Increased ceramide production occurs in many situations associated with cellular stress, such as ischemia, radiation, oxidative stress, chemotherapeutic agents, or exposure to oxidized low-density lipoproteins, and is also induced by proinflammatory mediators such as tumor necrosis factor (TNF), IL 1, and platelet-activating factor (PAF) (5–15). Ceramide may exert its actions via an alteration of the physical membrane properties (see 2.5. Membrane Microdomains) and/or by specific interaction with intracellular targets such as protein phosphatases, phospholipase A₂ (PLA₂), cathepsin D, or ion channels (see 2.5. Membrane Microdomains). Ceramide can be generated de novo, from sphingosine or by sphingomyelinase activity. The de novo synthesis, commencing with the condensation of serine with palmitoyl CoA, is catalyzed in the cytosplasmatic leaflet of the endoplasmatic reticulum. The hydrophobic ceramide is then transported by the ceramide-binding protein CERT to the Golgi apparatus (16). The synthesis of ceramide from sphingosine involves hydrolysis of glycosphingolipids and a retrograde activity of the enzyme ceramidase (17). The release of ceramide from sphingomyelin is catalyzed by several sphingomyelinase isoenzymes: the lysosomal acid sphingomyelinase (A-SMase), the secretory A-SMase, the Mg²⁺ dependent neutral sphingomyelinase isoenzymes (N-SMase), and the alkaline SMase (18).

The lysosomal A-SMase mediates sphingolipid turnover and degradation. The enzyme interacts with lipid membranes via its sphingolipid activator domain and also binds to the mannose-6-phosphate receptor. Both lysosomal and secretory A-SMases are derived from the same gene (SMPD1) and are Zn²⁺ dependent enzymes. However, whereas the lysosomal form is already tightly bound to the cation, the secreted form is stimulated by zinc (19). It appears likely that the extracellular A-SMase acts on plasmalemmal sphingomyelin because sphingomyelin is almost exclusively localized to the outer membrane leaflet (20). Endothelial cells are a wealthy source of A-SMase (21), A-SMase is present in most body fluids (22), and serum concentrations are raised in a number of disorders, many of them inflammatory (see Table 1).

The ubiquitously expressed N-SMase1 (SMPD2) is Mg²⁺-dependent, localizes to the endoplasmic reticulum (ER), and is reversibly inhibited by oxidized glutathione, but its role remains poorly defined; it was suggested to act mainly as a lysoPAF phospholipase C (PLC). N-SMase2 (SMPD3) is expressed in the Golgi, but probably also in the plasma membrane and appears to be the major N-SMase isoenzyme that becomes activated by oxidative stress (18, 23, 24). Recently, N-SMase3 (SMPD4) was cloned that is highly expressed in the heart, but also in the lungs and other organs (25).

Ceramide may further be converted by ceramide kinase (Cerk) to ceramide-1-phosphate that appears to have its own specific actions, such as activation of cytosolic phospholipase A₂ (cPLA₂) and regulation of apoptosis (26, 27). A recent study demonstrated a critical role of ceramide-1-phosphate in neutrophil homeostasis and showed strikingly reduced neutrophil numbers in the blood and spleen of Cerk^−/− animals (28).

Sphingosine cannot be synthesized de novo, but it is derived from N-deacetylation of ceramide by ceramidase or by hydrolysis from plasma membrane sphingolipids. While the initially suspected function of S1P as an intracellular messenger is still under debate (4), the discovery that the once orphaned G protein–coupled edg (endothelial developmental gene) receptors represent S1P receptors (S1P_1–5) has greatly stimulated research in this area. S1P is now recognized as a ubiquitous regulator of cell proliferation and survival, angiogenesis and endothelial barrier functions, cell mobility and chemotaxis, cytoskeletal organization, cellular calcium homeostasis, cell-to-cell contacts, and adhesion (4, 29, 30). There are several examples showing that the different S1P receptor subtypes may act antagonistically, such as in chemotaxis (stimulation by S1P₁ and S1P₃, inhibition by S1P₂) (31) or in endothelial barrier functions (see below). Another interesting antagonism seems to exist between S1P and ceramide, with ceramide being proapoptotic and S1P being antiapoptotic (4). In view of the consequences for cell growth and differentiation, and in consideration of the highly dynamic S1P and ceramide levels, this antagonism has been termed the sphingolipid “rheostat” (4). S1P₄ and S1P₅ are enriched in lymphoid tissues, and in brain and spleen tissue, respectively (30); these receptors have not been implicated in pulmonary (patho)physiology. It should be noted that the S1P receptors show a high homology to the receptors for lysophosphatidic acid, LPA_1/2/3, (32) and some receptors may even recognize both S1P and lysophosphatidic acid (33).

S1P is synthesized by two sphingosine kinase isoenzymes (SphK1, SphK2) and degraded by S1P phosphatases, lipid phosphate phosphatases, or S1P lyase (4). Platelets lack S1P lyase, implicating these cells as sources of the relatively high (0.5 μM) serum S1P levels (34), where it is bound to albumin and high-density lipoproteins (35). However, other cells such as erythrocytes, neutrophils, mast cells, mononuclear cells, and endothelial cells may also secrete S1P (36–38). In any case, control of extracellular S1P seems to be complex and apparently also involves export of SphK-1a to produce S1P outside the cells (39), degradation by lipid phosphate phosphatases (which are integral membrane proteins facing the extracellular side), and reuptake (40). SphK activity is stimulated by a variety of stimuli (41) including muscarinic M2 receptor agonists (leading to airway constriction [42]), activation of IgE receptors (43, 44), histamine, several growth factors, and various cytokines (29).

The study of sphingolipids is fraught with technical problems that need to be taken into account when following progress in this area. One particular problem is the significant hydrophobicity, particularly of ceramide, that makes it difficult to identify specific intracellular binding partners. Another shortcoming is the lack of quick and efficient ways to determine sphingolipid concentrations in biological samples. Although several ceramide antibodies are available, some need to be used with caution (45).

The nearly complete insolubility of long chain ceramides (>C16) in aqueous solutions also poses problems for studying the role of exogenously added ceramide. Thus, many studies using exogenous ceramide have used C2- or C6-ceramide, although their biophysical properties differ from those of long chain ceramides (46). In addition, it seems possible that specific ceramide species have specific roles, such as C16- and C24-ceramide in different stages of apoptosis (47). Therefore, the results from experiments with short chain ceramides need to be interpreted cautiously. The use of extracellular A-SMase is another approach to generate ceramide, at least in cell culture, although it is difficult to mimic transient changes of ceramide, to increase intracellular ceramide, and to control the amount of ceramide with this treatment.

Many growth factors induce S1P, whereas many agents that cause apoptosis invoke ceramide by several different mechanisms, as proapoptotic ceramide may be derived from A-SMase, N-SMase or even de novo synthesis (82, 83). Because ceramide is activated by so many different stimuli that do not cause apoptosis, it is important to question how cells can avoid apoptosis in the presence of increased ceramide levels. Some cells produce both S1P and ceramide, to the effect that S1P prevents ceramide-induced apoptosis that would occur otherwise (84). Other mechanisms that protect cells from ceramide-induced apoptosis are the strong expression of acid ceramidase in alveolar macrophages (85) or the inhibition of caspase 3 by interaction with A-SMase in airway epithelial cells (86).

Ceramide appears to cause apoptosis by several mechanisms such as the recruitment of death receptors to lipid rafts (87) as well as by the direct activation of protein phosphatase 2a and cathepsin (82, 88). The proapoptotic properties of ceramide are well illustrated by the fact that radiation increases pulmonary ceramide levels and that pulmonary endothelial cells in Smpd1^−/− mice are resistant to radiation-induced apoptosis as shown by TdT-mediated dUTP biotin nick end labeling (TUNEL) assay (15). TUNEL is a frequently-used apoptosis assay, although it is not entirely specific for apoptosis (89). With reference to its apoptotic properties, ceramide has been called the “tumor-suppressor lipid” (82). Of note, sphingosine also has proapoptotic properties that appear to be independent of ceramide (82).

The antiapoptotic actions of S1P are currently explained by its activation of nuclear factor (NF)-κB, Akt kinase and the intracellular inhibition of ceramide synthesizing enzymes (82, 90). For cancer cell lines (e.g., HEK293 and A549) S1P is an important survival factor and inhibition of S1P synthesis or action sensitizes these cells for some chemotherapeutics (91, 92). SphK1 is overexpressed in human tumor lung tissue (93). Thus, inhibition of sphingosine metabolism has been suggested as a means to sensitize resistant tumors. Interestingly, only S1P derived from SphK1 appears to be antiapoptotic, whereas overexpression of SphK2 suppresses growth and favors apoptosis (92). S1P is also mitogenic for transformed cell lines (92) and for human lung fibroblasts (94).

Unlike those suggested by the classical fluid mosaic model (95), cell membranes show a high degree of organization to which sphingolipids make an important contribution. Sphingolipids have a higher melting temperature than most other membrane lipids and thus increase membrane rigidity. Besides sphingolipids, cellular membranes are predominantly composed of cholesterol and (glycero)phospholipids. Sphingolipid headgroups interact with each other and the hydroxyl group in the cholesterol molecule via hydrophilic interactions, whereas the alkyl chains and the sterol ring system of cholesterol exhibit hydrophobic van der Waals interactions (96). The tight interactions result in a separation of sphingolipids and cholesterol from other phospholipids in the cell membrane and the formation of very small distinct domains in the cell membrane enriched in sphingolipids and cholesterol (96, 97), which are termed “rafts.” Cholesterol molecules seem to fill the voids between the bulky sphingolipids and thus promote the structure of these domains (98). The formation of these small membrane domains was concluded from biophysical experiments using artificial membranes and the detergent-resistant extraction of these domains in biochemical studies (96, 97, 99).

The generation of ceramide within lipid rafts was shown to result in the formation of larger membrane domains (100–102). In general, rafts serve as anchoring points of the cytoskeleton, as gathering places for receptors, as starting points of signaling cascades, and as locations for the uptake of microorganisms; although it should be pointed out that many protein–protein or protein–lipid interactions and receptor-mediated signaling processes are independent of rafts. In addition, many different rafts with varying lipid compositions exist, which may result in different modes of action. Although the exact functions of many rafts still require definition, it may be safe to state that ceramide-enriched membrane rafts seem to be predominantly involved in cellular stress responses. Caveolae are one particular form of lipid raft. In the lungs, caveolae are particularly abundant in endothelial cells (10,000–30,000 caveolae/cell, or ∼15% of the total cell volume), alveolar type II cells, fibroblasts, and smooth muscle (103, 104).

The rare Niemann-Pick disease (NPD) type A and B (recently reviewed in [131]), is the consequence of aberrant and at least partly inactive A-SMase variants. Type A NPD, which is lethal within the first 3 years of life, is characterized by neurological deficits, enlarged livers and spleens, and often by recurrent pulmonary infections. Patients with Type B NPD frequently show enlarged livers and spleens, proatherogenic alterations, and interstitial lung disease (132, 133), but otherwise the symptoms may vary. This variation is explained by over 100 known mutations of the SMPD1 gene (133, 134). In recent studies more than 90% of all patients with Type B NPD showed interstitial lung disease and more than 70% had abnormal lung function tests, although these clinical findings did not correlate well (132, 133, 135). In addition, foamy alveolar macrophages (Niemann-Pick cells) are frequently observed in patients with NPD (135) as they are in aged A-SMase–deficient animals (69). These findings show that in NPD pulmonary involvement is common. Currently, no specific therapy is available, but after a promising experimental study (136), a clinical trial (NCT00410566) examining A-SMase replacement therapy has been initiated.

Most studies in this area have focused on the role of A-SMase and S1P receptors on endothelial cells. These studies are reviewed below. More recently, further cell types have begun to be studied in models of ALI. Such studies showed that S1P₃ receptors play a particularly noteworthy role in promoting pulmonary inflammation in that they stimulate the late phase (18 h) of sepsis by propagating the release of IL-1β and tissue factor from dendritic cells (137). Thus, sphingolipids appear to regulate ALI at different levels, but this needs further exploration.

Allergic asthma is a chronic inflammatory disease characterized by a predominance of Th2 lymphocytes, recurrent activation of mast cells by activation of FcεR1 receptors, recurrent airway smooth muscle contraction, airway hyperresponsiveness, and airway remodeling, all of which appear to be regulated by sphingolipids (172), in particular by S1P, which is increased in asthmatic airways (173).

Ceramide-enriched membrane domains and rafts are critically involved in the infection of epithelial cells and the lung with several pathogens. One well-studied example is P. aeruginosa infection (112). P. aeruginosa induces, via unknown mechanisms, an activation of the A-SMase in epithelial cells and a translocation of intracellular localized A-SMase onto the extracellular leaflet of the cell membrane resulting in the formation of ceramide and ceramide-enriched membrane platforms in the outer leaflet (112). Genetic deficiency of the A-SMase or disruption of rafts by interference with cellular cholesterol not only prevented the formation of ceramide-enriched membrane platforms upon cellular infection with P. aeruginosa but also affected salient steps of the infection process such as P. aeruginosa internalization and induction of death in infected cells, whereas, at the same time, an uncontrolled cytokine release was observed.

Although the exact molecular mechanisms that are initiated in ceramide-enriched membrane domains during infection are unknown, it is possible that receptor clustering within these domains is critically involved in the effects induced by pulmonary P. aeruginosa infections. Thus, it was shown that CD95 receptor and cystic fibrosis transmembrane conductance regulator (CFTR) molecules cluster within ceramide-enriched membrane domains upon cellular infection with P. aeruginosa (112, 191). CD95 was previously shown to mediate apoptosis of cells infected with P. aeruginosa. Furthermore, clustering of the receptor in ceramide-enriched membrane platforms has been demonstrated to be required for the induction of cell death by CD95. Studies by G. Pier and colleagues (192, 193) demonstrated that CFTR functions as a receptor for P. aeruginosa, mediating internalization of the bacteria. P. aeruginosa lipopolysaccharide molecules bind to a short amino acid sequence in CFTR (AA 103–117) and deficiency of CFTR, or blockade of this binding motif, prevents P. aeruginosa internalization. Thus, the reorganization of cellular receptors such as CD95 and CFTR within P. aeruginosa–triggered ceramide-enriched membrane domains seems to be central for an adequate immune response of respiratory epithelial cells to P. aeruginosa. However, it is also possible that ceramide-enriched membrane domains have additional, yet undefined functions, during the infection of epithelial cells with P. aeruginosa. The mechanisms that result in an uncontrolled release of cytokines in the lung of Smpd1-deficient mice infected with P. aeruginosa are presently unknown. In vivo studies demonstrate that Smpd1-deficient mice, which are unable to form ceramide upon pulmonary infection with P. aeruginosa, are highly susceptible to an acute pulmonary P. aeruginosa infection, and a high percentage of the mice succumbed by a generalized sepsis within a few days after nasal infection with the bacteria, whereas wild-type mice were resistant (112).

It is unknown whether ceramide-enriched membrane domains mediate a negative effect on cytokine release or whether other effects of ceramide are required to balance cytokine release upon infection. The central role of membrane rafts for the response of epithelial cells to P. aeruginosa infections is also supported by studies on respiratory tract–derived epithelial cells (112, 194). These studies revealed the formation of large GM1-enriched membrane domains from small rafts, that colocalize with the previously described ceramide-enriched membrane domains (112). CFTR translocated into the raft fraction after infection, which was prevented by destruction of rafts by interference with the cholesterol metabolism (194). Most important, destruction of rafts prevented internalization of P. aeruginosa into epithelial cells, NF-κB nuclear translocation, and cellular apoptosis in vitro and in vivo in respiratory epithelial cells after pulmonary infection (194), supporting the notion that distinct membrane domains are critically involved in an adequate response of host cells to P. aeruginosa.

The notion that ceramide-enriched membrane domains are importantly involved in the infection with pathogens is consistent with previous findings on the infection of epithelial and endothelial cells with Neisseriae gonorrheae and Staphylococcus aureus (113, 114). These studies demonstrated that activation of the A-SMase and release of ceramide upon infection of epithelial cells with N. gonorrheae are required for internalization of the bacteria. Furthermore, expression of the A-SMase was a prerequisite for induction of death in endothelial cells infected with S. aureus.

S1P has also been implicated in infection. S1P is decreased in patients with pulmonary tuberculosis (195), possibly because mycobacteria inhibit macrophage SphK1 (196). This may be relevant for the disease because S1P stimulates the antimycobacterial activity of human macrophages and reduces mycobacterial growth, granuloma formation, and necrosis in the lungs of Mycobacterium tuberculosis–infected mice (197).

The role of rafts and ceramide-enriched membrane domains for pulmonary infections has been mainly investigated in the context of cystic fibrosis and P. aeruginosa infections. However, given the clinical significance of pneumonia, it will be important to define whether distinct membrane domains are also involved in pulmonary infections with other bacterial pathogens such as pneumococci or mycobacteria and even viral pathogens, for instance, influenza. Such studies may also help to explain the pulmonary infections that occur in approximately one-half of all patients with Niemann-Pick diseases, type B (133). Furthermore, it needs to be defined whether rafts can be targeted to prevent infections without altering normal lung physiology beyond a pathophysiological level.

Very elegant studies by I. Petrarche and colleages identified ceramide as a central mediator in the development of emphysema (198). These authors used blockade of vascular endothelial growth factor receptors as a model to investigate mechanisms of lung emphysema that typically are present in human chronic obstructive pulmonary disease (COPD) and emphysema caused by cigarette smoke. Vascular endothelial growth factor blockade resulted in the formation of ceramide predominantly in alveolar cells, which was mediated by an increased activity of the ceramide synthase pathway and a feed-forward activation of the A-SMase (198). A similar mechanism with two waves of ceramide has also been suggested for radiation-induced ceramide production (199). Increased ceramide was critical to trigger alveolar cell death, oxidative stress (200), and emphysema, events that were prevented by the inhibition of the ceramide synthase pathway in vivo using fumonisin B1 or myriocin. Intrapulmonary application of ceramide mimicked the effects of endogenous ceramide and also resulted in activation of caspase 3 in the lung, alveolar cell death, and emphysema. Most interestingly, cigarette smoke induced the production of ceramide in endothelial cells as well as the release of the proinflammatory mediator, TNF. Emphysema patients had increased ceramide levels that were localized to alveolar septal cells and macrophages. The authors concluded that pulmonary emphysema develops as a consequence of ceramide-induced oxidative stress and apoptosis of alveolar endothelial and/or epithelial cells with subsequent digestion of interstitial tissue by matrix metalloproteases (198, 200). In some sense, this mechanism is opposed to the classical view that one function of apoptosis is to diminish inflammation. In the emphysema model it would be the massive apoptosis stimulating an inflammatory response that finally leads to emphysema. Interestingly, also in this case, S1P antagonizes ceramide (198).

Additional studies extended these observations to the effects of ceramide on endothelial cells in the lung. A recent study by Castillo and colleagues (24) also implied the neutral N-SMase to be involved in airway epithelial cell death. This study indicated that reactive oxygen species, in particular H₂O₂, activate the N-SMase2, whereas peroxynitrate stimulated the A-SMase (24). Collectively, these studies strongly suggest endothelial ceramide as a novel target to prevent or treat emphysema and COPD. It will be interesting to determine whether sphingolipids also play a role in other cells relevant for the development of COPD, in particular, in macrophages and natural killer T cells thought to be critical for the development of COPD (201).

Children with cystic fibrosis (CF) very often develop infections with P. aeruginosa and once past childhood almost all patients with cystic fibrosis suffer from a chronic pneumonia with P. aeruginosa, Burkholderia cepacia, and/or S. aureus. Although the life expectancy of patients with CF has increased, these bacterial lung infections are key to the development of the disease and very often result in destruction of the lung. CF is caused by a mutation of CFTR and occurs with a frequency of 1:2,500 births, at least in Western countries. Several recent studies suggested a proinflammatory status in the lung, and possibly also other organs, of patients with CF that triggers chronic inflammation even without a bacterial or viral infection. Thus, it was shown that even noninfected Cftr-deficient mice suffer from increased IL-8 concentrations in the trachea (202, 203). Further, studies on aborted embryos with CF and on BAL fluids from patients with CF as young as 4 weeks with negative cultures for CF-related bacteria, virus, and fungi, revealed a significant increase of proinflammatory mediators in the lungs (204, 205). These studies suggest that patients with CF suffer from an uncontrolled inflammation in the lung that might be critical for the propensity of these patients to develop infections with P. aeruginosa and other bacteria.

Recent studies imply sphingolipids (206), and in particular ceramide, as critical regulators for the development of the high sensitivity of Cftr-deficient mice to P. aeruginosa infections (207). These studies demonstrated in different Cftr-deficient mouse strains that ceramide accumulation in respiratory epithelial cells and in the submucosal glands of uninfected Cftr-deficient mice is age dependent (207). Cftr deficiency alters the acidification of intracellular vesicles, most likely prelysosomes and lysosomes, in respiratory epithelial cells, which results in alkalinization of these vesicles. It should be noted that other vesicles, in particular endosomes, with a different pattern of ion channel expression than prelysosomes and lysosomes are not alkalinized in CF (208–210). The inappropriate alkalinization of (pre)lysosomes and other acidic vesicles may result in an imbalance of enzymes that release (acid sphingomyelinase) and consume ceramide (acid ceramidase) and an accumulation of ceramide in respiratory epithelial cells from Cftr-deficient mice. A similar accumulation was observed in respiratory epithelial cells and lung specimens from patients with CF, suggesting that the mouse data also apply to humans (207). The increased ceramide concentration in airway epithelial cells of Cftr-deficient mice triggered chronic pulmonary inflammation, death of respiratory epithelial cells, and deposition of DNA in bronchi and thus explains the proneness of Cftr-deficient mice to develop pulmonary P. aeruginosa infections. Extracellular DNA on epithelial cells may facilitate bacterial adherence in the respiratory tract, whereas an increased and chronic inflammatory response may affect lung functions, block the mucociliary clearance, alter the innate immune response, change the airway architecture, and finally result in pulmonary fibrosis. Partial inhibition of A-SMase that was achieved by heterozygosis of the A-SMase in Cftr-deficient mice (Cftr^−/−/Smpd1^+/− mice) or by pharmacological treatment of Cftr-deficient mice with the functional A-SMase inhibitor amitriptyline, normalized pulmonary ceramide, abrogated the above-mentioned consequences and, most important, prevented severe pulmonary infections with P. aeruginosa.

Although these studies indicate that ceramide also accumulates in CFTR-deficient epithelial cells from patients with CF, it remains to be determined whether the data obtained in a murine system can be transferred to the human situation. Furthermore, the role of sphingolipids in other cells that normally express Cftr—such as macrophages, endothelial cells, and neutrophils—for the development of inflammation and infection susceptibility in CF remains to be determined. Finally, one of the major problems of children with CF is the development of fibrosis along with the inflammation and infection. Fibrosis might be a consequence of chronic infection, but animal models suggest that fibrosis also occurs in Cftr-deficient mice without infection and may therefore be an independent part of the disease that needs further attention.

These results indicate a novel concept for explaining the pathogenesis of lung disease in CF, i.e., an increase of ceramide caused by an imbalance of enzyme activities upon intravesicular pH changes. On the other hand, the role of sphingolipids in infection, as described above (3.4. Sphingolipids in Pulmonary Infections), maintains that a normal defense against P. aeruginosa requires intact membrane rafts, at least some activity of the A-SMase, and normal ceramide values. Thus, any future treatment of CF with A-SMase blockers must likely be personalized and tightly controlled to prevent too strong an inhibition of the acid sphingomyelinase. We speculate that inhibition of the A-SMase to treat of CF might be facilitated by the development of inhalation drugs that block the A-SMase and, thus, will improve CF.

Several studies indicate that rafts are not only critically involved in the infection of respiratory cells with pathogenic bacteria, but also with viruses that affect the respiratory tract. Thus, the infection of human epithelial cells with pathogenic rhinoviruses that cause the common cold results in the formation of large membrane platforms derived from small rafts (115, 211). Destruction of these domains by extraction of cholesterol prevented internalization of the virus indicating the significance of rafts for the infection (115, 211). Further studies revealed a rapid activation of the A-SMase, release of ceramide, and formation of large ceramide-enriched membrane platforms upon infection with human rhinoviruses. Pharmacological or small interfering RNA-mediated inhibition of the A-SMase prevented infection of human epithelial cells with rhinovirus, supporting the notion that raft-derived domains and ceramide-enriched domains are important for infection of human epithelial cells with rhinoviruses.

Sphingolipids, in particular sphingomyelin, are present in pulmonary surfactant (212) and are, together with some phospholipids, increased in A-SMase–deficient mice (213). Of note, the BAL of patients with ARDS shows an elevated sphingomyelin:phosphatidycholine ratio (214). The relevance of this observation is not known, but interestingly, instillation of TNF leads to increased alveolar sphingomyelin levels and sphingomyelinase activity that may, in turn, raise ceramide concentrations that finally impair surfactant functions (139) and possibly also production (215).