The role of club cell phenoconversion and migration in idiopathic pulmonary fibrosis

Introduction

Idiopathic pulmonary fibrosis (IPF) is an age-related, chronic, and progressive lung disease of unknown etiology [1]. Notably, the key cellular and molecular events in early stage IPF are poorly understood [2]. Recent reports suggest that type II alveolar epithelial cell (AEC) dysfunction, caused by gene mutations, coupled with repetitive exposure to noxious stimuli contributes to IPF development [3,4]. As an example of such genetic predispositions related to pulmonary fibrosis, mutations in SFTPC, a gene encoding surfactant protein C (pro-SPC, a representative marker of type II AECs), have been associated with familial pulmonary fibrosis (FPF) kindreds. Patients with SFTPC mutations present with a histopathological pattern of usual interstitial pneumonia (UIP), a key pathological feature of IPF [5,6]. In the meantime, a particular minor allele of single-nucleotide polymorphism (SNP) in the putative promoter region of MUC5B, a gene largely expressed in bronchiolar epithelium has also been linked to familial interstitial pneumonia and IPF [7]. This indicates that not only type II AEC dysfunction, but also functional perturbation of the bronchiolar epithelial cells is a risk factor for pulmonary fibrosis.

Club cells (previously Clara cells) are non-ciliated bronchiolar epithelial cells with multiple functions including (i) xenobiotic metabolism, (ii) immunomodulation through secretion of club cell secretory protein (CCSP), and (iii) regeneration through progenitor activity [8]. The involvement of club cells in IPF or other lung diseases featuring pulmonary fibrosis is not clear, however, it has been continually suggested since the 1980s that there is a link between lung fibrosis and alveolar bronchiolization, a process where club cells and other bronchiolar epithelial cell types migrate and populate alveolar walls [9–13]. Intriguingly, a recent report provided novel insights into a pathological role for club cells in IPF, wherein the authors proposed that club cells accelerate IPF progression through promoting lung epithelial cell death [13]. Madala et al. demonstrated that club cell-specific overexpression of transforming growth factor alpha (TGF-α) activate mesenchymal cell migration and accumulation in lung fibrosis [14]. In spite of such rising attention of recent years being paid to club cells, the cumulative attention that club cells have garnered so far in the field of IPF is very little when it is compared to type II AECs. One of the reasons attributed to this is the relative sparsity of club cells, as defined and assessed by the expression of CCSP, in IPF lungs in comparison to type II AECs.

In most of the lung fibrosis studies published so far, CCSP expression was used to define and trace club cells. However, a recent study has identified an additional club cell markers [15]. Given the availability of newly established clubcell markers, no studies were initiated with these markers to investigate the potential contribution of club cells to IPF pathology. The newly identified club cell markers include, but are not limited to, Flavin monooxygenase 3 (Fmo3), paraoxonase 1 (Pon1), aldehyde oxidase 3 (Aox3) and Claudin-10 (Cldn10). Among these newly identified club cell markers, Claudin-10 (referred to as Cldn10 hereinafter) is a very unique protein. In the early developing lungs of mice, Cldn10 first appears throughout the developing airway epithelium, and as club cells mature and begin to express CCSP, Cldn10 expression converges to the lateral surface of club cells (Supplemental Figure S1A) [15]. This spatial expression pattern of Cldn10 in adult lung club cells is consistent with the fact that Cldn10 have functions in paracellular epithelial permeation as tight junction components. Most importantly, Cldn10’s expression in developing lungs appears earlier than all other club cell markers including CCSP. The expression of Cldn10 in mouse lungs reaches adult expression levels as early as embryonic day (E) 17.5, an earlier time-point than CCSP, of which expression at E17.5 is still very low and reaches adult expression levels only as late as postnatal day 7 (P7) (Supplemental Figure S1B) [15]. In addition, the re-establishment of Cldn10 in damaged bronchiolar epithelium after naphthalene-induced club cell ablation occurs more rapidly than CCSP [15]. Thus, Cldn10 is considered to be an early sign for the developing and repairing bronchiolar epithelium. In other words, Cldn10 is a sensitive tracer for immature and not fully differentiated club cells in both developing and repairing airways. Intriguingly, IPF has been repetitively associated with recapitulation of developmental pathways [16,17]. Further, impaired wound healing has been considered to be a key mechanism in the pathogenesis of IPF [18]. Thus, we surmised that pleiotropic evaluation of Cldn10 would shed light on the potential involvement of club cells in IPF pathogenesis in the context of aberrant activation of developmental and repair pathways.

Results

Discussion

In the current study, our compelling data suggest that club cell phenoconversion and migration are actively occurring in IPF lungs. Comparison between IPF and control lung sections double-labeled for club cell markers, CCSP and Cldn10, demonstrated relatively well-organized expression pattern of these two markers in control lungs. In contrast, Cldn10-positive clubcells in IPF lungs exhibited a wide variety of cell shapes and arrangements in both bronchiolar and alveolar areas. Most importantly, our results indicate that the transition of club cells from CCSP- and Cldn10-positive to a solely Cldn10-positive state in IPF bronchioles, is spatially associated with fibroblastic foci even though it still remains to be elucidated whether such phenoconversion of club cells is a cause or a result of myofibroblastic phenoconversion of the underlying fibroblasts. Semi-quantitative histopathological evaluation using IPF and control lung sections labeled for Cldn10 led us to develop a new concept that the transition of club cells from a quiescent to active state is associated with their acquisition of proliferative and migratory capacity. Immunogold labeling on IPF lung samples substantiated this concept of “migratory club cells” as a potential trigger for IPF. In addition, having observed a link between Cldn10-positive club cells and aberrant α-SMA signals in IPF lungs, we deduced mechanistic insights into how fibrotic lesions develop and expand in a temporally and spatially heterogeneous manner in IPF. We devised a plausible scenario wherein metaplastic club cells that have acquired proliferative and migratory capacity randomly translocate and eventually attach to nonfibrotic alveoli, and subsequently interact with and phenoconvert interstitial fibroblasts.

Alveolar bronchiolization, a process of proliferation and migration of club cells and other bronchiolar epithelial cell types, has been documented since the 1970s [19]. Its potential relevance to pulmonary fibrosis was examined in bleomycin-induced lung fibrosis mouse model [11]. Similar phenomena of bronchiolar cell migration to the alveoli were closely examined on paraquat-induced lung fibrosis model in monkeys [9,10]. A systematic lineage-tracing approach supports the notion that migration of club cells is an innate reparative process that is to replace damaged alveoli in the lung [12]. Recently, a novel pathological role for club cells in IPF was proposed, wherein the migratory club cells contribute to IPF progression by promoting epithelial cell death [13]. This concept of migratory club cells as pathogenic participants in IPF is compatible with our observations that Cldn10-positive club cells with morphological pleomorphism, supposedly immature or reverted club cells, are widely distributed in IPF lungs, i.e. in fibroblastic foci (Figure 1E & F; Figure 2G), areas with honeycomb change (Figure 2J & K), and fibrotic interstitium as well (Figure 2M; Figure 5F & I). Certainly, none of the previous studies using microarray analysis reported the upregulation of Cldn10 in IPF. Our preliminary studies using qPCR analysis did not show any significant upregulations of Cldn10 in IPF vs COPD or DIP lungs (data not shown). However, it is clear from our data that translocation of Cldn10, from lateral membrane or luminal edge of the club cells to their cytoplasm and/or nucleus, is closely associated with phenoconversion of club cells (Figure 3C & H; Figure 4G, H & I). So far, no papers have reported the presence of cytoplasmic or nuclear translocation of Cldn10 in any human diseases or cell types. However, nuclear translocation of Cldn2, another member of claudin family, has been implicated in lung cancer and associated with cell proliferation in lung epithelial cells [20]. In melanoma cells, Cldn1 localized in cytoplasm as well as nucleus, with its expression in the former correlating to increased migration [21]. From such examples of other members of the claudin protein family, it can be reasonably deduced that cytoplasmic and/or nuclear translocation of Cldn10 too is associated with a certain type of phenoconversion involving acquisition of proliferative and migratory abilities.

The emergence of metaplastic cuboidal cells is one of the key histopathological features of IPF [22]. Remarkably, the metaplastic epithelial cells overlaying myofibroblasts in fibroblastic foci exhibit cuboidal or oval shapes from our data (Figure 2G) as well as previous report [23]. Kawanami et al. (i) categorized cuboidal epithelial cells seen in fibrotic lung diseases including IPF, based on their subcellular structures under electron microscope observation, into two subtypes, type A and type B, and (ii) surmised, based on their ultrastructural similarities to well-characterized epithelial cell types, that type A and type B cuboidal epithelial cells are derived respectively from bronchiolar basal cells and respiratory bronchioles [22]. Our current study demonstrates that in IPF lungs Cldn10-positive cuboidal or oval cells widely populate fibrotic and nonfibrotic regions in varied arrangements, i.e. as a single cell (Figure 3G & H), in monolayer (areas designated by asterisks in Figure 2D & G; Figure 3) and in massive forms (arrow in Figure 2D; area surrounded by dot-dashed line in Figure 2O). Intriguingly, Cldn10-positive cellular masses were frequently observed in the airspace of IPF lungs and they are usually mosaic mixtures containing Cldn10-positive and negative cells with the latter being predominant (areas circled by dotted lines in Figure 2E & H; areas circled by dot-dashed lines in Figure 2I & O; area circled by dotted line in Figure 2 K; Supplemental Figure S2F, G and H). In addition, Cldn10 ^{neg or low} cuboidal cells were often encountered in overlaying epithelium in fibroblastic foci (Figure 2G; Supplemental Figure S2C and D). One possible explanation for these heterogeneous profiles of metaplastic cuboidal cells in IPF lungs is that they are derived from different cell types, i.e. club cells, and non-club progenitor cells such as basal cells. This hypothesis is compatible with the previous report by Kawanami et al. wherein 9 out 17 IPF samples tested contain both type A and type B cuboidal epithelial cells [22]. Another plausible explanation for this heterogeneity is “depending on the extent to which club cells dedifferentiate or revert back, their expression of Cldn10, as well as CCSP, decrease in a time-lagged manner (i.e., the decease of CCSP expression is followed by Cldn10 diminution).” This hypothesis is in agreement with our experience obtained through TEM observation wherein (i) IPF lungs encompass club cells with remarkably pleomorphic cell shapes and varying degrees of Cldn10 expression and (ii) clusters of Cldn10-negative cells with high nuclear/cytoplasmic ratio (N/C) and clear chromatin are often found in fibrotic interstitium (Supplemental Figure S4). Intriguingly, Watson et al. demonstrated that deficiency of CCSP causes alteration in club cell ultrastructure and increase in immune cell-specific transcripts during lung remodeling, suggesting that club cell phenoconversion-induced its acquisition of immune cell property and consequent decrease in CCSP secretion plays a pivotal role in lung remodeling [24].

Despite extensive studies that have been performed so far in the field of IPF, it still remains to be elucidated what cellular and molecular events cause perpetual phenoconversion of quiescent fibroblasts into a distinct α-SMA-expressing cell type, i.e. myofibroblasts [23]. Transforming growth factor beta 1 (TGF-β1) is a potent fibrogenic cytokine that plays a crucial role in IPF pathogenesis [25]. One of the most clinically relevant roles of TGF-β1 in IPF is to cause myofibroblastic phenoconversion [26]. In fibroblastic foci, high levels of TGF-β1 are spotted in hyperplastic epithelium overlaying myofibroblasts [27]. Given the fact that a certain proportion of the epithelial cells that overlay myofibroblasts in fibroblastic foci are Cldn10-positive (Figure 2G), it is plausible that dedifferentiated club cells trigger myofibroblastic phenoconversion through TGF-β1 secretion. This hypothesis is consistent with our data showing that in fibrotic areas of IPF lungs, Cldn10-positive cells were frequently juxtaposed with aberrant α-SMA signals (Figure 5F, I, L & O). In additional support of this hypothesis, our preliminary studies revealed that oxidative stress, an alleged risk factor for IPF, accelerates emergence of proliferative and TGF-β1-secreting club cells in certain mutant mice (separate manuscript). If such a profibrotic phenoconversion of club cells is a critical trigger for the appearance of myofibroblasts in IPF, then inhibition of club cell phenoconversion can reasonably be a novel therapeutic option for the treatment of IPF.

Type II AEC abnormality has long been implicated in IPF. Mutations in SFTPC, a representative gene expressed in type II AECs, have been linked to familial pulmonary fibrosis. Additionally, Type II AEC hyperplasia is an established histopathological finding in IPF, indicating the important contribution of type II AECs to IPF pathogenesis. In the present study, we did not perform quantitative or qualitative evaluation of the relative contributions of different epithelial cell types including type II AECs and club cells. However, our TEM observations revealed that a small number of dysmorphic lamellar bodies are frequently spotted in columnar epithelial cells lining bronchiolar epithelium, and occasionally encountered in metaplastic cuboidal cells in fibrotic interstitium (separate manuscript). These observations imply that in IPF lungs a certain proportion of epithelial cells with pluripotency, including CCSP-negative immature club cells, are preferentially committed to differentiate into type II AECs. If this is the case, pro-SPC-positive metaplastic type II AECs frequently observed in IPF lungs may be derived, at least in part, from club cells or non-club progenitor cells such as basal cells. Thus, it is plausible that the true nature of IPF lies in impaired differentiation of lung progenitor cells into normal and functional type II AECs.

Lineage tracing method using cell type-specific transgenic mice has provided access to otherwise unattainable information, especially regarding the cell types that participate in the development of pulmonary fibrosis [12,28,29]. Unfortunately, cell-specific markers can diminish upon cell differentiation or dedifferentiation, which makes it difficult to trace highly plastic cell lineage. Club cells have been known to change their profiles in response to environmental stressors [8], and our data also indicate that club cells are actively changing their epithelial properties in IPF lungs (Figure 1E & F). Further, while tracing of CCSP-positive club cells have long been available to researchers, a mutant strain enabling Cldn10 tracing is not yet available. Therefore, we reached a notion that relating subcellular ultrastructures of metaplastic epithelial cells to club cell markers in IPF lungs would provide us with otherwise unattainable information as to the origin of metaplastic lung epithelial cells. As expected, our studies have given us mechanistic insights about the origin of metaplastic cuboidal cells in IPF.

Little information is available on how club cells senesce during ageing, especially in the context of pulmonary fibrosis. Zhou et al. proposed that club cell senescence plays an important role in COPD pathogenesis by revealing a higher percentage of club cells in COPD lungs that are positive for CCSP as well as p16, a representative marker of ageing, compared to control lungs from asymptomatic nonsmokers [30]. They related senescent club cells to impaired alveolar regeneration and exacerbated inflammation in COPD. On the other hand, our data suggest that in IPF lungs club cell acquisition of proliferative as well as migratory ability plays a key role in impaired alveolar regeneration and consequent chronic fibrosis. It may well be that the balance between age-related club cell senescence and its survival adaptation determines the fate of the aged lungs.

In conclusion, our new approach combining immunohistochemical double staining and immunogold staining has allowed us to propose a new paradigm where club cells actively participate in IPF pathogenesis from early stages through perpetual phenoconversion (summarized proposal is shown in Figure 6). This new concept does not necessarily conflict with the current central paradigm of IPF where type II AECs play a critical role in IPF pathogenesis assuming that club cells under stressed conditions can lose their properties and gain type II AEC character as an innate reparative program. Further investigation is required to validate this new hypothesis. If the aberrant phenoconversion of club cells and impaired differentiation of club cells into functional type II AECs are established as essential triggers for myofibroblastic phenoconversion of normal fibroblasts, then club cell-targeted therapy might become one of the powerful strategic options for the treatment of IPF.

Materials and Methods

Supplementary Materials

Supplemental File