Lung epithelial stem cells and their niches: Fgf10 takes center stage

The adult lung epithelium is replaced over time, albeit very infrequently in comparison to organs exhibiting constant cellular turnover such as the skin and intestine. However, after injury, the lung harbors a remarkable capacity to regenerate and restore its function. This is dramatically illustrated after unilateral pneumectomy, which induces an expansion of stem cell populations and compensatory growth of the remaining lung to re-establish respiratory capacity [1]. The composition of the lung epithelium varies along a proximal-distal axis (Figure 1A), which is reflected in the diverse physiological functions of the lung. In the mouse, the pseudostratified epithelium of the trachea and main stem bronchi consists of ciliated cells, club (also known as Clara) cells, a few mucus/goblet cells, and relatively undifferentiated basal cells, which express the transcription factor transformation-related protein 63 (Trp63 or p63), cytokeratin (Krt) 5 and/or Krt14. In the smaller intralobar bronchioles, the pseudostratified epithelium now transitions into a simple single columnar to cuboidal epithelial layer devoid of basal cells and containing mostly club and ciliated cells interspersed with single or clustered neuroendocrine (NE) cells termed NE bodies (NEBs), which are most frequently located at airway bifurcations. Of note, the basal cell-containing pseudostratified epithelium in human lungs extends to the distal bronchioles [2]. In the most distal regions of the lung, approximately 90% of the alveolar epithelium is composed of flattened alveolar type (AT) I cells, which are in close apposition to the capillary endothelium, allowing for rapid and efficient gas exchange, and cuboidal ATII cells that express surfactant. It is now becoming clear that these different epithelial regions in the lung are maintained and repaired by distinct stem cell populations.

Lineage tracing experiments during normal homeostasis have identified three main stem cell populations responsible for maintaining the lung epithelium: basal cells, club cells and ATII cells. Their lineage relationships are depicted in Figure 1B. Basal cells in the proximal airways are a bona fide stem cell population that gives rise to club and ciliated cells [3‐6]. Club cells are also able to self-renew and give rise to ciliated cells and therefore meet the stem cell criteria as well. They are the predominant cell population responsible for maintaining the bronchiolar epithelium. In the trachea however, their contribution to epithelial self-renewal seems to be minimal, and as a population, they are replaced over time by new club cells derived from basal cells [3, 7]. NE cells self-renew but under normal homeostatic conditions do not give rise to other epithelial cell lineages [8]. The alveolar epithelium is maintained by ATII stem cells, which can self-renew and can give rise to ATI cells [9, 10].

The lung is directly exposed to the outside environment and must therefore be able to respond quickly and effectively to inhaled particles, pathogens and harmful gases. The conducting airway epithelium is therefore designed to be a crucial primary defense mechanism by mediating mucociliary clearance and forming a protective physical barrier, which maintains its structural integrity and is vital for normal lung function. By exposing mouse lungs to different types of injury it is thought that different stem cell populations are engaged, depending on the location and extent of injury. In the trachea, SO₂ or naphthalene injury kills most luminal cell types leaving behind a few surviving club cells and an intact basal cell layer (Figure 2A). Surviving club cells proliferate and contribute to epithelial restoration, but basal cells are the most important cell type mediating tracheal regeneration under these conditions [3, 7]. A Krt14⁺ basal cell-like stem cell population located in submucosal gland (SMG) ducts has been shown to drive epithelial regeneration after severe epithelial injury using a tracheal transplant model of hypoxic ischemia, which destroys nearly all epithelial cells (Figure 2A) [11, 12]. This stem cell population is still poorly characterized and its further study will require more advanced lineage tracing techniques. Together, these studies have led to the notion that basal stem cells can be placed at the top of stem cell hierarchy in the trachea, a concept which has recently been complicated by the finding that diphtheria toxin-mediated ablation of basal cells in the trachea results in the reprogramming of a subset of club stem cells into basal stem cells effectively restoring the basal stem cell population of the trachea [13] (Figure 2A,B). Moreover, very rare club cell-derived basal cells have also been observed after SO₂-mediated tracheal injury [7]. Interestingly, p63, a master regulator required for the development of basal cells [14], induces a basal cell phenotype and squamous metaplasia when ectopically expressed in club or ATII cells [15].

A widely used model to study epithelial regeneration of the bronchioles is naphthalene injury. In the bronchioles, naphthalene selectively depletes club cells except for the few naphthalene-resistant club cells (called variant club cells in some literature) located near NEBs [16, 17] and at bronchioalveolar duct junctions (BADJs) [18], which then expand and re-epithelialize the damaged airways (Figure 3A). This led to the hypothesis that this subpopulation of club cells is responsible for epithelial regeneration after injury. However, after injury, the majority of surviving club cells, regardless of their location, are capable of restoring the damaged lung epithelium [7]. Interestingly, NE cells have been shown to demonstrate some degree of plasticity and as such are capable of not only self-renewing but also to give rise to club and club cell-derived ciliated cells after airway epithelial injury [8] (Figure 3). However, elimination of pulmonary NE cells does not impair airway epithelial regeneration after naphthalene injury [8], suggesting that the contribution of NE cells to epithelial restoration is minimal. Bronchioalveolar stem cells (BASCs) are another population of naphthalene-resistant stem cells located at the BADJ and can be identified based on their coexpression of secretory (Scgb1a1) and alveolar (Sftpc) markers, [19] (Figure 1). BASCs are likely identical to the variant club stem cells located near BADJs (described above) and might also be referred to as distal airway stem cells (DASCs) [20, 21]. BASCs can self-renew and give rise to both bronchiolar and alveolar cell lineages in vitro and in vivo, the latter only after catastrophic alveolar epithelial injury [19, 22] (Figure 4A,B). However, lineage tracing of Scgb1a1⁺ cells during normal homeostasis or after hyperoxia injury (which selectively destroys ATI cells) did not reveal any contribution of BASCs to alveolar regeneration [7]. Using the same Scgb1a1 ^CreER mice however, it was shown that Scgb1a1⁺ club cells in the distal airway can give rise to both ATI and ATII cells following catastrophic bleomycin- or H1N1-mediated injury [20, 22‐25] (Figure 4A,B). Thus, it seems that the deployment of BASCs depends on the severity and type of injury. Interestingly, DASCs may dedifferentiate into basal cells after catastrophic bleomycin- or H1N1-mediated injury prior to regenerating ATI and ATII cells [20, 25] (Figure 4A,B). However, it remains possible that the observed increase in basal cells is due to the expansion of a minor basal cell population in the distal airways or by migration of basal cells originating from the proximal airways. Future experiments combining both lineage tracing and live imaging during the repair response will help to address this. Bronchiolization is often observed as honeycomb regions in lungs from patients with idiopathic pulmonary fibrosis (IPF). These bronchiolized areas have been shown to contain basal cells [26], suggesting that in humans with IPF, regeneration of the fibrotic lung may also be mediated in part by distal airway club or basal stem cells.

Importantly, although BASCs/DASCs can contribute to alveolar regeneration, ATII cells are the main stem cell population in the lung respiratory epithelium involved in replenishing ATII cells after their diphtheria toxin-mediated depletion [9], re-establishing the ATI population after hyperoxic injury [7, 10] and regenerating both ATI and ATII cells after bleomycin injury [9, 23] (Figure 4). In addition to ATII cells, an Sftpc⁻ and integrin (Itg) α6⁺/β4⁺ alveolar epithelial stem cell population has recently been characterized, having the potential to give rise to both ATII and club cells in vitro and in vivo after injury [27, 28] (Figure 4B). To what extent this cell population contributes to alveolar repair after injury is not clear and will require lineage tracing to answer this question.

Lung stem cells must give rise to the appropriate number of differentiated progeny to achieve homeostasis and to restore the functional organ after injury. Tissue damage can dramatically change the dimensions of an organ, and after regeneration tissue growth must halt once the original organ dimensions are restored. Therefore, checks and balances are in place to prevent unwanted stem cell responses, which could lead to pathological changes compromising tissue integrity and lung function. The behavior of virtually all stem cells, whether they are pluripotent or lineage-restricted, embryonic or adult, is controlled by the interplay between intrinsic transcriptional programs and extrinsic signals [29]. The extrinsic signals are provided by the niche, a local tissue environment that hosts and influences the behaviors or characteristics of stem cells [30] and that comprises other cell types and extracellular matrix (ECM) (Figure 5). An intimate association of stem cells with their niche is critical to ensure long-term maintenance of stem cell populations, as well as to direct stem cell differentiation into the appropriate lineages. As such, the niche can maintain stem cell quiescence, promote transient activation, and keep stem cells undifferentiated [31].

During lung development, a complex interplay between endoderm and mesoderm defines early developmental competence and cell fate. This cross-talk is orchestrated by key signals including Fgfs, Wnts, retinoic acid, Shh and Tgfβ (reviewed in [32]). Proximal-distal patterning of the lung is accompanied by the gradual restricted ability of developmental progenitors to generate the various epithelial lineages in the mature organ [33]. During early lung development, Fibroblast growth factor 10 (Fgf10), which is expressed in the distal mesenchyme and is regulated by Wnt signaling [34‐36], acts on the distal lung epithelial progenitors to maintain them and prevent them from differentiating into proximal (airway) epithelial cells by inducing Sox9 and repressing Sox2 expression [37, 38] (Figure 6). Fgf10 is presented as a dimer, bound to heparin sulfate (HS) proteoglycans in the basement membrane (BM), to its receptor Fgfr2b on distal epithelial progenitor cells [39‐42]. However, as the lung epithelium grows out, more proximally located cells become further displaced from this distal source of Fgf10 and gradually start to differentiate [33, 37, 38, 42, 43]. Precisely regulated Fgf10 expression and presentation via HS proteoglycans is necessary during early development so that high levels are achieved distally to promote rapid expansion of embryonic progenitors. Cell surface-tethered HS chains play pivotal roles in the local retention of Fgf ligands and can spread Fgf signaling to adjacent cells within a short range [44]. Vice versa, Fgf10 suppression around the developing airway, as well as during late gestation and postnatal development, is crucial to allow for proper maturation of the lung epithelium [39‐41, 43‐45]. In addition, ectopic Fgf10 overexpression at later stages of lung development prevents the alveolar differentiation program through the induction of Sox9 expression [43, 46, 47]. Interestingly, a subset of Fgf10-expressing cells in the distal mesenchyme during early lung development are progenitors for airway smooth muscle cells (ASMCs) [35, 48, 49] (Figure 6) and lipofibroblasts (LIFs) at later stages [50].

Fgf10 is also expressed in between tracheal cartilage rings during lung development [51‐53]. In the trachea, Fgf10 signaling plays an important role in basal cell differentiation and maintenance during mouse lung development [43]. Fgf10 knockout tracheas show a 50% reduction in basal cells compared to wild type tracheas [43]. In addition, overexpression of Fgf10 during lung development results in the ectopic differentiation of a large subset of Sox2⁺ airway epithelial cells into the basal cell lineage all along the conducting airway. Interestingly, inhibition of Fgf signaling from E15.5 until E18.5 resulted in a failure to maintain basal cells in the mouse trachea [43].

Much progress has been made in identifying the different stem cell populations involved in the maintenance and repair of the adult lung epithelium as well as their lineage relationships. However, much less is known about the nature of stem cell niches in the lung and how they influence stem cell behavior.

Cellular and molecular mechanisms important during development are often reactivated during adult injury repair and in disease. As such, we have recently identified ASMCs as a niche for club stem cells [54]. After naphthalene-mediated airway epithelial injury, surviving epithelial cells secrete Wnt7b in order to activate ASMCs by inducing their c-Myc-mediated proliferation and Fgf10 expression [54, 55], which recapitulates a progenitor-like state (Figure 3A). Fgf10 secreted by the niche then acts reciprocally on surviving club stem cells near NEBs and at the BADJ, imparting stem cell characteristics by inducing a transient epithelial-mesenchymal transition (EMT). As a result, club stem cells break quiescence, induce proliferation and initiate epithelial repair. Short-term Fgf10 overexpression enhances airway epithelial regeneration, whereas inhibition of Fgf10 signaling reduces regenerative capacity after naphthalene injury [54, 56], underscoring the central role of this signaling pathway in regeneration of the bronchiolar epithelium. Interestingly, this re-expression of Fgf10 by ASMCs is observed after different types of airway epithelial injury, including ozone and bleomycin injury [54].

In the mouse trachea, label-retaining basal stem cells preferentially reside in SMG ducts and in intercartilage regions [57, 58], two areas where Fgf10 expression is the highest [51‐53, 59]. Whether Fgf10 regulates SMG duct stem cells in the adult lung [11] remains to be investigated.

Interestingly, direct cell-cell contact between basal cells and club stem cells in the tracheal epithelium inhibits the reprogramming of club stem cells into basal stem cells [13]. In this context, basal stem cells themselves contribute to the club cell niche in the trachea. However, it is currently unclear what prevents club cells from reprogramming into basal cells in the lower mouse airways where basal cells are absent.

A novel subpopulation of Scgb1a1⁻ club cells, characterized by expression of N1ICD, Scgb3a2, Uroplakin 3a (Upk3a) and SSEA-1, has been identified juxtaposed to Ascl1-expressing NEBs and remain largely uncommitted during development but can give rise to Scgb1a1⁺ club and ciliated cells in the adult lung [60, 61]. This cell population is lost in Ascl1 ^−/− mice suggesting that the induction and/or maintenance of these club-like precursor cells is highly dependent on the presence of NEBs in the developing bronchi [60, 61]. These findings together with the observation that a naphthalene-resistant subpopulation of club cells is associated with NEBs [16, 62] led to the hypothesis that pulmonary NEBs form a stem cell niche essential for club cell regeneration during lung injury. However, lungs lacking NE cells regenerate normally after naphthalene injury [8]. Therefore, further studies are needed to establish whether NEBs indeed provide signals to keep the adjacent naphthalene-resistant club cell population in a more progenitor-like state.

A similar cross-talk between epithelial stem cells and their niche is essential to maintain and regenerate the distal respiratory epithelium [63‐65]. In the adult lung parenchyma, LIFs are found juxtaposed to ATII cells and are thought to contribute to the ATII stem cell niche, maintaining their stemness [9]. LIFs assist in the production of pulmonary surfactant by assimilating neutral lipids and transferring triglycerides to ATII cells for final processing of the surfactant [66‐71]. In addition, LIFs also express Fgf10[50, 54], which has been shown to act directly on ATII cells via Fgfr2b to drive the expression of Sftpc[43, 54, 56, 72, 73]. As such, Fgf10 acts as a protective and therapeutic agent against bleomycin-induced pulmonary fibrosis [56]. Similarly, clonal expansion of the Sftpc⁻ alveolar stem cells enriched for α6/β4 integrins relies on co-culture with Sca1⁺ mesenchymal cells, which can be substituted by exogenous Fgf10 [28]. An additional population of CD166⁻ and Fgf10⁺ adult lung mesenchymal stromal cells (MSC) exists, which functions as a progenitor for LIFs and can support lung epithelial stem cell growth in vitro due to its Fgf10 expression. As during lung development, Fgf10 expression in these cells can be inhibited by Tgfβ signaling [34, 74, 75].

Bmp4 secreted by BASCs after bleomycin injury acts on endothelial cells via Bmpr1a to trigger calcineurin/NFATc1-dependent expression of thrombospondin-1 (Tsp1). This endothelial-derived Tsp1 then drives the differentiation of BASCs into the alveolar lineage via a feedback loop [22]. Interestingly, after naphthalene injury, Bmp4 and Tsp1 expression is downregulated, which was shown to favor the differentiation of BASCs into the club cell lineage [22]. Considering that Bmp4 is a main target gene of Fgf10 signaling [73, 76‐78], it is currently unclear how Fgf10 might be involved in this process since Fgf10 expression is upregulated in ASMCs and Fgf10 signaling in BASCs is activated after both naphthalene and bleomycin injury [54] (Figure 4C).

Lastly, a recent study provides insight in how the human Lgr6⁺ alveolar epithelial stem cell niche is formed and maintained to effectively ensure stem cell self-renewal capacity [79]. The authors identified a paracrine circuitry in which basal levels of SDF-1 secreted by Lgr6⁺ stem cells recruit and prime fibroblasts to release TNFα. This TNFα then leads to the activation of a TGFβ/p38α-mediated autocrine loop in Lgr6⁺ stem cells, which further enhances SDF-1 production. These high SDF-1 levels then stimulate fibroblasts to produce angiogenic factors to promote angiogenesis [79].

As described above, crosstalk between stem cells and their niche crucially depends on paracrine and autocrine signaling molecules. However, the behavior of epithelial stem cells depends on the integration with several other signals as well, including direct contact with the underlying ECM as well as with neighboring epithelial cells (Figure 5).

Components of the ECM, including fibronectin, proteoglycans and collagens, anchor stem cells in their niche and provides them with mechanical signals, based in part on substrate rigidity [80, 81]. This allows stem cells to respond to physical stimuli such as mechanical stress. In addition, the ECM acts as a growth factor reservoir by its ability to bind locally secreted growth factors (for example, Fgf10 and Tgfβ) and other stem cell regulatory molecules (Figure 5), which can be released by proteases such as heparinase [82]. Integrin-mediated cell adhesion to the ECM in the niche guides stem cell fate decisions, including choices between quiescence or proliferation, self-renewal or differentiation, migration or retention, and cell death or survival [83]. For example, loss of contact with the ECM or reduced integrin expression in adherent cells triggers terminal differentiation of cultured epidermal stem cells [84, 85]. The pulmonary ECM is subjected to a continuous turnover of more than 10% of the total ECM per day [86]. Thus a dynamic equilibrium between synthesis and degradation of the pulmonary ECM maintains a physiological balance [87]. Alterations in its composition (for example, in response to injury) lead to changes in cell shape and behavior [88], altered binding affinity or cellular distribution of cell-surface receptors [89], and different cellular responses to growth factors [90]. The BM, a specialized form of ECM separating the epithelium from the mesenchyme, is a dynamic structure produced by collaboration between stromal fibroblasts and epithelial cells. Depending on the composition and physical characteristics of the BM, different growth factors can have completely different cellular outcomes, such as cell proliferation, growth arrest, differentiation or apoptosis [91]. Moreover, tissue engineering experiments suggest that a decellularized matrix is capable of dictating exogenous stem cell fate [92], which highlights the importance of integrin-mediated stem cell-ECM communication in controlling stem cell behavior. Bidirectional signaling between epithelial stem cells and the cellular and acellular components of their niche is essential for normal tissue homeostasis, and there is clear evidence that stem cell integrins and BM proteins are involved in this communication [93].

In contrast to our increasingly extensive knowledge of pathways that regulate stem cell properties in the lung during development and regeneration, not much is known about the molecular mechanisms that keep lung stem cells largely quiescent during homeostasis or in check after a regenerative response to prevent tissue overcrowding and dysplasia. Preservation of the quiescent state by the niche is an actively regulated process that maintains the number and function of stem cells [94]. Epithelial cells form polarized cell layers that function as barriers capable of interacting with the underlying matrix as well as with neighboring epithelial cells. Epithelial cell polarity is maintained through the combined action of three major regulatory complexes: Crumbs complex, Par complex and Scribbled complex [95‐97]. These complexes are important in organizing cell-cell and cell-matrix adhesion junctions, which have emerged as major signaling platforms that mediate the cross-talk between neighboring epithelial cells and the underlying ECM, thereby regulating epithelial cell differentiation and proliferation via the Hippo pathway [95, 96, 98‐105]. For example, cultured cells arrest proliferation and cell division when the culture becomes confluent. This contact inhibition mechanism is crucial in vivo to regulate organ size and its loss is a hallmark of solid tumors [106, 107]. An intriguing picture is emerging in which stem cell progeny are a critical niche component, providing essential feedback to their stem cell parents to control stem cell activity and behavior [30]. However, the exact mechanisms underlying (integrin-mediated) mechanotransduction by the ECM and contact inhibition in the lung remain poorly understood.

Chronic lung disorders, such as IPF, asthma and chronic obstructive pulmonary disease (COPD), are characterized by the progressive remodeling of the airways and/or parenchyma, which irreversibly leads to lung dysfunction. In most cases their etiology is not well understood and current therapeutics are aimed at ameliorating rather than curing these diseases. Tissue architecture remodeling involves pathological changes in the composition and physiological function of the epithelium. During normal homeostasis as well as after injury, lung epithelial stem cells need to give rise to the right amount of differentiated progeny while maintaining their self-renewing ability depending on the physiological context. Disrupting this balance can lead to excessive differentiation and stem cell exhaustion, hypoplasia or squamous metaplasia. Such abnormal changes in stem cell function may result from inherent (epi)genetic changes or a dysfunctional stem cell niche. Furthermore, pathological changes to the lung epithelium may drastically influence the crosstalk between stem cells and their niche, which further contributes to disease progression by inducing abnormal changes in mesenchymal components, such as ASMC hyperplasia and fibroblast activation leading to excessive ECM deposition and fibrotic scarring. Chronic injury and prolonged activation of repair pathways are thought to result in decreased repair potential by exhausting the stem cell pool, which leads to defective repair and progressive airway remodeling by promoting a fibrotic response [63].