Effects of the common polymorphism in the human aldehyde dehydrogenase 2 (ALDH2) gene on the lung

To detect the overall effect of the ALDH2 polymorphism on lung function in the general population, we conducted a cross-sectional association study on healthy volunteers (n = 967). Longitudinal data were also assessed for the annual decline of forced expiratory volume in 1 s (FEV1) over at least four years (n = 742). Additionally, to identify if the mutant allele is associated with the incidence of bronchial asthma or COPD, we genotyped a cohort of patients with bronchial asthma (n = 751) and COPD (n = 289) in comparison to the general population. Patients with bronchial asthma and COPD were recruited from the Tsukuba University Hospital and its affiliated hospitals [20‐22]. Recruitment of the healthy volunteers was carried out as described previously [23]. They were recruited from the general population who visited the hospital for annual health checkups. Individuals with no evidence of pulmonary disease were included as healthy volunteers. Genomic DNA was extracted from the peripheral blood samples of all participants by an automated DNA extraction system (QuickGene-610 L; Fujifilm, Tokyo, Japan). Genotyping of patients with bronchial asthma and COPD was carried out by using the TaqMan Drug Metabolism Genotyping Assays. Genotyping of healthy volunteers was carried out by using the Illumina Human-Hap550v3 BeadChip assay (Illumina, San Diego, CA, USA) and the genotyping data of ALDH2*2 (rs671) was extracted using PLINK version 1.07 [23, 24]. Associations of the ALDH2*2 genotype with lung function data in healthy volunteers were analyzed by linear regression models and were adjusted for age, sex and smoking status in PLINK version 1.07. Chi-square tests were used to analyze the association between the ALDH2*2 genotype and the development of asthma and COPD.

Human surgical samples were collected from patients with a lung pathology that was clinically indicated for surgical removal. Fifteen samples were from a “lobectomy”, three were from a “pneumonectomy”, and six were from a “lung segment” removal. An apparently healthy lung portion away from the tumor and a portion from the bronchial stump were excised, shipped overnight on ice and were processed immediately upon receipt.

Small portions were used for DNA and RNA extraction, and for paraffin embedding for histological assessment. The remaining tissues were used for epithelial cell retrieval. The lung tissue was finely minced and incubated at 37 °C in 10 U/mL elastase (Porcine Pancreatic, Elastin Products Company, Owensville, MO, USA) for 30 min. The suspension was passed through an 18G needle using a large syringe to help disperse the tissue pieces into single cells. The cell suspension was then filtered through a 100-μm strainer to remove the undigested clumps. RBCs were depleted by adding 4 mL of ACK lysis buffer. Up to 500 μL of DNase (Sigma Aldrich, St. Louis, MO, USA) was added to the cell pellet. Then cells were incubated in 0.25% trypsin/EDTA at 37 °C for 20 min to obtain a single cell suspension. Hematopoietic and endothelial cells were then depleted using anti-human CD45 and CD31 microbeads on the autoMACS cell separation system (Miltenyi Biotec, Cologne, Germany). The cells were stained with an anti-human EpCAM antibody (BD Biosciences, Franklin Lakes, NJ, USA) and the epithelial cells were sorted on a MoFlo sorter. Non-epithelial cells were plated on plastic flasks in DMEM for 2 h to isolate the fibroblasts. Unattached cells were discarded and the fibroblasts were allowed to proliferate to >90% confluence, and were then frozen for future use in co-cultures.

The portions from bronchi were cleaned under a dissecting microscope and were cut open, followed by incubation in 50 U/mL dispase for 2 h at room temperature. Detached epithelial sheets were scrubbed off the inner surface and were incubated in 0.25% trypsin/EDTA at 37 °C for 30 min to obtain a single cell suspension. These cells were then referred to as primary human bronchial epithelial cells (HBECs), and cultured as such or stained for a basal cell marker (ITGA6, NGFR, or CD44) followed by sorting for basal cells.

Bronchial epithelial cells were resuspended in MTEC/Plus medium and were mixed 2:1 with growth factor-reduced Matrigel (BD Biosciences). Lung cells were co-cultured with human lung fibroblasts in transwell inserts as previously described [9, 10].

Some culture wells were treated with the ALDH2 activator, Alda-1 [25] (Merck Millipore, Darmstadt, Germany), at 100–300 μM and/or H₂O₂ at 300–400 μM as a source of ROS. The growing colonies were imaged and were visually counted on day 14 and/or 21. Colonies in the inserts were either embedded in Histogel (Thermo Scientific, Waltham, MA, USA) followed by paraffin for histological assessment, or were digested with dispase and trypsin to single cells. Single cells were either passaged or were used for RNA extraction.

Genotyping of human surgical samples was conducted using 2 sets of PCR primers with DNA isolated from patient lung samples as previously described [26]. Genotypes were further confirmed by the TaqMan® genotyping assay (TaqMan Drug Metabolism Genotyping Assays -ALDH2, Drug Metabolism Genotyping Assay mix, Life Technologies Corporation, Waltham, MA, USA).

mRNA expression of other lung-expressed ALDHs and several antioxidant genes was assessed by qRT-PCR. The following primers were used on a Step-One ABI cycler using SYBR green.

5′-TCCACCATTCATTGACTCCA-3′,

5′-CGCTGATCTTGCTCATGG-3′,

5′-AGCTCCTGCAACTCCTCAAA-3′,

5′-CAGGCTTGATGGTATCACTGC-3′,

5′-GAGTGAG CCAGTACGATCAGTG-3′,

5′-GTTGGCAGATCCACTGGTTT-3′,

5′-TTGATTTTGGAGGGATCTCG-3′.

Small pieces of human lung samples were homogenized and lysed with cell lysis buffer and followed by measurement of protein concentrations. Equal amounts of protein were fractionated by electrophoresis and then transferred to PVDF membranes. The membranes were then incubated with antibodies against NRF2 (Santa Cruz), ALDH1A1 (Abcam), and ALDH3A1 (Santa Cruz) followed by incubation with secondary antibodies. For the detection of specific protein bands, the membranes were incubated in LumiGLO reagent and peroxide (Cell Signaling Technologies, Danvers, MA, USA), and then exposed to X-ray films.

We used two different types of mice with disturbed ALDH2 function; Aldh2*2 transgenic (Tg) mice were generated by the pronuclear injection of a plasmid carrying mouse Aldh2*2 cDNA with a single nucleotide mutation at the same position as that of the human ALDH2*2 polymorphism, under the control of CAG promoter [27] and Aldh2 ^−/− mice, in which the endogenous Aldh2 gene is replaced by the neomycin-resistance gene [28]. The wild type (WT) littermates were used as controls. Trachea and lung paraffin sections from WT, Aldh2*2 Tg, and Aldh2 ^−/− mice were collected from newborn, adult, and aged mice and stained with hematoxylin and eosin (H&E). Tracheas were examined for epithelial thickness at 3 different locations. Because of natural variation in epithelial thickness between tracheal non-cartilaginous posterior portion and the rest of the circumference, and between uppermost and lowermost parts, we restricted our measurements and comparisons to the supracartilaginous regions of the sides of tracheal portion between cartilaginous rings 4 and 8. Four to six mice were analyzed per group. Nuclear densities were calculated by counting number of nuclei per a 100 μm. Quantification for percentage of cellular types in all groups was performed by counting from immunostained sections.

Human surgical samples and mice samples were fixed with 4% paraformaldehyde and embedded in paraffin. Tissue sections (6-μm thickness) were prepared and immunostained as described previously [29]. The primary antibodies used included rabbit K5 (Covance, Princeton, NJ, USA), goat CC10 (SCGB1A1) and Pro-surfactant protein C (SPC) (Santa Cruz Biotechnology), TTF1 (Abcam), mouse MUC5AC (Thermo Scientific), and mouse acetylated β-tubulin (Sigma) for the identification of basal, club (Clara), alveolar type II, goblet, and ciliated cells. Interleukin (IL)-1β (Abcam), Gr-1 (BD), Peroxiredoxin 1 (Prdx1) (Abcam) and Hemagglutinin (Abcam) were used for characterization of the injury models. The appropriate Alexa-Fluor coupled secondary antibodies were used in double and triple stained sections. The nuclei were stained with DAPI and the slides were then examined by fluorescence microscopy using a Zeiss AxioImager microscope (Carl Zeiss).

Tracheas were cut between the 5^th and 8^th tracheal cartilaginous rings. All samples were fixed in 2.5% glutaraldehyde/0.1 M phosphate buffer (pH 7.4) for 2–4 h. After post-fixation in 1.5% osmium tetroxide/ 0.1 M phosphate buffer (pH 7.4) for 1 h, the samples were dehydrated in ethanol, treated with propylene oxide, then embedded in an epon resin, which was polymerized at 60 °C for 2–3 days. Semi-thin sections were stained with toluidine blue to select an optimal area of the epithelium with straight basement membrane and without artifacts. Ultra-thin sections were mounted on a nickel mesh and were electron-stained with uranyl acetate and lead citrate. The sections were then examined under a transmission EM, JEM 1400 (JEOE, Tokyo, Japan).

Collection of MTEC and whole lung epithelium, isolation of lung fibroblasts and their co-culture in Matrigel were essentially performed as described previously [29]. The medium used, treatment with H₂O₂ and/or Alda-1, quantification and processing of colonies at the end point were performed in a manner similar to the human organoid culture described above.

Data from all samples from each group are expressed as mean ± SD. The quantification of CFE and cell types was performed by visual counting under a microscope and digital images from at least 3 triplicates and 3 independent experiments. T test was used for pairwise comparisons and a P value less than 0.05 was considered significant.

Healthy volunteers (n = 967) were genotyped for the ALDH2 polymorphism then were compared for their lung function parameters based on their genotype. The longitudinal data for the annual decline of forced expiratory volume in 1 s (FEV1) over at least four years were available from 742 individuals. The incidence of the ALDH2 polymorphism was also studied in bronchial asthma (n = 751) and COPD (n = 289) patient cohorts in comparison to the general population. The frequency of the ALDH2*2 allele in these three populations ranged between 41.1–44.8%, consistent with previous reports for the Japanese population and in keeping with the Hardy–Weinberg equilibrium. The clinical characteristics of all participants are shown in Table 1.

The presence of an ALDH2*2 allele A is associated with a significantly lower FEV1/FVC than that of the WT ALDH2 allele G within the group of healthy individuals (GG vs. GA vs. AA, β = −0.621, p = 0.0164). However, the presence of the ALDH2*2 allele A was not associated with predicted FEV1% (GG vs. GA vs. AA, β = −0. 424, p = 0.486) or the rate of annual FEV1 decline over the examined period of 4 years (GG vs. GA vs. AA, β = −1.122, p = 0.496). Additionally, the ALDH2*2 allele was not associated with the development of bronchial asthma or COPD (p = 0.42 and 0.747, respectively).

The human surgical samples assessed, included 26 lung and 23 bronchial samples. DNA was isolated from all samples and genotyped for the ALDH2 polymorphism (with ALDH2/ALDH2 alleles i.e. has functional ALDH2, heterozygous ALDH2/ALDH2*2, or homozygous for the ALDH2*2 allele i.e. with non-functional ALDH2). All human samples examined were either ALDH2/ALDH2 or ALDH2/ALDH2*2 heterozygous and no ALDH2*2 homozygous samples. Fixed and embedded bronchial portions were histologically assessed for epithelial abnormalities. Of 13 bronchial samples with the ALDH2 alleles, three samples showed goblet cell hyperplasia. Of 10 ALDH2/ALDH2*2 heterozygous bronchial samples, two showed basal cell hyperplasia, one showed goblet cell hyperplasia, and one showed both goblet and basal cell hyperplasia (Fig. 1a-f). Smoking history was not associated with hyperplasia development in both genotypes. Then, to detect a potential functional impairment on stem cells, both bronchial and lung epithelial stem cells were examined using the in vitro 3D colony forming assay [9, 10]. The CFE of primary and passaged epithelial cells of ALDH2/ALDH2 and ALDH2/ALDH2*2 heterozygous bronchial and lung samples was compared (Fig. 1g-l). The average primary bronchial CFE from 8 ALDH2/ALDH2 samples was 2.33%, while that from 5 ALDH2/ALDH2*2 heterozygous samples was only 1.58% (Fig. 1m). When the colonies were passaged to examine clonality, CFE at passage 1 in 6 ALDH2/ALDH2 samples was 0.86%, and that in 4 ALDH2/ALDH2*2 heterozygous samples was 0.56% (Fig. 1n). The primary lung CFE in 11 ALDH2/ALDH2 samples was 0.3%, and that in 6 ALDH2/ALDH2*2 heterozygous samples was 0.12% (Fig. 1o). Although the average CFE of both bronchial and lung samples in samples with the ALDH2 genotype tended to be higher than that with the ALDH2*2 genotype, the difference was not statistically significant. ALDH2 activation using a small molecule activator, Alda-1 [25], successfully increased the CFE of both bronchial and lung ALDH2/ALDH2*2 heterozygous epithelium by approximately 50% and 34% of their original efficiency, respectively. However, Alda-1 treatment of ALDH2/ALDH2 bronchial and lung epithelium produced a similar respective induction of 51% and 32.5% (Fig. 1k, l, p, q). Analysis of the differentiation profile of bronchial and lung colonies from both genotypes revealed no difference.

To examine if the absence of ALDH2 activity in the lungs results in compensatory upregulation or downregulation, or shows no effect on other functionally relevant genes, we examined the expression of other lung-expressed ALDHs and antioxidant genes in all human samples and analyzed their levels in individuals with ALDH2/ALDH2 versus ALDH2/ALDH2*2 heterozygous genotypes using RT-PCR. The expression of antioxidant genes, NRF2 and PRDX1, were significantly lower in the ALDH2/ALDH2*2 heterozygous than in the ALDH2/ALDH2 (Fig. 2a-d). HMOX-1 showed a trend towards lower expression but the difference was not significant (Fig. 2e, f). ALDH1A1 and ALDH3A1, and the antioxidants, NQO1 and TXNRD1, showed no significant difference. ALDH1A1 and ALDH3A1 expression was confirmed at the protein level using western blotting. The expression of these proteins varied among samples and no significant difference was observed between the ALDH2/ALDH2 and ALDH2/ALDH2*2 heterozygous samples (data not shown).

To further characterize the effect of the ALDH2 loss-of-function polymorphism on lung epithelial development, stem cell function, and repair after injury, we used two different types of mice with disturbed Aldh2 function. An Aldh2 ^−/− mouse, which expresses no Aldh2 protein [28] and an Aldh2*2 Tg mouse, which possesses an intact endogenous Aldh2, but overexpresses the Aldh2*2 (mutant) gene and protein. This non-functional protein interferes with the functions of the endogenous Aldh2 in a “dominant negative effect”, mimicking the effect of the human ALDH2 polymorphism [27]. We histologically compared the naïve mouse tracheas among WT, Aldh2*2 Tg, and Aldh2 ^−/− mice at birth, adult, and old age (Fig. 3a-i). Interestingly, significant differences in the tracheal epithelia were observed only in adult mice, but not in the newborn or old aged mice. The tracheal epithelia were thinner and showed lower cellular density in Aldh2*2 Tg and Aldh2 ^−/− mice than in WT mice (Fig. 3d-f, quantified in m-n).

When the cellular subtypes of adult tracheas from all groups were analyzed to explain the decrease in epithelial thickness and cellular density in Aldh2*2 Tg and Aldh2 ^−/− mice, only the basal, but not the club or ciliated cells, showed a significantly lower ratio in Aldh2*2 Tg and Aldh2 ^−/− mice than in WT mice (Fig. 3j-l, quantified in o). The decrease in the basal cell ratio was compensated for by an increase in ciliated and unidentified (intermediate/undetermined) cells, but not by the club cells. Basal cells are the stem cells in mouse trachea, capable of both self-renewal and differentiation into club and ciliated cells; the role of the unidentified (intermediate/undetermined) cells is not well understood [33]. These findings suggest that Aldh2 is involved in regulating the self-renewal and differentiation of basal stem cells in vivo. On the other hand, histological examination of the lung parenchyma and intra-pulmonary airways from all the adult mouse groups showed no significant difference compared to those of WT mice (Fig. 3p-r). Intrapulmonary airways, which contain no basal cells and are lined mainly with club cells and ciliated cells, were immunostained and cellular subtypes were examined at the level of proximal bronchi, mid-level bronchi, and terminal bronchioles in all mice genotypes and no significant differences were detected among them (Fig. 3s-u, quantified in v). The alveolar compartments showed no abnormality as well.

The relation between mitochondrial morphology and its respiratory activity has been identified for a long time [34]. Morphological differences between the mitochondria of club cells and ciliated cells have been described and are attributed to the high energy requirement of ciliated cells, resulting in their mitochondria having more cristae and a scanty matrix [35, 36]. In contrast, pathological conditions like hypoxia or increased oxidative stress result in “reduced-density” of the mitochondrial matrix [37, 38]. As ALDH2 is a mitochondrial enzyme whose deficiency augments oxidative stress [12, 13], we examined the effect of its deficiency on the mitochondrial morphology in airway epithelial cells, using TEM. In the tracheas of WT adult mice, the previously described morphological differences between the mitochondria in club and ciliated cells were observed (i.e., fewer cristae and expanded matrix in club cells) [35, 36]. In this study, we detected the presence of a previously unknown minor population of “abnormal” mitochondria, which showed near complete lack of cristae and a marked decrease in matrix density in all groups (Fig. 4a-b). We found these abnormal mitochondria in the club cells and basal cells, but not in ciliated cells (Fig. 4c). In most club and basal cells of the WT trachea, these abnormal mitochondria represented less than 20% of total mitochondria per cell. However, some cells had most of their mitochondria showing this abnormal morphology. We quantified these abnormal mitochondria in a large number of cells from all groups to identify whether or not they were associated with the Aldh2 genotype. We set a threshold of 80% abnormal mitochondria per cell to count a cell as a “cell with abnormal mitochondria”. Interestingly, we observed a moderate increase in the percentage of these abnormal mitochondria per cell in the club and basal cells of the Aldh2*2 Tg and aged WT mice, but this increase was not as high as our set threshold of 80%. Surprisingly, many of the club and basal cells in Aldh2 ^−/− mice showed these abnormal mitochondria beyond the 80% threshold and the number of these cells was strikingly high (Fig. 4d) (Table 2).

The finding of a considerable number of morphologically abnormal mitochondria in the tracheal cells of Aldh2 ^−/− and, to a lesser extent, in Aldh2*2 Tg and WT old mice prompted us to conduct several mitochondrial functional assessments in these mice. We first assessed the amount of mitochondria in the tracheal and lung cells in the different mouse strains by immunostaining for the translocase of inner mitochondrial membrane, Tim23. We then examined the quantity of “functional” mitochondria using two different MitoTracker probes, which stain only the functional mitochondria with active membrane potential. Surprisingly, the tracheal and lung cells in Aldh2 ^−/− mice showed fewer mitochondria (total and functional) than those in the WT, whereas the Aldh2*2 Tg cells contained more mitochondria than the WT (functional) (Fig. 5a-h). Secondly, we measured mitochondrial ROS production using MitoSOX. Compared to WT control, ROS level was higher in tracheal epithelial cells from the Aldh2 ^−/− mice, but not from the Aldh2*2 Tg mice (Fig. 5i-j). Finally, we evaluated the energy production efficiency by exploiting the CellTiter-Glo Assay for measurement of ATP levels, and by measuring the overall contribution of the two major energy production pathways, the oxidative phosphorylation (mitochondrial respiration) pathway and glycolysis, in cellular respiration using the extracellular flux bioenergetics analyzer. No significant difference was observed in the ATP levels of the tracheal or lung cells among the groups (ATP levels in the Aldh2*2 Tg and Aldh2 ^−/− lungs were 90% and 103% and in their MTEC were 97% and 94% of that in the cells of WT mice, respectively). Comparing the OCR and ECAR in WT and Aldh2 ^−/− adult mice, MTECs showed no significant difference (Fig. 5k-l).

Next, we wanted to test whether the histological changes observed in the tracheal basal cells and the abnormalities in mitochondrial morphology and function that were detected in the tracheal basal and club cells of the Aldh2*2 Tg and Aldh2 ^−/− mice negatively influenced the ability of basal cells to act as stem cells in vitro. We therefore isolated MTECs from all groups and compared them using the 3D stem cell colony forming assay [29]. Parallel culture wells were treated with vehicle, H₂O₂ (as a source of ROS and oxidative stress), and/or the Aldh2 agonist, Alda-1. ROS treatment resulted in CFE reduction, while Alda-1 increased the CFE compared to that in the vehicle- and the ROS-treated wells (Fig. 6a-f). However, no significant difference in CFE was detected among the three groups of mice at the baseline or in response to ROS and/or Alda-1 treatments (Fig. 6g). Because fibroblasts from individuals with the ALDH2*2 polymorphism are known to be more vulnerable to injury than WT fibroblasts [39] and because fibroblasts are an essential component of the lung stem cell niche [29], we also investigated whether fibroblasts isolated from the lungs of WT, Aldh2*2 Tg, and Aldh2 ^−/− mice have a different influence when co-cultured with WT tracheal basal stem cells in the colony forming assay. Co-culture with fibroblasts increased the CFE of WT tracheal basal stem cells, but no significant difference was observed among the three fibroblasts groups (Fig. 6h-l). Analysis of the differentiation profile in the grown colonies also revealed no significant difference.

Because the disturbed Aldh2 function in our mouse model revealed some functional effects on mitochondria, but not on stem cells in naïve mice, we wanted to compare them when they were exposed to some clinically-relevant injuries. We exposed the mice to chronic injury by CS, to acute injuries by H1N1 influenza viral infection, and to chemical injury using polidocanol.

Aldh2*2 Tg mice and their WT littermates were exposed to CS for 16 weeks. Failure to gain weight compared to sham smoked controls is an established sign of the efficiency of the smoking protocol [40]. Indeed, both mice groups did not gain weight over the 16 weeks whereas the non-smoking control mice gained weight (Fig. 7a). All lungs were examined by micro-CT scanning followed by euthanasia and the lungs were collected and embedded in paraffin, to examine for digital and histological signs of lung damage and emphysema. Digital readings of the low attenuation area (LAA)%, CT values, and lung volumes were compared between the WT and Aldh2*2 Tg mice as previously described [31].

Although no change was detected in LAA%, the lung volume increased in WT mice (Fig. 7b) and mean linear intercepts (Lm) showed a significant increase in WT mice (Fig. 7c). These data suggest that mild emphysema developed at 16 weeks of CS exposure in WT mice as observed in the previous report [31]. On the other hand, we found that CS exposure neither increased the lung volume nor Lm in Aldh2*2 Tg mice (Fig. 7b-c) suggesting that the Aldh2*2 Tg mice are resistant to emphysema development in response to CS exposure.

As several mechanisms have been identified to mediate the pathogenesis of CS-induced emphysema, we examined several potential culprits. Mice without neutrophil elastase or IL-1 receptor did not develop CS-induced emphysema [41, 42] while mice with impaired expression of antioxidant genes developed more severe emphysema [43]. We used immunostaining to examine both naïve and CS-exposed WT and Aldh2*2 Tg mice for neutrophil infiltration, IL-1β expression and expression of the anti-oxidant Prdx1. In both WT and Aldh2*2 Tg mice, CS upregulated IL-1β and increased the number of neutrophils in the lungs. However, there were no significant differences between WT and Aldh2*2 Tg mice before or after CS-exposure (data not shown). Interestingly, naïve Aldh2*2 Tg showed higher expression of Prdx1, especially in airways. However, CS similarly induced Prdx1 expression in both WT and Aldh2*2 Tg mice (Fig. 7d-g).

WT and Aldh2 ^−/− mice were infected with H1N1 influenza through the nose and lungs were collected after 8 days. There was no difference in survival or body weight loss. All lungs were examined histologically, and extensive inflammatory cell infiltration and pneumonia were obvious in all mice with no significant difference between the WT and Aldh2 ^−/− (Fig. 8a-b). However, when we quantified the epithelial cell types by immunostaining, an average of only 70.6% of the lung area in WT mice, but 80.6% in Aldh2 ^−/− mice, showed complete loss of AT-II cells (Fig. 8c-d). Similarly, loss of club cells in the airways was more extensive in Aldh2 ^−/− mice (58%) than in WT mice (51.9%) (Fig. 8e-f). Viral load was confirmed using an immunostaining for hemagglutinin, a class I viral fusion protein from the influenza virus’ envelope proteins and we found no significant difference between WT and Aldh2 ^−/− mice (Fig. 8g-h).

We also optimized the polidocanol injury protocol as described in the methods section. We found that administration of 10 μL of 2% polidocanol using tracheal intubation through a neck incision under anesthesia by isoflurane inhalation with continuous O₂ resulted in the best survival and consistency. We administered polidocanol to the WT, Aldh2*2 Tg, and Aldh2 ^−/− mice, and then collected samples at 12 h, 24 h, 48 h, day 5, and day 7 (n = 3–4/time point/genotype) for histological analysis. At 12 and 24 h, most epithelial cells had sloughed, leaving only few basal cells attached to the basement membrane (Fig. 9a-b). At 48 h, the surviving basal cells were actively proliferating as indicated by their positive staining for PCNA (proliferating cell nuclear antigen) (Fig. 9c). At day 5 and more-clearly at day 7, the epithelium was returning to the pseudo-stratified arrangement with ciliated and club cells differentiating from the proliferating basal cells (Fig. 9d-e). There was no significant difference among the groups at any time points. Although the Aldh2*2 Tg and Aldh2 ^−/− mice had fewer tracheal epithelial basal cells than the WT mice in the steady state, the function of their basal cells in repairing the epithelia after polidocanol injury was as efficient as that in WT mice on day 7.