Antifibrotic effect of lung-resident progenitor cells with high aldehyde dehydrogenase activity

Background Aldehyde dehydrogenase (ALDH) is highly expressed in stem/progenitor cells in various tissues, and cell populations with high ALDH activity (ALDHbr) are associated with tissue repair. However, little is known about lung-resident ALDHbr. This study was performed to clarify the characteristics of lung-resident ALDHbr cells and to evaluate their possible use as a tool for cell therapy using a mouse model of bleomycin-induced pulmonary fibrosis. Methods The characteristics of lung-resident/nonhematopoietic (CD45−) ALDHbr cells were assessed in control C57BL/6 mice. The kinetics and the potential usage of CD45−/ALDHbr for cell therapy were investigated in bleomycin-induced pulmonary fibrosis. Localization of transferred CD45−/ALDHbr cells was determined using mCherry-expressing mice as donors. The effects of aging on ALDH expression were also assessed using aged mice. Results Lung CD45−/ALDHbr showed higher proliferative and colony-forming potential than cell populations with low ALDH activity. The CD45−/ALDHbr cell population, and especially its CD45−/ALDHbr/PDGFRα+ subpopulation, was significantly reduced in the lung during bleomycin-induced pulmonary fibrosis. Furthermore, mRNA expression of ALDH isoforms was significantly reduced in the fibrotic lung. When transferred in vivo into bleomycin-pretreated mice, CD45−/ALDHbr cells reached the site of injury, ameliorated pulmonary fibrosis, recovered the reduced expression of ALDH mRNA, and prolonged survival, which was associated with the upregulation of the retinol-metabolizing pathway and the suppression of profibrotic cytokines. The reduction in CD45−/ALDHbr/PDGFRα+ population was more remarkable in aged mice than in young mice. Conclusions Our results strongly suggest that the lung expression of ALDH and lung-resident CD45−/ALDHbr cells are involved in pulmonary fibrosis. The current study signified the possibility that CD45−/ALDHbr cells could find application as novel and useful cell therapy tools in pulmonary fibrosis treatment. Supplementary Information The online version contains supplementary material available at 10.1186/s13287-021-02549-6.

shown to exhibit pluripotency toward the nonhematopoietic cell lineage, can be isolated from various organs, including the bone marrow, adipose tissue, skeletal muscle, and the umbilical cord [1]. Bone marrow-derived MSCs, isolated from the most orthodox cell source of MSCs [2,3], have been shown to have immunomodulatory effects such as the inhibition of the proliferation of T-cells through secretion of anti-inflammatory cytokines and growth factors [4]. In a mouse model of bleomycin (BLM)-induced lung injury, administration of bone marrow-derived MSCs was reported to improve lung injury by exerting an anti-inflammatory effect [5]. With respect to lung resident stem cells, the Sca1 + /CD45 − /CD31 − cell population has been identified as lung tissue stem cells capable of differentiating into endothelial and lung epithelial cells in vitro. Moreover, when transferred into an elastase-induced lung injury mouse model, this population was demonstrated to significantly improve the survival rate and reverse lung damage [6]. Lung Hoechst 33342 dim side population (SP) cells are adult stem cells, which have also been identified to exhibit mesenchymal and epithelial potential [7]. Among the SP cells, the CD45 − /CD31 − fraction has been reported to have the characteristics of lung resident MSCs, due to their ability to differentiate into smooth muscle, bone, fat, and cartilage [8,9]. Furthermore, the number of lung resident SP cells was shown to be significantly reduced in mice with BLM-induced lung injury, and this reduction was correlated with the pathology of the lung injury. When administered intravenously into the lung, lung SP cell therapy was shown to reduce BLM-induced pulmonary fibrosis and pulmonary arterial hypertension [10]. These results suggest the existence of tissue-specific MSCs in the lung and their involvement in lung injury.
Aldehyde dehydrogenases (ALDH) are a group of enzymes that catalyze the oxidation of aldehydes to carboxylic acids, with 19 different isoforms in humans [11]. A cell population with high ALDH activity, called ALDH bright cells (ALDH br ), is associated with the stemness of various normal tissues and is involved in tissue repair [12]. Moreover, ALDH br isolated from the human bone marrow, reported to have a higher colony-forming capacity when compared to a cell population with low ALDH activity (ALDH dim ) [13], was shown to be a progenitor population for epithelial, endothelial, and mesenchymal lineages [14]. When administered in a mouse model of myocardial infarction, ALDH br collected from the human umbilical cord blood was demonstrated to enhance angiogenesis in the ischemic heart [15]. Given these findings, the existence of lung resident ALDH br and its contribution to tissue repair were speculated; however, little is known about lung resident ALDH br . The objectives of this study were to clarify the characteristics of lung-resident ALDH br and to evaluate its possible use as a tool for cell therapy in a mouse model of BLM-induced pulmonary fibrosis.

Animals and BLM-induced pulmonary fibrosis
This study, aimed at elucidating the characteristics of lung-resident ALDH br and exploring its usage in cell therapy, was performed in accordance with the protocols approved by the Animal Ethics Committee of Hiroshima University (A19-122 and 28-29-2). In this study, pulmonary fibrosis was induced as previously described [16] in C57BL/6J mice (6-8-week-old young female mice and 52 week old aged female mice) which were purchased from Charles River Laboratories Japan (Yokohama, Japan). The mice were maintained in a specific pathogen-free environment and randomly assigned to BLM or control groups. In experiments performed to confirm the localization of transferred cells, C57BL/6-Gt (ROSA)26Sor < tm1.1 (H2B-mcherry) Osb > heterozygotic mice (mCherry mouse, BRC No. RBRC06036, RIKEN, Tokyo, Japan) [17] systemically expressing the mCherry protein in their nuclei were used as a donor population. On day 0, after intraperitoneal injection of mixed anesthesia with medetomidine hydrochloride (0.3 mg/kg body weight; Kyoritsu Seiyaku, Tokyo, Japan), midazolam (4 mg/kg body weight, Sandoz K.K., Tokyo, Japan), and butorphanol tartrate (5 mg/kg body weight, Meiji Seika Pharma, Tokyo, Japan), pulmonary fibrosis was induced by endotracheal injection of BLM (2 mg/kg of body weight, Nippon Kayaku, Tokyo, Japan). Control mice received the same amount (2 mL/kg body weight) of phosphate-buffered saline (PBS, Nacalai Tesque, Kyoto, Japan) alone. For survival analysis, a higher dose of BLM (5 mg/kg) was used. At 7 and 14 days after BLM administration, both lungs were removed from each animal and the lung tissue was assessed for hydroxyproline, and mRNA expression and subjected to flow cytometry and histological analysis.

Cell isolation
The lungs were removed and minced in 1-mL Roswell Park Memorial Institute 1640 medium (Thermo Fisher Scientific, Waltham, MA, USA) supplemented with collagenase A (1 mg/mL, Roche, Basel, Switzerland), and incubated at 37 °C for 30 min. Following lysis of red blood cells with ACK Lysing Buffer (Life Technologies, Grand Island, NY, USA), the cells were resuspended in 2 mL of PBS containing 0.5% bovine serum albumin (Sigma-Aldrich, St. Louis, MO, USA) and 2 mM ethylenediaminetetraacetic acid (Sigma-Aldrich), and cell counting was performed.

Antibodies and ALDH staining
Antibodies (all purchased from BioLegend, San Diego, CA, USA) used for flow cytometry and cell sorting are shown in Additional file 1. After staining for cell surface proteins using the aforementioned antibodies, ALDH activity was expressed as fluorescent intensity using the ALDEFLUOR ™ Kit (STEMCELL Technologies Inc., Vancouver, Canada) according to the manufacturer's protocol, as previously reported [18]. A separate tube containing 5 µL of diethylaminobenzaldehyde (DEAB, provided in the ALDEFLUOR ™ Kit), a specific inhibitor of ALDH, was prepared to determine ALDH br gating.

Flow cytometry and cell sorting
Flow cytometric analysis of lung cells was performed using the following method, referring to a previous report [20]. Flow cytometry and cell sorting were performed using the FACS Aria II system (BD Biosciences, San Jose, CA, USA) and LSRFortessa X-20 (BD Biosciences). Data were analyzed using the FACS Diva (BD Biosciences) and the FlowJo (version 10.7.1, BD Biosciences) software. For the isolation of ALDH br , unnecessary cell populations were pre-depleted using magnetic cell sorting (MACS) cell separation using a Stem Cell Pre-Enrichment kit (Miltenyi Biotec, Bergisch Gladbach, Germany) prior to FACS according to the manufacturer's protocol. Cell sorting from mCherry-expressing donor mice and analysis of injected donor mCherry + cells was performed using the SORP Aria (BD Biosciences) and LSRFortessa X-20 (BD Biosciences) systems, respectively.

Cell culture and colony-forming assay
Sorted cells were seeded into 96-well plates at a density of 5-10 × 10 3 cells/well and cultured in Dulbecco's modified Eagle medium (DMEM, Thermo Fisher Scientific) and 10% fetal bovine serum (FBS, Sigma-Aldrich) supplemented with or without 20 ng/mL epidermal growth factor (EGF, BioLegend) or 20 ng/mL fibroblast growth factor-2 (FGF2, BioLegend) or both. The medium was changed every 3-4 days. For colony formation, 5.0 × 10 3 cells were seeded into 6-well plates using MethoCult (STEMCELL Technologies Inc.). Consecutively, 2 to 3 weeks after the start of culture, the number of proliferated colonies was counted.

Cell viability assay
Cells were seeded into 96-well plates at a density of 5.0 × 10 3 cells/well and the medium was changed every 3-4 days. After 3-4 weeks from the start of the culture, cell proliferation was evaluated using Cell Counting Kit-8 (Dojindo, Kumamoto, Japan).

Cell transfer to recipient mouse
Sorted 1.0 × 10 5 CD45 − /ALDH br and CD45 − /ALDHdim cells were dissolved in 100 µL PBS and administered intravenously via the tail vein to recipient BLM-pretreated mice on day 2 (2 days after treatment with BLM). To confirm the localization of transferred cells, 5.0 × 10 4 mCherry + CD45 − /ALDH br and CD45 − /ALDH dim cells were administered intravenously into recipient BLM-pretreated C57BL/6 mice on day 2. On the following day and 5 days after the injection (on days 3 and 7), the recipient mice were sacrificed, and lung samples were subjected to flow cytometry and histology analyses.

PCR and agarose gel electrophoresis
The sorted cells and the excised lungs were homogenized using 1 mL TRIzol reagent (Life Technologies) and total RNA was extracted using the RNeasy Mini Kit (QIAGEN, Venlo, Netherlands). The extracted RNA was reverse transcribed into cDNA using the High Capacity RNA-to-cDNA Kit (Applied Biosystems, Foster City, CA, USA). Real-time quantitative PCR was performed using the Applied Biosystems 7500 Fast Real-Time PCR System (Applied Biosystems) and the TaqMan Gene Expression Assays (Applied Biosystems) as previously described [16]. The expression of Actb (β-actin, Mm02619580_g1; Applied Biosystems) was used as an endogenous control. The TaqMan Gene Expression Assays were used as shown in Additional file 2. To distinguish the mCherry-heterozygotic mice the from wildtype mice, mouse-tail DNA was extracted using the DNeasy Blood and Tissue Kit (QIAGEN). The extracted DNA was subjected to PCR using the primers shown in Additional file 2. PCR conditions were as follows: 120 s at 94 °C, 10 s at 98 °C, 30 s at 60 °C, 120 s at 68 °C, repeated for 30 cycles. Amplified products were stained with SAFELOOK ™ (Fujifilm Wako Junyaku, Osaka, Japan), and bands were confirmed using electrophoresis on a 1% agarose gel.

Statistical analyses
All experiments were performed 2 or 3 times and the representative data are shown as median ± interquartile range except for mRNA data, which is shown as mean ± SEM to ensure the visibility of the graph. The Kruskal-Wallis test for median values was used to assess the statistical significance between groups. Correlation coefficients for parameters were calculated using the Spearman's rank correlation coefficient analysis. Kaplan-Meier analysis and log-rank test were used for survival analysis. A P value < 0.05 was considered significant. All statistical analyses were performed using JMP Pro 14 (SAS Institute Inc., Cary, NC, USA).

Detection of ALDH br in mouse lung
Following the determination of the appropriate ALDH br gating using ALDEFLUOR staining with the DEAB ALDH inhibitor, we observed a rare ALDH br population in the whole lung of mice (Fig. 1A). When we divided the whole lung cells into CD45 + hematopoietic cells and CD45 − nonhematopoietic cells (Fig. 1B), we noted that both fractions contained ALDH br (Fig. 1C, D). To assess lung resident ALDH br , we focused on the nonhematopoietic CD45 − /ALDH br fraction. Analysis of these nonhematopoietic cells, that is, the lung resident CD45 − / ALDH br fraction, revealed that this fraction was further divided into mesenchymal (platelet-derived growth factor receptor α positive, PDGFRα + ) and epithelial (epithelial cell adhesion molecule positive, EpCAM + ) phenotypes (Fig. 1E).

Characteristics of lung CD45 − /ALDH br
To determine the characteristics of lung CD45 − /ALD-H br , we collected CD45 − /ALDH br (n = 3) and CD45 − / ALDH dim (n = 3) cells using FACS. As shown in Additional file 3, pre-depletion of unnecessary cell populations prior to FACS resulted in the enrichment of the CD45 − /ALDH br fraction. To confirm if sorted CD45 − / ALDH br cells truly expressed high levels of ALDH mRNA and to determine the isoforms of ALDH that were mainly expressed in CD45 − /ALDH br cells, we performed realtime quantitative PCR. Our results showed that the levels of mRNA expression of ALDH1a, ALDH2, ALDH3a1, ALDH4a1, ALDH7a1, and ALDH18a were significantly higher in the CD45 − /ALDH br than in the CD45 − /ALD-H dim cells ( Fig. 2A, P = 0.049 for ALDH1a1, P = 0.049 for ALDH1a2, P = 0.049 for ALDH1a3, P = 0.037 for ALDH1a7, P = 0.049 for ALDH2, P = 0.046 for ALDH3a1, P = 0.049 for ALDH4a1, P = 0.037 for ALDH7a1, and P = 0.049 for ALDH18a). We further observed that when both cell populations were cultured, the CD45 − /ALD-H br population showed higher proliferative ability than the CD45 − /ALDH dim population (Fig. 2B, C, P = 0.009 between CD45 − and CD45 − /ALDH br and P = 0.009 between CD45 − /ALDH dim and CD45 − /ALDH br ). To examine whether CD45 − /ALDH br cells maintained a high ALDH activity in culture, sorted CD45 − /ALDH br cells were further cultured, harvested, and reexamined for ALDH activity. As shown in Fig. 3A, most proliferated cells were ALDH dim , with ALDH br cells accounting for approximately 5% of the total proliferative cells.
Next, we examined the colony-forming ability of CD45 − /ALDH br cells using a colony-forming assay. We found that CD45 − /ALDH br cells formed larger (Fig. 2D) and higher number (Fig. 2E) of colonies than the CD45 − / ALDH dim cells. Although a similar pattern of colony formation was observed for CD45 − , the size and the number of colonies were relatively small, suggesting that the colony-forming ability of the CD45 − population depended to a large extent on the CD45 − /ALDH br cells.
As the CD45 − /ALDH br population seemed to be a heterogeneous cell population and ALDH br is associated with stemness in various tissues, we evaluated the expression of surface antigens associated with the mesenchymal cells, fibroblasts, and the stem cells in the CD45 − / ALDH br population. As shown in Additional file 4, not all CD45 − /ALDH br cells expressed the representative markers of bone marrow-derived MSCs (CD44, CD73, CD90, and CD105). It was notable that the stage-specific embryonic antigen-4 (SSEA4) stem cell marker was solely expressed in CD45 − /ALDH br cells in the mouse lung.
Next, we examined the expression of ALDH in primary cultured lung fibroblasts (Fig. 3A). These primary cultured lung fibroblasts obtained from BLM-untreated wild-type C57BL/6 mice were shown to frequently express PDGFRα, but not CD45 or EpCAM, suggesting that these cells were truly fibroblasts (Fig. 3B). As shown in Fig. 3C, both fluorescent microscopy and flow cytometry revealed that the percentage of ALDH br cells in primary cultured lung fibroblasts was approximately 5%. Similarly, we noted that the percentage of ALDH br cells in fibroblast cell lines was also less than 5% (Fig. 3D).

Kinetics of CD45 − /ALDH br in BLM-induced pulmonary fibrosis
To investigate the kinetics of CD45 − /ALDH br in fibrotic lungs, we used endotracheal administration of BLM (2 mg/kg body weight) to generate a mouse model of pulmonary fibrosis. We found that the levels of hydroxyproline were significantly elevated in the BLM group (n = 9) 14 days after BLM administration (Fig. 4A, P = 0.005) compared with the PBS group (n = 4). On days 7 and 14, the percentage of total ALDH br cells in the lung was significantly elevated compared with that on day 0 (Fig. 4B), whereas the percentage of CD45 − /ALDH br cells and CD45 − /ALDH br /PDGFRα + cells, but not CD45 − /ALD-H br /EpCAM + cells, were significantly decreased (Fig. 4B). Among the ALDH br populations in the lung obtained on  (Fig. 4C). Real-time quantitative PCR analysis revealed that the mRNA expression of ALDH1a1, ALDH1a7, ALDH1l1, ALDH2, ALDH3a1, ALDH4a1, and ALDH7a1 was significantly lower in the fibrotic lung obtained on day 14 (Fig. 4D, P = 0.006 for ALDH1a1, P = 0.006 for ALDH1a7, P = 0.014 for ALDH1l1, P = 0.006 for ALDH2, P = 0.006 for ALDH3a1, P = 0.006 for ALDH4a1, and P = 0.009 for ALDH7a1). Consistent with the reduced number of CD45 − /ALDH br cells and the reduced expression of ALDH1a1 mRNA in the fibrotic lung, the expression of ALDH1a1 was reduced throughout the alveolar epithelia, especially in the areas of fibrosis, as demonstrated using immunostaining (Additional file 5).
The decrease in number of cells observed in CD45 − / ALDH br cells during BLM treatment was a feature observed in lung SP cells as well [10]. Therefore, we investigated the possibility of an overlap between ALDH br and lung SP cells. After Hoechst staining, ALDH staining was performed, followed by flow cytometry, which revealed that CD45 − /ALDH br population is completely different from CD45 − lung SP cells (Additional file 6).

Effect of CD45 − /ALDH br cell therapy on BLM-induced pulmonary fibrosis
In the preceding experiments, we presumed that CD45 − / ALDH br cells were depleted during pulmonary fibrosis; therefore, we assessed the possible usage of CD45 − / ALDH br cells in cell therapy for BLM-induced pulmonary fibrosis. Our results showed that both the levels of hydroxyproline (Fig. 5A, P = 0.023) and the degree of tissue fibrosis (Fig. 5B for HE staining, Additional file 7 for Masson's trichrome staining) in the lung obtained on day 14 were significantly lower in the CD45 − /ALD-H br i.v. group (n = 4) than in the CD45 − /ALDH dim i.v. group (n = 7). In the CD45 − /ALDH br i.v. group (n = 7-8), the mRNA expression of interleukin 6 (IL6) and transforming growth factor β1 (TGFb1) genes in lung tissues obtained on day 7 was significantly suppressed compared with the CD45 − /ALDH dim i.v. group (n = 9-10) (Fig. 5C, P = 0.042 for IL6, and P = 0.013 for TGFb1). Interestingly, we noted that the percentage of CD45 − /ALDHbr/ PDGFRα + cells, which was lowered, was recovered in the CD45 − /ALDH br i.v. group (n = 4) in the lung obtained on day 14 (Fig. 5D). Furthermore, the expression levels of ALDH1a1 and ALDH4a1 mRNAs, which were significantly reduced after treatment with BLM (Fig. 4D), were also recovered in the CD45 − /ALDH br i.v. group (n = 4) on day 14 (Fig. 5E, P = 0.008 for ALDH1a1, and P = 0.038 for ALDH4a1).
In addition, we assessed the effect of CD45 − /ALDH br cell therapy on survival using BLM-induced pulmonary fibrosis with a higher dose of BLM (5 mg/kg body weight). We observed that the higher dose of BLM led to approximately 80% mortality on day 14 in both the CD45 − /ALDH dim (n = 8) and the PBS i.v. groups (n = 8), whereas, surprisingly, no death was observed in mice that received CD45 − /ALDH br cell therapy (n = 5) (Fig. 5H).

Detection of transferred donor CD45 − /ALDH br in the recipient lung
To distinguish and trace the injected donor CD45 − / ALDH br cells in the lungs of recipient mice, mCherry knock-in mice were used as donors. After mCherry heterozygosity was confirmed using tail PCR (Additional file 8), donor CD45 − /ALDH br or CD45 − /ALDH dim cells were sorted from these mCherry-expressing mice using FACS (Fig. 6A) and transferred into wild-type C57BL/6 recipients pretreated with BLM. We observed that flow cytometry could detect donor mCherry-positive CD45 − / ALDH br (Fig. 6B) and CD45 − /ALDH br /PDGFRα + (Fig. 6C) cells in the recipient lungs more frequently in the CD45 − /ALDH br i.v. group than in the CD45 − /ALD-H dim i.v. group. Appropriate mCherry immunostaining conditions were determined using appropriate positive and negative controls (Fig. 6D), and we noted that mCherry-positive CD45 − /ALDH br and CD45 − / ADLH br /PDGFRα + cells were also found histologically in the recipient lung-transferred CD45 − /ALDH br (Fig. 6E).

Effects of aging on ALDH activity
Finally, we examined the role of aging on the CD45 − / ALDH br population. As shown in Fig. 7A, the percentages of whole CD45 − /ALDH br cell population and that of its CD45 − /ALDH br /PDGFRα + subgroup in the lung were not significantly different between aged and young mice that were not treated with BLM (on day 0). On the contrary, the percentage of CD45 − /ALDH br /PDGFRα + cells, but not CD45 − /ALDH br cells, in the lung obtained 7 days after treatment with BLM was significantly decreased in aged mice (n = 4-6) compared with that in young mice (n = 4-6). In a similar fashion, the percentage of ALDH br cells in cultured PDGFRα-predominant (as shown in Fig. 3B) primary lung fibroblasts obtained from the lung 7 days after treatment with BLM was significantly . D Histological analysis of mCherry immunostaining in lungs obtained from mCherry-expressing mice (positive control) and wild-type C57BL/6 mice (negative control). E Representative images of mCherry immunostaining in lungs obtained from recipient BLM-treated wild-type C57BL/6 mice transferred with mCherry + /CD45 − /ALDH br cells decreased in aged mice (n = 5-10) compared with young mice (n = 5-10) (Fig. 7B).

Discussion
The present study identified and characterized the nonhematopoietic/lung resident ALDH br cell populations in the mouse lung. The lung CD45 − /ALDH br population and, the CD45 − /ALDH br /PDGFRα + subpopulation are cell populations with high proliferative capacity. These population significantly reduced in pulmonary fibrosis. The high levels of expression of ALDH observed in CD45 − /ALDH br cells was mainly attributed to the ALDH1a subfamily, also known as RALDH, which was significantly reduced in BLM-treated lungs. When used as a tool for cell therapy, transferred CD45 − /ALDH br cells reached the site of lung injury and ameliorated BLM-induced pulmonary fibrosis. Thus, this study demonstrated CD45 − /ALDH br cells as a novel lung-resident stem cell population and suggested their potential therapeutic use in pulmonary fibrosis. Although ALDH br cells with stem cell properties have been detected in various normal tissues, including the bone marrow [14,23], umbilical cord blood [24,25], mammary glands [26,27], heart [28], and adipose tissue [29], little is known about lung-resident ALDH br cells. A study showed that isolated murine airway basal and submucosal gland duct ALDH br cells exhibited stem cell properties in normal/healthy lungs [30]. No previous study has investigated the significance of lung-resident ALDH br cells in respiratory diseases, such as pulmonary fibrosis. In the current study, lung-resident CD45 − /ALD-H br were rare and heterogeneous population with epithelial and mesenchymal lineages. The percentages of ALDH br in both primary cultured lung fibroblasts and fibroblast cell lines were low at approximately 5%, and hence, we assumed that ALDH br cells lost their activity during differentiation and proliferation, consistent with the findings of a previous report [24]. Similar to the CD45 − /ALDH br /PDGFRα + population in the current study, CD45 − lung SP cells have been reported to express mesenchymal markers and exhibit MSC properties [8], and have been shown to be decreased in BLM-induced pulmonary fibrosis [10]. However, in our study, we found that CD45 − /ALDH br is a novel population that is completely different from lung SP cells (CD45 − /CD31 − /Hoechst dim ). Therefore, it is reasonable that the expression of the surface markers of MSCs found in lung CD45 − / ALDH br cells differed from that in the SP cells. Instead, SSEA4, a marker for mesenchymal progenitors [31], was demonstrated to be solely expressed on CD45 − /ALDH br cells in the mouse lung. These results suggest that the CD45 − /ALDH br population might contain mesenchymal progenitors and CD45 − /ALDH br /PDGFR + cells maintained the ability to differentiate into the mesenchymal lineage.
During BLM-induced pulmonary fibrosis, we observed a downregulation in the expression of a broad spectrum of ALDH mRNAs in lung tissues. We also found that transferred CD45 − /ALDH br cells ameliorated BLMinduced pulmonary fibrosis by suppressing IL-6 and  Fig. 7 Effect of aging on ALDH activity. A Percentages of CD45 − /ALDH br and CD45 − /ALDH br /PDGFRα + cells in total lung cells during BLM-induced pulmonary fibrosis in young and aged mice (n = 4-6). *P < 0.05, **P < 0.01. ns, not significant. B ALDH activity in primary cultured lung fibroblasts obtained from young and aged mice (n = 5-10) before and 7 days after treatment with BLM. **P < 0.01. ns, not significant TGF-β. As an evidence, intravenously administered CD45 − /ALDH br cells were shown to reach the site of lung injury using mCherry-expressing mice as donors.
Additionally, these lung-protective effects of transferred CD45 − /ALDH br were accompanied by a recovery in the levels of ALDH, which had been decreased during fibrosis, suggesting that ALDH was involved in the mechanism of pulmonary fibrosis. Although little is known about the association of ALDH isoforms with lung diseases, ALDH1a1 and ALDH3a1 have been reported to be expressed in the human airway epithelium [32]. Jang and coworkers reported that the expression of ALDH3a1 was markedly increased in human airway epithelial cells exposed to cigarette smoke extract and that ALDH3a1 exerted protective action against smoking-induced airway epithelial damage [33]. In the current study, the expression of both ALDH1a1 and ALDH4a1 were upregulated in CD45 − /ALDH br cells and downregulated in the fibrotic lung after BLM administration, paralleling the reduction in the number of CD45 − /ALDH br cells. Likewise, intravenous administration of CD45 − /ALD-H br cells was shown to significantly recover the expression of ALDH1a1 and ALDH4a1 in the fibrotic lung. Therefore, we speculated that mesenchymal ALDH1a1 and ALDH4a1 might protect against BLM-induced pulmonary fibrosis. Indeed among ALDH family members, RALDHs (ALDH1a1, ALDH1a2, and ALDH1a3) catalyze the conversion of retinol to ATRA [22], supporting the self-renewal and cell differentiation of stem cells [34]. Several lines of evidence have suggested that ATRA exerted protective action against radiation pneumonitis and BLM-induced lung injury in mice through anti-inflammatory effects by activating protein kinase C δ (PKC-δ), inhibiting mitogen-activated protein kinase P38 α (p38MAPK) and nuclear factor kappa-light-chainenhancer of activated B-cells (NF-kB), and suppressing the production of IL-6 and TGF-β [35][36][37][38]. In the current study, we observed the upregulation of retinol-metabolizing pathway molecules, recovery of the expression of RALDH, and suppressed expression of IL-6 and TGF-β in BLM-induced pulmonary fibrosis treated with CD45 − / ALDH br cell therapy. On the other hand, the significance of ALDH4a1 in lung injury is currently unknown and further investigation is required.
In the fibrotic lung, after BLM administration, we observed a reduction in the number of cells in the CD45 − /ALDH br population, especially of its CD45 − / ALDH br /PDGFRα + subpopulation. This reduction was more remarkably observed in aged mice than in young mice. These results suggested that aging led to a decrease in the number of ALDH br cells in the lungs, especially in the lung PDGFRα + fibroblasts. As fibrotic lung diseases, especially idiopathic pulmonary fibrosis (IPF), commonly occur in the elderly [39] and stem cell senescence is one of the suggested causes of IPF [40], it is speculated that the decreased number of ALDH br cells in the lungs might accelerate fibrotic lung diseases in the elderly.
The limitation of this study is the difficulty in collecting sufficient number of cells. Because of the infrequency of existence of ALDH br cells, many donor mice lungs were necessary to acquire a sufficient number of ALDH br cells, signifying the challenge in applying the methods and the results of the present study to human lung diseases. If the collected cells could be proliferated while maintaining ALDH activity, the burden on donor could be minimized. To apply the current results to human translational studies in the future, development of less invasive methods for collecting ALDH br cells is required.

Conclusions
Our results strongly suggest that the lung expression of ALDH and lung-resident CD45 − /ALDH br are involved in pulmonary fibrosis. (Figure 8 summarizes the findings of the current study.) When administered intravenously, CD45 − /ALDH br ameliorated BLM-induced pulmonary fibrosis, signifying the possibility for CD45 − /ALDH br cells to find application as novel and useful cell therapy tools in pulmonary fibrosis treatment.
Abbreviations ALDH: Aldehyde dehydrogenase; ALDH br : Cell populations with high ALDH activity; ALDH dim : Cell population with low ALDH activity; ATRA : All-trans retinoic acid; BLM: Bleomycin; CRABP: Cellular retinoic acid-binding protein; EpCAM: Epithelial cell adhesion molecule; FACS: Fluorescence activated cell Lung injury (e.g., bleomycin-induced lung injury) triggers reduction of ALDH br cells in the lung, resulting in a suppressed retinol-metabolizing pathway, elevated concentrations of profibrotic cytokines (e.g., IL-6 and TGFβ1), and exacerbation of pulmonary fibrosis. Aging accelerates the injury-induced reduction in ALDH br cells. ALDH br cell therapy restores the impaired antifibrotic effects of ALDH br cells. Solid and dotted arrows indicate promotion and inhibition, respectively