Review

Pulmonary alveolar proteinosis in adults: pathophysiology and clinical approach

Anupam Kumar, Basem Abdelmalak, Yoshikazu Inoue, Daniel A Culver

Lancet Respir Med 2018; Pulmonary alveolar proteinosis (PAP) is a diffuse disease that results from the accumulation of lipoproteinaceous 6: 554–65 material in the alveoli and alveolar macrophages due to abnormal surfactant homoeostasis. Identification of the Published Online granulocyte–macrophage colony-stimulating factor (GM-CSF) as an indispensable mediator of macrophage maturation January 31, 2018 and surfactant catabolism was the key discovery leading to the current understanding of the pathogenesis of most http://dx.doi.org/10.1016/ S2213-2600(18)30043-2 forms of PAP. Impaired GM-CSF bioavailability due to anti-GM-CSF autoimmunity is the cause of approximately Division of Pulmonary & 90% of adult PAP cases. Abnormal macrophage function due to endogenous or exogenous triggers, GM-CSF receptor Critical Care Medicine, defects, and other genetic abnormalities of surfactant production account for the remainder of causes. The usual Spectrum Health–Michigan physiological consequence of PAP is impairment of gas exchange, which can lead to dyspnoea, hypoxaemia, or even State University College of respiratory failure and death. occurs occasionally in patients with PAP. For patients with moderate Human Medicine, Grand Rapids, MI, USA (A Kumar MD); to severe disease, whole lung lavage is still the first-line treatment of choice. Supplemental GM-CSF is also useful, but Departments of General details about indications, choice of agent, and dosing remain unclear. Other therapies, including rituximab, Anesthesiology and Outcomes plasmapheresis, and lung transplantation have been described but should be reserved for refractory cases. Research, Anesthesiology Institute (Prof B Abdelmalak MD), Introduction subfamily A member 3 (ABCA3), and thyroid transcription Department of Pulmonary Since its first description in 1958 by Rosen and factor-1.3,5–7 Clinical disease due to these typically Medicine, Respiratory colleagues,1 our understanding of the pathogenesis, manifests in neonates, often with concomitant pulmonary Institute, and Department of natural history, and treatment of pulmonary alveolar fibrosis, and will not be discussed further. PAP that does Pathobiology, Lerner Research Institute (D A Culver DO), proteinosis (PAP) has evolved considerably. PAP is not match any of these types is called unclassified PAP. Cleveland Clinic, Cleveland, OH, characterised by the accumulation of surfactant in Autoimmune PAP accounts for the vast majority (90–95%) USA; and Clinical Research alveolar macrophages and alveoli, resulting in impaired of adult cases, whereas secondary PAP comprises 5–10% of Center, National Hospital gas exchange. Clinically, patients with PAP can be adult PAP.8 Although large-scale multicentre studies Organization Kinki-Chuo Chest Medical Center, Sakai, Osaka, asymptomatic or present with progressive dyspnoea, focusing on epidemiology are scarce, data from a Japanese Japan (Prof Y Inoue MD) worsening oxygen requirements, and have an increased registry reveal estimated incidence of 0·24–0·49 per million Correspondence to: susceptibility to infections. PAP has a variable clinical population and an estimated prevalence 2·04–6·2 cases per Dr Anupam Kumar, Division of course, ranging from spontaneous resolution to death million population for autoimmune PAP.8 Pulmonary & Critical Care from infections or progressive respiratory failure. In this Medicine, Spectrum Health–Michigan State Review, we elucidate current knowledge of the University College of Human pathogenesis of PAP, and also discuss the diagnosis, Key messages Medicine, Grand Rapids, treatment, and future developments. MI 49503, USA • Pulmonary alveolar proteinosis (PAP) results from [email protected] alteration of surfactant homeostasis in the lung, usually Pathogenesis and classification of PAP due to impaired alveolar macrophage function. The principal abnormalities in adult PAP result from poor • Abnormalities involving granulocyte–macrophage surfactant clearance and altered surfactant homoeostasis in colony-stimulating factor (GM-CSF)-mediated alveolar the lung due to impaired alveolar macrophage function macrophage maturation constitute the most common (figure 1). Attenuated alveolar macrophage maturation is pathway in the pathogenesis of PAP. typically caused by inadequate granulocyte–macrophage • Autoimmune PAP (accounting for more than 90% of colony-stimulating factor (GM-CSF) signalling, which is adult PAP) occurs when anti-GM-CSF antibodies lead to crucial for development of the full range of alveolar deficiency of bioavailableGM-C SF. 2 macrophage metabolic and immune functions. Auto­ • Secondary PAP results from alveolar macrophage immune PAP, formerly known as idiopathic PAP, occurs dysfunction due to inhalational exposures (dust), when anti-GM-CSF antibodies lead to deficiency of haematopoietic disorders, infections, drugs, or disorders bioavailable GM-CSF. Rarely, genetic defects in the of immune regulation. GM-CSF receptor α or β chains (CSF2RA, CSF2RB) also • In patients with clinically significant lung disease, lead to impaired macrophage maturation in the absence of whole-lung lavage remains the standard initial treatment 3 autoimmunity (congenital or hereditary PAP). Finally, method. alveolar macrophage function might be abnormal because • Therapy with subcutaneous or inhaled recombinant of several haematological or environmental factors that GM-CSF as initial treatment or as sequel to whole-lung 4 cause macrophage dysfunction (secondary PAP). Other lavage could provide benefit for patients with PAP. genetic abnormalities involving surfactant or However, there are no large-scale studies directly metabolism can also cause PAP, including mutations comparing its efficacy to that of whole-lung lavage. in the surfactant proteins B or C, ATP-binding cassette

554 www.thelancet.com/respiratory Vol 6 July 2018 Downloaded for Anonymous User (n/a) at Cleveland Clinic Foundation from ClinicalKey.com by Elsevier on July 07, 2018. For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved. Review

Autoimmune PAP A Normal B Pulmonary alveolar proteinosis Anti-GM-CSF antibody Monocyte blocks macrophage GM-CSF Anti-GM-CSF antibody maturation GM-CSF mediates macrophage Hereditary PAP maturation by of α or β chain binding to normal of GM-CSF receptor receptors, thus Normal (CSFRA, CSFRB) facilitating Type II alveolar degradation of Secondary PAP cell surfactant Toxins and dust exposure Haematological Mature macrophage Dysfunctional/ disorders immature macrophage Infections

Transcription Degraded Transcription factors factors ↓ P-STAT5, ↓ PU.1 P-STAT5, PU.1 surfactant

Undegraded surfactant Surfactant production disorders Capillary Mutation of SFTPB, SFTPC, ABCA, NKX- (TTF-)

Figure 1: Pathogenesis of pulmonary alveolar proteinosis (A) The normal process of granulocyte–macrophage colony-stimulating factor-mediated maturation of a monocyte into a functional alveolar macrophage and degradation of surfactant. (B) The various mechanisms by which macrophage maturation or macrophage function is impaired, which results in development of pulmonary alveolar proteinosis; disorders of surfactant metabolism (blue box) result in accumulation of abnormal surfactant or . Copyright of The Center for Medical Art & Photography (Cleveland Clinic, Cleveland, OH, USA). P-STAT5=phosphorylated signal transducer and activator of transcription 5.

Pulmonary surfactant, which is composed of played by GM-CSF in the proper functioning of pulmonary phospholipids (90%) and proteins (10%), is synthesised macrophages.19 by type II alveolar epithelial cells.9,10 The surfactant Nakata and colleagues2 first identified GM-CSF- molecules are then secreted into the alveolar space, neutralising antibodies in the serum and bronchoalveolar where they line the walls of the alveoli. Phospholipids, lavage fluid of patients diagnosed with what was formerly which are the major component of the surfactant, play a known as idiopathic PAP—the majority of these cases are crucial role in reducing surface tension at the air–liquid now classified as autoimmune PAP.20,21 These neutralising interface, which prevents end-expiratory alveolar antibodies were not isolated in patients with congenital or collapse, stabilises alveolar size, and reduces elastic recoil secondary PAP.20,22 Furthermore, injection of GM-CSF- of the lung.10 Surfactant proteins A, B, C, and D are neutralising antibodies from human beings into healthy crucial for proper organisation of the surfactant, but are non-human primates reproduced a lung phenotype also important mediators of innate immunity of the similar to human PAP, thus substantiating the hypothesis lung.11,12 Surfactant is catabolised by uptake and recycling that autoimmune PAP results from auto-antibody-​ in the type II alveolar epithelial cells or alveolar mediated disruption of GM-CSF signalling.23,24 macrophages.13 All recognised forms of adult PAP Emerging data also show defective lipid component are thus a result of abnormal alveolar macrophage matur­ metabolism in alveolar macrophages as an important ation or function. pathogenic mechanism, linking GM-CSF signalling with cholesterol homoeostasis in the .25–27 Role of GM-CSF in PAP Deficiency of GM-CSF might also contribute to the One of the major breakthroughs in our understanding of development of pulmonary fibrosis, which occurs in a the pivotal role of GM-CSF for alveolar macrophage minority of patients with PAP. Studies on type II function was the near fortuitous discovery that GM-CSF alveolar epithelial cells of mice with bleomycin-induced knock-out mice developed intra-alveolar accumulation of fibrosis showed diminished expression of GM-CSF surfactants, remarkably similar in phenotype to human mRNA.28 Furthermore, increased fibrosis was noted PAP.14,15 Mice deficient in the GM-CSF receptor β chain also after administration of neutralising anti-GM-CSF IgG developed a lung phenotype similar to that of GM-CSF- to the bleomycin-treated rats.28 One of the proposed deficient mice.16 Replacement of the GM-CSF and mechanisms for development of fibrosis in GM-CSF treatment with aerosolised GM-CSF were found to correct deficiency is the reduced synthesis of the potent the pulmonary abnormalities in GM-CSF-deficient mice.17,18 antifibrotic prostaglandin, PGE2.29 However, proof that These and further studies established the crucial role these observations are directly relevant to fibrosis in www.thelancet.com/respiratory Vol 6 July 2018 555 Downloaded for Anonymous User (n/a) at Cleveland Clinic Foundation from ClinicalKey.com by Elsevier on July 07, 2018. For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved. Review

human beings with PAP is insufficient; development of Panel: Type of PAP and causes fibrosis is probably multifactorial. Auto-immune PAP (most common in adults; 90–95% of cases) Caused by anti-GM-CSF antibody-mediated disruption of GM-CSF signalling pathway Differentiation of alveolar macrophages GM-CSF plays a key part in the terminal differentiation Secondary PAP (5–10% of adult cases) of alveolar macrophages, a process mediated by Caused by: receptor ligation that induces phosphorylation of signal • Haemopoietic disorders (myelodysplasia [GATA 2 deficiency], plasma cell disorders, transducer and activator of transcription 5 (STAT5), myeloid leukaemia, lymphoma); which in turn stimulates multiple signalling pathways.30 • Immune dysregulation (severe combined immunodeficiency [SCID], GM-CSF also stimulates expression of PU.1, a myeloid hypogammaglobulinaemia, Fanconi anaemia, Behcet’s syndrome, bone-marrow maturation transcription factor that facilitates maturation transplantation, lung transplantation); and differentiation of alveolar macrophages.19 Alveolar • Infections (nocardia, pneumocystis, cytomegalovirus, HIV); macrophages from GM-CSF-deficient mice have reduced • Inhalation (dust-silica [acute silicoproteinosis], cellulose fibres, aluminium, titanium PU.1 expression.31 Similarly, PU.1 mRNA expression and dioxide, indium, agricultural, bakery flour, sawdust; chlorine, gasoline, nitrogen PU.1-dependent terminal differentiation markers (eg, dioxide, plastic, varnish, or paint fumes); CD32, m-CSFR) were deficient in alveolar macrophages • Lysinuric protein intolerance; from people with autoimmune PAP.19 • Drug-induced or iatrogenic (busulfan, chemotherapy or antineoplastics, ciclosporin, dasatinib, fentanyl, hydrofluoric acid [inhaled], imatinib, leflunomide, mycophenolate Lung immunity mofetil, sirolimus, silicone breast implant, status after haemopoietic stem-cell or Aside from its role in surfactant metabolism, maturation bone-marrow transplantation) induced by GM-CSF is required for immune function of Congenital or hereditary PAP (<1% of cases) myeloid cells. Despite possessing normal ultrastructure Caused by recessive mutations with disruption of GM-CSF signalling (CSF2RA, CSF2RB and differentiation markers, neutrophils from human mutations); surfactant production disorders (SFTPB, SFTPC, ABCA3, NKX2-1 [TTF-1] beings with autoimmune PAP have impaired basal 32 mutations) and antimicrobial functions. GM-CSF, via PU.1, also enhances phagocytosis by alveolar macrophages.31 Unclassified PAP (<1% of cases) GM-CSF-deficient mice have increased susceptibility to a Caused by idiopathic or acquired cases and negative for anti-GM-CSF antibodies, wide range of microbial pathogens, including bacteria, without background causes of secondary PAP and congenital or hereditary PAP viruses, fungi, and parasites.33–36 Human studies also show that GM-CSF therapy improves PU.1 expression and restores innate immune function.19 Clinical observations Correlation with biomarkers Remarks in human beings suggest that there is a higher than LDH Moderate correlation with PaO , excellent Concentration increased in up to 80% of all 2 expected incidence of unusual infections in patients with correlation with high A-a gradient4 patients with PAP; not specific1 PAP, including nocardiosis and mycobacterial infections. Anti-GM-CSF Serum titre: poor correlation with PaO2, Increased serum concentration is antibody (A-a gradient, pulmonary function; BAL titre: approximately 100% sensitive and specific GM-CSF appears to be particularly crucial for control of good correlation with high serum LDH, for the diagnosis of autoimmune PAP;51 not Nocardia spp, and the presence of antibodies to GM-CSF low PaO , high A-a gradient and DLCO50 standardised; availability limited to 2 have been identified as an independent risk factor for specialised centres disseminated nocardiosis, even in seemingly healthy KL-6 Elevated levels in BAL fluid and blood Concentration increased in PAP but limited 37 significantly correlates with higher serum availability and not specific for diagnostic individuals. 52,53 LDH, low PaO2, and high A-a gradient use Impaired alveolar macrophage function due to Surfactant Significant correlation with high A-a Concentrations increased in PAP but limited systemic or toxic causes is the hallmark of secondary proteins gradient, low vital capacity, and low DLCO54 availability and not specific for diagnostic PAP. Among the various causes of secondary PAP, (SP-A, SP-B, use haematological diseases (eg, myeloproliferative or SP-D) haematological disorders) tend to be the most common. CEA Significant correlation with high LDH and Concentration increased in PAP but test 55 Other triggers include solid malignancies, infectious low PaO2 specific for diagnostic use CYFRA 21-1 Significant predictor of disease severity and Concentration increased in PAP but limited diseases, inhalational exposures, autoimmune diseases, effectiveness of GM-CSF based therapy56 availability; sensitivity and specificity not and medications (panel). Occupational exposure to dust known inhalation is directly toxic to alveolar macrophages, Serum YKL-40 Increased in serum and BAL fluid of patients Concentration increased in PAP but limited causing functional impairment that could lead to PAP. with PAP; correlates with DLCO and marker availability and not specific for diagnostic One of the first recognised occupational exposures of disease progression57 use associated with PAP was in silica workers, in whom the

PAP=pulmonary alveolar proteinosis. LDH=lactate dehydrogenase. PaO2=partial pressure of arterial oxygen. A-a presentation can be fulminant, and is frequently referred gradient=alveolar-arterial gradient. GM-CSF=granulocyte–macrophage colony-stimulating factor. KL-6=Krebs to as acute silicoproteinosis.38 Since then, aluminium, van den Lungen-6. BAL=bronchoalveolar lavage. DLCO=diffusing capacity of the lungs for carbon monoxide. SP-A=. SP-B=surfactant protein B. SP-D=. CEA=carcinoembryonic antigen. titanium, and more recently, , have also CYFRA 21-1=cytokeratin-19 fragments. been implicated in development of PAP in exposed workers.39,40 A subset of these occupationally exposed Table: Correlation of biomarkers and serological tests with markers of PAP disease severity individuals might also have anti-GM-CSF antibodies,

556 www.thelancet.com/respiratory Vol 6 July 2018 Downloaded for Anonymous User (n/a) at Cleveland Clinic Foundation from ClinicalKey.com by Elsevier on July 07, 2018. For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved. Review

but the pathophysiological relationship between the tobacco use or environmental triggers and PAP has not exaggerated autoimmune response and the environ­ been established, there is a higher than expected proportion mental trigger is unclear. of smokers (50–70%) and dust exposure (20–50%) in reported cohorts.4,8,41 It is possible that the apparent higher Clinical features risk of PAP in men is due to historical differences in PAP typically presents in the third to sixth decade of life.4 exposure rather than intrinsic sex differences. The disease is more common in men, with a prevalence Symptoms of PAP are nonspecific. 50–90% of patients ratio of approximately 2:1.4,8 However, women might have with PAP report progressive dyspnoea.42 A substantial an earlier age of presentation than men.4 There is no known number of patients also develop cough, which can be racial predilection. Although a causal relationship between productive. Other symptoms include fatigue, weight loss, chest discomfort, and arthralgias. Because of their increased susceptibility to infections, fever in patients with PAP should prompt further investigation, but fever can be a manifestation of PAP itself. Up to a third of patients can A be asymptomatic.8,43 The physical examination is likewise nonspecific. Clubbing is found in up to 25% of individuals, and crackles might also be present but are not universal.8 Pulmonary function tests typically reveal restrictive impairment, reduction of the diffusing capacity of the lungs for carbon monoxide (DLCO), or both.4,8,44 Metrics of gas transfer (eg, partial pressure of arterial oxygen

[PaO2], alveolar-arterial gradient [A-a gradient]) are also B commonly used surrogate markers for disease severity.8 Inoue and colleagues45 used composite indices such as the Disease Severity Score, which categorises patients

Figure 2: High resolution CT scan images of a patient with pulmonary ordinally from 1 to 5 on the basis of PaO2 obtained on alveolar proteinosis room air in the prone position, as well as by symptoms. Images show diffuse ground-glass opacification (A) and the so-called crazy paving pattern (B). The Disease Severity Score has been a useful endpoint for routine surveillance of patients and for clinical trials

Figure 3: Characteristic milky or opalescent appearance of fluid return from whole-lung lavage in a patient with pulmonary alveolar proteinosis L1 and R1 show the drained fluid after the first lavage cycle in the respective left and right lung; L2 and R2 bottles were collected after 9 L of lavage; and L3 and R3 show the clear drained fluid at the last lavage cycle, in this case after a total of 30 L per lung. The authors have used up to 56 L per lung in more severe cases.

www.thelancet.com/respiratory Vol 6 July 2018 557 Downloaded for Anonymous User (n/a) at Cleveland Clinic Foundation from ClinicalKey.com by Elsevier on July 07, 2018. For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved. Review

46 in patients with PAP. A-a gradient has also been used as However, the concentration of serum anti-GM-CSF a sensitive primary endpoint in clinical trials.8 antibody does not correlate with disease severity.50 In individuals with CSF2RA or CSF2RB mutations, GM-CSF Serological tests and biomarkers concentrations in the blood might be increased because In the years after the first description of PAP, lactate of defective clearance by the mutated receptors.3 If there dehydrogenase was investigated as a potential bio­ are no obvious exposures to secondary PAP risk factors, marker.47,48 Concentrations can be increased in up to the combination of increased GM-CSF concentrations 80% of patients with PAP, but it is not specific.4 In cases and no measurable anti-GM-CSF antibodies should of suspected autoimmune PAP, antibodies to GM-CSF prompt genetic testing for variant receptor . Several can be identified in almost all patients, both in the serum other serum analytes have been investigated as surrogate and bronchoalveolar lavage fluid.8,49,50 By contrast, serum biomarkers of disease severity and progression, but none autoantibodies to GM-CSF are not increased in patients have become widely accepted, due in part to lack of with congenital or secondary PAP.8,20,22 The sensitivity and specificity, limited clinical availability, and longitudinal specificity of serum anti-GM-CSF antibody approaches insensitivity to change (table).52–57 100%, making it the most useful diagnostic test.51 Radiology Chest radiographs typically show bilateral alveolar A infiltrates. Classically, a bat-wing appearance with central hilar prominence is seen, which can progress to confluent infiltrates involving all five lobes. High-resolution chest CT (HRCT) scans usually reveal patchy ground-glass opacities.58 Interlobular septal thickening is present in up to 85% of patients (figure 2).59 The pattern of septal thickening frequently identified with PAP has been termed crazy paving, referring to the polygonal appearance of the secondary pulmonary lobules with interspersed ground glass attenuation.60 Crazy paving occurs in 83% of patients with PAP.58,61 HRCT characteristics could be useful for distinguishing secondary from autoimmune PAP: crazy paving is unusual in secondary PAP, and the distribution of ground-glass opacities is typically diffuse rather than geographically concentrated.58 Radiological findings similar to PAP can be seen in numerous other conditions, such as pulmonary oedema, alveolar haemorrhage, organising pneumonia, acute respiratory B distress syndrome, pneumocystis pneumonia, and lipoid pneumonia.60,62 Architectural distortion of the lungs suggesting fibrosis is unusual in patients with PAP, with a frequency of 7–20%.63 Fibrosis on HRCT during initial presentation or at follow-up also portends a poor prognosis.59,63 Pleural effusion, enlargement of mediastinal lymph nodes, and evidence of air trapping are unusual and should prompt consideration of an alternate or concomitant diagnosis. Pulmonary nodules are also atypical, and should be assessed for malignancy or infection as warranted. Quantitative assessment of the intensity and extent of opacification on CT scans (CT grade score) has shown promise for evaluating therapeutic response in autoimmune PAP.46

Diagnosis of PAP Although clinical presentation and HRCT pattern alone Figure 4: Histopathology of pulmonary alveolar proteinosis can be fairly suggestive in typical cases, the first step in (A) Haematoxylin and eosin stain of diffuse acellular eosinophilic material in the alveoli, respiratory bronchioles, and confirming PAP is usually a bronchoscopy. Broncho-​alveolar terminal bronchioles. Many of the alveoli and bronchioles are slightly expanded by the presence of the material lavage might demonstrate characteristic opalescent or (magnification × 12·5). (B) Haematoxylin and eosin stain of alveoli; finely granular material is apparent, and the interstitium is relatively unremarkable, with no substantial interstitial inflammation (magnification × 100). Images milky-appearing fluid (figure 3). The milky consistency is courtesy of Dr Carol Farver (Pathology and Laboratory Medicine Institute, Cleveland Clinic, Cleveland, OH, USA). due to the high lipoproteinaceous content of the amorphous

558 www.thelancet.com/respiratory Vol 6 July 2018 Downloaded for Anonymous User (n/a) at Cleveland Clinic Foundation from ClinicalKey.com by Elsevier on July 07, 2018. For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved. Review

material that accumulates in the alveolar spaces. Under light microscopy, characteristic­ acellular oval bodies that High diagnostic suspicion of PAP (symptoms, physical signs, characteristic HRCT findings, elevated LDH) are basophilic on May–Grünwald–Giemsa staining can be Some countries use markers–eg, KL-6, SP-D, SP-A, CEA, CYFRA observed. Histopathology shows diffuse dense acellular eosinophilic material in the airways with minimal interstitial inflammation (figure 4). The accumulated Bronchoscopy BAL alone can be diagnostic; transbronchial −−Surgical lung biopsy PAP ruled out material is characteristically periodic acid Schiff-positive biopsy should be considered if gross and oil red O-positive.4,64 Transbronchial biopsy increases appearance of BAL not characteristic the diagnostic yield and should be considered, particularly ++ if the radiology and gross appearance of bronchoalveolar 8,44 lavage fluid are not characteristic. Addition of trans­ PAP confirmed bronchial biopsy also obviates the need for a surgical lung biopsy in most situations; less than 10–20% of patients require surgical biopsy for confirmation.8,44,65 There are no Measurement of anti-GM-CSF autoantibody (where available) data about the safety and the role of transbronchial cryobiopsy specifically for the diagnosis of PAP. In patients Elevated anti-GM-CSF Normal anti-GM-CSF with suspected secondary PAP, workup should be targeted autoantibody concentration autoantibody level towards identifying the underlying cause. Figure 5 is a proposed algorithm for diagnosis of suspected PAP.

Treatment Patient has been Elevated serum GM-CSF concentration, exposed to risk family history of PAP, genetic mutation, − Does not There are no international consensus guidelines for factors for no secondary PAP, No other causes of fit any secondary PAP elevated GM-CSF such as infections, criteria treatment of PAP. Conventionally, therapeutic decisions (panel) malignancies, etc are largely based on disease severity and type of PAP. + Patients with mild symptoms and no evidence of substantial physiological impairment can be monitored Autoimmune Secondary Congenital/hereditary PAP-confirm Unclassifiable PAP PAP (GM-CSF α or β receptor mutation, etc) PAP by serial assessment of symptoms, pulmonary function testing, oxygenation, and chest radiography. Patients with Figure 5: Diagnostic algorithm for patients with suspected PAP moderate-to-severe disease warrant a more aggressive PAP=pulmonary alveolar proteinosis. HRCT=high-resolution chest CT. SP-D=surfactant protein D. SP-A=surfactant approach. Similar to recommendations for patients with protein A. CEA=carcinoembryonic antigen. CYFRA 21-1=cytokeratin-19 fragments. LDH=lactate dehydrogenase. other chronic lung diseases, smoking cessation and BAL=bronchoalveolar lavage. GM-CSF=granulocyte–macrophage colony-stimulating factor. Copyright of The preventive vaccination with influenza and pneumococcal Center for Medical Art & Photography (Cleveland Clinic, Cleveland, OH, USA). immunisation are important. Figure 6 depicts a suggested algorithm for management of adult PAP. USA) can help clear the proteinaceous material. Recently, investigators removed up to 40% more protein and Therapeutic whole-lung lavage prolonged the time to relapse by modifying the classic Whole-lung lavage (WLL) has been the cornerstone of lavage technique: they manually ventilated the lavaged therapy since it was first described in the 1960s.66 Since lung with 300 mL of air using a bag five times, after the first iterations of WLL, therapeutic WLL has evolved instilling 500 mL of saline solution before draining it.71 into the modern-day practice of single or sequential In a survey of the global practice of WLL in PAP,72 fever bilateral lavage by isolating each lung with a double-lumen was the most frequently reported adverse event (18%), endotracheal tube under general anaesthesia.67 Segmental followed by hypoxaemia (14%). Rates for pneumothorax lung lavage is also done through the bronchoscope were 0·8%. Despite the scarcity of rigorous prospective (repeated bronchoalveolar lavage) in some centres, but studies, data from the available literature generally show the inferior adequacy of this technique, as well as the improved oxygenation and pulmonary function testing arduous nature of the procedure, has relegated it to after WLL. In the comprehensive analysis by Seymour and situations in which WLL is not available.68 Presneill,4 data in the 3-month period after WLL showed a

WLL is done under general anaesthesia. On very rare mean increase in PaO2 of 20 mm Hg (n=41; 95% CI occasions, when patients do not tolerate single-lung 15·6–24·6, p<0·0001), forced expiratory volume in 1 s of ventilation, extra-corporeal oxygenation has been used to 0·26 L (n=33; 0·09–0·42, p=0·0034), forced vital capacity of facilitate the lavage procedure.69,70 In most situations, 0·50 L (n=40; 0·33–0·67, p<0·0001), and diffusing capacity lavage of the two lungs is done in separate sessions of the lungs for carbon monoxide of 4·4 mL/mm Hg × min (single-lung lavage) a few days to weeks apart. In select (n=25; 2·6–6·3, p<0·0001). There was a mean reduction in centres, lavaging both lungs (bilateral lung lavage) is A-a gradient of –30·6 mm Hg (SD 18·0) in 21 patients done in a single setting, with the most affected lung being (95% CI –38·8 to 22·4, p<0·0001). The median duration of lavaged first. The use of a chest percussion device during benefit, defined as disease recurrence requiring repeat the lavage process such as Vest (Hill-Rom, St Paul, MN, therapeutic lavage, was 15 months;4 66% of the patients www.thelancet.com/respiratory Vol 6 July 2018 559 Downloaded for Anonymous User (n/a) at Cleveland Clinic Foundation from ClinicalKey.com by Elsevier on July 07, 2018. For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved. Review

GM-CSF therapy Diagnosis of PAP confirmed The finding that deficient GM-CSF bioavailability is the main pathophysiological defect in autoimmune PAP generated interest in the therapeutic use of recombinant

Asymptomatic, FVC-preserved or mild Symptoms or reduced FVC, moderate to severely GM-CSF. In the primary single-patient experience of its reduction, normal-to-mild reduction in reduced DLCO, hypoxia efficacy, use of subcutaneous recombinant human DLCO, O saturation-normal 2 GM-CSF (rhGM-CSF) led to substantial improvement of oxygenation.74 Since then, several cohort studies, using Check for infection or pulmonary fibrosis subcutaneous or nebulised delivery, have generated the bulk of the available efficacy data.75–79 No spontaneous remission after careful observation In a prospective open-label trial in 25 adult patients If secondary PAP, treat underlying problem If exposure-related, avoid further exposure with PAP, subcutaneous administration of escalating Smoking cessation doses of exogenous GM-CSF improved oxygenation, Preventive vaccination Improvement Referral to ILD centre for therapeutic WLL Monitor every 3 to 6 months with procedure if indicated (or rhGM-CSF inhalation radiographic findings, and symptoms in 48% of the symptom assessment, FVC, DLCO, oximetry, if available) patients who completed the trial.76 During follow-up, only serum biomarkers such as LDH, or chest 33% of patients who responded required further lavage x-ray (monitor for spontaneous remission) Recurrence or worsening Obtain chest HRCT scan if change in or long-term oxygen supplementation, compared with symptoms, pulmonary function, or chest Improvement Repeat WLL and consider addition of recombinant 56% of those who did not respond.76 The regimen was GM-CSF therapy (subcutaneous or inhaled) x-ray appearance well tolerated with a low incidence of local and systemic Recurrence or worsening adverse effects. Although results from some earlier studies suggested that serum anti-GM-CSF antibody titre Refractory PAP could predict response to rhGM-CSF therapy, data Continue WLL with or without GM-CSF, as tolerated plus supporting this are limited.80 In a separate study, higher forced vital capacity and normal serum lactate dehydrogenase were found to be predictors of better Referral for lung transplantation Consider rituximab with or Enrolment in trials, if available response, although both probably reflect milder disease.75 without plasmapheresis Nebulised rhGM-CSF (sargramostim) was studied in a prospective trial in 35 Japanese patients with autoimmune Figure 6: Treatment algorithm for adult patients with PAP PAP.78 The study participants were observed for 12 weeks, PAP=pulmonary alveolar proteinosis. FVC=forced vital capacity. DLCO=diffusing capacity of the lungs for carbon monoxide. LDH=lactate dehydrogenase. HRCT=high-resolution chest CT. ILD=interstitial lung disease. and subsequently received nebulised rhGM-CSF over a rhGM-CSF=recombinant human granulocyte–macrophage colony-stimulating factor. WLL=whole-lung lavage. total period of 24 weeks, using six cycles of a regimen of Copyright of The Center for Medical Art & Photography (Cleveland Clinic, Cleveland, OH, USA). either high-dose therapy (250 μg inhaled twice a day for the first 8 days of each 2-week cycle, followed by once-daily followed-up for more than 1 year from diagnosis (median dosing for 4 of 14 days for another 12 weeks) or low-dose of 37 months) required more than one lavage.4 In more therapy (125 μg inhaled twice a day for 4 of 14 days over recent studies from China41 and France,73 the relapse rates six cycles). A majority (24 [62%] of 39) of the patients ranged from 30–56%. In the French study,73 except for showed an improvement in both subjective (dyspnoea)

PaO2, which improved by 6·4 mm Hg (p=0·0213) after and objective parameters (6-minute walk distance); WLL, there were no significant changes in other pulmonary 29 (83%) of 35 patients did not require further therapy for function parameters.73 The discrepancy in efficacy estimates 1 year. Clinically significant side-effects were not observed. is probably due to differences in techniques across regions, With extended follow up of 30 months from the end of centre experience, evolution of the technique over time, inhalation therapy, 23 (66%) of 35 patients did not require and the limitations of retrospective data collection. additional treatments.81 Data from a worldwide survey of 368 patients Limitations with use of GM-CSF-based therapies relate reported that the mean number of procedures per to the dearth of placebo-controlled randomised trials, and patient was 2·5 (SD 1·5) in a 5-year period.72 There are small sample sizes. The effect on A-a gradient is variable no large-scale studies that have identified predictors of but generally moderate, and the responder rate is likewise response to WLL. In the historic paper by Seymour and suboptimal (figure 7). Fortunately, two clinical trials Presneill,4 younger age was the only variable that comparing inhaled rhGM-CSF with placebo are ongoing, predicted poor response to WLL. More recent ex­ using molgramostim (NCT02702180) and sargramostim periences suggest that a percent predicted DLCO less (NCT02835742). than 42% might identify patients who are likely to require repeat WLL.4,41 Overall, WLL, despite being an Therapies targeting autoimmunity invasive procedure, remains the foundation treatment Corticosteroids for moderate to severe PAP. A detailed review of the By analogy with other autoimmune diseases, corticosteroid­ technical aspects of WLL can be found elsewhere.67 The therapy seems logical for treatment of autoimmune PAP. See Online for video video shows the WLL procedure. However, most data from patients with PAP treated with

560 www.thelancet.com/respiratory Vol 6 July 2018 Downloaded for Anonymous User (n/a) at Cleveland Clinic Foundation from ClinicalKey.com by Elsevier on July 07, 2018. For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved. Review

corticosteroids suggest more harm than benefit. In a A-a gradient (mm Hg) Inhaled rhGM-CSF 82 retrospective study, patients treated with steroids had a Response (%) dose-dependent worsening of their Disease Severity Scores. 20 100 Although the exact explanation is unknown, it is likely that the negative effect corticosteroids exert on macrophage 16 80 phagocytosis and catabolism account for the observations.83

Thus, corticosteroid therapy of PAP should generally be Responders* (%) reserved only for those with concomitant connective tissue 12 Subcutaneous rhGM-CSF 60 disease or other steroid responsive processes. 8 40 Rituximab A-a gradient (mm Hg) Pathological concentrations of the anti-GM-CSF auto­ antibody that triggers autoimmune PAP are an obvious 4 20 target for therapy directed at B cells. The anti-CD20 monoclonal antibody rituximab has been studied in 0 0 small trials and several case reports.84,85 In an open-label, Seymour75 Venkateshiah76 Wylam77 Tazawa78 Papiris79 proof-of-concept trial in 10 patients with autoimmune Figure 7: Increase in A-a gradient and rate of response in trials of GM-CSF based therapies in patients with PAP PAP, a single treatment regimen of two doses of 1000 mg rhGM-CSF=recombinant human granulocyte–macrophage colony-stimulating factor. PAP=pulmonary alveolar of rituximab, administered two weeks apart, improved proteinosis. A-a gradient=alveolar-arterial gradient. *Each study defined responder differently. gas exchange in seven patients.86 The therapy also led to an approximately 25% red­uction in bronchoalveolar Treatment of hereditary and secondary PAP lavage anti-GM-CSF IgG concentrations, with evidence For patients with secondary PAP, the initial therapeutic of improved lipid transport and surfactant catabolism.87 focus should generally be on removing offending Although the data are promising, larger randomised exposures or treating the underlying medical disorder. studies must be initiated to substantiate a role for For individuals with myelodysplastic syndrome, the rituximab in the routine therapy of autoimmune PAP. At development of PAP is associated with poor survival.99 In present, rituximab is employed mainly as salvage therapy patients with advanced respiratory compromise, WLL in patients with refractory PAP. An ongoing prospective could be useful.100 Bone-marrow transplantation has been trial of rituximab therapy for anti-cytokine autoantibody- considered in myelodysplastic syndrome complicated by associated diseases, which will include patients with secondary PAP, but the data are limited to case PAP, is underway in the USA (NCT01842386). reports.101,102 Bone-marrow transplantation, instead of lung transplantation, could also have a role in patients Plasmapheresis with abnormal GM-CSF signalling due to GM-CSF Plasma exchange to reduce autoantibody levels has been receptor variants.91 used with marginal success in a few patients with autoimmune PAP that is refractory to conventional Prognosis therapy. Experience in these cases has varied from using The clinical course of PAP is highly variable, ranging an extended protocol that spanned months, to an from spontaneous resolution to death due to progressive abridged version of 5 days, followed by rituximab.88,89 respiratory failure or infection.4 Although observations However, results with plasmapheresis in patients with from early cohorts suggested that spontaneous resolution PAP have not been consistent enough to warrant occurs in 25–30% of patients, more recent studies recommendation for routine use.90 observed resolution in only 7–8% of patients.4,103,104 There are no known reliable clinical features that can predict Lung transplantation which patients will have spontaneous remission. In a There is minimal experience with lung transplantation in large study starting in the 1950s by Seymour and patients with PAP. Recurrence in the lung allograft is one Presneill,4 the disease-specific 5-year survival was of the major concerns with transplantation in PAP.91–93 estimated to be 88% (SD 4), with survival rates of 79% at Recurrence in transplant recipients could be attributed to 2 years, 75% at 5 years, and 68% at 10 years. No difference the immune dysregulation that is persistent in patients was observed between sexes.4 The use of WLL was with PAP, particularly those with underlying genetic associated with better 5-year survival compared with no mutations.91 Secondary PAP can also develop de novo in treatment (94% vs 85%, p=0·04), but it is unclear whether some transplant recipients, and is probably related to the procedure itself or other patient factors were disruption of the native immune system by responsible for the difference.4 12 (18%) of 65 deaths immunosuppressant anti-rejection medications.94–98 In from PAP were attributable to uncontrolled infections.4 patients with PAP that is refractory to conventional However, a more contemporary Japanese cohort study treatment, lung transplantation might be offered but it reported 100% 5-year survival, with an infection rate of would be prudent to rule out underlying genetic PAP. only 5%, probably reflecting advances in diagnosis, www.thelancet.com/respiratory Vol 6 July 2018 561 Downloaded for Anonymous User (n/a) at Cleveland Clinic Foundation from ClinicalKey.com by Elsevier on July 07, 2018. For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved. Review

treatment, and prevention of complications such as infections.8 Pulmonary fibrosis might occur in up to Search strategy and selection criteria 63 20% of patients and is a poor prognostic factor. We searched Medline and PubMed for reviews and original Prognosis for patients with secondary PAP is influenced articles about pulmonary alveolar proteinosis published in by the underlying illness and generally tends to be poorer English from Jan 1, 1950, to July 1, 2017. We used the terms 44,99 than for patients with autoimmune PAP. “pulmonary alveolar proteinosis” in combination with “pathogenesis”, “diagnosis” and “treatment”. We also searched Future research individual terms such as “GM-CSF”, “alveolar macrophages”, One of the major impediments to estimating prevalence, “whole lung lavage”, “recombinant GM-CSF”, “rituximab”, ascertaining diagnosis, and defining the actual effect of “plasmapheresis”, and “lung transplantation”. We chose therapies in PAP is the rarity of the condition. A US studies mostly published within the past 20 years, although we registry for patients with PAP is currently recruiting and did not exclude earlier reports, especially if they were unique or should provide further insights into the prevalence and high-quality clinical trials. We also searched the reference lists genetic risk factors for development of the disease. The of articles identified by this search strategy and from our own registry also aims to demonstrate the feasibility of personal files. We used original reports when available. measuring GM-CSF antibodies with a diagnostic blood spot card (NCT02461615). The indications and benefits of therapy need to be macrophages, possibly paving the way for macrophage further delineated. WLL has remained the primary transplantation in humans. Thus, the concept can be therapy for patients with PAP, but there are no large-scale plausibly extended to human PAP when receptor defects randomised studies that have substantiated its long-term are the cause, opening the path for future use of benefits. Randomised trials that have compared WLL gene-corrected pluripotent stem-cell-derived monocytes with recombinant GM-CSF-based therapies have also not and macrophages.113 However, it is worth mentioning been done. Instances have been reported of use of one that genetic mutation-driven disturbances of GM-CSF method when the other one is unsuccessful, but they are signalling constitute a substantial minority of patients limited to case reports.105,106 Sequential GM-CSF inhalation with PAP. For congenital genetic defects, type II alveolar followed by WLL, and not the other way around, has also epithelial cells will be the requisite target. In this regard, been reported with remarkable results in a small number preliminary studies on human amnion epithelial cells of patients.107 Thus, further studies will need to answer have shown that, in appropriate culture media, they can questions regarding the identification of appropriate develop into alveolar epithelial phenotypes that might candidates for these therapies, and determining efficacy, be able to secrete surfactant proteins.114 Jacob and safety, and feasibility of combining these two treatment colleagues115 also reported successful generation of methods. type II alveolar epithelial cells from human pluripotent Impaired lipid metabolism, which is mediated by stem cells, again raising their potential use in lung GM-CSF, has also been proposed as a fundamental tissue regeneration. Although this finding is promising pathogenetic defect in PAP.25–27 This notion has spurred for restoration of normal homoeostasis, no such studies interest in the potential use of peroxisome proliferator- have been done specifically in patients with PAP. activated receptor-γ (PPAR-γ) agonists (eg, pioglitazone) in the treatment of PAP, given the crucial role of PPAR-γ in Conclusions facilitating cholesterol export from macrophages through Since the original description of PAP, our understanding stimulation of ATP-binding cassette subfamily A member of the pathogenesis and classification of the disorder have 1 and ATP-binding cassette subfamily G member 1.26,108–110 evolved substantially. Impaired alveolar macrophage Preliminary studies in Csf2-deficient mice support the maturation and surfactant clearance due to abnormal efficacy of PPAR-γ agonists in reducing cholesterol levels GM-CSF signalling are now recognised as the in alveolar macrophages and the turbidity of broncho­ fundamental defects in autoimmune PAP, which alveolar lavage fluid.27 constitutes the majority of adult cases. However, PAP can Transplanting Csf2rb gene-corrected pulmonary be secondary to other triggers, particularly bone-marrow macrophages in Csf2rb-deficient mice was shown to be disorders or inhalational exposures, and is rarely safe and effective.111 The administration of macrophages hereditary. Although WLL is an invasive treatment not only corrected the lung disease and secondary method, it seems to be safe, well tolerated, and confers systemic manifestations, but also normalised symptomatic and physiological benefits, at least in centres disease-related biomarkers.111 One study112 successfully with experience. GM-CSF-based therapies (subcutaneous used a lentiviral vector (expressing a codon-optimised and inhalation) are promising but there is a dearth of data human CSF2RA-cDNA) to mediate transduction of to support GM-CSF supplementation as first-line therapy cultured macrophage cell lines and primary human or how to optimally combine it with WLL. Future studies cells to reconstitute GM-CSF receptor α expression, and about PAP will focus on fine-tuning our knowledge of thus restore GM-CSF signalling in hereditary PAP disease pathogenesis, and on identifying newer pathways,

562 www.thelancet.com/respiratory Vol 6 July 2018 Downloaded for Anonymous User (n/a) at Cleveland Clinic Foundation from ClinicalKey.com by Elsevier on July 07, 2018. For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved. Review

mutations, and biomarkers. For genetic causes of PAP, 21 Uchida K, Nakata K, Trapnell BC, et al. High-affinity cell-based therapies offer hope for a cure, but much work autoantibodies specifically eliminate granulocyte-macrophage colony-stimulating factor activity in the lungs of patients with is necessary before their clinical application. idiopathic pulmonary alveolar proteinosis. Blood 2004; Contributors 103: 1089–98. All authors contributed equally to the design and writing of the 22 Kitamura T, Uchida K, Tanaka N, et al. Serological diagnosis of manuscript. idiopathic pulmonary alveolar proteinosis. Am J Respir Crit Care Med 2000; 162: 658–62. Declaration of interests 23 Sakagami T, Beck D, Uchida K, et al. Patient-derived granulocyte/ We declare no competing interests. macrophage colony-stimulating factor autoantibodies reproduce pulmonary alveolar proteinosis in nonhuman primates. References Am J Respir Crit Care Med 2010; 182: 49–61. 1 Rosen SH, Castleman B, Liebow AA. Pulmonary alveolar proteinosis. N Engl J Med 1958; 258: 1123–42. 24 Sakagami T, Uchida K, Suzuki T, et al. Human GM-CSF autoantibodies and reproduction of pulmonary alveolar proteinosis. 2 Tanaka N, Watanabe J, Kitamura T, Yamada Y, Kanegasaki S, Nakata K. N Engl J Med 2009; 361: 2679–81. Lungs of patients with idiopathic pulmonary alveolar proteinosis express a factor which neutralizes granulocyte-macrophage colony 25 Thomassen MJ, Barna BP, Malur AG, et al. ABCG1 is deficient in stimulating factor. FEBS Lett 1999; 442: 246–50. alveolar macrophages of GM-CSF knockout mice and patients with pulmonary alveolar proteinosis. J Lipid Res 2007; 3 Suzuki T, Sakagami T, Young LR, et al. Hereditary pulmonary 48: 2762–68. alveolar proteinosis: pathogenesis, presentation, diagnosis, and therapy. Am J Respir Crit Care Med 2010; 182: 1292–304. 26 Baker AD, Malur A, Barna BP, et al. Targeted PPARγ deficiency in alveolar macrophages disrupts surfactant catabolism. J Lipid Res 4 Seymour JF, Presneill JJ. Pulmonary alveolar proteinosis: 2010; 51: 1325–31. progress in the first 44 years. Am J Respir Crit Care Med 2002; 166: 215–35. 27 Sallese A, Suzuki T, McCarthy C, et al. Targeting cholesterol homeostasis in lung diseases. Sci Rep 2017; 7: 10211. 5 Nogee LM. Alterations in SP-B and SP-C expression in neonatal lung disease. Annu Rev Physiol 2004; 66: 601–23. 28 Christensen PJ, Bailie MB, Goodman RE, O’Brien AD, Toews GB, Paine R 3rd. Role of diminished epithelial GM-CSF in the 6 Wert SE, Whitsett JA, Nogee LM. Genetic disorders of surfactant pathogenesis of bleomycin-induced pulmonary fibrosis. dysfunction. Pediatr Dev Pathol 2009; 12: 253–74. Am J Physiol Lung Cell Mol Physiol 2000; 279: L487–95. 7 Hamvas A, Deterding RR, Wert SE, et al. Heterogeneous 29 Moore BB, Coffey MJ, Christensen P, et al. GM-CSF regulates pulmonary phenotypes associated with mutations in the thyroid bleomycin-induced pulmonary fibrosis via a prostaglandin-dependent transcription factor gene NKX2-1. Chest 2013; 144: 794–804. mechanism. J Immunol 2000; 165: 4032–39. 8 Inoue Y, Trapnell BC, Tazawa R, et al. Characteristics of a large 30 Hansen G, Hercus TR, McClure BJ, et al. The structure of the cohort of patients with autoimmune pulmonary alveolar proteinosis GM-CSF receptor complex reveals a distinct mode of cytokine in Japan. Am J Respir Crit Care Med 2008; 177: 752–62. receptor activation. Cell 2008; 134: 496–507. 9 Carey B, Trapnell BC. The molecular basis of pulmonary alveolar 31 Shibata Y, Berclaz PY, Chroneos ZC, Yoshida M, Whitsett JA, proteinosis. Clin Immunol 2010; 135: 223–35. Trapnell BC. GM-CSF regulates alveolar macrophage differentiation 10 Serrano AG, Perez-Gil J. Protein–lipid interactions and surface and innate immunity in the lung through PU.1. Immunity 2001; activity in the system. Chem Phys Lipids 2006; 15: 557–67. 141: 105–18. 32 Uchida K, Beck DC, Yamamoto T, et al. GM-CSF autoantibodies 11 Chiba H, Piboonpocanun S, Mitsuzawa H, Kuronuma K, and neutrophil dysfunction in pulmonary alveolar proteinosis. Murphy RC, Voelker DR. Pulmonary surfactant proteins and lipids N Engl J Med 2007; 356: 567–79. as modulators of inflammation and innate immunity. Respirology 33 LeVine AM, Reed JA, Kurak KE, Cianciolo E, Whitsett JA. 2006; 11 (suppl): S2–6. GM-CSF-deficient mice are susceptible to pulmonary group B 12 Kingma PS, Whitsett JA. In defense of the lung: surfactant streptococcal infection. J Clin Invest 1999; 103: 563–69. protein A and surfactant protein D. Curr Opin Pharmacol 2006; 34 Paine R 3rd, Preston AM, Wilcoxen S, et al. 6: 277–83. Granulocyte-macrophage colony-stimulating factor in the innate 13 Wright JR. Clearance and recycling of pulmonary surfactant. immune response to Pneumocystis carinii pneumonia in mice. Am J Physiol 1990; 259: 1–12. J Immunol 2000; 164: 2602–09. 14 Dranoff G, Crawford AD, Sadelain M, et al. Involvement of 35 Carey B, Staudt MK, Bonaminio D, van der Loo JC, Trapnell BC. granulocyte-macrophage colony-stimulating factor in pulmonary PU.1 redirects adenovirus to lysosomes in alveolar macrophages, homeostasis. Science 1994; 264: 713–16. uncoupling internalization from infection. J Immunol 2007; 15 Stanley E, Lieschke GJ, Grail D, et al. Granulocyte/macrophage 178: 2440–47. colony-stimulating factor-deficient mice show no major 36 Riopel J, Tam M, Mohan K, Marino MW, Stevenson MM. perturbation of hematopoiesis but develop a characteristic Granulocyte-macrophage colony-stimulating factor-deficient mice pulmonary pathology. Proc Natl Acad Sci USA 1994; 91: 5592–96. have impaired resistance to blood-stage malaria. Infect Immun 2001; 16 Robb L, Drinkwater CC, Metcalf D, et al. Hematopoietic and lung 69: 129–36. abnormalities in mice with a null mutation of the common beta 37 Rosen LB, Rocha Pereira N, Figueiredo C, et al. Nocardia-induced subunit of the receptors for granulocyte-macrophage granulocyte macrophage colony-stimulating factor is neutralized by colony-stimulating factor and interleukins 3 and 5. autoantibodies in disseminated/extrapulmonary nocardiosis. Proc Natl Acad Sci USA 1995; 92: 9565–69. Clin Infect Dis 2015; 60: 1017–25. 17 Huffman JA, Hull WM, Dranoff G, Mulligan RC, Whitsett JA. 38 Xipell JM, Ham KN, Price CG, Thomas DP. Acute silicoproteinosis. Pulmonary epithelial cell expression of GM-CSF corrects the alveolar Thorax 1977; 32: 104–11. proteinosis in GM-CSF-deficient mice. J Clin Invest 1996; 97: 649–55. 39 Miller RR, Churg AM, Hutcheon M, Lom S. Pulmonary alveolar 18 Reed JA, Ikegami M, Cianciolo ER, et al. Aerosolized proteinosis and aluminum dust exposure. Am Rev Respir Dis 1984; GM-CSF ameliorates pulmonary alveolar proteinosis in 130: 312–15. GM-CSF-deficient mice. Am J Physiol 1999; 276: L556–63. 40 Cummings KJ, Nakano M, Omae K, et al. Indium lung disease. 19 Bonfield TL, Raychaudhuri B, Malur A, et al. PU.1 regulation of Chest 2012; 141: 1512–21. human alveolar macrophage differentiation requires 41 Zhao YY, Huang H, Liu YZ, Song XY, Li S, Xu ZJ. Whole lung granulocyte-macrophage colony-stimulating factor. lavage treatment of Chinese patients with autoimmune pulmonary Am J Physiol Lung Cell Mol Physiol 2003; 285: L1132–36. alveolar proteinosis: a retrospective long-term follow-up study. 20 Kitamura T, Tanaka N, Watanabe J, et al. Idiopathic pulmonary Chin Med J 2015; 128: 2714–19. alveolar proteinosis as an autoimmune disease with neutralizing 42 Suzuki T, Trapnell BC. Pulmonary alveolar proteinosis syndrome. antibody against granulocyte/macrophage colony-stimulating factor. Clin Chest Med 2016; 37: 431–40. J Exp Med 1999; 190: 875–80.

www.thelancet.com/respiratory Vol 6 July 2018 563 Downloaded for Anonymous User (n/a) at Cleveland Clinic Foundation from ClinicalKey.com by Elsevier on July 07, 2018. For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved. Review

43 Shah PL, Hansell D, Lawson PR, Reid KB, Morgan C. 67 Abdelmalak BB, Khanna AK, Culver DA, Popovich MJ. Pulmonary alveolar proteinosis: clinical aspects and current Therapeutic Whole-lung lavage for pulmonary alveolar proteinosis: concepts on pathogenesis. Thorax 2000; 55: 67–77. a procedural update. J Bronchology Interv Pulmonol 2015; 22: 251–58. 44 Bonella F, Bauer PC, Griese M, Ohshimo S, Guzman J, Costabel U. 68 Cheng SL, Chang HT, Lau HP, Lee LN, Yang PC. Pulmonary Pulmonary alveolar proteinosis: new insights from a single-center alveolar proteinosis: treatment by bronchofiberscopic lobar lavage. cohort of 70 patients. Respir Med 2011; 105: 1908–16. Chest 2002; 122: 1480–85. 45 Inoue Y, Nakata K, Arai T, et al. Epidemiological and clinical 69 Chauhan S, Sharma KP, Bisoi AK, Pangeni R, Madan K, features of idiopathic pulmonary alveolar proteinosis in Japan. Chauhan YS. Management of pulmonary alveolar proteinosis with Respirology 2006; 11 (suppl): S55–60. whole lung lavage using extracorporeal membrane oxygenation 46 Tokura S, Akira M, Okuma T, et al. A semiquantitative computed support in a postrenal transplant patient with graft failure. tomographic grading system for evaluating therapeutic response in Ann Card Anaesth 2016; 19: 379–82. pulmonary alveolar proteinosis. Ann Am Thorac Soc 2017; 70 Hasan N, Bagga S, Monteagudo J, et al. Extracorporeal membrane 14: 1403–11. oxygenation to support whole-lung lavage in pulmonary alveolar 47 Martin RJ, Rogers RM, Myers NM. Pulmonary alveolar proteinosis: proteinosis: salvage of the drowned lungs. shunt fraction and lactic acid dehydrogenase concentration as aids J Bronchology Interv Pulmonol 2013; 20: 41–44. to diagnosis. Am Rev Respir Dis 1978; 117: 1059–62. 71 Bonella F, Bauer PC, Griese M, Wessendorf TE, Guzman J, 48 Fountain FF, Jr. Lactate dehydrogenase isoenzymes in alveolar Costabel U. Wash-out kinetics and efficacy of a modified lavage proteinosis. JAMA 1969; 210: 1283. technique for alveolar proteinosis. Eur Respir J 2012; 40: 1468–74. 49 Bonfield TL, Russell D, Burgess S, Malur A, Kavuru MS, 72 Campo I, Luisetti M, Griese M, et al. Whole lung lavage therapy for Thomassen MJ. Autoantibodies against granulocyte macrophage pulmonary alveolar proteinosis: a global survey of current practices colony-stimulating factor are diagnostic for pulmonary alveolar and procedures. Orphanet J Rare Dis 2016; 11: 115. proteinosis. Am J Respir Cell Mol Biol 2002; 27: 481–86. 73 Gay P, Wallaert B, Nowak S, et al. Efficacy of whole-lung lavage in 50 Lin FC, Chang GD, Chern MS, Chen YC, Chang SC. Clinical pulmonary alveolar proteinosis: a multicenter international study of significance of anti-GM-CSF antibodies in idiopathic pulmonary GELF. Respiration 2017; 93: 198–206. alveolar proteinosis. Thorax 2006; 61: 528–34. 74 Seymour JF, Dunn AR, Vincent JM, Presneill JJ, Pain MC. 51 Uchida K, Nakata K, Carey B, et al. Standardized serum GM-CSF Efficacy of granulocyte-macrophage colony-stimulating factor in autoantibody testing for the routine clinical diagnosis of autoimmune acquired alveolar proteinosis. N Engl J Med 1996; 335: 1924–25. pulmonary alveolar proteinosis. J Immunol Methods 2014; 402: 57–70. 75 Seymour JF, Presneill JJ, Schoch OD, et al. Therapeutic efficacy of 52 Lin FC, Chen YC, Chang SC. Clinical importance of granulocyte-macrophage colony-stimulating factor in patients with bronchoalveolar lavage fluid and blood cytokines, surfactant idiopathic acquired alveolar proteinosis. Am J Respir Crit Care Med protein D, and Kerbs von Lungren 6 antigen in idiopathic 2001; 163: 524–31. pulmonary alveolar proteinosis. Mayo Clin Proc 2008; 83: 1344–49. 76 Venkateshiah SB, Yan TD, Bonfield TL, et al. An open-label trial of 53 Bonella F, Ohshimo S, Miaotian C, Griese M, Guzman J, granulocyte macrophage colony stimulating factor therapy for Costabel U. Serum KL-6 is a predictor of outcome in pulmonary moderate symptomatic pulmonary alveolar proteinosis. Chest 2006; alveolar proteinosis. Orphanet J Rare Dis 2013; 8: 53. 130: 227–37. 54 Seymour JF, Doyle IR, Nakata K, et al. Relationship of anti-GM-CSF 77 Wylam ME, Ten R, Prakash UB, Nadrous HF, Clawson ML, antibody concentration, surfactant protein A and B levels, and Anderson PM. Aerosol granulocyte-macrophage colony-stimulating serum LDH to pulmonary parameters and response to GM-CSF factor for pulmonary alveolar proteinosis. Eur Respir J 2006; therapy in patients with idiopathic alveolar proteinosis. Thorax 27: 585–93. 2003; 58: 252–57. 78 Tazawa R, Trapnell BC, Inoue Y, et al. Inhaled granulocyte/ 55 Fang SC, Lu KH, Wang CY, Zhang HT, Zhang YM. Elevated tumor macrophage-colony stimulating factor as therapy for pulmonary markers in patients with pulmonary alveolar proteinosis. alveolar proteinosis. Am J Respir Crit Care Med 2010; Clin Chem Lab Med 2013; 51: 1493–98. 181: 1345–54. 56 Arai T, Inoue Y, Sugimoto C, et al. CYFRA 21-1 as a disease severity 79 Papiris SA, Tsirigotis P, Kolilekas L, et al. Long-term inhaled marker for autoimmune pulmonary alveolar proteinosis. granulocyte macrophage-colony-stimulating factor in autoimmune Respirology 2014; 19: 246–52. pulmonary alveolar proteinosis: effectiveness, safety, and lowest 57 Bonella F, Long X, He X, et al. Serum YKL-40 is a reliable biomarker effective dose.Clin Drug Investig 2014; 34: 553–64. for pulmonary alveolar proteinosis. Respirology 2017; 22: 1371–78. 80 Bonfield TL, Kavuru MS, Thomassen MJ. Anti-GM-CSF titer 58 Ishii H, Trapnell BC, Tazawa R, et al. Comparative study of predicts response to GM-CSF therapy in pulmonary alveolar high-resolution CT findings between autoimmune and secondary proteinosis. Clin Immunol 2002; 105: 342–50. pulmonary alveolar proteinosis. Chest 2009; 136: 1348–55. 81 Tazawa R, Inoue Y, Arai T, et al. Duration of benefit in patients with 59 Holbert JM, Costello P, Li W, Hoffman RM, ogersR RM. CT autoimmune pulmonary alveolar proteinosis after inhaled features of pulmonary alveolar proteinosis. AJR Am J Roentgenol granulocyte-macrophage colony-stimulating factor therapy. 2001; 176: 1287–94. Chest 2014; 145: 729–37. 60 Rossi SE, Erasmus JJ, Volpacchio M, Franquet T, Castiglioni T, 82 Akasaka K, Tanaka T, Kitamura N, et al. Outcome of corticosteroid McAdams HP. “Crazy-paving” pattern at thin-section CT of the lungs: administration in autoimmune pulmonary alveolar proteinosis: radiologic-pathologic overview. Radiographics 2003; 23: 1509–19. a retrospective cohort study. BMC Pulm Med 2015; 15: 88. 61 Mehrian P, Homayounfar N, Karimi MA, Jafarzadeh H. Features of 83 Linden M, Brattsand R. Effects of a corticosteroid, budesonide, on idiopathic pulmonary alveolar proteinosis in high resolution alveolar macrophage and blood monocyte secretion of cytokines: computed tomography. Pol J Radiol 2014; 79: 65–69. differential sensitivity of GM-CSF, IL-1 beta, and IL-6. Pulm Pharmacol 1994; 7: 43–47. 62 Frazier AA, Franks TJ, Cooke EO, Mohammed TL, Pugatch RD, Galvin JR. From the archives of the AFIP: pulmonary alveolar 84 Amital A, Dux S, Shitrit D, Shpilberg O, Kramer MR. proteinosis. Radiographics 2008; 28: 883–99. Therapeutic effectiveness of rituximab in a patient with unresponsive autoimmune pulmonary alveolar proteinosis. 63 Akira M, Inoue Y, Arai T, et al. Pulmonary fibrosis on Thorax 2010; 65: 1025–26. high-resolution CT of patients with pulmonary alveolar proteinosis. AJR Am J Roentgenol 2016; 207: 544–51. 85 Borie R, Debray MP, Laine C, Aubier M, Crestani B. Rituximab therapy in autoimmune pulmonary alveolar proteinosis. 64 Burkhalter A, Silverman JF, Hopkins MB, 3rd, Geisinger KR. Eur Respir J 2009; 33: 1503–06. Bronchoalveolar lavage cytology in pulmonary alveolar proteinosis. Am J Clin Pathol 1996; 106: 504–10. 86 Kavuru MS, Malur A, Marshall I, et al. An open-label trial of rituximab therapy in pulmonary alveolar proteinosis. Eur Respir J 65 Goldstein LS, Kavuru MS, Curtis-McCarthy P, Christie HA, 2011; 38: 1361–67. Farver C, Stoller JK. Pulmonary alveolar proteinosis: clinical features and outcomes. Chest 1998; 114: 1357–62. 87 Malur A, Kavuru MS, Marshall I, et al. Rituximab therapy in pulmonary alveolar proteinosis improves alveolar macrophage lipid 66 Ramirez J, Kieffer RF Jr, Ball WC Jr. Bronchopulmonary lavage in homeostasis. Respir Res 2012; 13: 46. man. Ann Intern Med 1965; 63: 819–28.

564 www.thelancet.com/respiratory Vol 6 July 2018 Downloaded for Anonymous User (n/a) at Cleveland Clinic Foundation from ClinicalKey.com by Elsevier on July 07, 2018. For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved. Review

88 Luisetti M, Rodi G, Perotti C, et al. Plasmapheresis for treatment of 103 Kariman K, Kylstra JA, Spock A. Pulmonary alveolar proteinosis: pulmonary alveolar proteinosis. Eur Respir J 2009; 33: 1220–22. prospective clinical experience in 23 patients for 15 years. Lung 1984; 89 Garber B, Albores J, Wang T, Neville TH. A plasmapheresis protocol 162: 223–31. for refractory pulmonary alveolar proteinosis. Lung 2015; 104 Prakash UB, Barham SS, Carpenter HA, Dines DE, Marsh HM. 193: 209–11. Pulmonary alveolar phospholipoproteinosis: experience with 90 Jezequel A, Kerjouan M, Lederlin M, Laine-Caroff C, Camus C, 34 cases and a review. Mayo Clin Proc 1987; 62: 499–518. Delaval P, et al. Plasmapheresis failure in the treatment of 105 Yu H, Sun X, Wang Y, Xu Z, Huang H. Whole lung lavage auto-immune pulmonary alveolar proteinosis. Rev Mal Respir 2017; combined with Granulocyte-macrophage colony stimulating factor 34: 240–43. inhalation for an adult case of refractory pulmonary alveolar 91 Takaki M, Tanaka T, Komohara Y, et al. Recurrence of pulmonary proteinosis. BMC Pulm Med 2014; 14: 8 7. alveolar proteinosis after bilateral lung transplantation in a patient with 106 Yamamoto H, Yamaguchi E, Agata H, et al. A combination therapy a nonsense mutation in CSF2RB. Respir Med Case Rep 2016; 19: 89–93. of whole lung lavage and GM-CSF inhalation in pulmonary alveolar 92 Santamaria F, Brancaccio G, Parenti G, et al. Recurrent fatal proteinosis. Pediatr Pulmonol 2008; 43: 828–30. pulmonary alveolar proteinosis after heart-lung transplantation in a 107 Ohkouchi S, Akasaka K, Ichiwata T, et al. Sequential GM-CSF child with lysinuric protein intolerance. J Pediatr 2004; 145: 268–72. inhalation after whole lung lavage for pulmonary alveolar proteinosis: a 93 Parker LA, Novotny DB. Recurrent alveolar proteinosis following report of five intractable cases. Ann Am Thorac Soc 2017; 14: 1298–304. double lung transplantation. Chest 1997; 111: 1457–58. 108 Malur A, Baker AD, McCoy AJ, et al. Restoration of PPARγ reverses 94 Albores J, Seki A, Fishbein MC, et al. A rare occurrence of lipid accumulation in alveolar macrophages of GM-CSF knockout pulmonary alveolar proteinosis after lung transplantation. mice. Am J Physiol Lung Cell Mol Physiol 2011; 300: L73–80. Semin Respir Crit Care Med 2013; 34: 431–38. 109 Yvan-Charvet L, Wang N, Tall AR. Role of HDL, ABCA1, and 95 Philippot Q, Cazes A, Borie R, et al. Secondary pulmonary alveolar ABCG1 transporters in cholesterol efflux and immune responses. proteinosis after lung transplantation: a single-centre series. Arterioscler Thromb Vasc Biol 2010; 30: 139–43. Eur Respir J 2017; published online Feb 23. 110 Bonfield TL, Thomassen MJ, Farver CF, et al. Peroxisome DOI:10.1183/13993003.01369-2016. proliferator-activated receptor-gamma regulates the expression of 96 Kadikoy H, Paolini M, Achkar K, et al. Pulmonary alveolar alveolar macrophage macrophage colony-stimulating factor. proteinosis in a kidney transplant: a rare complication of sirolimus. J Immunol 2008; 181: 235–42. Nephrol Dial Transplant 2010; 25: 2795–98. 111 Suzuki T, Arumugam P, Sakagami T, et al. Pulmonary macrophage 97 Tokman S, Hahn MF, Abdelrazek H, et al. Lung transplant recipient transplantation therapy. Nature 2014; 514: 450–54. with pulmonary alveolar proteinosis. Case Rep Transplant 2016; 112 Hetzel M, Suzuki T, Hashtchin AR, et al. Function and safety of 2016: 4628354. lentivirus-mediated gene transfer for CSF2RA-deficiency. 98 Hasan A, Ram R, Swamy T. Pulmonary alveolar proteinosis due to Hum Gene Ther Methods 2017; 28: 318–29. mycophenolate and cyclosporine combination therapy in a renal 113 Lachmann N, Happle C, Ackermann M, et al. Gene correction of transplant recipient. Lung India 2014; 31: 282–84. human induced pluripotent stem cells repairs the cellular phenotype 99 Ishii H, Seymour JF, Tazawa R, et al. Secondary pulmonary alveolar in pulmonary alveolar proteinosis. Am J Respir Crit Care Med 2014; proteinosis complicating myelodysplastic syndrome results in 189: 167–82. worsening of prognosis: a retrospective cohort study in Japan. 114 Moodley Y, Ilancheran S, Samuel C, et al. Human amnion epithelial BMC Pulm Med 2014; 14: 3 7. cell transplantation abrogates lung fibrosis and augments repair. 100 Chaulagain CP, Pilichowska M, Brinckerhoff L, Tabba M, Erban JK. Am J Respir Crit Care Med 2010; 182: 643–51. Secondary pulmonary alveolar proteinosis in hematologic 115 Jacob A, Morley M, Hawkins F, et al. Differentiation of human malignancies. Hematol Oncol Stem Cell Ther 2014; 7: 127–35. pluripotent stem cells into functional lung alveolar epithelial cells. 101 Nishinakamura R, Wiler R, Dirksen U, et al. The pulmonary Cell Stem Cell 2017; 21: 472–88. alveolar proteinosis in granulocyte macrophage colony-stimulating factor/interleukins 3/5 beta c receptor-deficient mice is reversed by bone marrow transplantation. J Exp Med 1996; 183: 2657–62. 102 Tabata S, Shimoji S, Murase K, et al. Successful allogeneic bone marrow transplantation for myelodysplastic syndrome complicated by severe pulmonary alveolar proteinosis. Int J Hematol 2009; 90: 407–12.

www.thelancet.com/respiratory Vol 6 July 2018 565 Downloaded for Anonymous User (n/a) at Cleveland Clinic Foundation from ClinicalKey.com by Elsevier on July 07, 2018. For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.