Role of Surfactant and Hyperoxia

To the Editor: The recent companion papers by Huang et al. and Piantadosi et al. at Duke (13, 15) represent a most interesting study into the pathophysiology of hyperoxic lung injury and its attenuation by exogenous surfactant. Their finding that “the interstitial PMN (polymorphonuclear leukocyte) volume in the O₂ + surfactant lungs was not significantly different than the value for the O₂ group” might suggest a physical rather than abiological cause, especially considering the accompanying fluid shifts, consistent with their morphological evidence for compromised fluid barriers and the “better preservation of epithelial integrity” imparted by exogenous surfactant. It would therefore seem unfortunate that the authors should dismiss physicalexplanations for their findings because “all these surfactant preparations effectively reduce surface tension” (13), as though this were the only physical role for surfactant in the lung.

From the acknowledgments listed, it would appear that the authors (13) might have been advised of only the conventional Hawgood-Clements (H-C; so-called “bubble”) model (15) for the role of surfactant in the alveolus, the debate on the validity of which has been largely suppressed until recently (10, 16). This model is based on the assumption that surface-active phospholipid (SAPL) locates onlyat the air-aqueous interface of a continuous aqueous hypophase assumed to line the alveolus where, for reasons of surface curvature, it must reduce surface tension to “near zero” at end expiration, at least. Such a concept has been described as “absurd” (2) for reasons of basic physics, whereas other shortcomings of this model have been discussed in detail elsewhere (8, 10, 16).

Exosurf has been well designed to spread rapidly across an air-aqueous interface, reducing surface tension in doing so, thus making it well suited for assisting air inflation of a neonatal lung, which starts as essentially “wet.” However, clinical success in treating respiratory distress syndrome does not necessarily validate the H-C model. Surfactants are attracted to many interfaces, including those of solids, and have been much studied in the physical sciences for the many highly desirable properties that they can impart when reversibly bound, i.e., adsorbed to those surfaces in the true physicochemical sense of that word, sometimes termed “chemisorption” (1). Few synthetic surfactants, if any, that reduce the surface tension of water do not adsorb to other interfaces, and this would also appear to apply to SAPL on tissue.

Ultrastructural evidence for a surfactant layer adsorbed to pulmonary epithelium takes three forms: electron microscopy, epifluoresence microscopy, and an appreciable degree of hydrophobicity imparted to pulmonary epithelium when this tissue surface is exposed to surfactant (11). Moreover, at the alveolar level in adult lungs, any surface fluid tends to collect as “pools” at the septal corners, or “pits” elsewhere, rather than wet the surface and form a continuous bubble lining (4). This distribution of water would seem far more likely to arise in the normal alveolus under the influence of homeostatic forces maintaining fluid balance than the continuous liquid lining of the bubble (H-C) model, for which no mechanism for control of its thickness has ever been offered by its proponents. Our efforts, and especially the superb electron micrographs of Ueda and co-workers (17, 19), clearly demonstrate an oligolamellar layer of SAPL lining on epithelium with no intervening aqueous hypophase. Such a coating is confirmed by epifluorescence microscopy (18). Proponents of the bubble model, however, could argue that these are artifacts, because Ueda et al. (19) and ourselves substitute tannic acid for much of the conventional gluteraldehyde as fixative, since aldehydes destroy hydrophobic surfaces (20). Even when using more conventional fixatives for his long and illustrious career devoted to lung morphology, Weibel (21) concludes that “the osmiophilic lining (of surfactant)follows the epithelial surface rather than the interface with air.” Recognized by its characteristic interlaminar spacing of 45 Å, oligolamellar SAPL has also been demonstrated “bridging” intercellular junctions in various tissues, as though rendering them “tight” by “caulking” (3, 9).

Adsorbed layers of synthetic surfactants are widely used industrially as barriers to water or hydrated ions, and this would appear to hold for SAPL, where a monolayer can increase resistance to ions as small as hydrogen ions by an order of magnitude (12). There is also much indirect evidence for barrier properties imparted by truly adsorbed SAPL in vivo (7). Pulmonary SAPL can impart water repellency to a surface otherwise exposed to air (6), whereas, in the absence of air, e.g., in the peritoneal cavity, exogenous SAPL has been shown to improve ultrafiltration when added to the dialysate in continuous ambulatory peritoneal dialysis (14).

Hence, there would seem to be a good case for Huang et al. and Piantadosi et al. (13, 15) to consider the fluid barrier derived from the adsorption of SAPL from Exosurf as a possible physicalmechanism explaining their results on hyperoxic lung injury. Speaking generally, in considering the vast wealth of knowledge of surfactants available in the physical sciences, it seems unfortunate that only those few aspects related to an air-aqueous interface should have been permitted to enter the respiratory literature.