To the Editor:
I read with interest the report by Davis et al. [1] on epithelial injury-shedding in vivo caused by tobacco smoke. Their model involves airways equipped with human-like subepithelial microvessels carrying systemic blood. By contrast, mice lack a bronchial microcirculation. Severe bronchial epithelial damage is induced with widespread cell loss from an apparently intact epithelial basement membrane. By weeklong pre-treatment with simvastatin, the smoke-induced epithelial derangement and associated inflammation are largely prevented. Although a high dose was used, the drug effect is impressive. The authors apparently excluded: 1) involvement of oxidative stress, as reflected by haem oxidase-1, in the smoke-induced epithelial desquamation; and 2) mevalonate-dependent mechanisms in the “therapeutic” effect of simvastatin. They discuss the possibility that simvastatin inhibits granulocyte-induced epithelial injury but also mention that the drug may target epithelial cells directly. Most previously reported anti-inflammatory actions of simvastatin are dependent on the mevalonate pathway, making the authors’ observations stand out. Incidentally, mevalonate independence is compatible with direct epithelial effects of simvastatin involving inhibition of interferon regulatory factor 3 [2]. By this mechanism, simvastatin reduces epithelial production of select cytokines/mediators, but there is no known link with epithelial cytoprotection. Perhaps the bronchoalveolar fluid collected by Davis et al. [1] can provide further information on released epithelial factors? For example, to what extent did the drug preserve epithelial innate immunity function?
During recent decades the epithelial field has, to a large extent, been addressed by cell culture approaches, hiding the in situ epithelium’s interaction opportunities. Particularly important is the supply of bioactive molecules and cells from a profuse subepithelial microcirculation. The in vivo approach taken by Davis et al. [1] is, therefore, commendable. However, they fail to mention that mere loss of epithelial lining cells, without inflammatory insults, will initiate a series of significant acute, as well as more sustained, effects in vivo [3–6]. A discussion of effects of epithelial injury-repair processes and their sequelae is justified in view of the dramatic denudation produced in the in vivo study by Davis et al. [1]. This comment is strengthened by potential involvement of epithelial shedding-induced effects in inflammatory and remodelling features of asthma and chronic obstructive pulmonary disease (COPD) [3, 6–8]. Indeed, the damaged epithelium illustrated by Davis et al. [1] may be more reminiscent of asthmatic than COPD bronchi, perhaps making their model one of severe asthma overlapping with COPD.
My arguments are underpinned by original work carried out by Ingrid and Jonas Erjefält and others in the mid-1990s. Unfortunately, there appears to have been limited development since then regarding epithelial shedding-restitution events and sequelae in vivo in human-like airways [7, 8]. I acknowledge that leading laboratories have advanced asthma and COPD research immensely with sophisticated culture methods revealing intriguing differences between health and disease, and I apologise for not reviewing in vitro experiments here. I also appreciate that the contribution from the microcirculation is like opening a Pandora’s Box, potentially blurring favoured molecular in vitro mechanisms. Hence, in view of prevailing priorities in biological–medical research, the aspects discussed here are not frequently entertained in the current literature. Yet, in a discussion of the in vivo findings of Davis et al. [1], they may merit a revisit.
If sloughing of epithelium leaves an intact basement membrane, epithelial repair starts immediately and proceeds rapidly (over a damaged basement membrane, repair is severely delayed as it may be in cell culture). All cells bordering the denuded area in vivo participate in the repair [4]. Thus, ciliated and secretory cells, once considered end-differentiated cells, internalise their cilia and lose their secretory granules, respectively; this occurs within few minutes after denudation, as these cells promptly transform into flattened repair cells that speedily (several microns per minute) migrate to produce a new cell cover [4]. There is thus little need for recruitment of epithelial progenitor cells from the circulation. The speed with which patchy denuded areas can be covered in vivo may explain, in part, why increased bronchial absorption permeability has been so difficult to demonstrate in asthma and COPD patients [3, 5, 7].
The basement membrane appears to be unharmed in the study by Davis et al. [1], yet the authors do not report signs of epithelial repair in their sections of bronchial tissue. A problem may be that sectioning itself causes denudation in airways with a fragile epithelium [8]. Hence, whole-mount tissue preparation, rather than sectioning, is probably a preferable method for demonstrating sites of epithelial damage repair [9]. Observations of whole mounts have revealed that the injury caused by inhaled noxious agents is exceedingly patchy. This is so even when it is ensured that the mucosal surface area is equally exposed in vivo to a culprit challenge [7, 9]. In an inflamed airway, actual patches of epithelial sloughing thus occupied <1% of the surface area. However, this limited injury was seen as ≥20% frank denudation in cryostat sections; no denudation occurred in sections of uninflamed airways [9]. It should thus be of interest to see the distribution of epithelial sloughing induced by tobacco smoke in whole-mount preparations of the exposed mucosal surface.
Denudation occurs without any bleeding [3]. However, signals from the denudation site go to the profuse subepithelial microcirculation, causing prompt and lasting extravasation of bulk plasma from post-capillary venules [10, 11]. Similarly, extravasation of leukocytes is also prominent. Activated neutrophils may dominate, but in an already eosinophilic mucosa, the eosinophils join in and become activated [11]. Eosinophils degranulate by the disease-relevant mode of primary cytolysis [3, 6] and the resulting free eosinophil granules abound at patchy sites of epithelial injury/repair [12]. The early phases of epithelial restitution after denudation thus occur in a dynamic fibrin–fibronectin gel replete with activated granulocytes (epithelial cells in the area of interest apparently do not produce fibrin–fibronectin [10]). Epithelial hotspots of shedding restitution processes also send out signals causing hypersecretion from surrounding mucosal goblet cells (fig. 1) [4].
As soon as the denuded basement membrane has become well covered by a layer of migrating repair cells (note that these undifferentiated cells may mistakenly be reported as normal basal cells), the leukocyte-rich plasma protein-derived gel ceases to be fed from the microcirculation. It becomes degraded and is shed into the airway lumen. A granulocytic inflammatory picture, involving deranged epithelium, agrees with features of severe states of asthma and COPD. An inflammatory exudate, probably produced in part by the events discussed above, also correlates well with severity of asthma and COPD [13]. Indeed, Hogg et al. [14] reported that the presence of an inflammatory exudate best correlated with the severity of COPD. In summary, it is possible that injury repair processes themselves evoke part of the inflammatory responses to the smoke-induced epithelial injury [1]. The inhibition of this inflammatory response by simvastatin [1] would then reflect epithelial cytoprotection as a primary drug action.
Davis et al. [1] did not examine, nor did they discuss, effects on airway remodelling in their epithelial injury model. However, epithelial denudation repair involves a significant second phase with proliferation, metaplasia and differentiation of the epithelial lining cells. Epithelial metaplasia may characterise COPD bronchi but little evidence is really available. The epithelial restitution process further induces proliferation of subepithelial fibroblast and smooth muscle cells [4]. In addition, by repeated shedding restitution events, the epithelial sub-basement membrane may thicken [5]. However, this feature is not typical in COPD; moreover, it can be produced, along with several other remodelling features in asthma, by noninflammatory bronchoconstriction [15]. The structural proliferation changes associated with epithelial repair outlast responses such as inflammatory cell recruitment/activation and plasma exudation [4, 11]. This observation is of interest because severe asthma can exhibit epithelial–mucosal histopathology without “inflammation” [15]. Even more interesting perhaps, Creola bodies (clusters of desquamated epithelial cells) in aspirated sputum, independent of other factors, were reported to predict development of asthma in children [16]. As observed by Naylor [17], Creola bodies in sputum have also been significantly associated with exacerbations of asthma.
The epithelium holds the stage in current asthma/COPD research. The focus is on roles of epithelial cells as drivers of immunopathogenic mechanisms. This well-justified interest in molecular capacities of the strategically located epithelium has probably eclipsed a primary interest in the shedding–restitution aspects discussed above. Perhaps the predominance of cell culture studies over more demanding, but relevant, in vivo approaches has contributed to this balance. Further in vivo discoveries of mechanisms and pharmacological control of epithelial shedding and its sequelae are needed, I think, both to validate current notions and to develop novel research paradigms in this field of interest. The in vivo approach by Davis et al. [1] may turn out to be instrumental for therapeutic developments. There is an unmet medical need for “cytoprotective” drugs that reduce occurrence of epithelial injury in diseased bronchi.
Footnotes
Conflict of interest: None declared.
- Received June 13, 2013.
- Accepted September 4, 2013.
- ©ERS 2014