Increased Regeneration Following Stress-Induced Lung Injury in Bleomycin-Treated Chimeric Mice with CD44 Knockout Mesenchymal Cells

CD44, an adhesion-molecule promoting cell-migration, is shown here to increase in stress conditions following bleomycin-induced apoptosis in alveolar epithelial cells (AECs), a main target of lung injury. In vivo, it inhibits tissue regeneration and leads to fibrosis. We show that some AECs survive by the ataxia-telangiectasia mutated kinase/ATM pathway, and undergo a CD44-mediated epithelial-mesenchymal transdifferentiation (EMT) with migratory capacities in vitro, and in vivo. We assessed apoptosis vs. proliferation of AECs following bleomycin, ATM/P53 signaling pathway in AECs, and CD44 involvement in EMT, cell motility and tissue regeneration in vitro and in vivo. Expression of survival genes, CD44, and ATM/p53 pathway was elevated in AECs surviving bleomycin injury, as were the markers of EMT (downregulation of E-cadherin, upregulation of N-cadherin and vimentin, nuclear translocation of β-catenin). Inhibition of CD44 decreased AECs transdifferentiation. Bleomycin-treated chimeric CD44KO-mice had decreased EMT markers, ATM, and mesenchymal cells (α-SMA+) accumulation in lung, increased surfactant-b, diminished lung mesenchymal cell motility, and increased lung tissue regenerative capacity following bleomycin injury, as indicated by lung collagen content and semiquantitave morphological index scoring. Thus, AECs surviving lung injury are plastic and undergo ATM-mediated, CD44-dependent transdifferentiation, preventing tissue regeneration and promoting fibrosis. Synthetic or natural compounds that downregulate CD44 may improve tissue regeneration following injury.


Introduction
We previously detected bleomycin-induced reactive oxygen species (ROS) resulting in oxidative stress, mitochondrial leakage, DNA damage response (DDR), and apoptosis, in normal alveolar epithelial cells (AECs) [1]. Interestingly, the main inducers of apoptosis can also promote EMT, suggesting that EMT could potentially serve as a pathway for escape from cell death [2]. EMT occurs during embryonic development and organogenesis and represents an obstacle to tissue regeneration following injury. Thus, EMT may also cause diseases such as fibrosis, cancer, rheumatoid arthritis, tissue degradation, and others [3,4]. Many EMT inducers act directly or indirectly by repressing

Survival Genes
A mouse apoptosis pathway gene array kit (GEArray; SuperArray Inc., Bethesda, MD, USA) was used to determine the expression levels of multiple genes involved in epithelial cell apoptosis as we previously detailed [19]. Briefly, cDNA probes were denatured and hybridized with the SuperArray membrane, which was then scanned by the X-ray. A pool of four cDNA spots for each gene was used and digitized using ScanAlyze software (shareware, available at http://eisenlab.org/software.html), and normalized by subtraction of the background average intensity value of three spots containing plasmid DNA (PUC18). The average of two RPL13A spots was used as a positive control and set as the baseline value compared signal to other spots intensity using the GEarray analyzer program.

Data Analysis and Representation
The nonparametric Kruskall-Wallis test was applied to compare variables measured at different time intervals or following different treatments. Multiple pairwise comparisons were performed using the Mann-Whitney non-parametric test with the Bonferroni correction for significance. Two-way ANOVA was used to assess time and treatment effects. All statistical tests were two-tailed, and a p-value of 0.05 or less was considered significant.

A Portion of AECs, Mouse Lung Epithelial (MLE) Cells, Exposed to Bleomycin Resist Apoptosis and Upregulate Cell-Survival Genes
We have previously shown that MLEs' viability decreases and apoptosis increases with growing doses of bleomycin [1]. Here we demonstrate, that despite being exposed to high bleomycin (BLM) concentration (200 mU, 24 h), some MLE cells resist apoptosis ( Figure 1A left panel), note annexin-V (−) cells, indicating non-segregated area of the annexin-V in dot-blot.
To further characterize the subset of MLEs resisting cell death, cell sorting by flow cytometry and expansion in culture (HITES) of annexin-V-negative MLEs (BLM-annexin-V (−) ), was performed ( Figure 1A, right panel).
Concomitantly, annexin-V (−) cells showed the upregulation of survival genes which are known to be expressed in response to oxidative-stress and DNA-damage, such as GADD45, P21 and others, but also of genes involved in apoptosis regulation, such as ATM, IAP's and survivin (Table 1). Concomitantly, annexin-V (−) cells showed the upregulation of survival genes which are known to be expressed in response to oxidative-stress and DNA-damage, such as GADD45, P21 and others, but also of genes involved in apoptosis regulation, such as ATM, IAP's and survivin (Table 1). Table 1. Mouse lung epithelial (MLE) cells exposed to bleomycin upregulate cell-survival genes. Apoptosis superarray membranes spotted with gene specific cDNA. Fragments of 96 apoptosis and survival related genes were with cDNA probes synthesized from three total RNA samples corresponding to cultured Annexin V (−) MLE cell sorted population vs. control/saline-treaded MLE cells (Ctrl). The relative expression levels of the various genes were estimated by comparing signal intensity to RPLA 13A housekeeping genes and then quantitated by densitometry after background subtraction. p < 0.05, n = 3.   Table 1. Mouse lung epithelial (MLE) cells exposed to bleomycin upregulate cell-survival genes. Apoptosis superarray membranes spotted with gene specific cDNA. Fragments of 96 apoptosis and survival related genes were with cDNA probes synthesized from three total RNA samples corresponding to cultured Annexin V (−) MLE cell sorted population vs. control/saline-treaded MLE cells (Ctrl). The relative expression levels of the various genes were estimated by comparing signal intensity to RPLA 13A housekeeping genes and then quantitated by densitometry after background subtraction. p < 0.05, n = 3. ATM is known to function in cell survival [27,28] and to phosphorylate a high number of proteins during double-strand DNA breaks repair. Bleomycin (BLM) induces, ATM in Annexin-V (−) MLEs, (Figure 2A, with graphical presentation, left panel and insert), ATM activation ( Figure 2B) and an array of phosphorylated proteins (fold ratio to saline-treated MLEs) of ATM cascade; Erk, ATM, BRCA1, RAD50, GADD45 and chk1 in BLM-Annexin-V (−) MLE cells ( Figure 2C). Concomitantly we show here increased proliferation in BrdU uptake from a percentage of 33% in saline-treated MLEs (Ctrl) to 44% ( Figure 2D with graphical presentation), in bleomycin-treated MLEs.

BLM-Annexin V (−) AECs Upregulate Ataxia Telangiectasia Mutated (ATM) Protein Expression, Activation and Phosphorylation with other Linked Cell-Survival Proteins, and Proliferate
ATM is known to function in cell survival [27,28] and to phosphorylate a high number of proteins during double-strand DNA breaks repair. Bleomycin (BLM) induces, ATM in Annexin-V (-) MLEs, (Figure 2A, with graphical presentation, left panel and insert), ATM activation ( Figure 2B) and an array of phosphorylated proteins (fold ratio to saline-treated MLEs) of ATM cascade; Erk, ATM, BRCA1, RAD50, GADD45 and chk1 in BLM-Annexin-V (-) MLE cells ( Figure 2C). Concomitantly we show here increased proliferation in BrdU uptake from a percentage of 33% in saline-treated MLEs (Ctrl) to 44% ( Figure 2D with graphical presentation), in bleomycin-treated MLEs.

Bleomycin-Injured Chimeric-CD44 KO Mice Decrease EMT Markers in Their Lungs, and Isolated α-SMA + Cells Decrease their Motility Phenotype In Vitro
We have previously shown [21] that young mice injured with bleomycin (0.03-0.05 mU/mouse) evolve pulmonary fibrosis by day 14 but as opposed to aged mice [10], we show that lung tissue regenerates by day 56 [21]. Here we found CD44 overexpression at fibrosis apex (day 14) in WT C57BL/6 mice compared to control saline-treated mice lung fibroblasts (58% vs. 8%, correspondingly)

Bleomycin-Injured Chimeric-CD44 KO Mice Decrease EMT Markers in Their Lungs, and Isolated α-SMA + Cells Decrease their Motility Phenotype In Vitro
We have previously shown [21] that young mice injured with bleomycin (0.03-0.05 mU/mouse) evolve pulmonary fibrosis by day 14 but as opposed to aged mice [10], we show that lung tissue regenerates by day 56 [21]. Here we found CD44 overexpression at fibrosis apex (day 14) in WT C57BL/6 mice compared to control saline-treated mice lung fibroblasts (58% vs. 8%, correspondingly) ( Figure 4A, BLM-day 14 vs. Ctrl). CD44 levels returned to almost normal levels (17.5%) at regeneration ( Figure 5A, BLM-day 56 vs. 14 and Ctrl; see also graphical presentation ( Figure 4A, BLM-day 56 vs. 14 and Ctrl). ( Figure 4A, BLM-day 14 vs. Ctrl). CD44 levels returned to almost normal levels (17.5%) at regeneration ( Figure 5A, BLM-day 56 vs. 14 and Ctrl; see also graphical presentation ( Figure 4A, BLM-day 56 vs. 14 and Ctrl). Chimeric CD44 KO mouse model, with CD44 KO mesenchymal cells and WT hematopoietic cells, generated by us as detailed previously [19] allows the assessment of CD44's role, specifically in mesenchymal cells in tissue regeneration in vivo, thus excluding the possible contribution of CD44negative immune cells. We exposed chimeric-CD44 KO -and control WT-mice to 0.05 mU bleomycin ( Figure 4B) or saline ( Figure 4B, inserts), and analyzed, at day 14, EMT in IHC, ( Figure 4B vs. inserts). SP-b expression was significantly reduced in bleomycin chimeric CD44 KO mice ( Figure 4B, SP-b in CD44 KO vs. WT), compared to controls ( Figure 4B, inserts of SP-b in CD44 KO vs. WT inserts). Moreover, αSMA + cells, isolated from the lungs of bleomycin-injured chimeric CD44 KO mice, compared to WT (day 14), show significantly decreased motility almost equal to those from control saline-treated mice lungs (Ctrl: WT or CD44 KO    Chimeric CD44 KO mouse model, with CD44 KO mesenchymal cells and WT hematopoietic cells, generated by us as detailed previously [19] allows the assessment of CD44's role, specifically in mesenchymal cells in tissue regeneration in vivo, thus excluding the possible contribution of CD44-negative immune cells. We exposed chimeric-CD44 KO -and control WT-mice to 0.05 mU bleomycin ( Figure 4B  ATM, the pro-survival protein, activated in AECs in vitro following bleomycin exposure (Figures 2, 3) was upregulated in vivo in lungs of bleomycin-injured WT mice (WT in Figure 5A and insert) but downregulated in chimeric CD44 KO mice, (Figure 5A, WT vs. CD44 KO ). At 14 days of bleomycin exposure, chimeric CD44 KO mice had increased lung tissue regeneration (less fibrosis), as determined using Sircol-red assay by significant decreases in lung collagen deposition ( Figure 5B). From an average of 250 microgram per lobe in WT mice lungs to only 110, in chimeric CD44 KO mice ( Figure 5B, WT vs. CD44 KO ). We used a semi morphological quantitative index (SMI) to grade tissue regeneration vs. fibrosis development ( Figure 5C, WT vs. CD44 KO ) in bleomycin-injured mouse lungs (from 1 to a maximum of 5), of H&E-stained sections ( Figure 5D, WT vs. CD44 KO ). There was a significant decrease (p < 0.05) in the average of SMI grading from 2.6 in WT mice lungs to only 0.5 in mice with CD44 KO mesenchymal cells ( Figure 5C, WT vs. CD44 KO ). H&E staining show increased regeneration in lung tissue sections of bleomycin-compared to saline-treated chimeric CD44 KO mice ( Figure 5D, CD44 KO vs. WT vs. inserts).  KO . Bars (▬) represent median score. Three lung sections were scored from each animal, using a grading scheme as follows: grade 0 -normal tissue; grades 1-5 were used for animals with regenerated tissue (grade 0 to 1) or pulmonary inflammation and fibrosis, with higher grades (2 to 5) indicating more severe lesions [24]. Representative results from three independent experiments. For all comparisons, p < 0.05. (D) Hematoxylin-eosin (H&E) staining at day 14 post-BLM, in WT and CD44 KO murine lung tissue sections. Representative of 4-5 sections in 3 separate experiments with similar results, n = 7-10.

Discussion
Previously we demonstrated that AECs exposed to bleomycin enter a ROS-mediated stress state and undergo cell death [1]. Here we evaluated AECs survival and lung tissue regeneration vs. lack of regeneration and evolution of fibrosis with a specific aim at ATM, a DNA repair factor, and the role of CD44 in EMT, in vitro, and in vivo, in the murine bleomycin-induced lung injury as an experimental model of ROS-mediated lung injury [10]. Our results focus on a subset of AECs that do not undergo bleomycin apoptosis in vitro (annexin-V (−) , Figure 1A) and display an array of survival genes ( Figure 1B). ATM expression, and signaling, were significantly increased in AECs/annexin-V (−) subset (Figure 2A-C) and subsequently, their proliferative activity ( Figure 2D). Moreover, AECs Bars (-) represent median score. Three lung sections were scored from each animal, using a grading scheme as follows: grade 0 -normal tissue; grades 1-5 were used for animals with regenerated tissue (grade 0 to 1) or pulmonary inflammation and fibrosis, with higher grades (2 to 5) indicating more severe lesions [24]. Representative results from three independent experiments. For all comparisons, p < 0.05. (D) Hematoxylin-eosin (H&E) staining at day 14 post-BLM, in WT and CD44 KO murine lung tissue sections. Representative of 4-5 sections in 3 separate experiments with similar results, n = 7-10.

Normal Lung Regeneration Following Bleomycin-Injury in Chimeric-CD44 KO Mice, Compared to Congenic-Controls
ATM, the pro-survival protein, activated in AECs in vitro following bleomycin exposure (Figures 2 and 3) was upregulated in vivo in lungs of bleomycin-injured WT mice (WT in Figure 5A and insert) but downregulated in chimeric CD44 KO mice, ( Figure 5A, WT vs. CD44 KO ). At 14 days of bleomycin exposure, chimeric CD44 KO mice had increased lung tissue regeneration (less fibrosis), as determined using Sircol-red assay by significant decreases in lung collagen deposition ( Figure 5B). From an average of 250 microgram per lobe in WT mice lungs to only 110, in chimeric CD44 KO mice ( Figure 5B, WT vs. CD44 KO ). We used a semi morphological quantitative index (SMI) to grade tissue regeneration vs. fibrosis development ( Figure 5C, WT vs. CD44 KO ) in bleomycin-injured mouse lungs (from 1 to a maximum of 5), of H&E-stained sections ( Figure 5D, WT vs. CD44 KO ). There was a significant decrease (p < 0.05) in the average of SMI grading from 2.6 in WT mice lungs to only 0.5 in mice with CD44 KO mesenchymal cells ( Figure 5C, WT vs. CD44 KO ). H&E staining show increased regeneration in lung tissue sections of bleomycin-compared to saline-treated chimeric CD44 KO mice ( Figure 5D, CD44 KO vs. WT vs. inserts).

Discussion
Previously we demonstrated that AECs exposed to bleomycin enter a ROS-mediated stress state and undergo cell death [1]. Here we evaluated AECs survival and lung tissue regeneration vs. lack of regeneration and evolution of fibrosis with a specific aim at ATM, a DNA repair factor, and the role of CD44 in EMT, in vitro, and in vivo, in the murine bleomycin-induced lung injury as an experimental model of ROS-mediated lung injury [10]. Our results focus on a subset of AECs that do not undergo bleomycin apoptosis in vitro (annexin-V (−) , Figure 1A) and display an array of survival genes ( Figure 1B). ATM expression, and signaling, were significantly increased in AECs/annexin-V (−) subset (Figure 2A-C) and subsequently, their proliferative activity ( Figure 2D). Moreover, AECs surviving bleomycin, downregulated epithelial cell marker, E-cadherin and upregulated those of mesenchymal cells, N-cadherin and vimentin ( Figure 3A and C) and increased nuclear translocation of β-catenin ( Figure 3B) indicating EMT. The proportion of AECs/annexin-V (−) , expressing CD44 adhesion molecule was increased ( Figure 3C) and KM81, a specific CD44 mAb, decreased markers of EMT in these cells ( Figure 3D). Using the bleomycin-induced lung injury in WT mouse lungs, we show, that myofibroblasts isolated from the lungs at times of tissue degradation and lack of regeneration (day 14), increased CD44 expression and decreased it, to normal levels, at times when the pulmonary tissue regenerates (see our previous report on fibrosis kinetics [21] ( Figure 4A). Bleomycin-vs. saline-treated WT mice decreased surfactant-b, and increased αSMAcells in their lungs while chimeric CD44 KO mice, with specific knockout mesenchymal cells, show the opposite ( Figure 4B). Moreover, myofibroblasts isolated from the lungs of chimeric CD44 KO mice, compared to those from WT, lost their migratory capacity ( Figure 4C). Concomitantly, these mice show decreased ATM staining in cells of the alveolar interstitium ( Figure 5A) with decreased fibrotic pathology ( Figure 5B-D), thus evidencing improved regeneration of the lung tissue after bleomycin-induced injury.
The capacity of AECs to undergo EMT under oxidative stress is well established [9,29]. The role CD44 plays in EMT makes it an important target of numerous fields [4,30], including when the tissue fails to regenerate and fibrosis evolves [7,8] and differences in regulation of several known CD44 isoforms, may function differentially and bidirectionally upregulated or downregulated in EMT and, consequently, in fibrosis [8,31]. Our results accordant with others detecting CD44 upregulation in AECs subjected to hypoxia, with corresponding decrease in E-cadherin and increase in vimentin [32]. CD44-mediated signaling in pulmonary fibrosis has been addressed in a number of studies [33,34]. This study corroborated CD44's role, specifically in mesenchymal cells and in AECs acquiring mesenchymal properties ( Figure 3) and in reducing propensity for tissue regeneration following injury. In vivo we found, for the first time that knocking out CD44, specifically in mesenchymal cells, excluding the hematopoietic lineage, leads to significantly improved outcome in the regenerative capacity of the tissue following bleomycin injury, particularly reducing EMT ( Figure 4) and pulmonary fibrosis ( Figure 5). Moreover, as others [32], we show that myofibroblasts isolated from bleomycin-injured WT mice lungs demonstrated CD44 overexpression and increased motility, while those from chimeric-CD44 KO , did not ( Figure 4). In our model CD44 kinetics in lung myofibroblasts (Figure 4) is in good correlation with BLM-induced pulmonary fibrosis kinetics, which in young mice reaches maximum by day 14-21 and then regenerate the pulmonary tissue by day 28-56 [10]. Although similarly Li et al. showed that fibroblasts isolated from bleomycin-treated CD44-null mice or from WT-mice treated with CD44 mAb, impair invasive capacity, their CD44-null mice are not protected from pulmonary fibrosis and lack tissue regeneration [33]. This could be explained by the fact that Li et al. used CD44-null C57BL/6 mice which does not allow to exclude the contribution of CD44 hematopoietic cells in lung fibrosis [31]. In contrast, we used CD44 KO chimeric mice with WT hematopoietic cells.
ATM, involved in stress response, cell cycle regulation, tissue regeneration, apoptosis and survival [35], appears to influence AECs response to injury and fibrosis outcome [36]. We show here (Figures 1-3) that apoptosis-resistant AECs undergo adaptation, involving increased ATM expression with phosphorylation of key regulators of cell cycle and survival such as Erk, Brca1, RAD50, GADD45 and chk1, thus barring regeneration of normal lung tissue. Concomitantly, ATM was demonstrated to promote liver fibrosis [37]. Our findings further reveal ATM downregulation in chimeric-CD44 KO mice, who attenuate lung fibrosis and thus increase tissue regeneration (Figures 4 and 5). Others, nevertheless, associated ATM deficiency with increased pro-fibrosis cell's proliferation. For example, myocytes following heart injury [38], pancreatic cells undergoing EMT [39] and, even myofibroblasts in bleomycin-treated ATM-deficient mice [36]. The latter, used ATM-knockout or deficient mice, without excluding hematopoietic-lineage. Moreover, in our study, ATM absence results in regeneration of the lungs of bleomycin-treated mice lacking CD44 + myofibroblasts. The role of ATM in fibrosis may also depend on the particular splice variant in question [40] and appears to be strongly depended on the particular condition, tissue, or model investigated [15,36,39,41].
To conclude, our results demonstrate the key role of CD44 in transdifferentiation and motility of AECs surviving stress induced by bleomycin and in subsequent suppression of lung tissue regeneration and promotion of pulmonary fibrosis. This survival of AECs is associated with activation of ATM and ATM-signaling. EMT may represent a possible ATM-mediated cytoprotective mechanism of alveolar epithelial cells under oxidative stress. These results bear further significance for the tackling of the questions of promoting of regenerative capacity and rejuvenation of damaged tissue in aging and aging-associated disease as fibrosis.