Activation of aldehyde dehydrogenase 2 protects ethanol‐induced osteonecrosis of the femoral head in rat model

Abstract Objectives Osteonecrosis of the femoral head (ONFH) is a devastating disease characterized by destructive bone structures, enlarged adipocyte accumulation and impaired vascularization. The aldehyde dehydrogenase 2 (ALDH 2) is the limiting enzyme for ethanol metabolism with many physiological functions. The aim was investigated the potential protective role of activated ALDH 2 by Alda‐1 for ethanol‐induced ONFH. Materials and Methods The ethanol‐induced ONFH in rat was performed to explore the protective of Alda‐1 by various experimental methods. Subsequently, the effect of Alda‐1 and ethanol on the osteogenic and adipogenic differentiation was investigated via multiple cellular and molecular methods. Finally, the effect of Alda‐1 and ethanol on the neo‐vascularization was detected in Human umbilical vein endothelial cells (HUVECs) and ONFH model. Results Firstly, radiographical and pathological measurements indicated that alda‐1 protected ethanol‐induced ONFH. Moreover, ethanol significantly inhibited the proliferation and osteogenic differentiation of BMSCs, whereas Alda‐1 could distinctly rescue it by PI3K/AKT signalling. Secondly, ethanol remarkably promoted the lipid vacuoles formation of BMSCs, while Alda‐1 significantly retarded it on BMSCs by AMPK signalling pathway. Finally, ethanol significantly inhibited proliferation and growth factor level resulting in reduced angiogenesis, whereas Alda‐1 could rescue the effect of ethanol. Additionally, Alda‐1 significantly reduced the occurrence of ONFH and promoted vessel number and distribution in alcoholic ONFH. Conclusions Alda‐1 activation of ALDH 2 was highly demonstrated to protect ethanol‐induced ONFH by triggering new bone formation, reducing adipogenesis and stimulating vascularization.


| INTRODUCTION
Osteonecrosis of the femoral head (ONFH) is the death of bone which leads to the collapse of the bone structure of head, joint pain, destructive bone and articular cartilage, and lose of normal functions. 1 About three quarters of ONFH patients are in the third to sixth decades of their life span. 2 Although a mass of studies has been extensively explored the aetiology, pathogenesis, diagnosis and therapy of ONFH, no unambiguous consensus of its exact physiological mechanism and origin is been made. 3,4 Epidemiologic studies show that alcohol consumption is the most common factor for adult patients, but the clear mechanism of how alcohol leading to the blood supply interruption and death of bone cells of femoral head is unclear until now. 5,6 Chronic and excessive alcohol consumption has become a major health event with rising mortality and morbidity in the whole world, which lead to a vast of abnormal biochemical, physiological and clinical matters that come from the toxic effect of alcohol on human bodies, like liver, gonads, heart, brain, marrow and bone. 7,8 Alcohol consumption is a vital factor for more than 60 diseases and injures, and leads to about 2.5 million deaths per year (World Health Organization 2011). The detoxifying type of alcohol in human mainly occurs in liver by two enzymatic steps. Firstly, alcohol is quickly converted to acetaldehyde by alcohol dehydrogenase (ADH). Secondly, acetaldehyde is limited metabolized to acetate by the mitochondrial aldehyde dehydrogenase 2 (ALDH 2). 9,10 Notably, the first step of alcohol metabolism is a reversible reaction, but the second step of alcohol metabolism is a limiting step. Moreover, ALDH 2 is possible the only enzyme for acetaldehyde metabolism. 9 Therefore, ALDH 2 exerts a vital protective, preventive and therapeutic enzymatic function against alcohol toxicity on human. Such as, chronic alcohol-induced insulin insensitivity and cardiac dysfunction was alleviated by ALDH 2. 11 ALDH 2 ameliorates ethanol-induced atrial fibrillation by detoxifying 4-HNE. 12 Active ALDH 2 prevents ethanol-induced cell death and hepatic steatosis in mice. 13 Alda-1 is a small molecule, a chemical chaperone and agonist of ALDH 2, which specifically binds to ALDH 2 and is capable of elevating the catalytic ALDH 2 enzyme activity more than tenfolds. 14 Accumulative research has indicated that Alda-1 exerts a lot of protective effect by activating ALDH 2. For example, Alda-1 accelerates cutaneous wound healing by Akt/GSK-3β/β-Catenin signalling pathway via activating ALDH 2. 15 Alda-1 activation of ALDH 2 shows the therapeutic effect in myocardial ischaemia-reperfusion injury. 16 Activation of ALDH 2 by Alda-1 decreases lung oedema, alveolar epithelial tissue permeability and lung inflammatory via accelerating the clearance of aldehydes. 17 Moreover, Alda-1 protects alcohol-induced DNA damage in oesophagus via activation of ALDH 2. 18 Alda-1 pharmacologically activating ALDH 2 prevents and protects ethanol-induced hepatic cell death and steatosis by reactivating transcription factors and upregulating fatty acid oxidation enzymes in mice. 13 Alda-1 pharmacological activation of ALDH 2 effectively reducing the maintenance and acquisition of chronic alcohol intake in rats. 19 Therefore, we hypothesized that Alda-1 has a potential protective effect on alcoholinduced ONFH by activating ALDH 2. In the present study, our results indicated that Alda-1 prevented alcohol-induced decreased bone formation and vascularization, and increased adipogenesis in alcoholinduced rat ONFH model.

| Animal model
With the approval of the Animal Committee at the The First Affiliated Hospital of Xiamen University, 38-week male Sprague-Dawley rats were performed to build the ethanol-induced ONFH experiment. The whole rats were randomly divided into the control group, ethanol (ET) group and ET + Alda-1 (50 mg/kg, i.p.) group. The normal Lieber-DeCarli liquid diet (without ethanol) was adapted for rats 1 week.
Then, the ethanol-containing Lieber-DeCarli liquid diet was fed for the ET group and the ET + Alda-1 group, but the ET + Alda-1 group was intraperitoneally injected with Alda-1 (50 mg/kg/d). At the same time, the control group was fed with the normal Lieber-DeCarli liquid diet. The diet was refreshed every day and all rats had free availablality to food. In order to reflect the dynamic new bone formation, fluorescence staining (20 mg/kg tetracycline, 10 mg/kg calcein-AM and 30 mg/kg alizarin red S) (Aladdin) was intraperitoneally injected at week 0, 2 and 4.

| Histological staining
The decalcified femoral heads were embedded in paraffin, and cut

| Cell culture
The femurs and tibias of 4-week-old SD rats were performed to obtained bone mesenchymal stem cells (BMSCs) according to the previous reports. 20,21 α-minimum essential medium (α-MEM) (Gibco, USA) with 10% foetal bovine serum (FBS, Gbico) and 1% penicillin/ streptomycin (Gbico) were used to incubated with BMSCs in a humidified atmosphere at 37 C with 5% CO 2 . The third to the seventh pas-

| Osteogenic differentiation assay
Alizarin red staining (ARS, Solarbio) and alkaline phosphatase (ALP, Solarbio) staining were performed to investigate the effect of ethanol (100 mM), alda-1 (20 μM) and the PI3K antagonis LY294002 (500 nM) on the osteogenic differentiation of BMSCs. Briefly, 1 Â 10 5 cells in 500 μl medium per well were seeded in 24-well plates. Then osteogenic medium (Cyagen, China) was added to every well for inducing osteogenic differentiation when the confluence of cells was 80%, and the osteogenic medium was refreshed by every 2 days. The ARS and ALP staining were conducted at day 21 and day 7 according to the protocols, respectively, and a Leica microscope was used to capture all images. Then, cells were washed and incubated with the secondary antibodies (1:500, CST), DAPI and phalloidin. Finally, a Leica microscope was used to capture images.

| Adipogenic differentiation assay
Oil red O staining (Solarbio) was performed to investigate the effect of ethanol (100 mM), alda-1 (20 μM) and the AMPK antagonis DMP (10 μM) on the adipogenic differentiation of BMSCs. In briefly, each well of six-well plates was seeded with 2 Â 10 5 cells in 2 ml medium.
Then adipogenic medium (Cyagen, China) was added to every well for inducing adipogenic differentiation when the confluence of cells was 80%, and the adipogenic medium A and B were refreshed by the protocol. Oil red staining was performed at day 21.

| Cell migration assay
The effect of ethanol (100 mM), alda-1 (20 μM) and the AMPK antagonis DMP (10 μM) on the migration of HUVECs was investigated by wound healing assay and transwell assay. As for wound healing assay, each well of sixwell plates was seeded with 2 Â 10 5 cells in 2 ml medium. Then a sterile pipette tip was used to scratch the confluent cell monolayer, and different treatment of medium was added to each well. Finally, images were captured at the specific time-point and Image J was performed to measure the rates of wound healing.

| Quantitative polymerase chain reaction
Quantitative polymerase chain reaction (QPCR) was performed to the expression level of relative mRNA in BMSCs or HUVECs. After cells were incubated in sixwell plates with different treatment of medium for 48 h, total mRNA of cells was obtained following the protocols (EZBioscience, USA). Equal purified RNA was used for the reverse transcriptase reaction, and the SYBR premix ex taq kit (EZBioscience) was used for QPCR. GAPDH was used as the internal gene. The primers were as Table S1.

| Western blot assay
Western blot (WB) assay was performed to investigate the relative target protein level in BMSCs or HUVECs. Briefly, cells were incubated in sixwell plates with different treatment of medium for specific time, cells were lysed by RIPA lysis buffer (1 mM PMSF, Beyotime, China) for the extracting the total. Then the protein concentration was tested, equal protein was electrophoresed, transferred, blocked and incubated with the primary antibodies for GAPDH, COL I, OCN, AKT, p-AKT, AMPK and p-AMPK (1:1000, CST). Next, PVDF membrane was washed and incubated with the secondary antibodies (1:5000). Finally, the membrane was reacted with ECL kit (Thermo Scientific, USA), and the signal was quantified by a scanning densitometry.

| ELISA
The absolute quantification VEGF content in HUVECs was detected by a VEGF ELISA kit (Neobioscience, China). The liquid supernatant of cells was obtained after HUVECs were incubated with different treatment for 48 h. The supernatant was centrifuged, diluted and detected of VEGF following the manufacturer's protocol. The absorbance values were gained at 450 nm used for calculating the VEGF level according to the standard curve.

| Statistical analysis
All data were shown in means ± SEM. The SPSS 20.0 (IBM, USA) was used to analyse the statistically difference by one-way ANOVA or Student's test. *p < .05 was considered to have statistical significance.

| Alda-1 protected ethanol-induced ONFH in rats
Alda-1 activation of ALDH2 exerts a lot of protective effect in alcoholic animal models, so we hypothesis that Alda-1 might also have protective effect in ethanol-induced ONFH. It is consistent with our hypothesis that intraperitoneal injection of Alda-1 alleviated ethanolinduced ONFH in rat model because of the following reasons and evidences. All group rats had the gradual and similar weight gain in the whole experiment ( Figure S1). Firstly, micro-CT scanning of the rat femoral heads indicated that there were remarkably bone defect and signs of osteonecrosis in the ethanol group (8 of 10 rats) ( Figure 1A).
However, only one rat of the ethanol+Alda-1 group exerted mild signs of bone defect and osteonecrosis, and there was no signs of bone defects or osteonecrosis in the Control group ( Figure 1A). Moreover, we performed the selective antagonist of AMPK pathway Dorsomorphin (DMP) to clarify that AMPK pathway was actually involved in ethanol-induced inhibition of adipogenesis. The increased expression of p-AMPK when BMSCs were co-treated with ethanol and Alda-1 was reduced by DMP treatment (Figure 5G,H). What is more, the involvement of AMPK pathway was further demonstrated by the morphological performance in vitro. The inhibitive effect of Alda-1 on ethanol-induced adipogenesis were dramatically abolished by DMP co-treatment ( Figure 5I). Taken together, these results indicated that Alda-1 retarded the ethanol-induced adipogenesis by AMPK pathway.

| DISCUSSION
ONFH is a devastating disease and been called the deathless cancer, because it seems like not fatal, but it will cause long-term pain, walking disorders in patients, or even unable to walk normally, whole life needs people to take care of. There has been no way to completely cure ONFH in clinical, except patients can only achieve the purpose of self-care through total hip arthroplasty (THA). 2,22 Notably, alcoholism is the main factor of ONFH in the worldwide, especially for the adult patients. 23 Therefore, it is urgent to discover and understand the mechanism of alcoholic ONFH, and develop new therapeutic drugs and methods for clinical usage.
After ethanol is absorbed into the body, it is oxidized into acetaldehyde in the mitochondria by the alcohol dehydrogenase; then acetaldehyde is further oxidized into non-toxic acetate by the limited metabolized ALDH 2. 24 What is more, both ethanol and acetaldehyde are toxic for many cells and organs leading to more than 60 diseases. 10,25 Since the expression level of ALDH 2 in the body is basically constant or difficult to change, so the key to the prevention and treatment of alcoholic diseases is to increase the activity of ALDH 2. Therefore, the ALDH 2 activators are to be highly promise drugs for pharmacotherapy for alcoholic diseases, Alda-1 is a highly selective agonist of ALDH and capable of greatly elevating the ALDH 2 activity. 14 A pile of reports has indicated the protective effect of Alda-1 against numerous diseases via activating ALDH 2.
Accumulative previous studies have shown that obviously impaired osteogenesis capacity, the number and the osteogenic potential of osteoblasts of BMSCs in ONFH patients. Therefore, BMSCs are the key factor for the development and treatment of alcoholic ONFH. In the present study, ethanol significantly inhibited the proliferation and osteogenic differentiation of BMSCs, which was consistent with our previous research. 21,26 Moreover, Alda-1 could obviously reverse the inhibition of ethanol-induced proliferation and osteogenic differentiation of BMSCs. What is more, our work indicated that Alda-1 alleviated the ethanol-induced inhibitory osteogenesis of BMSCs by PI3K/AKT signalling pathway. PI3K/AKT signalling is a central pathway for the osteogenic differentiation of BMSCs. Alda-1 accelerates cutaneous wound healing by Akt signalling pathway. 15 Intriguingly, Alda-1 significantly reduced the occurrence of ONFH and promoted osteogenic differentiation and new bone formation in the rat model of alcohol-induced ONFH. Taken together, Alda-1 might alleviate the ethanol-induced ONFH by PI3K/AKT pathway via regulating the osteogenesis of BMSCs.
The vital pathological features of ONFH progress includes intramedullary fat deposition, increased number of adipocytes and enlarged lipid droplets. 27,28 It is well known that fat is mainly derived from adipogenic differentiation of BMSCs, and only a little of BMSCs would differentiate into adipocytes under normal conditions. Therefore, furthest inhibition the adipogenic capacity of BMSCs is the key point for preventing the development of ONFH. Accumulated studies indicate that ethanol triggers the transformation of BMSCs predominantly into adipocytes. 29 Our results also demonstrated that ethanol significantly promoted the adipogenic differentiation, fortunately, Alda-1 significantly retarded alcohol-induced adipogenesis of BMSCs via AMPK pathway. Our previous studies indicated that ethanol promoted adipogenesis leading to ONFH by AMPK pathway. 21 Moreover, Alda-1 activation of ALDH 2 protected against ethanol-induced cardiac defect and apoptosis via AMPK signalling. 30 In summary, Alda-1 might alleviate the ethanol-induced ONFH via retarding the ethanol-induced adipogenesis by AMPK pathway.
The femoral head is a highly vascularized tissue, blood vessels supply oxygen, minerals and nutrients for bone homoeostasis and remodelling, carry away the metabolites. 31 Moreover, another name of ONFH is Aseptic ischaemic necrosis of the femoral head. Therefore, vascularization of the femoral heads is vital for prevention and treatment of ethanol-induced ONFH. The previous research indicated that high dose of ethanol significantly inhibited the proliferation, migration and angiogenesis of HUVECs by terminally regulating growth factors (like VEGF and PDGF). 32,33 What is more, Alda-1 activation of ALDH 2 increased angiogenesis via upregulation of VEGF receptors. 34 Similarly, high dose of ethanol significantly inhibited the expression level of growth factors resulting in reduced angiogenesis, whereas Alda-1 could rescue the inhibitory effect of ethanol on angiogenesis in the present study. Intriguingly, Alda-1 significantly reduced the occurrence of ONFH and promoted vessel number and distribution in the rat model of alcohol-induced ONFH.
In summary, Alda-1 activation of ALDH 2 was highly demonstrated to protect ethanol-induced ONFH by triggering new bone formation, reducing adipogenesis and stimulating vascularization.
Therefore, ALDH 2 may be a potential therapeutic target, and small molecule Alda-1 may be a promising pharmacotherapeutic for ONFH in the future.

AUTHOR CONTRIBUTIONS
Hongping Yu and Xiaoyi Lin designed the study, analysed, interpreted the data and drafted the paper. Xiaoyi Lin, Daoming Zhu, Kaiyang Wang, Pengbo Luo, and Gang Rui participated in acquiring the data.
Hongping Yu, Xiaoyi Lin, Daoming Zhu, Fuan Liu, and Youshui Gao designed the work and drafted the paper.