Surfactant Composition and Homeostasis

Surfactant is a complex mixture of 90% phospholipids and 10% surfactant proteins that functions at the air-liquid–alveolar wall interface to reduce surface tension, thereby preventing alveolar collapse. It is produced in type II alveolar epithelial cells, secreted into the alveolar space, and cleared by either recycling or catabolism by type II cells or uptake and catabolism by alveolar macrophages. Surfactant composition, production, and homeostasis are reviewed in Chapter 8 . Studies in animals and humans in which the pathogenesis of PAP has been elucidated have identified GM-CSF as a critical regulator of surfactant homeostasis ( Fig. 70-1 ).

Schematic of pulmonary surfactant homeostasis within an alveolus of a healthy individual.

Surfactant lipids and proteins are synthesized in type II alveolar epithelial cells and secreted into the alveolar space. Extracellular surfactant contributes to the surfactant layer present at the air-liquid interface, which plays a critical role by reducing surface tension within the alveolus. Surfactant lipids and proteins are removed from the extracellular surfactant by uptake and recycling in type II cells or by uptake and catabolism in alveolar macrophages. Granulocyte-macrophage colony-stimulating factor (GM-CSF), a critical regulator of surfactant catabolism in alveolar macrophages, functions by binding to heterologous receptors on alveolar macrophages and stimulating surfactant catabolism via the transcription factor PU.1. GM-CSF is also critical for stimulating the terminal differentiation of precursors into mature alveolar macrophages.

Granulocyte/Macrophage Colony-Stimulating Factor

GM-CSF is a 23-kD cytokine that signals via cell surface receptors comprising a GM-CSF–binding, low-affinity α chain (CDw116) and a nonbinding, affinity-converting β chain (CD131). Although neither chain has intrinsic signaling activity, the β chain constitutively binds Janus kinase 2 (JAK2), a tyrosine kinase involved in cytokine signaling. Ligand binding causes the formation of αβJAK2 multimers and activation of JAK2, resulting in phosphorylation of α and β chains, activation of other kinases, and initiation of signaling via multiple pathways including the signal transducer of activation and transcription-5 (STAT5).

Murine Models of Disordered Surfactant Homeostasis

GM-CSF–deficient ( Csf2 ^KO ) mice provided the first clue to the pathogenesis of PAP. Although neither production of surfactant by type II cells nor its uptake by alveolar macrophages was altered, clearance of surfactant by alveolar macrophages was impaired. Pulmonary GM-CSF, acting via the transcription factor PU.1, coordinately regulated surfactant catabolism; cell adhesion; expression of Fc receptors, mannose receptors, and other receptors; phagocytosis; bacterial killing; and Toll-like receptor-4 signaling. GM-CSF was identified as a critical regulator of the terminal differentiation of alveolar macrophages in mice.

GM-CSF receptor β chain–deficient ( Csf2rb ^KO ) mice develop PAP indistinguishable from that of GM-CSF–deficient mice. Because transplantation of wild-type bone marrow corrected PAP in these mice, myeloid cells, and not lung resident epithelial cells, were identified as the cellular site of pathogenesis. A biomarker of PAP in these mice (increased serum GM-CSF) suggested the means that led to identification of hereditary PAP caused by CSF2RA and CSF2RB mutations in children. Importantly, Csf2rb ^KO mice and children with CSF2RA or CSF2RB mutations develop a lung disease that is identical in every respect including lung pathology (surfactant-filled alveoli, preserved architecture); cytopathology (oil red O–stained macrophages, cell debris); alveolar macrophage biomarkers ( reduced messenger RNA [mRNA] for PU.1, peroxisome proliferator–activated receptor -γ [PPAR-γ], ABCG1); and BAL biomarkers (increased surfactant proteins/lipids, turbidity, cytokines [GM-CSF, M-CSF, MCP-1]) and clinical course (progressive surfactant accumulation). Studies in these mice have identified pulmonary macrophage transplantation (PMT) as a novel therapeutic approach for hereditary PAP.

Mice deficient in ABCG1, a transmembrane protein mediating cellular cholesterol efflux, accumulate cholesterol in alveolar macrophages and type II cells and develop pulmonary lipidosis, indicating ABCG1 may be important in surfactant homeostasis. Importantly, ABCG1 expression is decreased in alveolar macrophages from GM-CSF–deficient mice, patients with autoimmune PAP, GM-CSF receptor–deficient mice, and patients with hereditary PAP (unpublished data). Conditional disruption of PPAR-γ (a known transcriptional regulator of ABCG1) specifically in alveolar macrophages also disrupts surfactant clearance.

Together, murine studies suggest the pathogenesis of primary PAP, both autoimmune and hereditary, may involve disruption of a pathway including GM-CSF, PU.1, PPAR-γ, and ABCG1 that is critical for surfactant clearance by alveolar macrophages.

In secondary PAP, several models have shed light on the pathogenesis. For example, reduction of alveolar macrophage numbers by chemical depletion increases surfactant pool size and inhalation of respirable silica results in alveolar proteinosis.

In addition to these models of surfactant clearance, murine models of abnormal surfactant production called PSMD disorders develop lung diseases faithfully recapitulating their respective human diseases. They are clinically, histologically, and biochemically distinct from PAP in GM-CSF–deficient mice and have improved the understanding of surfactant homeostasis. For example, SP-B–deficient mice develop respiratory failure at birth. Studies in these mice showed that SP-B is critical in the post-translational processing of SP-C, organization of surfactant phospholipids in lamellar bodies, formation of tubular myelin in alveoli, generation of surfactant films capable of reducing surface tension, and lung function during the early postnatal period. ABCA3-deficient mice develop respiratory failure at birth, and studies in these mice established ABCA3 as critical to lamellar body formation and pulmonary surfactant biogenesis.

Other murine models have been shown to develop PAP and are providing insights to surfactant homeostasis including murine models of severe combined immunodeficiency, pulmonary overexpression of interleukin (IL)-4 or IL-13, and SP-D–deficiency.

Role of GM-CSF Autoantibodies in Primary PAP

In 1999, Nakata’s group first identified neutralizing GM-CSF autoantibodies in the BAL fluid and serum of idiopathic, now called autoimmune, PAP patients. Multiple subsequent studies confirmed that high levels of GM-CSF autoantibodies are associated with idiopathic PAP but not with secondary PAP, other lung diseases, PSMD disorders, or healthy individuals ( Fig. 70-2 ). In one study, the concentration of GM-CSF autoantibodies in 158 idiopathic PAP patients was 113 ± 7 µg/mL, while levels were less than 1 µg/mL in patients with secondary PAP, PSMD disorders, or other lung diseases. GM-CSF autoantibodies are composed of IgG, predominantly of the IgG1 and IgG2 subclasses; are highly specific for GM-CSF; have a high binding affinity (20 ± 7.5 pM); and are capable of neutralizing GM-CSF at levels up to many thousands times higher than the physiologic concentrations of GM-CSF in vivo. Notwithstanding these important observations, GM-CSF autoantibodies have been reported at low frequency in healthy people and in patients with autoimmune diseases without evidence of PAP; in addition, GM-CSF comprises the dominant anti-cytokine activity in pharmaceutical immunoglobulin prepared from healthy individuals. In a more recent study utilizing multiple detection methods, GM-CSF autoantibodies were demonstrated in all of 72 healthy individuals with no prior exposure to exogenously administered GM-CSF, albeit at low levels (median [interquartile range] = 1.04 [0.63 to 1.7] µg/mL serum). Another confusing observation was that in patients with PAP, serum GM-CSF autoantibodies did not correlate with disease severity.

Serum antibodies to GM-CSF are elevated in autoimmune pulmonary alveolar proteinosis (PAP).

Serum GM-CSF autoantibody levels in individuals with autoimmune PAP ( n = 223), secondary PAP ( n = 33), pulmonary surfactant metabolic dysfunction ( n = 5), other lung diseases ( n = 24), and in healthy controls ( n = 13). The detection limit of the assay was 0.5 µg/mL of serum.

(Some of these data were reported in Inoue Y, Trapnell BC, Tazawa R, et al: Characteristics of a large cohort of autoimmune pulmonary alveolar proteinosis patients in Japan. Am J Respir Crit Care Med 177:752–762, 2008.)

To prove that GM-CSF autoantibodies were critical to the pathogenesis of idiopathic PAP and not a related but unimportant epiphenomenon, GM-CSF autoantibodies from these patients were isolated in pure form (i.e., a single band on SDS gels) and used to immunize healthy nonhuman primates. Results demonstrated unequivocally that GM-CSF autoantibodies reproduced the cardinal molecular and cellular features of PAP, including surfactant-filled alveoli with preserved wall architecture, foamy alveolar macrophages, alveolar macrophage biomarkers (reduced PU.1, PPAR-γ, and ABCG1 mRNA), and BAL biomarkers (increased SP-D and turbidity). Evaluation of GM-CSF autoantibodies isolated from the immunized primates or directly from patients has similar bioactivity. Together with the clinical studies in humans, these data provided strong evidence that GM-CSF autoantibodies were in fact the pathogenic driver in idiopathic PAP, which is now recognized as autoimmune PAP. Neutrophils from the immunized primates also had impaired phagocytosis similar to defects reported in these patients, in GM-CSF deficient mice, and in normal human neutrophils incubated with purified GM-CSF autoantibodies in vitro. Using the CD11b stimulation index assay (see later), GM-CSF signaling in immunized primates was found to decrease inversely with GM-CSF autoantibody concentration at values below 5 mcg/mL and was zero at higher values, thus defining the critical threshold concentration needed to block GM-CSF signaling. In a translational study using this assay, a similar relationship and threshold were observed in humans. Thus seemingly discrepant observations (i.e., the critical role of GM-CSF autoantibodies in PAP pathogenesis and the lack of correlation between their concentration and disease severity) were reconciled by the hypothesis that a critical threshold level of GM-CSF autoantibodies was needed to reduce GM-CSF bioactivity sufficiently to impair surfactant clearance.

GM-CSF autoantibodies are polyclonal and directed at epitopes throughout the GM-CSF molecule. Evaluation of 19 monoclonal autoantibodies from 6 patients demonstrated that they use multiple immunoglobulin V genes and targeted at least four non-overlapping epitopes on GM-CSF, suggesting that GM-CSF is driving the autoantibodies and not a B-cell epitope on a pathogen cross-reacting with GM-CSF.

The observations that GM-CSF stimulation of neutrophil functions is impaired in autoimmune PAP and GM-CSF deficient mice provides a further explanation for the increased infection risk in PAP caused by disruption of GM-CSF signaling. Further, the inverse correlation between GM-CSF autoantibody levels and neutrophil function (phagocytosis) in healthy people suggests a potential physiologic role for these autoantibodies, such as scavenging free GM-CSF, which can function as a proinflammatory cytokine. Supporting this concept, using methods capable of detecting GM-CSF, whether bound by GM-CSF autoantibody or not, serum GM-CSF concentrations in healthy people (3048 ± 484 pg/mL serum, n = 11) are much higher than previously reported, and more than 99% is bound to GM-CSF autoantibodies both in healthy individuals and in PAP patients.

Genetic Factors

A series of reports have now established hereditary PAP as a newly described genetic disease caused by genetic mutations in CSF2RA or CSF2RB , which encode GM-CSF receptor α or β chains, respectively ( Fig. 70-3 ). In some cases, the genes have been cloned, the defects reproduced in vitro, the signaling abnormalities studied in detail, ^23-,25 and the results used to develop several novel biomarker-based diagnostic tests (see later). One particularly useful test, serum GM-CSF, was used to identify a number of patients with hereditary PAP due to mutations in the GM-CSF receptors and is capable of identifying patients with defects in either chain of the receptor. Of importance to the pathogenesis of PAP, in families with two siblings carrying the same function-impairing GM-CSF receptor gene mutation, the older sibling had minimal disease while the younger sibling had more extensive disease, suggesting another factor besides loss of GM-CSF signaling is important in determining disease severity in hereditary PAP. Several genetic factors associated with hematologic disease and secondary PAP have been identified including mutations in the gene encoding the transcription factor GATA2, as well as in GMCSF receptor chain genes.

Genetic mutations in CSF2RA and CSR2RB and corresponding GM-CSF receptor α and β chain abnormalities associated with the development of hereditary pulmonary alveolar proteinosis.

For each gene, a schematic of the normal protein is shown above to indicate regions corresponding to the signal peptide, extracellular domain, transmembrane spanning domain, and intracellular domain. Shown below each normal protein are the abnormal proteins caused by genetic mutations (indicated) known to cause hereditary PAP.

(Some data are from Suzuki T, Sakagami T, Young LR, et al: Hereditary pulmonary alveolar proteinosis: pathogenesis, presentation, diagnosis, and therapy. Am J Resp Crit Care Med 182:1292–1304, 2010.)

Mutations in three genes encoding surfactant proteins, SFTPB, SFTPC, and ABCA3, disrupt surfactant production and function and cause respiratory disease in neonates, children, and older individuals. See Chapter 8 for additional information on pulmonary surfactant. SP-B and SP-C are hydrophobic peptides that reside within the surfactant phospholipid layer, which functions to reduce surface tension at the alveolar air/liquid interface, whereas ABCA3 likely transports lipids, including phosphatidylcholine, cholesterol, sphingomyelin, and phosphatidylglycerol, into lamellar bodies in alveolar type II cells, where the surfactant complex is assembled, processed, and stored.

SFTPB is a small gene (9.7 kb) encoding a 79-amino acid, hydrophobic protein. Infants homozygous for recessive loss-of-function mutations in SFTPB develop respiratory failure and die shortly after birth. In contrast, individuals heterozygous for SP-B deficiency alleles have normal lung function. More than 30 mutations have been identified to date in fewer than 100 infants with SP-B deficiency. Two thirds have an insertional frameshift mutation caused by the replacement of a single nucleotide by three nucleotides at codon 121 (designated 121ins2). Among infants with SP-B deficiency, approximately 60% are homozygous for the 121ins2 mutation, 25% are heterozygous for this and another function-altering allele, and 15% have other mutations. SP-B deficiency is associated with the aberrant processing and secretion of an immature SP-C peptide, which may contribute to the respiratory failure associated with hereditary SP-B deficiency.

SFTPC is a small gene (3 kb) encoding a 35-amino acid hydrophobic protein. Known mutations in SFTPC are expressed in a dominant fashion and are associated with interstitial lung disease in neonates, children, and adults. Twenty-five percent of individuals with SP-C–associated lung disease have a point mutation in SFTPC causing the substitution of threonine for isoleucine at codon 73 (I73T). Immunohistochemical analysis of lung tissue from an infant with the I73T mutation demonstrated normal staining patterns for pro-SP-B, SP-B, and pro-SP-C. Although infants with this mutation can present with respiratory distress syndrome, most children with the mutation become symptomatic with interstitial lung disease beyond the newborn period.

ABCA3 is a large gene (80 kb) encoding a 1704-amino acid protein. Autosomal recessive mutations in ABCA3 can result in fatal surfactant deficiency in neonates and chronic respiratory insufficiency in older children. Function-altering mutations in ABCA3 result in surfactant that is deficient in phosphatidylcholine and has decreased function, suggesting that ABCA3 has an important role in pulmonary surfactant phospholipid homeostasis. A relatively common missense point mutation in ABCA3 causing the substitution of a valine for glutamic acid at codon 292 (E292V) is associated with chronic interstitial lung disease in older children who are heterozygous for this and a different function-altering mutation on the other ABCA3 allele.

NKX2.1 is essential for expression of SP-B, SP-C, and ABCA3. Haploinsufficiency of the gene encoding for the transcription factor NKX2.1 (also known as thyroid transcription factor 1 or TTF1) causes a complex phenotype in neonates that includes, albeit variably, hypothyroidism, brain abnormalities, and acute and chronic lung disease.

Although the lung diseases associated with mutations in SP-B, SP-C, ABCA3, and NKX2.1 disrupt surfactant homeostasis, cause varying degrees of surfactant accumulation (i.e., PAP), and are often cited as hereditary forms of PAP in the medical literature, they are distinguished from disorders of surfactant clearance by their surfactant dysfunction (i.e., functionally abnormal surfactant), histopathologic abnormalities (gross parenchymal derangement and interstitial disease), and clinical course. Consequently, the term pulmonary surfactant metabolic dysfunction (PSMD) disorders has been proposed for this group of genetic disorders.

Disease Associations

PAP has been reported in association with various medical conditions believed to cause the disorder, resulting in use of the term secondary PAP. Systemic disorders associated with secondary PAP include malignant and other hematologic diseases, nonhematologic malignancies, immune deficiency syndromes, chronic inflammatory syndromes, and chronic infections ( Table 70-1 ). In the past, chronic myelogenous leukemia and myelodysplastic syndrome were the most frequently reported hematologic disorders associated with PAP. In one single-institution study, PAP was reported in 5.3% of patients with hematologic malignancies and in 8.8% of those who were neutropenic. Of possible pathogenic significance, 9 of 10 patients in one study were neutropenic when PAP developed and, in most, PAP recovered spontaneously after bone marrow engraftment. In another, PAP developed during a period of neutropenia that resolved after treatment with granulocyte colony-stimulating factor. In a third report, PAP developed in three patients with acute myeloid leukemia with GM-CSF receptor abnormalities, including reduced levels of the β chain (three of three) and undetectable α chain (two of three). PAP has been found in about 10% of patients with lysinuric protein intolerance, an ultra-rare inherited disorder caused by defective cationic amino acid transport. Secondary PAP has been reported in association with various forms of immunodeficiency, including thymic alymphoplasia, immunoglobulin A deficiency, solid organ transplantation, and acquired immunodeficiency syndrome.

Although the mechanism(s) responsible for secondary PAP are poorly studied, two common themes have emerged when considered in the context of the critical role of alveolar macrophages in surfactant homeostasis. Conditions reducing either the numbers or functions of alveolar macrophages would be expected to reduce the capacity of resident alveolar macrophages to clear surfactant from the lung surface.

Environmental Factors

PAP has been associated with inhalation exposure to environmental agents including cigarette smoke, a wide variety of organic and inorganic dusts, fumes, and gases ( Table 70-2 ). Epidemiologic studies (see later) document that PAP is more common in smokers. Five of the 27 original patients reported by Rosen and colleagues had been exposed to wood dust. Most reports implicating other environmental exposures are single cases or small retrospective series and, in general, have not established a causal relationship between the exposure and development of PAP. One early report on 138 PAP patients found that half had significant exposure to dust and fumes, of which the most common exposure was to inhaled silica, noted in 10 individuals. With the advent of occupational protective equipment in the workplace, this form of “acute silicoproteinosis” is rare. Other case reports document the association of PAP with exposure to aluminum dust, cellulose fibers, cement dust, titanium dioxide, indium-tin oxide, and nitrogen dioxide. The link between silica and PAP is supported by studies in rats demonstrating that inhalation of silica particles disrupts surfactant homeostasis, resulting in increased surfactant accumulation. In five separate large series of patients with PAP, a history of dust exposure, when data were available, was present in 26% to 54% of patients. Further research is necessary to determine what role environmental inhalation exposure plays in the pathogenesis of PAP.

Mechanisms of Disruption of Surfactant Homeostasis

In mice, a series of observations has established GM-CSF as a critical regulator of pulmonary surfactant homeostasis, including the following: (1) surfactant homeostasis requires the presence of GM-CSF in the lung, and PAP develops in the absence of either pulmonary GM-CSF or its receptor; (2) surfactant accumulation in GM-CSF–deficient mice is caused by decreased catabolism by alveolar macrophages and not by increased production; (3) pulmonary GM-CSF is required for normal expression of transcription factor PU.1 in alveolar macrophages, and enforced (i.e., retroviral-mediated) expression of PU.1 restores surfactant catabolism in alveolar macrophages from GM-CSF knockout mice; and (4) surfactant homeostasis can be restored in GM-CSF receptor β-deficient mice by transplantation of bone marrow from a wild-type mouse.