Basal-like Progenitor Cells: A Review of Dysplastic Alveolar Regeneration and Remodeling in Lung Repair.

Despite the central importance of the respiratory system, the exact mechanisms governing lung repair after severe injury remain unclear. The notion that alveolar type 2 cells (AT2s) self-renew and differentiate into alveolar type 1 cells (AT1s) does not fully encompass scenarios where these progenitors are severely affected by disease, e.g., H1N1 influenza or SARS-CoV-2 (COVID-19). Intrapulmonary p63+ progenitor cells, a rare cell type in mice but potentially encompassing more numerous classic basal cells in humans, are activated in such severe injury settings, proliferating and migrating into the injured alveolar parenchyma, providing a short-term "emergency" benefit. While the fate of these cells is controversial, most studies indicate that they represent a maladaptive repair pathway with a fate restriction toward airway cell types, rarely differentiating into AT2 or AT1 cells. Here, we discuss the role of intrapulmonary basal-like p63+ cells in alveolar regeneration and suggest a unified model to guide future studies.

Main Text

Introduction

The lung is unique among mammalian vital organs; while it exhibits low cellular turnover during homeostasis, the lung can regenerate through widespread proliferation and differentiation of resident progenitor cell populations upon injury. Epithelial insults and respiratory infections such as influenza can disrupt the gas exchange units of the lung, the alveoli, destroying alveolar epithelial type 2 (AT2) and type 1 (AT1) populations, which in extreme cases results in diffuse alveolar damage (DAD) and acute respiratory distress syndrome (ARDS). While likely an oversimplification, speaking generally two distinct regenerative responses to lung injury can occur in parallel within the alveoli: functionally beneficial regeneration that typically follows mild-to-moderate pulmonary injury, and maladaptive epithelial remodeling that follows severe lung injury exemplified by pandemic strains of H1N1 influenza and likely SARS-CoV-2 (COVID-19). During functionally beneficial regeneration, surviving AT2 cells self-renew and differentiate into oxygen-exchanging AT1 cells (Barkauskas et al., 2013; Evans et al., 1973), providing the damaged lung with a source of essential surfactant and gas-exchanging epithelial surface. In more severely injured regions of alveolar parenchyma, maladaptive repair occurs in which injury-associated hypoxia activates a population of rare, airway-resident intrapulmonary p63 progenitors that migrate into and proliferate within the alveoli (Kumar et al., 2011; Vaughan et al., 2015; Zuo et al., 2015), forming an ectopic bronchiolar-like epithelium expressing markers of basal stem cells, such as cytokeratin 5 (Krt5) and Trp63 (referred to hereafter as p63, indicating both protein and gene expression) and, later, differentiated airway epithelial markers, including Scgb1a1 and Scgb3a2 (Kanegai et al., 2016; Vaughan et al., 2015) (Figure 1). Notably these intrapulmonary progenitor cells in humans likely encompass the much more abundant “classic” basal cells, already expressing both Krt5 and p63, present throughout the airways. Considering the ongoing COVID-19 pandemic, which similarly induces ARDS accompanied by DAD and a sustained decrease in pulmonary function in some individuals (Cheung, 2020; Salehi et al., 2020), defining the response of intrapulmonary p63+ progenitors to injury and the mechanisms underlying their sustained activation, persistence, and differentiation is critical in understanding and treating severe pulmonary insults. Here, we attempt to provide a comprehensive and unified overview of this phenomenon, which, we posit, ultimately represents maladaptive epithelial remodeling founded upon a short-term benefit.

Figure 1

Overview of Activation, Expansion, and Differentiation of Intrapulmonary p63+ Progenitor Cells Upon Lung Injury

Upon injury and subsequent hypoxia, rare intrapulmonary p63+ progenitor cells are activated, upregulate other basal cell markers, including Krt5, and proliferate and migrate distally into the small airways and alveoli. These cells have been demonstrated to readily generate airway cell types (left), but only rarely give rise to normal alveolar cell types (right), thus representing a dysplastic alveolar repair response. This may result in persistent loss of lung function where a significant fraction of the alveolar epithelium has been replaced by dysplastic, “bronchiolized” epithelium. Figure created with BioRender.com.

Foundations of Basal-like Cell Expansion

While injury-induced expansion of intrapulmonary p63+ cells has now been best characterized in mouse models, to our knowledge the first actual demonstration of this phenomenon was in humans. After the 1957 influenza pandemic, it was observed in explanted patient tissue that the lung epithelium was able to apparently repair starting 5 to 7 days after infection by replacing the damaged cells with a “stratified pseudo-metaplastic layer” in the respiratory bronchioles and alveolar ducts and a monolayer of epithelial cells in alveolar regions, “giving the impression of neoplastic growth” (Hers et al., 1958). Almost two decades later, this finding was confirmed in mice: after experimental infection of mice, Loosli et al. (1975) demonstrated that H1N1 (PR8) influenza partially destroyed alveolar epithelial cells and profoundly disrupted lung function, followed later by the onset of fibrosis. Provocatively, the authors also described the formation of cystic lesions that replaced AT2s, were adjacent to fibrotic areas, and did not resolve with time, ultimately proposing that these “epithelial nodules” had a cellular origin in the bronchiolar epithelium (Loosli et al., 1975). Interpretation of these early studies is nonetheless limited by the lack of immunohistochemical staining and lineage-tracing analysis, and these reports were largely forgotten for nearly 40 years.

Many years later, alveolar epithelial injury after murine infection with the same H1N1 strain used by Loosli et al. was shown to induce the dramatic appearance of proliferating epithelial cells expressing cytokeratin 5 (Krt5) and the transcription factor p63 (Trp63), bearing markers of trachea basal cells but appearing ectopically in small airways and alveoli. Given that basal cells are normally restricted to the trachea in mice, the development of what appeared to be airway basal cells in the alveoli was an overtly dramatic and seemingly bizarre observation, capturing the immediate attention of the pulmonary research community. These cell clusters, often called “pods,” eventually organized into structures resembling alveoli (Kumar et al., 2011), similar to the cysts noted by Loosli et al. (1975). It has been suggested that pod formation represents a bona fide regenerative response to lung injury, i.e., that Krt5+ p63+ cells ultimately differentiate into AT2s and AT1s, although the initial report did not definitively demonstrate this transdifferentiation via Cre-mediated fate mapping (Kumar et al., 2011). As discussed later, there is in fact scant evidence that these Krt5+ pods contribute significantly to normal alveolar cells types in most contexts (Kanegai et al., 2016; Ray et al., 2016; Vaughan et al., 2015; Xi et al., 2017; Yang et al., 2018). Nonetheless, many subsequent reports confirmed the appearance of ectopic Krt5+ cell expansion into the alveoli (Ray et al., 2016; Xi et al., 2017; Yang et al., 2018; Zuo et al., 2015), and the occurrence of this phenomenon is no longer considered controversial.

Spurred by these studies in mice, recent work has confirmed that basal cell-like expansion occurs readily in human cases of severe lung injury (Taylor et al., 2018). The authors of this study specifically examined biopsies and autopsy samples of ARDS patients diagnosed with DAD, finding that 39% of these cases exhibited peribronchiolar basaloid pods staining for KRT5 and p63, which the authors remarked “mirrored data from murine models of lung injury/regeneration.” These data also seem to mirror the original 1958 report of Hers et al., acknowledging of course that without immunohistochemistry one cannot be certain they are describing the same phenomenon. While bronchiolization/honeycomb cysts featuring basal-like cells is a well-described histopathological feature of interstitial lung disease (Plantier et al., 2011; Seibold et al., 2013), it appears that, as in mice, it is also a common feature of more acute lung injury. Importantly, DAD has been frequently observed in COVID-19-related ARDS (Xu et al., 2020), suggesting KRT5+ basal-like cell expansion may play an important role in that setting as well, although this has not yet been directly demonstrated.

The ectopic appearance of Krt5+ pods in the small airways and alveoli after lung injury seems to be correlated to the severity of the disease. It has been exhaustively demonstrated that mild-to-moderate injury induces AT2 proliferation and differentiation into AT1s (Barkauskas et al., 2013; Evans et al., 1975), re-establishing the gas-exchanging alveolar epithelial surface without the formation of bronchiolization/dysplasia. However, there appears to be a threshold for depletion of the alveolar epithelium after which barrier restitution can no longer be achieved solely by the surviving AT2s. One very informative report took advantage of a relatively mild H3N2 influenza virus, HK-x31, which causes airway injury but only minor AT2 loss and no Krt5+ cell expansion. Only by first ablating ∼50% of AT2s (via diphtheria toxin) and then infecting with HK-x31 was Krt5+ cell expansion observed (Yee et al., 2017). Although the numerical threshold of alveolar epithelial cell death required to observe Krt5+ expansion is unclear, it appears that in extreme cases of AT2 loss an alternate epithelial repair pathway must be engaged to maintain barrier function and prevent death of the organism (Vaughan et al., 2015). It is important for future studies to assess whether Krt5+ dysplastic expansion is a function only of a threshold loss of alveolar epithelium or whether injury-associated cues, perhaps particular inflammatory cytokines, might be required for basal-like cell expansion to occur.

The Dysplastic Cell-of-Origin in Mice and Humans

While expansion of Krt5+ cells after severe injury is dramatic, there are nonetheless no pre-existing Krt5+ or p63+ cells present in the alveoli in either mice or humans, so the origin of these cells was not immediately obvious. Extensive lineage-tracing studies have since elucidated the cell-of-origin of post-injury Krt5+ pods in the murine lung. Utilizing the p63-CreERT2 mouse (Lee et al., 2014), it has been established that virtually all Krt5+ pods arise from pre-existing intrapulmonary p63+ progenitors (Xi et al., 2017; Yang et al., 2018). These cells are absent in the alveoli at baseline and instead reside in the intrapulmonary airways before their injury-induced activation and migration (Vaughan et al., 2015; Yang et al., 2018; Zuo et al., 2015). Despite their localization to the airways, similarities in cellular markers, and common embryonic origin (Yang et al., 2018), intrapulmonary p63+ progenitors in mice constitute a distinct stem cell population from tracheal basal cells, as bona fide basal cells from the trachea do not appear to migrate into the alveoli after injury as demonstrated by transplant of fluorescent tracheas into non-fluorescent isogenic mice followed by influenza infection (Vaughan et al., 2015). Instead, these rare intrapulmonary p63+ cells become activated by injury and proliferate and migrate significantly, to result in the expansive “blooms” of Krt5+ pods in injured alveoli.

One point of contention surrounding the cell-of-origin of Krt5+ pods concerns their markers during homeostasis aside from p63. It was originally thought that intrapulmonary p63+ progenitors expressed the basal cell marker Krt5 and its heterodimerization partner Krt14 before injury, similarly to basal cells in other epithelia (e.g., the trachea, skin, mammary gland). As such, two of the original studies documenting the origin of Krt5+ pods utilized both Krt5-CreERT2 and Krt14-CreERT2 lineage-tracing methods to isolate and mark intrapulmonary p63+ progenitors (originally called distal alveolar stem cells [DASCs]) before and after injury (Kumar et al., 2011; Zuo et al., 2015). Later studies indicated that the majority of intrapulmonary p63+ progenitors do not express appreciable Krt5 at baseline, and that the abundance of Krt5/Krt14-traced p63+ progenitors originally noted was due to tamoxifen persistence (Vaughan et al., 2015). Indeed, tamoxifen persistence and prolonged recombination have been reported in other systems and can severely confound interpretations of lineage-tracing experiment results (Reinert et al., 2012). In this scenario, inadequate chase periods between administration of the final tamoxifen dose and subsequent injury results in residual tamoxifen inducing recombination in intrapulmonary p63+ progenitors as they respond to injury and upregulate Krt5, explaining initial reports that Krt5+ pods are derived from pre-existing Krt5+ cells. Instead, lineage tracing indicates that most pods are derived from non-Krt5-expressing progenitor cells and that only 15%–20% of post-injury Krt5+ pods arise from cells expressing appreciable Krt5 (Ray et al., 2016; Vaughan et al., 2015). Considering that most Krt5+ pods come from a Krt5negative source and all Krt5+ pods are derived from intrapulmonary p63+ progenitors, it was surmised that most of these cells adopt Krt5 expression in response to injury (Vaughan et al., 2015). These findings have been confirmed through reports that pre-existing Krt5-expressing cells make minimal contributions to post-injury Krt5+ cell expansion (Ray et al., 2016; Yang et al., 2018; Zheng et al., 2014).

There is also debate as to whether intrapulmonary p63+ progenitors express the club cell marker Scgb1a1 (also known as CC10). Club cells themselves exhibit stem/progenitor properties in the airways, demonstrated by their ability to dedifferentiate into basal cells (Tata et al., 2013) and their ability to regenerate the distal airway epithelium after chemical injury (Rawlins et al., 2009). Indeed, one report suggested that Krt5+ cells arise from transdifferentiated club cells (Zheng et al., 2014). In Vaughan et al., 2015, the CC10-CreERT2 (aka Scgb1a1-CreERT2) mouse was used to label club cells before injury to determine if they contributed to alveolar regeneration. As long as proper chase time was used, regenerated cells in the alveoli that arose from pre-existing club cells were found to be entirely separate from post-injury Krt5+ pods, suggesting that pre-existing CC10+ club cells do not give rise to Krt5+ dysplastic cysts (Vaughan et al., 2015). This too has been confirmed by reports of contribution of club cells to ≤1% of Krt5+ pods after injury (Ray et al., 2016). However, others refute this finding, reporting variable contribution of pre-existing CC10-labeled cells to post-injury Krt5+ pods (Yang et al., 2018; Yuan et al., 2019; Zheng et al., 2014). It should be noted that the tamoxifen washout periods used in these publications differs. Tamoxifen persistence from the relatively short chase period of 1 week could explain the higher percentage of CC10-labeled Krt5+ pods described in some studies (Zheng et al., 2014) as opposed to lower percentages found with a longer chase period of 21 days (Yang et al., 2018). Possibly contributing to this ambiguity, the CC10-CreERT2 line is notable for labeling cells outside of the club cell lineage, exhibiting recombination in up to 8% of AT2s (especially those residing near terminal airways) (Rawlins et al., 2009) and a significant number of ciliated cells (Kathiriya et al., 2020). Other reasons for the discrepancies in the reported contribution of club cells to post-injury Krt5+ areas remain unclear, especially since all aforementioned publications utilize the same CC10-CreERT2 mouse (Rawlins et al., 2009). Still, the uncertainty of CC10 expression in some intrapulmonary p63+ progenitors does not alter their fundamental identity as an intrapulmonary basal-like progenitor cell which can be uniformly characterized by p63 expression.

Importantly, bronchiolization/dysplastic remodeling appears histologically quite similar in both mice and humans (Taylor et al., 2018), indicating that this response is likely evolutionarily conserved among mammals. That said, recognizing that important cellular and anatomical differences exist across species (nicely reviewed in Basil and Morrisey, 2020), the cell-of-origin for alveolar Krt5+ expansion may not itself be conserved. While classic basal cells (Krt5+ p63+) persist only as far distally as the mainstem bronchi in mice, larger mammals, including humans, possess basal cells persisting distally all the way to the terminal bronchioles (Evans et al., 2001). It therefore seems quite possible that post-injury bronchiolization in humans is derived from classic basal cells rather than the rare intrapulmonary p63+ KRT5negative progenitors present in mice. However, several groups have also observed p63+ cells lacking KRT5 and KRT14 staining present in distal airways of human lungs (Smirnova et al., 2016; Vaughan et al., 2015), which could possibly preferentially expand into the alveoli upon injury akin to the intrapulmonary p63+ cells in mice. Further complicating this question is the recent identification of additional “basaloid” cells present in bronchiolized epithelium identified by single-cell RNA sequencing of human idiopathic pulmonary fibrosis samples. These aberrant basal-like cells express some basal cell markers (p63, KRT17) but lack KRT5 (Adams et al., 2020; Habermann et al., 2020). Whether this KRT5negative basaloid expansion occurs after DAD, especially that induced by respiratory viruses, is unknown. Finally, it is also possible that, in humans, AT2s upregulate basal cell genes to give rise to KRT5+ dysplasia, as has been observed in in vitro hypoxia experiments (Xi et al., 2017), although this does not occur in mice (Vaughan et al., 2015). While lineage-tracing studies cannot be performed in humans, it may be possible to begin to interrogate the cell-of-origin of human lung bronchiolization utilizing cell barcoding in human precision-cut lung slices ex vivo or to at least make inferences from pseudotime analysis of single-cell transcriptomic analyses. Ultimately, while the cell-of-origin may differ, dysplastic expansion of basal-like cells occurs quite similarly across species, or at least in mice and humans.

The Acute Role and Ultimate Fate of Intrapulmonary p63+ Progenitors

In light of the apparent destruction of the native alveolar epithelium, it has been assumed that intrapulmonary p63+ progenitors are important immediately after alveolar injury to re-epithelialize the denuded alveolar epithelium and restore barrier function. Perhaps the best evidence for this is the demonstration that diphtheria toxin-mediated ablation of expanding Krt5+ cells results in mice with permanent, dramatic reduction in oxygen saturation and the development of profound, non-resolving immune infiltrates and fibrosis (Zuo et al., 2015). Moreover, as discussed earlier, mice in which AT2s are experimentally depleted have increased levels of Krt5+ pod expansion after influenza infection, suggesting that intrapulmonary p63+ progenitors can “sense” disturbances in the alveolar epithelium and act quickly to “remedy” epithelial defects (Yee et al., 2017). These studies imply that expansion of p63+ cells into Krt5+ dysplastic epithelium serves to rapidly restore epithelial barriers in acute injury settings. But what then becomes of these alveolar Krt5+ cells?

Despite the growing consensus that intrapulmonary p63+ progenitors primarily give rise to Krt5+ pods and ultimately bronchiolized epithelium as opposed to alveolar cell types (Table 1), some studies disagree, instead indicating that they can readily differentiate into AT1s and AT2s (Kumar et al., 2011; Zuo et al., 2015). For example, some reports utilize podoplanin (Pdpn) expression as a specific marker of AT1s and observe Krt5 lineage-traced PDPN+ cells after recovery from injury. This is interpreted as Krt5+ cells differentiating into AT1 cells to regenerate normal alveolar epithelial structures (Zuo et al., 2015). However, Pdpn is also expressed by basal cells, thus providing no definitive distinction between dysplastic tissue and regenerated AT1s (Rock et al., 2011; Van de Laar et al., 2014). This same study indicated that ∼70% of Krt5 “lineage-labeled cells ultimately express type I (1H8+ and PDPN+) or type II (SPC+) pneumocyte markers,” but the relative proportion of cells stained by each marker are not indicated (Zuo et al., 2015). Earlier studies used proprietary monoclonal antibodies that marked alveoli in humans and rats to identify supposed alveolar cells derived from intrapulmonary p63+ progenitor precursors (Kumar et al., 2011). While these antibodies appear to label regenerated alveolar regions, the ambiguity of which antigens are recognized by these antibodies makes precise interpretation of intrapulmonary p63+ progenitor contribution to alveolar epithelial regeneration difficult (Kumar et al., 2011).

Table 1

Differentiation of Intrapulmonary p63+ Basal-like Cells after Influenza Injury

Intrapulmonary p63+ Basal-like Cell Fate after Influenza Injury	Endogenous Differentiation Frequency	Notes and Evidence
AT2 cells	very rare∗	∗Differentiation potential of intrapulmonary p63+ cells into AT2 cells is higher after bleomycin injury, up to ∼30% of Krt5-traced cells.
AT1 cells	very rare	PDPN expression is shared by basal cells, so reports of Krt5-traced AT1 cells are inconclusive
Club cells	common	CC10 and SCGB3A2 staining
Goblet cell	common	MUC5AC and TFF2 staining
Ciliated cells	unknown
Tuft cells	common	DCLK1 staining and Trpm5-GFP reporter
Neuroendocrine cells	unknown
Ionocytes	unknown

All this said, several studies have observed some, albeit quite rare (typically less than 10%), differentiation of Krt5+ cells into pro-surfactant protein C+ (SPC+) AT2s after influenza utilizing well-validated rabbit anti-sera. This appears to be dependent on the nature of injury, as ∼30% of Krt5 lineage-traced cells exhibit AT2 differentiation upon resolution of bleomycin injury (Vaughan et al., 2015), although many fewer Krt5+ cells arise in the first place with bleomycin. So, while most studies dispute the idea that intrapulmonary p63+ progenitors readily differentiate into alveolar cell types, many agree that some plasticity is possible. Manipulating this plasticity may prove useful for future strategies aimed at enhancing functional lung regeneration.

Signals Controlling Intrapulmonary p63+ Basal-like Cell Activation and Fate

The signaling pathways upstream of intrapulmonary p63+ progenitors' activation, mobilization, and differentiation offer critical insights into how cell fate trajectories might be experimentally modulated, especially to promote transdifferentiation into AT2 and AT1 cells. Many studies into perturbations of these pathways utilize the Sox2-CreERT2 mouse to induce signaling changes before injury, so Sox2 expression in the adult lung warrants careful description here. Sox2 is expressed in all airway cells, including intrapulmonary p63+ progenitors, which give rise to Krt5+ pods, but also p63negative Sox2+ progenitors that appear more capable of AT2 differentiation (Kathiriya et al., 2020; Vaughan et al., 2015; Xi et al., 2017). Both intrapulmonary p63+ progenitors and AT2-biased Sox2+ progenitors have been grouped under the catch-all term lineage-negative epithelial progenitors (LNEPs), and studies using the Sox2-CreERT2 mouse effectively target this entire heterogeneous LNEP population (Xi et al., 2017). Moreover, so-called bronchoalveolar stem cells present in distal airways are almost certainly Sox2 lineage labeled as well (Kim et al., 2005; Liu et al., 2019), not to mention the possibility that Sox2-labeled club cells may also differentiate into AT2s (Guha et al., 2017; Yuan et al., 2019). While these additional progenitor lineages are of great interest, existing reviews provide an excellent overview of these other cell types, and further discussion here is outside the scope of our focus on intrapulmonary p63+ progenitors.

Accordingly, studies utilizing Sox2-CreERT2 mice can be difficult to interpret since changes to signaling pathways will occur in all airway cells, making it impossible to deconvolute which specific cell type was primarily affected by the signal modulation. When studying pathway modulation after the onset of injury, many groups use Krt5-CreERT2 mouse, which specifically targets post-injury intrapulmonary p63+ progenitors and is far easier to interpret since Krt5 specifically marks injury-activated intrapulmonary p63+ progenitors (see “Cell-of-Origin” section).

Severe pulmonary insults cause widespread alveolar epithelial cell death, stripping some of the alveoli of their critical oxygen-exchanging cells and promoting edema, generating sporadic hypoxic microenvironments within injured lungs. Intriguingly, post-injury Krt5+ pods are found only in hypoxic regions of the damaged lung, which suggested hypoxia-inducible factor (HIF1α) as a likely mediator of cellular responses to severe injury (Xi et al., 2017). When the Sox2-CreERT2 mouse is used to both label airway progenitors and delete HIF1α (Hif1a) before injury, Krt5+ pod expansion overall is not only significantly reduced, but the Sox2-labeled Hif1a−/− cells that do expand differentiate into highly proliferative AT2s instead (Xi et al., 2017). Again, it is unknown whether intrapulmonary p63+ progenitors, p63negative Sox2+ progenitors, or both are reacting to Hif1a deletion in this study due to the broad labeling of the Sox2-CreERT2 mouse. At the molecular level, Notch signaling and the basal-like differentiation program are highly induced in hypoxic in vitro culture of Sox2-enriched populations; both are abolished with Hif1a deletion, suggesting that HIF1α-dependent Notch signaling is a driver of the basal-like phenotype, either promoting this fate pathway in intrapulmonary p63+ progenitors or AT2-biased Sox2+ progenitors (Xi et al., 2017).

Notch signaling is critical for Krt5+ expansion both during the course of injury and after its resolution, and influenza-injured lungs harbor intense expression of the canonical Notch target Hes1 in Krt5+ cysts (Vaughan et al., 2015). Notch inhibition shortly after influenza injury (5 days post-infection) significantly attenuates Krt5+ expansion, while Notch inhibition both in vitro and at later time points in vivo (31 days post-infection) increases the differentiation of Krt5+ cells into SPC+ AT2s, again demonstrating that some plasticity toward alveolar cell types is present in these cells (Vaughan et al., 2015). This was foreshadowed by developmental studies years earlier demonstrating that persistent Notch signaling is incompatible with alveolar differentiation (Guseh et al., 2009). It should be noted, however, that high-dose γ-secretase inhibition combined with known alveolar differentiation factors IBMX and dexamethasone still only resulted in ∼10% of Krt5+ cells transdifferentiating into AT2s, indicating that Notch inhibition alone is insufficient to induce widespread alveolar conversion.

Recognizing that intrapulmonary p63+ progenitors can differentiate into AT2s (albeit rarely) under certain conditions, the role of Wnt signaling, a potent pro-differentiation and proliferation signal for AT2s (Nabhan et al., 2018; Zacharias et al., 2018), was investigated. β-Catenin stabilization in vivo before injury using Sox2-CreERT2 increases the percentage of Sox2-traced SPC+ cells, while the amount of Krt5+ cells decreases, indicating that increased Wnt signaling is sufficient to drive an airway cell-to-AT2 transition (Xi et al., 2017). However, as with HIF1α deletion, it is unclear which Sox2+ progenitor responds to β-catenin stabilization to promote AT2 differentiation (Xi et al., 2017).

Importantly, “activated” intrapulmonary p63+ progenitors post-influenza are refractory to many of these pro-alveolar differentiation signals. When the Krt5-CreERT2 mouse was used to delete Hif1a or stabilize β-catenin in injured intrapulmonary p63+ progenitors, the fraction of Krt5-traced cells that differentiate into AT2s remains unchanged (Xi et al., 2017). This finding suggests at least two possibilities: (1) a change—possibly transcriptional, epigenetic, or both—takes place in intrapulmonary p63+ progenitors after injury that narrows their fate potential dramatically, rendering them unaffected by signaling pathways that could have otherwise altered their fate trajectory, or (2) intrapulmonary p63+ progenitors cannot be induced to differentiate into AT2s via the HIF1α or Wnt signaling pathways at all, and the changes in Sox2-traced AT2s observed using the Sox2-CreERT2 mouse to delete Hif1a or stabilize β-catenin primarily affected the Sox2+ p63negative progenitor subpopulation and not intrapulmonary p63+ progenitors. The outcome of increased AT2 generation in these experiments can therefore be interpreted as competition between heterogeneous progenitor cell types (p63+ and p63negative) as opposed to true fate conversion of intrapulmonary p63+ progenitors.

Akin to Wnt signaling, FGF10, which signals through the receptor FGFR2B, is known to be an important niche-specifying signal for AT2s and was investigated as a potential fate-modifying signal in the airway progenitor population. Fgfr2b deletion before bleomycin injury using the Sox2-CreERT2 mouse abrogates both Krt5+ pod and AT2 expansion from the Sox2+ progenitors, implicating a role for FGF signaling in airway contribution to alveolar regeneration as a whole (Yuan et al., 2019). Conversely, sustained FGF10 signaling in all Sox2-traced cells beginning 2 weeks after bleomycin injury resulted in a significant shift favoring AT2 fate over Krt5+ fate (Yuan et al., 2019). Again, it is unclear in which Sox2+ cell type this FGF10-induced AT2 fate promotion occurs. To this end, the Krt5-CreERT2 mouse was utilized to delete Fgfr2b after injury, resulting in similarly lowered Krt5+ area outgrowth (Yuan et al., 2019). These data, along with reports that FGF10 activation in vitro augments intrapulmonary p63+ progenitor proliferation and that administration of recombinant FGF10 to influenza-injured mice results in more Krt5+ pod areas (Quantius et al., 2016), confirms that FGF10 signaling indeed promotes intrapulmonary p63+ progenitor expansion after injury.

In addition to external signals that influence intrapulmonary p63+ cell fate, an understanding of cell-autonomous transcriptional control underlying cell fate and activation is essential toward modulating lung repair responses. The essential basal cell transcriptional regulator p63 (specifically the ΔN isoform) has been implied as an important player in intrapulmonary p63+ progenitor maintenance largely due to the finding that p63-CreERT2 mice have fewer intrapulmonary p63+ progenitors before injury and exhibit a dampened Krt5+ outgrowth response to influenza (Yang et al., 2018). The authors attribute this attenuated response to p63 haploinsufficiency due to the CreERT2 knock-in allele at the Trp63 locus, lending credence to the possibility that loss of even a single p63 allele can disrupt maintenance of the this cell population at baseline and impede dysplastic outgrowth. These data suggests that dysplasia and the maintenance of resting and activated intrapulmonary p63+ progenitors relies on p63 expression. Similarly, Sox2 has been implicated as exerting transcriptional control over the basal-like program of intrapulmonary p63+ cells. As with p63-CreERT2, attenuated Krt5+ pod expansion after influenza has been noted in Sox2-CreERT2 mice (in which CreERT2 replaces the Sox2 coding region) compared with non-transgenic controls, suggesting that functional copies of both Sox2 alleles are necessary for normal intrapulmonary p63+ cell response to injury (Ray et al., 2016). It has not been determined if Sox2-CreERT2 mice have fewer intrapulmonary p63+ progenitors at baseline, and so it is unknown if Sox2 haploinsufficiency results in a defective response to injury or if Sox2 regulates intrapulmonary p63+ progenitor maintenance as well.

Tuft Cells—an Elusive Cell with a “Sense” of Purpose

While much focus has been placed on AT2/AT1 differentiation (or lack thereof) from intrapulmonary p63+ progenitors, less attention has been directed to the airway-related cell types they more readily give rise to. One such cell arising within dysplastic regions of the lung post-influenza are tuft cells (Rane et al., 2019), also called brush or solitary chemosensory cells, that were shown to be derived from intrapulmonary p63+ progenitors (Rane et al., 2019). The function of these cells and their downstream effects in the lungs are currently being investigated, but it has been demonstrated that tuft cells are responsible for activation of Th2-polarized immune response through the secretion of interleukin-25 in the intestines (Gerbe et al., 2016), as well as the trachea (Bankova et al., 2018; McGinty et al., 2020) and nasal epithelium (Kohanski et al., 2018). Given their clear role in epithelial remodeling in the intestine, it is possible that crosstalk between Krt5+ cells and tuft cells contributes to chronic inflammation and persistent pathological changes in the lung parenchyma (Keeler et al., 2018; Rane et al., 2019). Stimulation of tuft cells in the mouse lung results in increased permeability and edema (Rane et al., 2019) and promotes generation of even more tuft cells in a “feedforward” loop, as has been described for intestinal tuft cells. The physiological impact of ectopic tuft cells in post-influenza lungs and their ultimate relationship to intrapulmonary p63+ progenitor-derived dysplastic epithelium remains to be elucidated.

Functional Characterization of Post-injury Keratinization/Dysplasia—“Band-Aid” or Permanent Maladaptation?

Because intrapulmonary p63+ progenitor cells and their progeny occupy alveolar space where AT2/AT1 restoration and functional regeneration might otherwise occur, and given that differentiation of these cells into AT2/AT1s is infrequent in most cases (Kanegai et al., 2016; Vaughan et al., 2015), it must be considered whether p63+ progenitors and their derivatives are potential contributors to the long-term compromise of pulmonary function seen in some human survivors of severe pulmonary injury (Koppe et al., 2016; Liu et al., 2015). The development of chronic inflammation (Keeler et al., 2018; Rane et al., 2019) and bronchiolized epithelium within previously gas-exchanging areas of the lung almost certainly have deleterious effects on long-term lung health. Physiological impairment and decreased oxygen saturation would be one of the most obvious and immediate results of the substitution of AT2/AT1 cells by Krt5+ dysplastic epithelia in the distal airspace, assuming that this dysplastic epithelium does not perform effective gas exchange. In addition to the Krt5+ dysplasia itself, in mice chronic inflammation persists in the lung in areas where viral remnants are detectable (Keeler et al., 2018). Moreover, some club cells appear capable of surviving infection but exhibit transcriptomic alterations (likely epigenetic) that enable broad, temporary antiviral immunity but may also contribute to long-term dysfunction (Hamilton et al., 2016). Indeed, several smaller studies and case reports in influenza survivors demonstrate long-term dysfunction, which may be at least partially attributable to epithelial dysplasia. For instance, nearly 50% of patients who survived infection with an avian-origin H7N9 influenza A strain in 2013 exhibited long-term (≥2 years) lung pathology as judged by imaging (fibrosis, ground-glass opacities) and lung function defects (carbon monoxide diffuse capacity, FEV) (Chen et al., 2017). Similar but more minor chronic pathology has been observed in survivors of H1N1-associated ARDS (Luyt et al., 2012). All this said, it is true that most ARDS patients do seemingly show full recovery over time, although even this does not necessarily mean Krt5+ dysplasia has resolved. Instead, less-injured regions of the lung may undergo compensatory regrowth, similar to what occurs after pneumonectomy, essentially resulting in the generation of new alveoli via expansion and neo-alveolarization of remaining (or in this case, uninjured) lung lobes to compensate for the loss of the gas exchange area (Buhain and Brody, 1973). Still, if the persistence of intrapulmonary p63+ progenitor-derived dysplasia accounts for even a fraction of observed long-term dysfunction, future studies focused on resolving dysplasia would be expected to have significant impacts on the restoration of lung function.

Blind Men and an Elephant

Despite the many studies focused on injury-induced Krt5+ cell expansion, there is a conspicuous lack of shared terminology, making a unified model of this unique regenerative phenomenon difficult. Adoption of a naming convention would unify the disparate terminology of DASCs and the p63+ subset of LNEPs. While there is some disagreement as to whether these intrapulmonary p63+ basal-like cells express additional markers (Krt5, Scgb1a1) at baseline, many other characteristics are agreed upon: their scarcity in mice, presence in intrapulmonary airways, p63 expression, and the fact that these cells are the sole source of post-influenza Krt5+ dysplastic cell expansion. Although not unanimous, most studies reviewed here agree that intrapulmonary p63+ progenitors' inherent capacity to generate alveolar cell types is limited. Nonetheless, other p63negative airway cell types seem quite capable of transdifferentiating into AT2s, as demonstrated by both Sox2- and CC10-lineage tracing, recent studies utilizing dual SPC- and CC10-lineage tracing, and orthotopic transplantation of LNEPs. In retrospect, the fact that LNEPs are a heterogeneous pool of both intrapulmonary p63+ progenitors and less well-defined Sox2+ p63negative/H2-K1high (Kathiriya et al., 2020) cells has only confused the issue of p63+ cell fate potency. For the time being, referring to these cells explicitly as intrapulmonary p63 progenitors should help clarify that these emergency responders are fundamentally distinct from other airway progenitor cells which, unlike intrapulmonary p63+ progenitors, exhibit higher potential for AT2/AT1 differentiation.

Conclusions—Is Something Better than Nothing?

While the appearance of dysplastic, keratinized areas after severe influenza has been described extensively, it remains an entity of much debate. Though some authors argue that p63+ Krt5+ cells that compose these “epithelial scars” can resolve and differentiate into AT2s and AT1s, this view is not supported by most subsequent studies. Most studies instead support the view that honeycomb cysts, or keratinized pods, are a near-permanent and indelible result of alveolar progenitor cell (AT2) depletion and, while beneficial in the acute phase of injury, are maladaptive in the long term (Kanegai et al., 2016; Rane et al., 2019).

Careful analysis and interpretation of these studies might lead to the conclusion that, in the absence of able progenitors, any kind of epithelial repair is better than none. The general recruitment of airway cells into the alveoli for the purposes of repair (Guha et al., 2017; Kathiriya et al., 2020; Liu et al., 2019; Ray et al., 2016; Strunz et al., 2020; Yee et al., 2017) further supports this hypothesis. In this view, albeit simplistic, the propagation of p63+ Krt5+ cells in the immediate aftermath of distal lung injury with significant loss of AT2 progenitors represents a much-needed temporary solution to a grave problem. The initial presence and proliferation of intrapulmonary p63+ progenitors would not be the problem, per se, but rather their persistence. Still, the fact that these cells seem to possess some, if limited, plasticity to generate alveolar cell types presents a strong rationale for further investigation into the signaling and epigenetic status of intrapulmonary p63+ progenitors. We posit that identification of means to reprogram/transdifferentiate intrapulmonary p63+ progenitors into AT2s and AT1s represents a promising therapeutic target, possibly converting detrimentally remodeled tissue into truly regenerated, functional lung.

Author Contributions

A.I.W. and M.F.d.M.C. contributed equally to the writing of the manuscript, under the supervision of A.E.V.