed according to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health.

Consent to participate

Not applicable.

Consent for publication

Not applicable.

Human ethics

Not applicable.

Declaration of competing interest

The authors declared that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Appendix A. Supplementary data

The following is/are the supplementary data to this article: