Hereditary pulmonary alveolar proteinosis caused by recessive CSF2RB mutations

To the Editors:

Pulmonary alveolar proteinosis (PAP) is a syndrome characterised by accumulation of surfactant in alveoli resulting in respiratory insufficiency 1. Surfactant homeostasis is critical for lung function and is tightly regulated, in part, by pulmonary granulocyte-macrophage colony-stimulating factor (GM-CSF), which is required for surfactant clearance by alveolar macrophages 2 and alveolar macrophage maturation 1. The effects of GM-CSF are mediated by cell-surface receptors composed of GM-CSF-binding α-chains and affinity-enhancing β-chains (encoded by CSF2RA and CSF2RB, respectively) 3. Ligand binding activates signalling via multiple pathways including the signal transducer and activator of phosphorylation (STAT)5 4. Disruption of GM-CSF signalling causes PAP by impairing surfactant catabolism in alveolar macrophages 1. In 90% of patients, PAP is caused by neutralising GM-CSF auto-antibodies 5, 6. Through the Rare Lung Diseases Network global PAP detection programme, we identified PAP caused by recessive CSF2RA mutations and developed novel diagnostic methods to identify patients with PAP caused by GM-CSF receptor dysfunction 4, 7. Herein, we report a case of hereditary PAP caused by disruption of GM-CSF receptor β-chain function.

A previously healthy 9-yr-old female presented with bilateral pneumonia, followed 3 months later by progressive dyspnoea of insidious onset. The diagnosis of PAP was suggested by chest radiograph findings, high-resolution computed tomography and bronchoalveolar cytology, and was confirmed by surgical lung biopsy. Pulmonary histopathology was typical of primary PAP (fig. 1) and she was successfully treated by serial whole lung lavage therapy. Details of the case history are included in the online supplement. A GM-CSF auto-antibody test was negative and the serum GM-CSF level was increased (25.9 pg·mL⁻¹) suggesting GM-CSF receptor dysfunction as the molecular basis of PAP 4, 7. A molecular evaluation was undertaken and included GM-CSF receptor detection, STAT-5 phosphorylation, CSF2RA and CSF2RB nucleotide sequencing, and cloning as reported previously 4, 7. GM-CSF receptor α- and β-chains were detected on blood leukocytes from the patient and all family members by flow cytometry and Western blotting (data not shown) 4, 7. Nucleotide sequencing of leukocyte mRNA and genomic DNA revealed a normal CSF2RA sequence for the patient and all family members (data not shown). A single CSF2RB point mutation (c.812C>T) in exon 7 was identified in both mRNA and DNA from the patient (who was homozygous for the mutation) and from her parents (both heterozygous) but not from her brother (fig. 2a). This mutation caused a single amino acid change (p.Ser271Leu) in the GM-CSF receptor β-chain that impaired STAT5 phosphorylation in blood leukocytes following stimulation by GM-CSF and interleukin (IL)-3, but not IL-2 (fig. 2b). Since IL-3 stimulates STAT-5 phosphorylation via the β-chain common to GM-CSF, IL-3 and IL-5 receptors 8, we used IL-3 as an alternative means to demonstrate β-chain dysfunction in our patient (fig. 2b). Since IL-2 stimulates STAT5 phosphorylation independent of the β-chain, we used IL-2 as a positive control to demonstrate STAT5 phosphorylation in the patient's cells (fig. 2b). Consistent with disruption of GM-CSF receptor function, cell-mediated clearance of GM-CSF was impaired in the patient's leukocytes compared to those from a healthy control (fig. 2c). These results demonstrate that homozygous but not heterozygous CSF2RB^S271L mutations impair GM-CSF receptor function in parallel with the occurrence of PAP (fig. 2d).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1–

Radiographic and histopathological appearance of the lungs of the patients with pulmonary alveolar proteinosis caused by recessive CSF2RB^S271L mutations. a) Posterior-anterior chest radiograph at the time of diagnosis, 9 yrs of age. Alveolar infiltrates are present throughout both lung fields. b) High-resolution computed tomography of the chest. Ground-glass opacification is superimposed on thickened interlobular and septal lines. c) Chest radiograph prior to an annual whole lung lavage therapy at 17 yrs of age. d) Bronchoalveolar lavage (BAL) fluid cytology at diagnosis (Papanicolaou stain). e) BAL fluid cytology at diagnosis (Periodic acid-Schiff (PAS) stain). f) Surgical lung biopsy obtained at diagnosis. Note the presence of PAS-staining material filling alveoli and also present in terminal airways. g) High-power view showing that alveolar wall architecture is normal (haematoxylin and eosin stain). d, e, g) Scale bars = 50 μm. f) Scale bar = 200 μm.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2–

Nucleotide sequence of CSF2RB, granulocyte-macrophage colony-stimulating factor (GM-CSF) receptor signalling, GM-CSF clearance and pedigree of the patient's family. a) PCR-based nucleotide sequencing of genomic DNA using CSF2RB-specific primers (5′-TCTCGGAGCTGTTGGACACA-3′, 5′-AAGATGCTCACCCTGCATCTG-3′). Amino acid sequence numbers are based on the human CSF2RB cDNA with first base of the ATG codon labelled +1 (Genbank accession number NM_000395). A single C>T substitution at nucleotide 812 results in a single amino acid substitution Ser>Leu at amino acid 271. The patient was homozygous and both parents were heterozygous for this mutation (CSF2RB^S271L) while the brother did not carry the mutation. b) GM-CSF receptor signalling analysis. Fresh heparinised blood from the patient, a healthy individual (positive control) or a patient with recessive CSF2RA^R217X mutations (negative control) was incubated alone (−) or with GM-CSF (GM), interleukin (IL)-2 or IL-3 (each at 10 ng·mL⁻¹) for 15 min followed by Western blot analysis to detect phosphorylated signal transducer and activator of transcription (pSTAT5), total STAT5 or actin (to ensure equal loading of cell lysates). The IL-3 control demonstrates that the β-chain can function independently of the GM-CSF receptor α-chain. c) Evaluation of cell/receptor-mediated GM-CSF clearance. Blood leukocytes isolated from the patient (•) or a healthy individual (□) or dishes of culture media without cells (○) were incubated with GM-CSF (1 ng·mL⁻¹) added at time 0. At subsequent times, GM-CSF concentration was measured in culture media by ELISA as described previously 4. ^#: significant differences in levels of GM-CSF in media in plates containing control cells compared to plates with no cells. d) Segregation of the CSF2RB^S271L allele in the patient's family. •: homozygous individuals; ▓: heterozygous individuals; □: non-carriers of the CSF2RB^S271L allele. The health status of each member is indicated and the arrow indicates the propositus.

Gene cloning and expression of GM-CSF receptors in human 293 cells 4, 7 reproduced the signalling defect caused by the CSF2RB^S271L mutation (fig. 3a). Interestingly, use of increased concentrations of GM-CSF for stimulation demonstrated partial functioning GM-CSF receptors derived from the CSF2RB^S271L allele (fig. 3b). Consistent with this, cell-mediated clearance of GM-CSF by receptors harbouring this mutation was reduced but not absent (fig. 3c and d).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3–

Reproduction and function of granulocyte-macrophage colony-stimulating factor (GM-CSF) receptors harbouring the CSF2RB^S271L mutation. Human embryonic kidney epithelial cells (293 cells) were evaluated alone (negative (-ve) control) or after transfection-mediated expression of a normal CSF2RA allele together with either a normal CSF2RB allele (positive (+ve) control), the patient's allele (CSF2RB^S271L) or a previously proposed mutant CSF2RB allele (CSF2RB^P603T; previously named P602T using a different naming convention 9). a) Cells were incubated with (+) or without (−) GM-CSF (10 ng·mL⁻¹, 15 min) and evaluated by immunoprecipitation and Western blotting as described previously 4 to detect phosphorylated signal transducer and activator of transcription (pSTAT)5 or total STAT5. Note, the CSF2RB^P603T -derived receptors resulted in signalling equal to the normal GM-CSF receptors. b) The same experiment was performed as in a) but increased GM-CSF concentrations were used for stimulation. Note, the partial signalling of CSF2RB^S271L-deirved GM-CSF receptors. c) Evaluation of the effects of the CSF2RB^S271L (•) or CSF2RB^P603T (▾) alleles on cell-mediated GM-CSF clearance. GM-CSF (1 ng·mL⁻¹) was added at time 0 and subsequently measured in culture media at the indicated times by ELISA. Note, the CSF2RB^S271L-derived receptors cleared GM-CSF at an intermediate rate compared to healthy control (CSF2RB^Normal; □) and that CSF2RB^P603T-derived receptors had normal GM-CSF clearance. ○: no cells. d) Evaluation of the effects of CSF2RB^S271L and CSF2RB^P603T mutations on cell-mediated GM-CSF clearance. Results are similar to those in c) except that measurements were made 24 h after initiating GM-CSF exposure (n = 3 independent determinations in each group). p<0.01 for comparison to cell-free control. ^#: GM-CSF was undetectable.

Previously, a CSF2RB-point mutation causing a substitution at amino acid 603 (P603T) was proposed as the molecular basis of PAP 10. We evaluated this mutation using the gene cloning approach described previously. The P603T mutation did not affect GM-CSF receptor function (fig. 3a) or cell-mediated GM-CSF clearance (fig. 3c and d). Furthermore, population studies showed it is present in ∼6% of the general population (NCPI single nucleotide polymorphism Database ID rs1801122; www.genecards.org). Based on these results, CSF2RB^P603T represents a sequence polymorphism rather than a genetic cause of PAP as proposed 10. Thus, this is the first reported case of hereditary PAP due to CSF2RB mutations.

Our findings demonstrate that CSF2RB is critical for surfactant homeostasis in humans and that homozygous but not heterozygous CSF2RB mutations impairing GM-CSF receptor function cause a hereditary form of primary PAP. They extend prior findings that PAP is caused by the absence of GM-CSF receptor β in mice 10. The age at onset, presentation, biomarkers, pulmonary histopathology, natural history and response to whole lung lavage therapy are similar to those of hereditary PAP caused by CSF2RA mutations 4, 7. Except for the earlier age at onset and some biomarkers (increased serum GM-CSF and absence of GM-CSF auto-antibodies), many features are similar to those of patients with autoimmune PAP 1. The remarkable similarities in pathogenic, radiographic, histopathological and clinical findings in PAP caused by CSF2RA 4, 7, CSF2RB mutations or high levels of neutralising GM-CSF auto-antibodies in humans 1, by injection of patient-derived human GM-CSF auto-antibodies in healthy non-human primates 6 or by CSF2RB mutations or GM-CSF deficiency in mice 2, suggest the pathogenesis of each is similar and support the usefulness of grouping them together as primary PAP. The partial functioning of CSF2RB^S271L-derived receptors suggests that aerosolised GM-CSF therapy may be of clinical benefit for this patient.

• ©ERS 2011