The lavage effluents from 15 children without lung disease and 19 children with chronic obstructive bronchitis were used as controls or disease controls, for comparison with the lavage effluents that were available from our previously described cohort of neonatal, pediatric or juvenile patients with respiratory distress of unknown cause. These children were seen in western European medical hospital centers (mainly from France and Germany) and were analyzed for a genetic defect leading to deficiency in SP-B and SP-C 17 18 .

The lavage effluents from the children without lung disease were aliquots obtained previously in a study that assessed inflammation in children with chronic tracheostoma in comparison to these controls 19 . The lavage effluents were obtained during anesthesia for elective surgery for minor conditions. The usage of this material and that of the children with chronic bronchitis for this study was approved by the ethics committee at the University of Munich. Written informed consent was obtained from the patients where appropriate from age and from the caregivers.

Children with chronic obstructive bronchitis in whom anomalies of the airways, cystic fibrosis, primary ciliary dyskinesia, gastro-esophageal reflux, immuno deficiencies, allergic asthma and passive smoke exposure were excluded as causes and in whom a lavage was performed during the diagnostic work up, were used as a disease control group. The obstruction was determined by chest auscultation during the course of the disease. Details of these patients are given in table 1.

From the cases with SP-B deficiency we initially described, sufficient BAL material for analysis was available from 6 neonates (URD 6-II.1, 2-II.1, 7-II.1, 4-II.1, 3-II.1, 9-II.4), now labeled no-SP-B 1–6. All these babies had respiratory distress, and alveolar infiltrates with various degrees of interstitial involvement. A congenital heart disease or a lung disease due to mycoplasma, chlamydia, and viruses had also been ruled out. Details on the subjects are given in table 1. All but 2 subjects had mutations of SP-B as the cause for the SP-B deficiency.

From the cases with pulmonary alveolar proteinosis, sufficient BAL material for analysis was available from 15 children (URD 10-II.1, 11-II.3, 17-II.2, 25-II.3, 19-II.1, 20-II.2, 21-II.1, 16-II.2, 27-II.3, 22-II.1, 26-II.1, 23-II.3, 13-II.1, 13-II.2, 18-II.2), now labeled PAP 1–15. Most of these cases were less severely affected, had dyspnea and progressive cough, sometimes accompanied by cyanosis, finger clubbing, failure to thrive in the younger ones, and asthenia or weight loss in the others. Chest x-ray showed typical alveolar as well as interstitial infiltrates (table 1). In all these patients mutations of SP-B were excluded, 3 patients (PAP 04, PAP 10 and PAP 12) had heterozygous mutations in SFTPC. None of these children was investigated for ABCA3 mutations. All known secondary causes of PAP were excluded.

In addition, 7 subjects with chronic respiratory distress of unknown cause, in the absence of SP-B deficiency or alveolar proteinosis were investigated. BAL was available from 6 (URD 31-II.3, 40-II.1, 36-II.2, 30-II.1, 39-II.1, 37-II.2) of the initial 15 patients and from another infant born at 36 wks of gestation, with acute respiratory distress and development of chronic respiratory distress of unknown cause, after exclusion of SP-B, SP-C deficiency, and pulmonary alveolar proteinosis. None of these children was investigated for ABCA3 mutations. The children were labeled cRD 1–7 and their outcomes are given in table 1.

Routinely, the fluid recovered from BAL (4 × 1 ml 0.9% NaCl/kg body weight, b.w.) was pooled and the cells separated before analysis. Alternatively, in very sick neonates, repetitive tracheal aspirates after the instillation of 1 ml 0.9%NaCl/kg b.w. were collected over time periods of several hours up to a week, pooled and used for biochemical analyses.

All antisera used were polyclonal and raised in rabbits. The antibodies against SP-B (c329) and SP-C (22/96) were gifts from Dr W. Steinhilber, Altana AG, Konstanz, FRG and were used at a dilution of 1:10,000 20 . The antisera against pro-SP-B were raised against peptides of pro-SP-B, which were also used to determine the specificity of the signals on the immunoblots in all cases. The abbreviations and location of these peptides in the pro-SP-B sequence is indicated in figure 1. NFPROX was raised against SRQPEPEQEPGMSDPL, NFLANK against QARPGPHTQDLSEQQ, both were used at 1:2000 dilution. CFLANK was raised against GPRSPTGEWLPRDSECHLCMS, used at 1:1000 dilution and CTERMB was raised against LDREKCKQFVEQHTPQLLTL, used at 1:5000 dilution. Pro-SP-C was detected by anti-serum used at 1:5.000 dilution and raised previously against ESPPDYSTGPRSQ, i.e. Glu10–Gln23 of the amino acid sequence in pro-SP-C. The characteristics of all these antibodies has been described previously in detail 21 22 23 .

Total protein content of the samples was determined with the Biorad Protein Assay Kit (Biorad, Richmond, CA), which is based on the method by Bradford 24 . Ten to twenty-five μg of total protein were separated under reducing conditions on NuPage10% Bis-Tris gels using a NOVEX X-cell II Mini-Cell system (Novex, San Diego, CA). At least two sets of gels were prepared in parallel for each patient. Following electrophoresis the gels were either silver stained 25 , or subjected to Western transfer. For immunodetection, the proteins in the gels were transfered onto a PVDF membrane (ImmobilonP, Millipore, Bedford, MA) with a NuPage Blot module (Novex, San Diego, CA) according to the manufacturers recommendations.

Surfactant proteins and their pro-forms were detected on the PVDF membrane by immunoblot using the polyclonal rabbit antisera described in detail above, and the enhanced chemiluminescence assay (Amersham Biosciences, Buckinghamshire, UK) with horseradish peroxidase conjugated goat anti-rabbit polyclonal anti-IgG (1:10,000; Dianova, Hamburg, FRG).

To verify the specificity of the antibodies used to probe the pro-forms of SP-B and SP-C, a duplicate blot was prepared in each case and probed with an antibody solution containing 1 μM of the peptide, against which the antibody was raised. Antigen specific bands on the blot disappeared under these conditions. The blots were developed by exposure of X-ray film (Hyperfilm ECL, Amersham Biosciences, Buckinghamshire, UK) to the blots.

In the group of controls blots were first incubated with antibody against CTERMB and after that with the SP-B antibody respectively first with antibody against SP-C and after that with the pro-SP-C antibody, with and without competing peptide. In the other groups there were separate blots for each incubation with antibodies against SP-B and SP-C and their proforms, with and without competing peptide. Under these conditions the assay could detect about 2.5 ng of SP-B or SP-C per lane. In several experiments, aliquots of a patient with pro-SP-C forms were run in parallel as a positive control for pro-SP-C forms.

Immunoblots and silver stained gels were scanned with the Fluor-S MultiImager (Biorad, Richmond, CA) gel documentation system, and the resulting images were analyzed with the Software MultiAnalyst (Biorad, Richmond, CA).

To determine if the proteins that reacted with the CTERMB antibody on the immunoblots were glycosylated, the samples were deglycosylated before applying them on the gel (4). In brief, 1 unit of recombinant N-glycosidase F (Roche Molecular Biochemicals, Mannheim) was added to 500 μl incubation buffer (100 mM Na-phosphate, 25 mM EDTA, 1% β-mercaptoethanol, 0.5% Triton X-100, 0,1% SDS, pH 7.2). The vacuum dried sample was resuspended in 20 μl of this solution and incubated for 15 h at 37°C. The sample was then vacuum dried and analyzed by Western immunoblot.

For SFTPB mutation screening, first the 121ins2 frame-shift mutation was searched using the restriction enzyme cleavage SfuI endonuclease by PCR. In 121ins2-negative patients, SFTPB exons 1–11 and the promoter region were PCR-amplified and the purified PCR products served as templates in the sequencing reaction using Ready Reaction Dye Terminator Cycle Sequencing Kit With AmpliTaq^® DNA Polymerase, FS (PEBiosystems, Foster City, CA) with forward and reverse PCR oligonucleotides used as extension primers. Extension products were analyzed using the ABIPRISM™ 310 Genetic analysis System (PEBiosystems), as previously reported in detail 18 . Similarly, SFTPC exons 1–6 were analysed 17 .

Statistical calculations were performed with the Software GraphPad Prism 4.0 (GraphPad Software, San Diego, CA). Differences in nonparametric values were calculated with the Kruskal-Wallis test. For pair wise comparisons of groups we used Dunn's test (2). Differences in frequencies were calculated with the Fisher exact test. Correlation coefficients were determined according to Pearson. Results with a p ≤ 0.05 were considered significant.

The children with chronic obstructive bronchitis had a slight increase in neutrophils (3% (2; 15)(data are median and (25.; 75. percentile)) compared to children without lung disease (1% (1; 2); p = 0.035) and a somewhat lower viability (80% (70; 90) and 90% (80; 97) in children without lung disease; p = 0,035). The other variables did not differ and were within the normal range, i.e. children with chronic obstructive bronchitis: total cell count 150/μl (82; 275), macrophages 80% (69; 90) of total cells, lymphocytes 10% (4/14), eosinophils 0% (0; 2) and recovery was 54% (39; 70) and the children without lung disease: total cell count 115/μl (82; 180), macrophages 87% (82; 92) of total cells, lymphocytes 11.5% (7; 14.5), eosinophils 0% (0; 0.5) and recovery was 48% (42; 62).

Mature SP-B was regularly detected in all lavages from normal children and from those with chronic bronchitis at a median molecular weight of 7 kDa (Tab. 1, Fig. 2). Similarly, pro-SP-B forms with a molecular weight of 25–26 kDa were commonly observed using an antibody against the C-terminal flanking propeptide of pro-SP-B (Tab. 2, Fig. 2). Those bands never reacted with NFPROX, but showed reactivity with NFLANK, demonstrating that this was a processing intermediate generated by removal of the proximal N-terminal amino acids. A similar, but somewhat more truncated, 19–21 kDa pro-SP-B fragment was detected sporadically in these children (Tab. 2, Fig. 2). The pro-SP-B forms at 25–26 and 19–21 kDa were glycosylated as treatment with N-glycosidase F resulted in a significant drop in size for both peptides (not shown). A 40–42 kDa form and a 34–36 kDa form of pro-SP-B were rarely detected. Except for a single case when a 9 kDa C-terminal cleavage fragment was observed, in these children no other cleavage products of pro-SP-B processing were identified. Mature SP-C with M_r of 5.0 kDa was present in both controls and children with chronic bronchitis, whereas pro-SP-C forms were never detected in BAL (Tabs. 1 and 2, Fig. 2).

6 of all children investigated did not have SP-B in their lavages. Of these, 4 had lethal mutations of the SFTPB gene, i.e. SP-B deficiency (Tab. 3). Pro-SP-B processing products were not found in patient 5, having a 457delC/121ins2 compound heterozygote mutation (Fig. 3, Tab. 3). Unexpectedly, patients 3 and 6, homozygous for 121ins2, and patient 2 homozygous for 496delG had small but specific (competitive) pro-SP-B bands at about 19–21, 25–26 or 34–36 kDa (Tab. 3). Aberrant pro-SP-C bands previously thought to be diagnostic of SP-B mutations were only detected in 121ins2-mutations but not with 457delG 17 26 or with 496delG mutations.

In the other two infants with no SP-B in the lavages, SFTPB and SFTPC mutations were excluded 17 18 . These patients had significant amounts of pro-SP-B at 25–26 kDa, similar to that observed in the comparison groups. They also did not have pro-SP-C forms in their lavages, providing additional indirect evidence against SP-B processing defects. However, one of these two patients, i.e. patient 4 (Tab. 1), also lacked mature SP-C. This infant died at the age of 1 month from respiratory failure. This case suggests the presence of SP-B and SP-C processing defects arising by means other than from mutations of these genes, i.e. alterations in the protein processing machinery or in the lipid transporters, like ABCA3, as recently shown 27 . The other child (patient 1, Tabs. 1 and 3) is still alive with corticosteroids.

In all subjects with PAP, except patient 5, antibodies against GM-CSF in their sera or lavages were excluded in the pathogenesis of their disease. Although SP-B was abundantly present and mutations of SFTPB were excluded 18 , alterations of SP-B processing from other causes have not been excluded. In general, the same pro-SP-B processing products were observed as in the control and the chronic bronchitis group, however, the 25–26 kDa band was stained by NFLANK at increased frequency (Tab. 4, Fig. 4, lanes 4 and 5). In addition, 15 kDa and 13 kDa bands were present that were only stained by NFPROX. These peptides represent the N-terminal cleaved processing fragments, which were detected only in these patients and not in the respective control group (Tab. 4, Fig. 4, lanes 6 and 7). Three of the PAP patients (PAP 14, PAP 05 and PAP 10) had bands reacting merely with CFLANK or NFLANK. These bands were at 8, 9, 11 and 12 kDa. These may represent imprecisely processed SP-B, still having not completely removed small N- or C-terminal peptide stretches (Figs. 1, Tab. 4).

Among the PAP patients, only 2 had consistent pro-SP-C bands (Tab. 4). Subject PAP 08, a patient with a heterozygous SFTPC mutation and previously described in detail, had 3 bands, and subject PAP 04, in whom no SP-C mutation was detected, had one band at 6 kD 17 . Those 2 patients with the SFTPC mutation g.2125G>A 17 had no pro-SP-C bands with this antibody.

The infants with chronic respiratory distress of unknown cause had no mutations of SFTPB or SFTPC, and normal SP-B and SP-C in their lavages (Tab. 1). Nevertheless, aberrant pro-SP-C was detected in one of these infants at 9 kDa (Tab. 5). Concerning the processing of pro-SP-B significant deviations from the pattern observed in the control groups were observed in some of these children with cRD. Indeed a pro-SP-B precursor at 40–42 kDa was observed more frequently in these patients (Fig. 5, Tab. 5). Similarly, as in PAP, bands reacting with NFPROX, representing fragments of the cleaved N-terminus, were detected (Tab. 5, Figs. 1 and 5).