Detection and quantification of epithelial progenitor cell populations in human healthy and IPF lungs

Lung function, barrier integrity, and epithelial homeostasis are maintained, in large parts, by its inner lining, which is constituted by lung epithelial cells. Different types of lung epithelial cells have been identified and characterized in the past, including ciliated, goblet, or basal cells in proximal airways, and ciliated, club (Clara), neuroendocrine, or basal cells in distal airways. The pseudostratified airway epithelial layer continues into a flat epithelial layer composed of alveolar type 1 (AT1) or type 2 (AT2) cells in the alveolar regions. Lung injury induced by injurious agents (e.g. cigarette smoke, particulate matter, viruses) primarily affects the integrity of this epithelial compartment, either by inducing barrier dysfunction or epithelial progenitor cell perturbation [1‐4].

Recently, lung epithelial progenitor cells have attracted significant interest, as they serve as stem cells for differentiated lung epithelial cell types, such as secretory, ciliated, or AT1 cells in mice [3]. Basal cells are relatively undifferentiated epithelial cells, which are located attached to the basal lamina of the stratified and pseudostratified airway epithelium [5]. Basal cells participate in anchoring of the airway epithelium to the basement membrane and thus separate the underlying stroma from the epithelial tubes. Airway basal cells also maintain the homeostasis of the normal epithelial layer by giving rise to the differentiated airway epithelial cells during postnatal growth and in the adult under steady state, as previously shown by us and others using a variety of in vivo and in vitro assays [6‐8].

Airway basal cell dysfunction contributes to a number of pathologies, as disruption in the balance between basal cell proliferation and differentiation leads to epithelial hyperplasia/hypoplasia [5] or abnormal differentiation patterns [4, 9], e.g. in lung cancer. Hyperplastic basal cell lesions are also observed in chronic obstructive pulmonary disease (COPD) [5, 10] and cystic fibrosis (CF) [11], whereas the epithelial injury and loss in bronchiolitis obliterans syndrome (BOS) may indicate stem cell exhaustion or a loss in epithelial progenitor populations [5].

In mice, recent studies using a number of reporter and fate-mapped mouse strains have demonstrated that lung basal cells contribute to regeneration of the lung in response to injury, based on their capacity to proliferate and differentiate into mature cell types [12, 13]. Evidence of the regenerative capacity of basal cells in human results from the observation that basal cells isolated from different regions of the human lung proliferate and differentiate to mature ciliated, goblet, or club (Clara) cells in air-liquid interface (ALI) cultures or bronchosphere assays ex vivo [14]. In vivo, injury/repair models have demonstrated that disruption of the basal cell layer is associated with an uncontrolled proliferation of the underlying stroma, resulting in an accumulation of fibroblasts and immune cells that subsequently obliterate the airways [15].

Emerging evidence shows that basal cells are composed of multiple heterogeneous subpopulations, under physiological as well as pathological conditions. As an example, mouse tracheal basal cells characteristically express cytokeratin 5 (KRT5), while only a limited subset expresses cytokeratin 14 (KRT14). Interestingly, KRT14 is upregulated in mouse lung basal cells in response to naphtalene-injury [16]. As such, ongoing evidence highlights a role for KRT5⁺KRT14⁺ basal cells in post-injury regeneration of the mouse lung [6, 12‐14].

Details about definitive basal cell subpopulations, however, remain to be elucidated, in particular in the human lung. In this context, basal cell subsets expressing distinct keratin (KRT) isoforms have been described [17] and recent evidence suggests alterations in KRT abundance and expression in lung disease with features of diffuse alveolar damage [18, 19]. Increased KRT5 and KRT14 expression has also been reported in the alveolar regions in idiopathic pulmonary fibrosis (IPF) [19]. Yet, the distinct quantitative and spatial abundance of KRT5⁺ and KRT14⁺ cells to IPF is unknown.

To this end, we sought to investigate and quantify the distribution of KRT5⁺ and KRT14⁺ cell populations in human lungs, obtained from healthy donors or IPF patients. We provide here, for the first time, a quantitative analysis of the distribution of KRT5⁺ and KRT14⁺ single- and double-positive cell populations in the healthy human lung. Importantly, we describe dramatic changes in the distribution and morphology of these cells in IPF. Finally, we seek to characterize their differentiation potential by fluorescent co-staining of these populations with well-accepted epithelial differentiation markers, such as acetylated tubulin, Mucin 5B, or Clara Cell 10 kDa Protein (CC10) in IPF.

We show, for the first time, a proximal-to-distal quantification of bronchial basal cell subtypes in the human lung, characterized by various combinations of p63, KRT5, and KRT14 expression. While several combinations of KRT5^+/-KRT14^+/-p63^+/- basal cells populate the human airways in distinct locations, the total absence of a KRT5⁺KRT14⁺p63⁺ population in the healthy distal lung is remarkable. In IPF, the overall amount and distribution of the KRT5⁺p63⁺ cells is dramatically modified, and a frequent KRT5⁺KRT14⁺p63⁺ population emerges in the distal IPF lung. While KRT14^- cells co-expressed alveolar and bronchial epithelial differentiation markers, KRT14⁺ cells exhibit a metaplastic phenotype and failed to costain with any of these markers. This strongly suggests that the KRT5⁺KRT14⁺p63⁺ cell type identified in the distal IPF lung is functionally different from the KRT5⁺KRT14^-p63⁺ cell population.

The decrease in basal cell numbers along the proximal-distal axis reported herein is consistent with the profile reported 27 years ago by Boers et al. [21]. In modification to those observations, we used the well-accepted basal cell marker p63 for cell quantification, which gives a more precise estimation of total basal cell numbers compared with morphological measures alone [4]. We sought to characterize and quantify airway epithelial progenitors in the human lung, as such data is missing up-to-date, while quantitative investigations have been performed for KRT5⁺KRT14^- and KRT5⁺KRT14⁺ populations in the mouse trachea [22]. Evidence of the expression of KRT5, but not KRT14, by human basal cells came from colony forming-basal cells, isolated from the proximal or distal human lung [14]. The expression of KRT14 in human basal cells was observed in basal layers of bronchospheres formed by basal cells from human lungs [6]. In addition, early studies also described KRT14 expression in basal cells in conducting healthy human airways, although without any evidence of KRT5 expression [17]. Thus, our study now provides evidence that 1) KRT5⁺KRT14^-p63⁺ and KRT5⁺KRT14⁺p63⁺ populations coexist in similar proportions in the conducting airways, 2) KRT5⁺KRT14⁺p63⁺ cells are absent from distal human healthy lungs, 3) some KRT5⁺ cells are p63^-, which suggests a sequential loss of basal cells markers during steady state cell renewal in the conducting airways; and 4) a rare population of KRT5^-p63⁺ exists in the distal human lung, previously observed also by Vaughan et al. [12]. This KRT5^-p63⁺ population of the distal airways could either express KRT17 [17] or KRT15, which is described as a marker for all basal cells in the mouse trachea [22].

IPF is a progressive, irreversible, and fatal lung disease, in which the lung epithelium is replaced by scarring tissue rich in collagens and extracellular matrix, honeycombing, and myofibroblast-rich regions. Available therapeutic approaches slow down progression of the disease, but regenerative medicine approaches would represent a promising novel option for the treatment of IPF [23]. As such, one potential option is to target basal cells, endogenous lung epithelial cells likely to have regenerative potential. While the accumulation of p63⁺ cells has been previously reported in IPF [24], distinct subpopulations were undefined. Similarly, an accumulation of KRT5⁺ cells [12], as well as KRT5⁺-KRT14⁺ cells [20] has been observed, but neither quantified nor characterized. Here, we show that, in striking contrast to normal lungs, distal IPF lungs present 1) an increased abundance of KRT5⁺ cells, 2) the emergence of a frequent KRT5⁺KRT14⁺ subpopulation, 3) morphological changes in these basal cells, and 4) the existence of characteristic patterns of KRT5⁺KRT14^- and KRT5⁺KRT14⁺ structures. Moreover, the quantitative analysis of these different progenitor subpopulations revealed a regional heterogeneity in the distal IPF lung. While in healthy-like non fibrotic, or in fibrotic areas, the basal cell progenitors are rarely found, they are abundant in areas of bronchiolization, suggesting the possibility of an ongoing regenerative process in these specific regions.

Current knowledge in this field is limited by the lack of an irreversible animal model more closely mimicking these features of IPF, since bleomycin-treated mice do not exhibit basal cell accumulation in the lungs [12]. To date, only the tracheal epithelium-injury model using naphtalene reliably increases the proportion of KRT14⁺ cell populations in the mouse [22]. Interestingly, upon infection with H1N1, KRT5⁺p63⁺ cells also accumulate in the distal parts of mouse lungs, forming clusters or pods [14]. In our study, we detected, amongst others, similar patterns of KRT5⁺ cells in IPF lungs. Moreover, regions of pronounced KRT5⁺ cell metaplasia in IPF samples were frequently characterized by KRT14 coexpression, suggesting hyperplastic potential of the KRT5⁺KRT14⁺ basal cell population. This is consistent with the hyperproliferative potential attributed to KRT14+ cells. Indeed, the knockdown of KRT14 in basal epithelial cells of the skin leads to cell cycle arrest [25].

In mice, reporter lines have enabled to describe the different origins of the KRT5⁺ basal cell population that arises in the mouse lung after injury. Vaughan et al. showed that the KRT5⁺ cells accumulating after H1N1 injury are mostly not of a KRT5⁺, but a lineage-negative origin. Another study showed that Clara cells dedifferentiate into p63⁺KRT5⁺ basal cells [26]. Interestingly, we observed a surprising coexpression of KRT5 or KRT14 with the Clara cell marker CC10 in the basal layers of typical bronchiolar structures in IPF. Therefore, although we cannot provide direct evidence of the origin of basal cells in IPF, a dedifferentiation of CC10⁺ cells cannot be excluded.

The dedifferentiation and transdifferentiation capacities of basal cells are strongly dependent on the type of injury [10] and their proximal-to-distal origin [14]. In the mouse, tracheal KRT5⁺ cells give rise to club (Clara) and ciliated cells under steady state or epithelial injury conditions [6]. After selective ablation of KRT5⁺KRT6⁺p63⁺ distal airway stem cells (DASC), mice failed to regenerate AT1⁺ structures after H1N1 injury [13]. The differentiation potential of human basal cells is tightly dependent on their origin and assay-specific environment [14]. In IPF, Vaughan et al. could not identify KRT5⁺SPC⁺ coexpressing cells. In contrast, in the human lung, KRT5 was frequently coexpressed with ProSPC, in particular in hyperplastic regions. Moreover, further costainings between KRT5 and alveolar/bronchial epithelial cell markers suggested a multidirectional differentiation capacity of these cells in IPF.

The fate of these basal progenitor cells in IPF is likely to be dependent on the dysregulation of numerous growth factors and cytokines, which can affect the differentiation potential of basal cells [27], but also on the adjacent ECM [28], all of which is dramatically altered in IPF. The dysregulated signaling environment changes the intracellular signaling cascades in the IPF epithelium [29], in particular with respect to aberrant activation of developmental pathways in IPF [30, 31]. During lung development, SHH, the most broadly expressed Hedgehog ligand, controls proper branching morphogenesis and patterns the lung [32]. SHH is upregulated in IPF, where it localizes to alveolar and bronchiolar epithelial cells, as previously reported by Bolaños et al. [30]. In agreement, we observed SHH staining in the distal IPF lung, but not in KRT5⁺ basal cells. Interestingly, the SHH-induced transcription factor GLI1, which is known for its involvement in stem cell renewal in non-small cell lung cancer [33], localized to the nuclei of most KRT5⁺ cells in the distal IPF lung, arguing for activation of the SHH pathway in these cells. HES1, a transcription factor activated by canonical Notch signaling, induces Clara cell differentiation in the developing lung [34]. HES1 is expressed in KRT5⁺ cells in areas of honeycombing in IPF, suggesting impaired differentiation potential into AT2 cells [12]. Our co-immunostainings for KRT5 and HES1 confirmed patchy Notch signaling activation in the KRT5⁺ subpopulation. In addition, SOX9 participates in bronchial development. Tight control of SOX9 levels is required for proper epithelial cell proliferation, since SOX9 inhibits epithelial differentiation in the developing mouse lung [35]. Upregulation of SOX9 is associated with lung cancer [36]. We observed significant numbers of KRT5⁺ cells in the IPF lung with prominent nuclear SOX9⁺ staining. Altogether, these and other findings thus suggest abnormal activation of at least three signaling pathways involved in cell proliferation and differentiation in IPF basal cells, and further indicate that the differentiation potential of the KRT5⁺ population in IPF is modified compared with quiescent healthy cells.

While KRT14⁺ cells have been reported to coexpress AT2 cell markers in DAD [18], we could not detect any costaining of KRT14⁺ with differentiation markers. The function of the KRT5⁺KRT14⁺ population therefore remains unclear. Under steady-state conditions, both KRT5⁺KRT14^- and KRT5⁺KRT14⁺ subpopulations are mitotic in the mouse trachea [22]. Interestingly, the increase in the number of KRT14+ cells and their contribution to the mitotic pool suggests that this subpopulation gives rise to a highly mitotic reparative pool after naphtalene injury in the mouse [22]. The hypothesis of a proliferative function of KRT14⁺ cells is reinforced by the tumorigenic potential demonstrated for KRT14⁺ cells from the skin [25], but also in the lung, where tumors spontaneously develop in knock-in mice for human KRT14 [37]. Yet, some evidence also suggests a differentiation potential for the KRT5⁺KRT14⁺ population. Indeed, after naphtalene injury, some KRT14⁺ cells had a luminal morphology in the mouse trachea, whereas the differentiating cells were mostly KRT14^-, which indicates a quick shift in the KRT14 expression during differentiation.

Lately, new pathways determining basal cell fates after injury have been emerging in the mouse. Pardo-Saganta et al. showed that depending on the intracellular pathways activated in the basal cells, they were committed either to a ciliated or a secretory cell lineage [38]. These discoveries, combined with an accurate description of epithelial progenitor subpopulations in human healthy and diseased lung, are promising approaches for further use of lung epithelial progenitor cells for therapeutic manipulation and epithelial regeneration.

This work provides a quantitative analysis of the regional distribution of KRT5⁺KRT14⁺p63⁺ and KRT5⁺KRT14^-p63⁺ basal cell subpopulations in the healthy human lung. Dramatic changes in the amount and distribution of these cell populations are observed in IPF. In IPF, these populations express differentiated epithelial cell markers, showing that in a context of injury/disease, they are likely attempting to regenerate the epithelium. This work provides a basis for further descriptions of the molecular pathways that may control the fate of progenitor basal cells in human lung disease.

acTUB, acetylated tubulin; AQP5, aquaporin 5; AT1, alveolar Type 1; AT2, alveolar type 2; CC10, clara cell 10 kDa-protein; IPF, idiopathic pulmonary fibrosis; KRT, keratin; MUC5B, mucin 5B; ProSPC, pro-surfactant protein C