Normal bone maintenance relies on the formation of new bone by osteoblasts (OBs) and the resorption of aged bone by osteoclasts (OCs), a process known as “bone remodeling” (1). Imbalances in this process can result in bone diseases, such as osteoporosis, characterized by increased bone resorption and decreased bone formation. A reduction in bone mineral density caused by the imbalanced bone remodeling process can eventually lead to an elevated risk of bone fractures (1). Various factors, including sex, aging, and genetic variations, contribute to the bone remodeling process as well as environmental factors such as lifestyle, diet, smoking, and alcohol consumption (2, 3).

Alcohol is metabolized by alcohol dehydrogenase into acetaldehyde, which is further broken down to acetic acid by acetaldehyde dehydrogenase 2 (ALDH2) (4). ALDH2 is a mitochondria enzyme, and is the most efficient enzyme for removing toxic acetaldehyde after alcohol consumption (5). ALDH2 deficiency leads to the accumulation of acetaldehyde, resulting in physiological responses, known as the Asian alcohol flush reaction and other uncomfortable feelings such as nausea, headache, and rapid heart rate (6). Roughly 40% of East Asians possess a single G-to-A point mutation in the ALDH2 gene (7). This single nucleotide polymorphism, referred to as rs671, causes an amino acid change from ALDH2Glu504Lys (ALDH2*2), which acts as a dominant-negative protein (7). The combination of ALDH2*1 and ALDH2*2 alleles results in three genotypes with varying enzymatic activity: wild-type (ALDH2*1/*1), heterozygous (ALDH2*1/*2), and homozygous (ALDH2*2/*2) (8). An estimated 560 million Easy Asians are carriers of ALDH2*2 and the prevalence varies from ∼30% to as high as 45% in the populations of different regions (7).

Several studies have associated ALDH2 genetic mutations with bone formation and related diseases such as osteoporosis. An investigation into the association between hip fractures and ALDH2 mutations revealed that ALDH2 rs671 increases the risk of hip fracture (2). Additionally, an ALDH2 polymorphism was found to be linked to an increased risk of osteoporosis in a Japanese cohort (3, 9). A significant difference was observed in the stiffness index, an indicator of osteoporotic fractures, and the absence of the ALDH2 Glu allele was shown to negatively impact the risk of osteoporosis, particularly in women (9). Aceta-ldehyde treatment significantly inhibits early osteoblastogenesis in cultures of mouse- and human-derived bone marrow cells (3, 9). This inhibitory effect is particularly toxic to OB precursors at low concentrations as low as 0.004%∼0.020% and inhibits cell proliferation in two human OB cell lines (3, 9). The presence of ALDH2*2 can induce oxidative stress due to the accumulation of undegraded acetaldehyde during alcohol consumption, which can lead to inflammation or apoptosis, ultimately negatively affecting osteogenesis and increasing the incidence of osteoporosis (10, 11). These previous reports suggest that alcohol consumption and ALDH2 mutation together might also affect osteogenesis and bone formation.

In this study, we investigated the effects of ALDH2 mutations on bone formation. We generated human induced pluripotent stem cells (hiPSCs) with and without heterozygous ALDH2 mutations and induced OB differentiation. Different tendencies during OB differentiation with or without acetaldehyde treatment were confirmed and compared between the wild type and mutation groups. These findings indicate that groups with the ALDH2 mutation show a different response to acetaldehyde treatment and may be more susceptible to impaired bone formation.

Generation and culture of hiPSC

ALDH2*1/*1- and ALDH2*1/*2-hiPSCs were generated from peripheral blood mononuclear cells (PBMCs) using a previously described reprogramming method, which involved serial centrifugation and Sendai viruses (Fig. 1A) (12, 13). The cells were cultured in vitro nectin-coated dishes (Thermo Fisher Scientific). The culture medium was replaced daily with fresh Essential 8 (E8) medium (Thermo Fisher Scientific).

Figure 1. Characterization of ALDH2*1/*1-hiPSCs and ALDH2*1/*2-hiPSCs. (A) Schematic representation of hiPSC generation. (B) DNA sequencing results of ALDH2*1/*1- and ALDH2*1/*2-hiPSCs confirming ALDH2 mutation. (C) Measurement of ALDH2 activity in ALDH2*1/*1-hiPSCs (n=3) and ALDH2*1/*2-hiPSCs (n=3). (D) Bright field microscope images and alkaline phosphatase (AP) staining of ALDH2*1/*1- and ALDH2*1/*2-hiPSCs. Scale bar=200 μm. (E, F) Gene expression of pluripotency markers in ALDH2*1/*1-hiPSCs (n=3) and ALDH2*1/*2-hiPSCs (n=3), determined by RT-PCR and qRT-PCR. GAPDH was used as a loading control. Relative gene expression values were calculated using the delta-delta Ct (2^(-ΔΔCt)) method. (G) Immunocytochemical staining for pluripotency markers OCT4, SOX2, KLF4, TRA-1-60, TRA-1-81, and SSEA4. Scale bar=100 μm. Data are presented as the mean±SEM. ^##p<0.01 (t-test) indicates statistical significance. PBMC: peripheral blood mononuclear cell, ALDH2: acetaldehyde dehydrogenase 2, hiPSCs: human induced pluripotent stem cells, RT-PCR: real-time polymerase chain reaction, qRT-PCR: quantitative RT-PCR.

OB differentiation of hiPSCs

The in vitro osteogenic differentiation assay conducted in this study followed the protocol outlined in previous research (14). Briefly, hiPSCs were seeded at a density of 3.75×10⁴ cells/cm² on a culture plate coated with 0.1% gelatin. The cells were cultured in E8 medium supplemented with 10 μM rho-associated kinase inhibitor. Over a period of 2∼3 days, the cells reached 80% confluence with daily medium changes. Subsequently, the culture medium was replaced with osteogenic differentiation medium (ODM). The ODM was prepared by supplementing Dulbecco’s modified Eagle’s medium-low glucose (Thermo Fisher Scientific) with 15% fetal bovine serum (Thermo Fisher Scientific), dexamethasone (100 nM, D4902; Sigma-Aldrich), β-glycerol-2-phosphate disodium salt (10 mM, G9422; Sigma-Aldrich), and ascorbic acid (50 μg/mL; Sigma-Aldrich). The medium was changed daily, and the cells were cultured for a duration of 7 days (Fig. 2A). In certain experiments, 4 mM acetaldehyde (402788; Sigma-Aldrich) was added to the culture medium for a period of 3 days (Fig. 3A).

Figure 2. Differentiation of ALDH2*1/*1 hiPSCs and ALDH2*1/*2 hiPSCs-derived OBs. (A) Schematic representation of the osteogenic differentiation process of ALDH2*1/*1-hiPSCs and ALDH2*1/*2-hiPSCs. (B) Bright-field microscopy images showing the mineralized cell morphology. Scale bar=40 μm. (C) Calcium deposition and hydroxyapatite quantification in ALDH2*1/*1-OBs and ALDH2*1/*2-OBs on day 7, visualized using Alizarin Red S and OsteoImage staining. Scale bar=40 μm. (D) Quantitative assessment of Alizarin Red S staining by absorbance measurement at 405 nm. (E) Verification of OsteoImage staining using the ImageJ software. (F) Quantified expression levels of pluripotent marker OCT4 and osteogenic marker genes, including RUNX2, COL1A1, and OCN, in ALDH2*1/*1-hiPSCs (n=3) and ALDH2*1/*2-hiPSCs (n=3) after 7 days of osteogenic differentiation. Data are presented as the mean±SEM. Statistical significance: ^#p<0.05, ^##p<0.01, ^###p<0.001 (t-test), **p<0.01 (two-way ANOVA). E8: Essential 8, ALDH2: acetaldehyde dehydrogenase 2, hiPSCs: human induced pluripotent stem cells, OBs: osteoblasts.

Figure 3. Effects of acetaldehyde on OB differentiation. (A) Schematic representation of OB differentiation in ALDH2*1/*1-hiPSCs and ALDH2*1/*2-hiPSCs with or without 4 mM acetaldehyde. (B) Cell viability of the ALDH2*1/*1- and ALDH2*1/*2-hiPSCs derived OBs against acetaldehyde. (C) Measurement of ALDH2 activity in acetaldehyde-treated OBs. (D) Alizarin Red S staining confirming calcium deposition in OBs, measured by absorbance at 405 nm. (E) Quantitative data showing the expression of osteogenic markers, including RUNX2, COL1A1, OCN, and ALP, in OBs treated with or without 4 mM acetaldehyde for 3 days. (F) Protein expression of RUNX2, OCN, and ALDH2 in OBs with or without 4 mM acetaldehyde. GAPDH was used as a loading control. (G) Quantification of western blot results using the Adobe Photoshop. Data are presented as the mean±SEM. Statistical significance is indicated as ^#p<0.05, ^##p<0.01, ^###p<0.001 (t-test), *p<0.05, **p<0.01, ***p<0.001 (determined by two-way ANOVA). E8: Essential 8, ALDH2: acetaldehyde dehydrogenase 2, OBs: osteoblasts.

Cell viability assay

To confirm the ALDH2*1/*1- and ALDH2*1/*2-hiPSCs derived OBs cell viability against acetaldehyde, Cell Coun-ting Kit-8 assay (CCK-8; Dojindo Molecular Technologies) was performed. We added 0, 2, 4, 8, 16, or 32 mM acetaldehyde to the culture medium; then, CCK-8 solution was added a 1：10 ratio. Absorbance was measured 450 nm using a microplate reader.

Alizarin Red S staining

To assess the extent of calcium deposition, we utilized the Alizarin Red S Staining Quantification Assay (#8678; ScienCell Research Laboratories). Staining was conducted following the manufacturer’s protocol. Briefly, differentiated cells were rinsed with 1X phosphate-buffered saline (PBS) and fixed with 4% paraformaldehyde (PFA) for 20 minutes. Subsequently, the cells were exposed to Alizarin Red S solution and stained at room temperature (RT) for 30 minutes. Stained samples were observed under an optical microscope. To quantify the level of staining, the samples were treated with a 10% acetic acid solution at RT with shaking for 30 minutes. The cells were gently scraped from the plate and transferred to a microcentrifuge tube containing 10% acetic acid. After vortexing for 30 seconds, the samples were heated at 85℃ for 10 minutes, followed by a 5-minute incubation on ice. The samples were then centrifuged at 20,000 ×g for 15 minutes and 200 μL of the supernatant was transferred to a fresh microcentrifuge tube. Next, 10% ammonium hydroxide was added to neutralize the acid. Finally, the absorbance of the supernatant was measured at 405 nm using a microplate reader.

OsteoImage Mineralization Assay

The hydroxyapatite, a calcium phosphate mineral was identified using the OsteoImage Mineralization Assay (PA-1503; Lonza). In brief, the cells were rinsed with 1X PBS and fixed with 4% PFA at RT for 20 minutes. Following the wash with 1X wash buffer, the staining reagent was added, and the samples were incubated at RT shielded from light, for 30 minutes. Subsequently, the samples underwent three 5-minute washes with 1x wash buffer. Stained samples were observed using an inverted fluorescence microscope (Axio Observer.Z1; Carl Zeiss).

Quantitative real-time polymerase chain reaction and real-time polymerase chain reaction

In this experiment, total RNA was isolated from all cells using TRIzol reagent (Thermo Fisher Scientific) and cDNA was synthesized using the RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific). Quantitative real-time polymerase chain reaction (qRT-PCR) was performed with a SYBR Green Mix (04707516001; Roche) and Light-Cycler 480 Instrument II (Roche Diagnostics). The expression levels of genes were normalized to GAPDH and calculated using the delta delta cycle-threshold (ΔΔCt) method. Real-time polymerase chain reaction (RT-PCR) was performed using i-TaqTM DNA Polymerase (iNtRON BIOTECHNOLOGY). The primer sequences for PCR are provided in Table 1.

Table 1 . Primer sequences for RT-PCR and qRT-PCR analysis

Name	Direction	Primer sequence (5’-3’)	Target size (bp)
Human OCT4	Forward	ACCCCTGGTGCCGTGAA	190
	Reverse	GGCTGAATACCTTCCCAAATA
Human SOX2	Forward	CAGCGCATGGACAGTTAC	321
	Reverse	GGAGTGGGAGGAAGAGGT
Human NANOG	Forward	AAAGGCAAACAACCCACT	270
	Reverse	GCTATTCTTCGGCCAGTT
Human LIN28	Forward	CTTCGGCTTCCTGTCCAT	122
	Reverse	CTGCCTCACCCTCCTTCA
Human RUNX2	Forward	GTGCCTAGGCGCATTTCA	78
	Reverse	GCTCTTCTTACTGAGAGTGGAAGG
Human COL1A1	Forward	TCTGCGACAACGGCAAGGTG	146
	Reverse	GACGCCGGTGGTTTCTTGGT
Human OCN	Forward	CGCTACCTGTATCAATGGCTGG	123
	Reverse	CTCCTGAAAGCCGATGTGGTCA
Human ALP	Forward	GACCCTTGACCCCCACAAT	68
	Reverse	GCTCGTACTGCATGTCCCCT
Human RANKL	Forward	ACATATCGTTGGATCACAGCACAT	100
	Reverse	CAAAGGCTGAGCTTCAAGCTT
Human OPG	Forward	TGCTGTTCCTACAAAGTTTACG	433
	Reverse	CTTTGAGTGCTTTAGTGCGTG
Human TNFα	Forward	GAGGCCAAGCCCTGGTATG	91
	Reverse	CGGGCCGATTGATCTCAGC
Human GAPDH	Forward	CTGTTGCTGTAGCCAAATTCGT	101
	Reverse	ACCCACTCCTCCACCTTTGA

RT-PCR: real-time polymerase chain reaction, qRT-PCR: quantitative RT-PCR.

Immunocytochemistry

The cells were fixed with 4% PFA for 30 minutes, after which the cells were rinsed twice with 1x PBS and incu-bated with 50 mM NH4Cl for 10 minutes. After additional washes, 0.1% Triton X-100 was added, and the cells were incubated at RT for 30 minutes. Subsequently, the cells were blocked with 2% bovine serum albumin (BSA) in PBS for 30 minutes, incubated with the primary antibody, and diluted in 2% BSA, at RT for 2 hours. Next, the cells were washed with 2% BSA and incubated with Alexa Fluor 594-conjugated goat anti-rabbit IgG (H＋L) antibody (A11037; Molecular Probes) and Alexa Fluor 488-conjugated goat anti-mouse IgG (H＋L) antibody (A11029; Molecular Probes) at RT for 1 hour. After washing, the cells were treated with 4,6-diamino-2-phenylindole (10236276001; Roche) for 10 minutes. Following another wash with 2% BSA and 1x PBS, the cells were mounted using an antifade antibody (H-1700; Vector Laboratories Inc.). Images were captured using an upright fluorescence microscope (Axio Imager.M2; Carl Zeiss).

ALDH2 activity assay

ALDH2 activity was assessed using the Mitochondrial Aldehyde Dehydrogenase (ALDH2) Activity Assay Kit (ab115348; Abcam) according to the manufacturer’s instructions. Cells were scraped and incubated with 1x extra-ction buffer supplemented with a Phosphatase Inhibitor Cocktail Set IV (1：50, 524628; Merck Millipore) and PMSF protease inhibitors (100 mM). Subsequently, the samples were shaken at 4℃ for 20 minutes and centrifuged at 16,000 ×g for 20 minutes at 4℃. The extracted proteins were quantified using a bicinchoninic acid protein assay. The proteins were diluted in 1x incubation buffer. After incubation at RT for 3 hours with shaking at 0.02268 ×g, the samples were washed by aspirating or decanting from the wells and then 300 μL of 1X wash buffer was dispensed into each well. Next, 200 μL of 1X activity solution was added to each well and the absorbance was measured at 450 nm using an absorbance microplate reader after 30 minutes.

Western blot analysis

Cellular proteins were extracted using RIPA buffer (R0278; Sigma-Aldrich) supplemented with the PMSF protease inhibitor (100 mM). Proteins were quantified by the method described in the ALDH2 activity assay. Protein samples were separated on 8%, 10%, and 12% sodium dodecyl sulfate-polyacrylamide gels through electrophoresis, and subsequently transferred onto nitrocellulose membranes. Then 1 hour of blockade with 3% BSA diluted in 1X TBS supplemented with Tween-20, the membranes were incubated overnight at 4℃ with the following primary antibodies: anti-RUNX2 (1：1,000, ab23981; Abcam), anti-osteocalcin (anti-OCN, 1：500, sc-30044; Santa Cruz Biotechnology Dallas), anti-osteoprotegerin (anti-OPG, 1：500, sc-390518; Santa Cruz Biotechnology), anti-RANKL (1：1,000, sc-377079; Santa Cruz Biotechnology), anti-tumor necrosis factor α (anti-TNFα, 0.2 μg/mL, ab9739; Abcam), anti-4-hydroxynonenal (anti-4HNE, 1：1,000, bs-6313R; Bioss Antibodies), anti-ALDH2 (1：1,000, PA5-29717; Thermo Fisher Scientific), and anti-GAPDH (1：5,000, ab8245; Abcam). The membranes were washed with 1X TBST and then incubated for 1 hour with the secondary antibodies. Finally, the protein was visualized using the WESTSAVE Gold ECL Solution (LF-QC0103; Ab Frontier) and a bio-image analysis system (Amersham Imager 600; Fuji Photo Film Co., Ltd.). Protein levels were normalized to GAPDH. Band intensity was quantified by analyzing the pictures using Adobe Photoshop.

Enzyme-linked immunosorbent assay

The recombinant human TNFα protein (210-TA; R&D Systems) and cultured medium was diluted in 5X enzyme-linked immunosorbent assay ELISPOT DILUENT (248227-000; Invitrogen) and were coated on the 96-well plate overnight at 4℃. The plate was washed three times with 1x PBS and blocked with 5% skim milk at RT for 1 hour. Then the plate was again washed three times with 1x PBS and incubated with anti-TNFα (0.5 μg/mL) diluted in 5% skim milk at RT for 2 hours. The plate was washed three times with 1x PBS and then incubated at RT for 1 hour with the secondary antibody diluted in 5% skim milk. After washing the plate three times with 1x PBS, add 1X TMB SUBSTRATE SOLUTION (249156-000; Invitrogen) and incubated at RT for 10 minutes. Finally, stop solution was added and absorbance was measured using a microplate reader.

Human cytokine array

The cytokine array was assessed using a human XL Cyto-kine Array Kit (ARY022B; R&D Systems) according to the manufacturer’s protocol. The membranes were blocked at RT for 1 hour and then incubated overnight at 4℃ with culture medium that cultured for 7 days. Then, the detection antibody cocktail was diluted in 1X array buffer and incubated at RT for 1 hour. Finally, the membrane incubated with 1X Streptavidin-HRP at RT for 30 minutes. Images were acquired using a bio-image analysis system (Amersham Imager 600) and were quantified.

Statistical analysis

All experiments were conducted a minimum of three times. Statistical analyses were performed using the Prism 9.0 software (GraphPad Inc.). The results are presented as the mean±SEM. Statistical significance between groups was determined using Student’s t-test (^#p<0.05, ^##p<0.01, ^###p<0.001 indicate statistical significance). For nonpa-rametric quantitative datasets, a t-test was employed, and a one-tailed p-value was calculated. Two-way ANOVA was used for several analyses (*p<0.05, **p<0.01, ***p<0.001 indicate statistical significance).

Pluripotency of ALDH2*1/*2-hiPSCs

We generated the ALDH2-hiPSC lines from peripheral PBMCs using Sendai virus expressing Yamanaka factors (Oct4, Sox2, Klf4, and c-Myc) (Fig. 1A). In ALDH2*1/*2-hiPSCs, we confirmed a single nucleotide alteration from guanine (G) to adenine (A) at exon 12 of the ALDH2 gene. This alteration replaced the 504th amino acid (glutamate) with lysine, resulting in a missense variant (Table 2). Based on the DNA sequencing results, we categorized the generated hiPSC lines into two groups: hiPSCs without mutations (ALDH2*1/*1-hiPSCs, n=3) and those with a heterozygous mutation (ALDH2*1/*2-hiPSCs, n=3) in the ALDH2 gene (Fig. 1B). ALDH2 activity was analyzed in both ALDH2*1/*1-hiPSCs and ALDH2*1/*2-hiPSCs. The activity of ALDH2 was lower in ALDH2*1/*2-hiPSCs compared with that of the ALDH2*1/*1-hiPSC group (Fig. 1C). However, both ALDH2*1/*1-hiPSCs and ALDH2*1/*2-hiPSCs exhibited morphological characteristics typical of hiPSC colonies and stained positive for alkaline phosphatase, confirming their undifferentiated state (Fig. 1D, Supplementary Fig. S1A). The expression of pluripotency-related genes, including OCT4, SOX2, NANOG, and LIN28, showed no significant difference between the groups (Fig. 1E). The gene expression of pluripotency markers OCT4, SOX2, and NANOG in ALDH2*1/*1-hiPSCs and ALDH2*1/*2-hiPSCs were further quantified (Fig. 1F). Interestingly, only SOX2 expression significantly decreased in ALDH2*1/*2-hiPSCs. Additionally, we observed no differences in the protein expression of pluripotency markers OCT4, SSEA4, TRA-1-60, SOX2, TRA-1-81, and KLF4 between ALDH2*1/*1-hiPSCs and ALDH2*1/*2-hiPSCs (Fig. 1G, Supplementary Fig. S1B). Our results demonstrate that generally the genetic mutation in the ALDH2 gene did not affect the pluripotency of the generated ALDH2*1/*2-hiPSCs and that there was no difference in pluripotency marker expression between ALDH2*1/*1-hiPSCs and ALDH2*1/*2-hiPSCs.

Table 2 . ALDH2*1/*2-hiPSCs with ALDH2 mutations

Gene	Chromosome position	Nucleotide change	Protein mutation
ALDH2	chr12:111803962-111803962	c.1510G>A	p.Glu504Lys

Osteogenic differentiation using ALDH2*1/*2-hiPSCs

Previously, we confirmed that ALDH2 mutation did not alter the general characteristic of hiPSCs. To confirm the affect of ALDH2 mutation on in vitro osteogenesis, OB differentiation was induced in each hiPSC lines for 7 days (Fig. 2A). Successful differentiation was indicated by the presence of calcium deposits in the differentiated cells (Fig. 2B). Calcium deposits in OBs derived from ALDH2*1/*1- and ALDH2*1/*2-hiPSCs was assessed on day 7 using Alizarin Red S and OsteoImage staining (Fig. 2C). The staining results confirmed that ALDH2*1/*2-OBs generated fewer calcium deposits compared to ALDH2*1/*1-OBs, and particularly a significant difference was confirmed through OsteoImage staining (Fig. 2D, 2E). The expre-ssion of OB-specific markers was confirmed in ALDH2*1/*1-OBs and ALDH2*1/*2-OBs on day 7 and the levels were compared to that of each undifferentiated hiPSCs (Fig. 2F). The levels of the pluripotency marker OCT4 were significantly reduced in the differentiated cells. On the other hand, increased expression of osteogenic markers such as RUNX2, COL1A1, OCN, and ALP was confirmed in the differentiated OBs. The expression of RUNX2, a critical transcription factor in OB differentiation was significantly increased in both groups. Yet, the increased rate of COL1A1 expression compared with the control hiPSCs in ALDH2*1/*2-OBs were significantly lower than that of ALDH2*1/*1-OBs. In summary, we observed decreased calcium deposition in ALDH2 mutated OBs, and less reduced COL1A1 expression was less induced compared with that of wild type ALDH2*1/*1-OBs.

Alteration in OB differentiation induced by acetaldehyde treatment

As previously mentioned, several studies have reported the inhibitory effect of acetaldehyde on osteogenesis (3, 9). The range of acetaldehyde concentration was first determined based on prior research (3, 9), and we further confirmed that 4 mM acetaldehyde was the highest concentration that did not affect the viability of cells (Fig. 3B). In ALDH2*1/*1-OBs, acetaldehyde treatment significantly increased ALDH2 activity; however, no changes were observed between the ALDH2*1/*2-OB groups (Fig. 3C). Alizarin Red S staining revealed a significant reduction in calcium deposition in both groups when treated with acetaldehyde and ALDH2*1/*2-OBs showed a more significant reduction (Fig. 3D). The effect of acetaldehyde on osteogenic specific gene expression was also confirmed in cultured ALDH2*1/*1- and ALDH2*1/*2-OBs (Fig. 3E). In ALDH2*1/*1-OBs, acetaldehyde significantly reduced the expression of COL1A1 and OCN and the expression levels of OCN and ALP were significantly lower in ALDH2*1/*2-OBs compared with ALDH2*1/*1-OBs. The expression of COL1A1 and OCN was significantly reduced in ALDH2*1/*2-OBs when treated with acetaldehyde. The protein expression of osteogenic marker, RUNX2 and OCN, along with ALDH2 was confirmed in the differentiated OBs (Fig. 3F, 3G). Both ALDH2*1/*1- and ALDH2*1/*2-OBs showed decreased RUNX2 expression after acetaldehyde treatment (Fig. 3F). The protein expression of OCN was significantly lower in ALDH2*1/*2 without acetaldehyde treatment, which showed a similar tendency to that of the gene expression results (Fig. 3G). The expression of OCN in ALDH2*1/*2-OB did not alter with or without acetaldehyde treatment. ALDH2 protein expression did not significantly change after acetaldehyde in all groups; however, ALDH2*1/*2-OBs had a significantly lower expression of ALDH2 protein compared with that of ALDH2*1/*1-OBs. These results indicate that acetaldehyde affects in vitro osteogenesis in both ALDH2*1/*1- and ALDH2*1/*2-OBs, but with different tendencies between the groups.

Impact of ALDH2 mutation and acetaldehyde on the RANKL/OPG axis during in vitro osteogenesis

In comparison with cytotoxicity, inflammation is a quicker and more sensitive toxic response to chemical exposure. Hence, our investigation aimed to determine if acetaldehyde triggers an inflammatory response in ALDH2*1/*1- and ALDH2*1/*2-OBs. The RANKL/OPG axis plays a critical role in inflammation in various tissues including bone (15) and TNFα is one of the factors that is reported to regulate this axis. These factors are also critical in bone turnover. We first confirmed the gene expression of TNFα, OPG, and RANKL in the differentiated OBs with or without acetaldehyde (Fig. 4A). The gene expression of TNFα showed an increasing tendency in both groups; however, this was only significant in ALDH2*1/*1-OBs. Interestingly, the gene expression of OPG was significantly decreased in ALDH2*1/*2-OBs compared with that of ALDH2*1/*1-OBs without acetaldehyde treatment, and both groups exhibited a decrease in OPG expression when cultured with 4 mM acetaldehyde. The gene expression of RANKL, however, was significantly increased in the acetaldehyde treated ALDH2*1/*2-OBs. The protein expression of inflammatory markers was confirmed in the differentiated cells (Fig. 4B). 4HNE is an aldehyde product of lipid peroxidation associated with pathological changes resulting from oxidative stress; moreover, it is also reported to be linked to TNFα regulation (16). While ALDH2*1/*1-OBs did not show a significant change after acetaldehyde treatment, the significantly lower expression of 4HNE in ALDH2*1/*2-OBs showed a significant increase after acetaldehyde treatment. ALDH2*1/*1-OBs did not show a consistent TNFα (Fig. 4C), yet, all samples of the ALDH2*1/*2-OBs showed a significant increased expre-ssion of TNFα. OPG showed the same tendency in both groups, yet, RANKL expression did not. Interestingly, RANKL expression showed an increased expression in ALDH2*1/*2-OBs without acetaldehyde treatment, and the expression did not significantly change after treatment with acetaldehyde. This resulted in a decreased tendency of OPG/RANKL ratio in ALDH2*1/*2-OBs (Fig. 4D). Soluble TNFα (sTNFα) was measured in the cultured sup of each group and both groups showed increasing TNFα after acetaldehyde treatment (Fig. 4E). The soluble cytokine release in the cultured sup was further confirmed (Supplementary Fig. S2A). We confirmed that acetaldehyde treatment reduced the expression of CD30, FGF2, and FGF19 in both groups (Supplementary Fig. S2B). However, the expression of CD147, Cripto-1, IGFBP-2, and PDGF-AB/BB showed a different expression pattern between the groups. While, the expression of CD147 and IGFBP-2 did not alter with or without acetaldehyde treatment in ALDH2*1/*1-OBs, the expression of both cytokines was downregulated in ALDH2*1/*2-OBs after acetaldehyde treatment. Moreover, Cripto-1 expression was totally absent in ALDH2*1/*2-OBs, while, PDGF-AB/BB was significantly increased.

Figure 4. Impact of acetaldehyde on ALDH2*1/*2-OBs: oxidative stress and altered TNFα, OPG, and RANKL levels. (A) Quantification of TNFα, OPG, and RANKL expression in OBs treated with or without 4 mM acetaldehyde. (B) Protein expression of 4HNE, TNFα, OPG, and RANKL in OBs with or without 4 mM acetaldehyde. GAPDH was used as a loading control. (C) Quantification of western blot results using the Adobe Photoshop. (D) Calculation of OPG and RANKL ratio based on quantified western blot results. (E) Measurement of TNFα in the cultrued medium of each group. Data are presented as the mean±SEM. ^#p<0.05, ^##p<0.01, ^###p<0.001 (t-test), *p<0.05, ***p<0.001 (two-way ANOVA) indicate statistical significance. TNFα: tumor necrosis factor α, OPG: osteoprotegerin, 4HNE: 4-Hydroxynonenal, OBs: osteoblasts.

These findings suggest that acetaldehyde triggers an inflammatory environment, and might also induce an imbalanced bone remodeling process, as suggested by a reduced OPG/RANKL ratio.

The ALDH2 gene plays a crucial role in the breakdown of acetaldehyde into acetic acid during alcohol metabolism (4). Genetic variation in ALDH2 results in the accumulation of acetaldehyde, which can have detrimental effects on the brain, liver, heart, and bones, leading to various diseases (2, 3). Both acetaldehyde and ALDH2 polymor-phisms are associated with osteoporosis, with ALDH2 mutations having a more significant impact on fracture risk and reduced bone density (1, 3, 9). Acetaldehyde impairs the viability and differentiation of OBs and osteoprogenitor cells, leading to inhibited bone formation and lower bone mass (17). Additionally, acetaldehyde hampers embryonic bone development in rats (18). Surprisingly, acetaldehyde primarily disrupts carbohydrate and protein metabolism in the liver, but also, surprisingly suppresses bone protein synthesis by 70%∼90%, resulting in DNA damage and cell death in the bone marrow (19). However, the specific effects of ALDH2 mutations and acetaldehyde on bone remodeling have not been thoroughly elucidated.

In this study, we initially confirmed that there were almost no discernible differences in the characteristics of hiPSCs except for SOX2 expression, between the wild-type and ALDH2*1/*2 groups, corresponding to the genetic mutation (Fig. 1D-1G). We confirmed that ALDH2 enzymatic activity was lower in the ALDH2*1/*2 group; however, no statistically significant difference was observed, which indicates that ALDH2 activity changes between the group did not affect the pluripotency of the generated hiPSCs (Fig. 1C). Subsequently, we proceeded with the differentiation of OBs and verified the presence of calcium deposits (Fig. 2C-2E). While it was indistinguishable through alizarin red S staining results, less formation of calcium deposits was confirmed through OsteoImage staining. Gene expression analysis further confirmed a decrease in the expression of pluripotency markers, accompanied by an increase in the expression of osteogenic markers in the differentiated OBs (Fig. 2F). COL1A1 encodes type I collagen, which is the most abundant extracellular matrix protein in bone. However, the relative expression of COL1A1 was less increased in ALDH2*1/*2-OBs compared with that of ALDH2*1/*1-OBs. These results suggest that ALDH2 genetic polymorphisms do exert an effect on OB differentiation.

To further compare the difference between ALDH2*1/*1- and ALDH2*1/*2-OBs, acetaldehyde was administered during the first 3 days of the differentiation process (Fig. 3). Unlike the results in the hiPSCs, ALDH2 in the differentiated OBs showed a significantly different ALDH2 activity (Fig. 3C). The activity of ALDH2 in ALDH2*1/*2-OBs showed roughly only half of the normal activity (i.e., wild type OBs). Moreover, while ALDH2 activity spiked after acetaldehyde treatment in ALDH2*1/*1-OBs, no significant changes were observed in the ALDH2*1/*2-OBs. This shows that ALDH2 enzyme activity is increased after OB differentiation in the wild type group, and that ALDH2 enzyme activity is impaired in differen-tiated ALDH2*1/*2-OBs. We expected that acetaldehyde treatment would slightly affect OB differentiation in the ALDH2*1/*1-hiPSCs; however, we also observed a steep reduction in the wild type cells (Fig. 3D).

The gene expression of COL1A1 and OCN was significantly decreased in the ALDH2*1/*1-OBs, but a more significant reduction was observed in the ALDH2*1/*2-OBs. Interestingly, the protein expression of OCN was already significantly decreased in the mutated ALDH2*1/*2-OBs and did not alter after acetaldehyde treatment (Fig. 3F). OCN encodes osteocalcin, which is one of the key players in bone biology, and a factor expressed and secreted solely by OBs (17). In the early phase of OCN discovery (20), OCN was hypothesized to play a role in extracellular matrix mineralization and used as a serum marker for bone formation. However, OCN depletion showed only minor effects on bone density and mineralization in mice (21). Since this discovery, the role of OCN has been reported based on murine and in vitro studies, mostly as a regulator in physiological pathways in an endocrine manner. Hoshi et al. (1) reported a similar study using transgenic mice expressing ALDH2*2 and an OB cell line (MC3T3-E1); when osteogenic differentiation was induced in the OB cells with acetaldehyde treatment, OCN was also significantly inhibited, which is consistent with the results in this study. There are also studies that report the critical role of OCN in liver diseases as OCN levels are inversely associated with liver disease in human (22). Mice treated with OCN during a high-fat and cholesterol diet had reduced scores in nonalcoholic fatty liver disease activity. An active undercarboxylated form of OCN reduced fat accumulation and inflammatory response in chicken embryonic hepatocytes by inhibiting the ROS-JNK signaling pathway (23). Therefore, we suggest that the ALDH2 mutation does not directly affect the bone, but the altered characterization of OBs might affect the liver condition of the ALDH2*2 population. In future studies, co-culture with ALDH2*1/*2-OBs and liver organoids might be interesting to confirm this issue.

The presence of 4HNE is detected in nearly all tissues experiencing oxidative stress (16). Treatment of cultured OBs with 4HNE induces increased oxidative stress and reduced bone formation. 4HNE is a common marker for oxidative stress and its possible pathogenesis in the nervous, respiratory, cardiovascular system has been verified in various studies (24). 4HNE is also thought to be a crucial factor for liver injury and liver cirrhosis (25). Furthermore, mice with ALDH2 genetic mutations exhibit osteopenia due to dysfunctional OBs, particularly impaired differentiation caused by the accumulation of acetaldehyde and subsequent elevation of 4HNE levels in OBs (1). In this study, we detected the protein expression of 4HNE in OBs with ALDH2*1/*1 and ALDH2*1/*2 genotypes to confirm if acetaldehyde exposure increased 4HNE in ALDH2*1/*2-OBs in particular. The expression of 4HNE in ALDH2*1/*2-OBs significantly increased after acetaldehyde treatment (Fig. 4C). 4HNE expression in ALDH2*1/*1-OBs did not increase after acetaldehyde treatment; however, unex-pectedly 4HNE was highly increased in non-treated ALDH2*1/*1-OBs, and almost absent in non-treated ALDH2*1/*2-OBs. Recently, Xiao et al. (26) confirmed the role of 4HNE in the progression of steroid-induced osteonecrosis of the femoral head (SIONFH) in 36 patients and 28 healthy volunteers. The expression level of 4HNE in the SIONFH group serum was higher than that in the normal control group; however, it decreased with the progression of SIONFH. Moreover, it was suggested that high expression of 4HNE in the early stage of SIONFH aggravates the oxidative stress of osteocytes, leading to disrupted balance in the bone remodeling process, and eventually to femoral head collapse. After collapse, the decreased levels of 4HNE are thought to be associated with nonvital bone with a large portion of empty osteocyte lacunae. However, our results also suggest that 4HNE might be a critical factor for in vitro osteogenesis, which requires further confirmation.

TNFα is a cytokine that affects bone metabolism in various inflammatory diseases and pathological processes, including rheumatoid arthritis, bone fractures, and ankylosing spondylitis (27). TNFα is a type 2 transmembrane protein and exists as transmembrane TNFα (tmTNFα) and sTNFα (28). tmTNFα is a precursor of sTNFα and is cleaved by metalloproteinase, TNFα converting enzyme (TACE), and released as sTNFα. Both tmTNFα and sTNFα are biologically active, and both can bind to TNF receptor I (TNFRI, p55) and TNF receptor II (TNFRII, p75) (29). Individuals with ALDH2 mutations have a higher risk of hip fracture and osteoporosis, which holds significant immunological implications as it has been shown to correlate with TNFα (3). Alcohol and acetaldehyde have been demonstrated to decrease bone mineral density by triggering oxidative stress and elevating TNFα levels, consequently inhibiting osteogenic differentiation and bone formation (30). In accordance with these findings, as depicted in Fig. 4, we verified the gene and protein levels of TNFα in ALDH2*1/*1- and ALDH2*1/*2-OBs cultured with or without acetaldehyde treatment. In our study, TNFα was highly increased in the lysate of differentiated OBs, especially in ALDH2*1/*2-OBs (Fig. 4C). sTNFα was not detected in the cultured sup when confirmed using a cytokine array; however, increased levels of TNFα were confirmed in both groups after acetaldehyde treatment without significance (Fig. 4E). This might suggests that tmTNFα might be a crucial factor in ALDH2 mutation-related symptoms rather than sTNFα. In the case of spondyloarthritis (SpA), sTNF levels in the synovial fluid were significantly decreased compared with rheumatoid arthritis despite similar levels of joint inflammation (31). In the synovial tissue of SpA, the expression of TACE was downregulated, which is thought to be responsible for the increased levels of tmTNFα in SpA. In tmTNFα transgenic mice knocked out with either TNFRI or TNFRII, it was confirmed that TNFRI is essential for inflammation and TNFRII for pathological new bone formation. These pathways might be responsible for the maintained RUNX2 and OPG expression in ALDH2*1/*2-OBs, and the reason why ALDH2 mutation has less impact on bone compared with other tissues such as the liver. Therefore, in future studies, it might be interesting to confirm the expression of TNFRs in cells to confirm which pathway is activated and altered by the mutation.

TNFα also plays a crucial role in stimulating osteoclastogenesis along with receptor activation of RANKL. RANKL, a member of the TNF family, is expressed by OBs and binds to the receptor RANK on the surface of OCs and OC precursors, initiating various signaling pathways that ultimately lead to bone resorption (32). A previous study demonstrated that alcohol promotes bone resorption in female rats by inducing osteoclastogenesis through increased expression of RANKL in OBs (33). The interaction between RANKL and its receptor RANK, which regulates the balance between bone formation and resorption, is inhibited by OPG. OBs produce OPG, which acts as a decoy receptor, binding to RANKL and preventing its interaction with RANK, thereby inhibiting the activation of OCs and bone resorption (34). After confirming induced oxidative stress and increased TNFα levels, we evaluated the expression of OPG and RANKL in ALDH2*1/*1- and ALDH2*1/*2-OBs treated with or without acetaldehyde (Fig. 4). Treatment with acetaldehyde reduced the mRNA levels of OPG in both groups of OBs, while RANKL expression was significantly increased only in the ALDH2*1/*2 group (Fig. 4A). However, western blotting revealed a significant reduction in OPG levels only in acetaldehyde-treated ALDH2*1/*2-OBs (Fig. 4B). Moreover, the protein level of RANKL was significantly higher in both acetaldehyde-treated and untreated ALDH2*1/*2-OBs compared with untreated ALDH2*1/*1-OBs. The OPG/RANKL ratio, based on the protein expression levels of OPG and RANKL, was significantly decreased in acetaldehyde-treated ALDH2*1/*2-OBs compared with untreated ALDH2*1/*1-OBs (Fig. 4D).

Other soluble cytokines were also released in the cultured sup (Supplementary Fig. S2A). Acetaldehyde treatment reduced the expression of CD30, FGF2, and FGF19 in both groups, while the expression of CD147, Cripto-1, IGFBP-2, and PDGF-AB/BB showed a different expre-ssion pattern between ALDH2*1/*1- and ALDH2*1/*2-OBs (Supplementary Fig. S2B). Interestingly, Cripto-1 was totally absent in ALDH2*1/*2-OBs compared with ALDH2*1/*1-OBs. While Cripto-1 is a fetal oncoprotein that plays critical roles in stem cell differentiation, embryogenesis, and tissue remodeling, it is also well known to contribute in cancer development and progression (35). From other perspectives, ALDH2 also represents a tumor suppressor in multiple cancer entities (36). This finding might suggest crypto-1 as a critical factor for ALDH2*2 allele carrier individuals with high risk for alcohol-related cancers. On the other hand, several studies have reported an inhibitory effect of PDGF on mesenchymal stem cell (MSC)-based osteogenesis (37). While the role of PDGF on osteogenesis is still controversial, further studies confirming the increased PDGF-AB/BB in ALDH2*1/*2-OBs might be a potential factor for better understanding the ALDH2 polymorphism characteristics.

Our study had several limitations that should be ack-nowledged. Firstly, we did not clearly demonstrate the exact mechanism through which acetaldehyde impairs osteogenic differentiation. However, recent research has indica-ted that alcohol consumption impairs the lineage differentiation of bone marrow-derived MSCs via the PI3K/AKT/mTOR/Sp7 pathway. This impairment leads to bone loss by inhibiting osteogenic differentiation and promoting fat differentiation (38). Additionally, the study highlighted that alcohol’s impact on osteogenesis is associated with increased expression of TNFα and interleukin (IL)-1β, as well as WNT signaling (28). Secondly, we did not confirm whether acetaldehyde-induced osteoclastogenesis in OCs triggers the secretion of OPG and RANKL from OBs. TNFα and IL-1β, cytokines known to influence bone loss, are involved in the bone resorption process and induce osteoclastogenesis by up-regulating NF-κB and JNK signaling pathways (39). Moreover, increased osteoclastogenesis induced by alcohol may be mediated by elevated expression of IL-6 and RANKL in mice subjected to liquid diets (40). Therefore, the impairment of bone formation and resorption caused by alcohol and acetaldehyde involves various cytokines and signaling pathways, and further studies are necessary to elucidate the role of ALDH2 in bone metabolism.

Supplementary data including two figures can be found with this article online at https://doi.org/10.15283/ijsc23151

There is no potential conflict of interest to declare.

All of the data used to support the findings of this study are included within the article.

Conceptualization: JL, JHJ. Data curation: JL, YAR, JHJ. Formal analysis: JL, HH, YAR. Funding acquisition: YAR, JHJ. Investigation: JL. Methodology: JL, SIJ, HH, YAR. Project administration: YAR, JHJ. Resources: YAR, JHJ. Software: JL, SIJ, HH, YAR. Supervision: YAR, JHJ. Validation: JL, HH, YAR, JHJ. Visualization: JL, HH, YAR. Writing – original draft: JL. Writing – review and editing: JL, HH, SIJ, YAR, JHJ.

This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (NRF-2020R1A2C3004123, NRF-2019R1A5A2027588, and NRF-2021R1C1C2004688). This research was also supported by the Catholic Institute of Cell Therapy in 2024 and by the Basic Medical Science Facilitation Program through the Catholic Medical Center of the Catholic University of Korea funded by the Catholic Education Foundation.