**1. Introduction**

The intra-alveolar surfactant is a mixture of 90% lipids (mainly phospholipids) and 10% proteins including surfactant proteins which are produced, stored and secreted by alveolar epithelial type 2 (AE2) cells [1]. Among the protein component the hydrophobic surfactant proteins B (SP-B) and C (SP-C) are both of high relevance for the surface tension lowering properties of alveolar surfactant at the air-liquid interface within the alveolar space. Underneath this interface, a very thin layer of liquid, also referred to as the hypophase, covers the alveolar epithelium [2,3]. This reduction in surface tension is critical to prevent end-expiratory collapse of distal airspaces and to reduce the work of breathing. Comparing quasi-static pressure-volume (PV) loops of air and liquid-filled healthy lungs it becomes obvious that the hysteresis, e.g., the area within the PV-loop, is mainly a feature of the air-liquid interface. The corresponding surface phenomena such as surface tension, moreover, result in larger pressures needed for lung inflation and a loss of energy during the PV-loop as characterized by the hysteresis [4]. The surface active molecules, such as phospholipids, form a dynamic layer at the air-liquid interface. This layer is mechanically challenged by the cyclic intra-tidal changes in the geometry of alveoli and interalveolar septa which expand during inspiration and form pleats at the end of expiration [5]. Accordingly, the surfactant layer at the air-liquid interface is compressed at end-expiration and reduces the surface tension close to 0 mN/m. During inspiration the surface active film of phospholipids is expanded and the surface tension increases at larger lung volumes [6,7]. SP-B, due to its properties to generate phospholipid membrane-membrane contacts, has been suggested to be involved in the formation and stabilization of surface active films so that surface tension lowering properties are critically dependent on the biophysical properties of SP-B [2,8].

The acute respiratory distress syndrome (ARDS) is characterized by a severe failure of the lungs' central function in gas exchange resulting from an injury of the blood-gas barrier. Accordingly, inflammation, alveolar flooding with surfactant inactivation and alveolar collapse are typical features of ARDS [9,10]. The high surface tension itself has effects on the acinar microarchitecture including its dynamic changes during respiration, also known as alveolar micromechanics. Based on the Wilson-Bachofen model, surface tension at the air-liquid interface results in forces which would induce a piling up of interalveolar septal walls and therefore a reduction in surface area [11]. These surface tension-related forces are counter-balanced by the axial system of elastic fibers which usually surround the alveolar entrance. High surface tension induced by lavage of the lung with tween has been shown to result in a diversity of abnormalities in alveolar micromechanics such as intratidal alveolar recruitment/derecruitment [12] but also asynchronous alveolar dynamics such as inverse alveolar ventilation, alveolar stunning or alveolar pendelluft [13].

Clinical studies demonstrated that a reduction in the SP-B level in the alveolar space represents an early event during the time course and can even precede the development of the complete clinical picture of ARDS [14,15]. In established ARDS, moreover, the level of SP-B in broncho-alveolar lavage fluid (BAL) correlated convincingly with the impairment in surfactant function as characterized by the minimum surface tension [16]. However, in the context of clinical ARDS, the relevance of this early SP-B reduction has not been investigated in detail although it has been suggested that surfactant inactivation plays a central role in the development of ARDS [17]. It is well known that high surface tension leads to edema formation [18] and expiratory alveolar derecruitment [19]. In the conditional SP-B knockout mouse model high surface tension results in pulmonary inflammation, respiratory failure and death [20,21]. However, the mechanisms leading to respiratory failure in SP-B deficiency-induced high surface tension are not entirely understood. Several models of pulmonary micromechanics offer realistic scenarios by which SP-B deficiency induced high surface tension can result in progressive lung injury [22–25]. The failure of surfactant to reduce surface tension during expiration results in alveolar instability with derecruitment of alveolar surface area or even complete alveoli which can be recruited if transpulmonary pressure gradients increase again during inspiration [26]. This repetitive intratidal alveolar recruitment/derecruitment has been observed in lavage models of acute lung injury during mechanical ventilation [12]. There is evidence that repetitive recruitment/derecruitment of distal airspaces can be harmful to the alveolar epithelium thereby contributing to ventilation-induced lung injury (VILI) via a mechanism known as atelectrauma [27]. Computational modelling combined with experimental validation in cell culture model systems provided evidence that the opening of a fluid-occluded distal airspace can, in the presence of high surface tension, be associated with potential harmful pressure gradients acting on the epithelial lining [22]. Accordingly, restoring surfactant function and reducing surface tension protected the epithelial lining in models of fluid occluded distal airspace recruitment [22,28]. It has also been shown that the properties of the air-liquid interface are critical for the function of AE2 cells because the surface tension forces exert deforming mechanical

stresses (e.g., shear stress and tensile strain) on the AE2 cells [23,24,29]. As a result AE2 cells change gene expression profiles in a way that resembles VILI, cyclic alveolar stretch, and pulmonary fibrosis [23].

Transferring these micromechanical models of atelectrauma [22] and interfacial stresses [23] into the context of SP-B deficiency-induced high surface tension it can be hypothesized that repetitive opening of distal airspaces containing fluid (= hypophase) during breathing might be an initial trigger for injury of the alveolar epithelium. These injuries may represent the initial injurious event that occurs prior to vascular leak and alveolar edema accumulation. On the other hand, high surface tension has also been suggested to result in alveolar edema [18] so that the initial consequence of SP-B deficiency could be alveolar fluid accumulation and heterogeneous alveolar ventilation. Discrete alveolar flooding has been shown to be sufficient to induce epithelial injury characterized by vascular leakage due to overdistension of neighboring alveolar airspaces even during ventilation with quite low tidal volumes [30]. Therefore, injury of the alveolar epithelium and progressive respiratory failure might be the consequence, and not the cause, of edema formation.

Based on these considerations the goal of the present study was to understand the relationship between impaired acinar micromechanics, injury of the alveolar epithelium and the alveolar fluid properties in SP-B deficiency induced high surface tension. For this purpose, the time course of these different pathologies was investigated after depletion of SP-B in a mouse model expressing SP-B under control of a doxycycline dependent promotor [21]. Doxycycline-containing food was withdrawn and animals were investigated 1 (group: Dox off d1) and 3 (group: Dox off d3) days thereafter using lung mechanical and BAL parameters and quantitative morphology based on design-based stereology. The latter included a protocol of vascular perfusion fixation of the lungs at airway opening pressures (Pao) on expiration of 2 and 10 cmH2O in order to investigate the recruitability of distal airspaces. Alveolar microarchitecture was described by stereological parameters such as the total volume of alveolar airspaces, the number of open alveoli, the total surface area of alveoli and the mean thickness of interalveolar septal walls. For the assessment of lung injury, the surface area of the epithelial basal lamina covered by injured cells (based on ultrastructural criteria) was determined. Finally, the intra-alveolar fluid was characterized by its absolute volume per lung, the mean thickness and the surface area of the alveolar epithelial cells covered by fluid. As additional parameters lung mechanical data and protein levels in BAL were determined. The data of lungs from mice after withdrawal of doxycycline containing food (Dox off) were compared to those of lungs from mice having been fed with the doxycycline containing food (Dox on).

#### **2. Results**

#### *2.1. Lung Mechanics*

Data describing lung mechanical function are detailed in Figure 1. The quasi-static compliance (Cst) is the slope of the deflation limb of a quasi-static PV-loop at an airway opening pressure of 5 cmH2O and was significantly reduced in Dox off d3 compared to both the Dox on and the Dox off d1 groups (Figure 1A). However, a difference regarding this parameter was not observed between Dox on and Dox off d1. Quasi-static PV-loops were further investigated and the hysteresis (= area within the PV loop) was determined (Figure 1B). While hysteresis in Dox off d1 was significantly increased compared to Dox on this was not the case considering Dox off d3. Inspiratory capacity (IC) is defined as the volume of displaced air into the lung during a ramp inflation from 3 to 30 cmH2O over a period of 6 s. Unlike Cst, IC showed a significant decrease in Dox off d1 compared to Dox on (Figure 1C). Dox off d3 was characterized by an even more dramatic reduction in IC compared to Dox on and Dox off d1 (Figure 1C).

As a forth parameter, the tissue elastance H, was determined using the forced oscillation technique (FOT) fit to the constant phase model [31] during ventilation at positive end-expiratory pressures (PEEP) of 2 and 10 cmH2O. The tissue elastance H reflects the lung mechanics at the different PEEP levels because the onset pressure of the FOT corresponded to the PEEP level and since the FOT volume variations were quite small (3ml/kg bodyweight). As such, tissue elastance H takes the degree of end-expiratory airspace collapse into consideration [19,26,32,33]. At both PEEP levels (2 and 10 cmH2O) there was no significant difference between Dox on and Dox off d1 Figure 1D). Dox off d3, however, demonstrated a significant increase in H compared to both other groups at both PEEP levels (Figure 1D). The Newtonian resistance (Rn) is also determined from the constant phase model fit to the FOT measurements. This parameter reflects pathologies of the conducting airways and shows a clear dependence on the PEEP at which it was determined (Figure 1E). The decrease of Rn with increasing PEEP can be explained by interdependence of conducting airways and surrounding lung parenchyma [34]. Outward tethering forces of the elastic fiber system which connects the conducting airways and the pleura increase with lung volume (and PEEP) so that the resistance of the conducting airways is reduced. At PEEP = 2 cmH2O there were no significant differences in Rn between study groups. However, increasing the PEEP from 2 to 10 cmH2O provided a smaller reduction in Rn for Dox off d3 so that a significant difference became apparent at PEEP 10 cmH2O compared to Dox on (Figure 1E).

**Figure 1.** Lung mechanical properties. Quasi-static compliance (**A**), and hysteresis of quasi-static PV loops (**B**), inspiratory capacity (**C**), tissue elastance (**D**), and Newtonian resistance (**E**). PEEP: positive end-expiratory pressure.
