INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Pulmonary surfactant is the mixture of lipids and specific proteins that lowers alveolar surface tension and prevents atelectasis at end-expiration. Deficiency of pulmonary surfactant because of immaturity is the principal cause of respiratory distress syndrome (RDS) in prematurely born infants (1). Recently it has been recognized that a specific inability to produce surfactant protein B (SP-B) causes lung disease in full-term infants (2). SP-B is a hydrophobic protein that enhances the surface tension-lowering properties of surfactant phospholipids and is derived from proteolytic processing of a larger precursor protein (proSP-B) (3). Hereditary SP-B deficiency is an autosomal recessive disorder in which affected infants develop severe respiratory disease that clinically and radiographically resembles RDS in premature infants (4). However, while premature infants with RDS may be successfully treated with surfactant replacement and supportive care, infants with hereditary SP-B deficiency develop progressive respiratory failure that is refractory to all treatment modalities except lung transplantation (5). Besides the absence or marked reduction of SP-B, other abnormalities of surfactant metabolism, including aberrant processing of surfactant protein C (SP-C), have also been observed in the lungs of SP-B-deficient infants (6). The mechanisms causing these abnormalities are unknown, although they appear to be secondary to the absence of either mature SP-B or proSP-B. Similar abnormalities have been observed in the lungs of mice genetically engineered to be homozygous for a null mutation in the SP-B gene, which also develop respiratory disease and die in the newborn period (7). The incidence of hereditary SP-B deficiency is unknown.

SP-B is encoded by a single gene on human chromosome 2 containing 11 exons, of which the 11th is untranslated (8). The 79 amino-acid, mature SP-B peptide is encoded in exons 6 and 7, corresponding to codons 201 to 279 of the 2 kb SP-B mRNA (9). A net 2 base pair (bp) insertion into codon 121 of the SP-B mRNA (121ins2), causing a frame-shift and premature termination signal that results in complete absence of proSP-B and mature SP-B, has been identified in unrelated infants with hereditary SP-B deficiency (10). The finding of a common mutation may be useful for diagnosis and in population studies designed to estimate the incidence of disease by determining mutation frequency in the general population (6). However, the 121ins2 mutation is not the exclusive cause of SP-B deficiency, as other mutations have been identified, including mutations resulting in single amino acid substitutions that have been associated with partial or transient deficiency (11). The spectrum and nature of disease-causing mutations in SP-B deficient infants are unknown. The goals of the present study were as follows. (1) To identify the mutations responsible for hereditary SP-B deficiency in a cohort of affected infants in order to determine the relative contribution of the 121ins2 mutation in causing disease. (2) To characterize novel SP-B gene mutations and determine the ranges of expression of SP-B, proSP-B, and proSP-C associated with different types of mutations.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patient Samples

The study population consisted of infants from 32 unrelated families cared for at 25 medical centers in the United States, Europe, Asia, and Australia, in whom hereditary SP-B deficiency was established as the basis for their lung disease. Children from two families (A, B) were identified retrospectively with the diagnosis established by immunohistochemical examination of archived paraffin-embedded lung tissue. One was a full-term infant who died at 3 mo of age from respiratory failure who was identified previously as having the 121ins2 mutation on one allele and an unknown mutation on the other allele (15). Blood samples from the parents in Family B were referred to the investigators because two full-term infants born to these parents died from persistent pulmonary hypertension (16). Blood samples from Family C were referred for genetic studies because of a family history of multiple neonatal deaths caused by respiratory failure; DNA was first analyzed from the parents and then from a child who subsequently died from respiratory failure in the newborn period. The clinical and identical molecular findings in other members of this kindred have recently been reported independently by another group of investigators (13).

The remaining SP-B-deficient infants were identified as part of a prospective program designed to identify affected infants. Samples from infants suspected of having SP-B deficiency as the cause of their lung disease were referred to us for evaluation from January 1, 1994 through October 31, 1998. Criteria for evaluation included: (1) gestational age equal to or greater than 35 wk; and (2) lung disease severe enough to be treated with high frequency ventilation, extracorporeal membrane oxygenation, or conventional ventilation for more than 1 wk; (3) no clear etiology for the lung disease. Initial evaluation included analysis of DNA for the 121ins2 mutation as described below (all patients) and evaluation of SP-B expression in tracheal aspirates (n = 24) using an ELISA assay as previously described (17). Both alleles were characterized if the 121ins2 mutation was present or if SP-B was undetectable in tracheal aspirates by ELISA assay. When available blood was obtained from the parents of affected infants and their DNA prepared and analyzed. A diagnosis of SP-B deficiency was considered established if mutations were identified on both alleles in association with reduced or absent SP-B expression. Using these criteria, children from 29 families were identified as having hereditary SP-B deficiency, and they make up the remainder of the study group. The clinical course and molecular and genetic findings from one family (I) have been reported (14). Informed consent for genetic studies was obtained from the parents of all affected infants under a protocol approved by the institutional review boards of the institutions involved.

DNA samples from subjects of Northern European descent (generously provided by Drs. Garry Cutting and Harry C. Dietz) were analyzed anonymously for each mutation identified to determine whether they represented common polymorphisms. These samples had been obtained from adults with a family history of an unrelated genetic disorder (cystic fibrosis or Marfan's syndrome) who were generally in good health without a known history of lung disease, although detailed clinical information was unavailable because of the anonymous nature of the testing. For protein blotting studies, control lung tissue was obtained from patients undergoing lung transplantation, including normal lung tissue from an unused donor lung, a 15-yr-old male with pulmonary hypertension, and a 2-yr-old female with bronchopulmonary dysplasia, as previously described (2).

Antisera

Polyclonal antisera were prepared in rabbits against the surfactant proteins as listed below; production and characterization of these antisera have been described previously (6, 18). Antisurfactant protein A (SP-A) (R84030) was generated to deglycosylated, beta-eliminated, human SP-A purified from human alveolar proteinosis material. This antibody recognizes both glycosylated and deglycosylated forms of SP-A (18). Anti-SP-B (R28031) was generated to the mature, active, airway SP-B peptide (79 amino acid, M_r = 8,700) isolated from bovine lung extracts (17). Anti-proSP-B antisera were generated to either the full-length (381 amino acid), human recombinant SP-B (pre)proprotein (R4633 or R55522) or against a recombinant peptide containing 102 residues of the C-terminal fragment of the human SP-B proprotein (R96189). These antisera are specific for proSP-B in that they detect either the N- or C-terminal propeptides but do not recognize mature SP-B (20). Anti-proSP-C antisera were generated to a fusion protein containing glutathione-s-transferase (GST) and amino acids 1 to 20 from the amino terminus of the human SP-C proprotein and absorbed against GST (R68514) or against a recombinant fusion protein containing GST and amino acids 1 to 197 for the human SP-C proprotein and absorbed against GST (R76983). These antisera immunoprecipitate the 21 kD proprotein, as well as processing intermediates of M_r = 26,000, 21,000, 16,000 and 14,000. Preincubation of antisera with the fusion proteins results in ablation of immunostaining (6).

Protein Blotting

Homogenates of frozen lung tissue were prepared, separated by SDS-PAGE electrophoresis, transferred to nitrocellulose membranes, and immunoreacted for the surfactant proteins as previously described (2). The primary antisera used were R55522 (proSP-B) at a 1:1,000 dilution; R96189 (proSP-B) at a 1:1,000 dilution, R28031 (mature SP-B) at a 1:500 dilution, and R76983 (proSP-C) at a 1:1,000 dilution. Proteins were detected using either a colorimetric assay with an alkaline phosphatase-conjugated second antibody and nitro blue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate (Promega, Madison, WI), or enhanced chemiluminescence using a horseradish peroxidase-conjugated second antibody (Amersham Life Sciences, Arlington Heights, IL), followed by exposure to Kodak X-ray film for 30 s to 5 min. Results shown are representative of at least three separate experiments, each utilizing lung tissue from different parts of the lung, except where limited amounts of lung tissue permitted only duplicate analyses from a single piece of lung tissue (infants from Families E, G, and N).

Immunohistochemistry

Formalin-fixed, paraffin-embedded biopsy and/or autopsy samples were received for evaluation. Antigen retrieval methods were employed initially to insure that a negative result was due to a lack of protein and not due to a loss of antigenicity in older blocks, and they subsequently were employed on a routine basis. Sections 5 µm thick were cut on a rotary microtome and loaded onto polysine-coated slides (Fisher, Atlanta, GA). A Vectastain ABC Peroxidase Elite Rabbit IgG kit (Vector Laboratories, Inc., Burlingame, CA) was used to detect the antigen:antibody complexes as previously described (6). The enzymatic reaction product was enhanced with nickel cobalt to give a black precipitate, and the sections were counterstained with nuclear fast red. Sections selected for antigen retrieval were heated in sodium citrate buffer at pH 6.0 for 15 min at 90° C (21). Endogenous peroxidase was quenched, and then the sections were blocked with 2% normal goat serum before incubating overnight at 4° C with the primary antibody at dilutions ranging from 1:2,000 to 1:64,000, depending upon the specific antibody used. Controls included: (1) formalin-fixed, paraffin-embedded, surgical samples of human pediatric lung obtained from the Division of Pathology (Children's Hospital Medical Center, Cincinnati, OH), and (2) omission of the primary antibody to check for endogenous biotin and peroxidase activity, as well as nonspecific binding of the secondary antibody.

DNA Preparation and Analysis

Genomic DNA was prepared from blood leukocytes using a commercially available kit (Puregene; Gentra Systems, Minneapolis, MN), analyzed for the presence of the 121ins2 mutation by PCR amplification and restriction analysis using SfuI, and for possible novel SP-B gene mutations by use of heteroduplex analysis as previously described (10). Briefly, PCR products corresponding to the first 10 exons of the SP-B gene were generated with the polymerase chain reaction using primers and conditions listed in Table 1. The amplimers were heated to 95° C for 5 min, allowed to slowly cool to 37° C, and then separated on a 40-cm nondenaturing polyacrylamide gel (MDE; FMC Bioproducts, Rockland, ME). PCR products from infants known to have the 121ins2 mutation on one allele were analyzed directly, along with PCR products from their parents when available. PCR products from infants in whom the genotype was completely unknown were analyzed directly, and by mixing an equal amount of a PCR product from the corresponding exon from a patient in whom the SP-B sequence was identical to the published sequence. DNA bands were visualized by UV illumination after staining in 0.5 µg/ml ethidium bromide for 30 min. PCR products were sequenced by end-labeling a specific primer with ³³P-dATP using T₄ polynucleotide kinase (New England Biolabs, Beverly, MA), and cycle sequencing was performed using the Circumvent kit from New England Biolabs. Sequencing reaction products were separated on a 6% polyacrylamide gel, the gel dried, and exposed to Kodak X-ray film overnight. For patients in whom only genomic DNA was available, the remainder of the SP-B coding sequence (exons 1 through 10) was determined by direct sequencing of PCR products from the remaining exons. For patients in whom suitable tissue for RNA preparation was available, SP-B cDNA clones were generated and sequenced as described below. Patient SP-B genomic and/or cDNA sequences were compared with published SP-B DNA sequences (8, 9).

RT-PCR

RNA was prepared from frozen lung tissue by an acid-phenol extraction method using a commercially available reagent according to the manufacturer's directions (Trizol, BRL Life Technologies, Gaithersburg, MD) (10). Suitable tissue was available only from Patients A, D, E, H, M, and O. Five micrograms were reverse-transcribed using Superscript II (BRL Life Technologies) with an oligo dT primer (Boehringer Mannheim Biochemicals, Indianapolis, IN) and reagents and conditions supplied by the manufacturer. SP-B-specific cDNA was generated using primers and conditions as previously described (10). SP-B cDNA was subcloned into a plasmid vector (PCR 2.1; Invitrogen, La Jolla, CA) and sequenced as previously described (8, 9). Manual sequencing of plasmid clones using the dideoxy chain termination method was performed using a kit purchased from Amersham Life Sciences; automated sequencing was performed through the DNA analysis facility of the Johns Hopkins University School of Medicine.

Mutation Specific Analysis in Control Subjects

DNA sequence changes identified were analyzed for altered restriction sites with the aid of DNA sequence software (Gene Runner, Hastings, NY). Restriction enzymes specific for sequences altered by mutations were purchased from New England Biolabs and used to digest PCR products for the exon containing the specific mutation using reagents supplied by and conditions suggested by the manufacturer. Restriction products were analyzed by agarose gel electrophoresis; the concentration ranged from 1.5% to a 3:1 mixture of Nusieve agarose (FMC Bioproducts) to regular agarose (BRL Life Technologies), depending upon the resolution required. In some cases, larger PCR products spanning adjacent exons were generated so as to incorporate existing restriction sites as internal controls to assess completeness of digestion. For the C100G mutation, a PCR primer was synthesized corresponding to genomic nucleotides 2482 to 2509 (5'-TCTGGTC TCCCAGGACACGATAAGGAAG-3'), which eliminated 2 existing MnlI sites so as to facilitate the analyses. Exon 9 PCR amplification products from control DNA samples were analyzed on 3:1 nusieve: regular agarose gels, as these conditions resolved the heteroduplex bands from the native PCR products for the 1043ins3 and 1048del12 mutations.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Identification of SP-B Gene Mutations

Sixteen infants prospectively evaluated for hereditary SP-B deficiency were homozygous for the 121ins2 mutation, establishing the diagnosis in these children. In 13 of these families both parents were confirmed to be carriers of the 121ins2 mutation; DNA was unavailable from both biologic parents of two infants, and one infant was homozygous for the 121ins2 mutation as a result of apparent maternal isodisomy.

An additional 16 SP-B-deficient infants were identified who by restriction analysis had the 121ins2 mutation on only one (n = 10) or neither (n = 6) allele (Table 2). In 15 of 16 families, heteroduplex bands consistent with a possible mutation were observed in the PCR products from a single exon, as well as in exon 4 from subjects with the 121ins2 mutation (data not shown). DNA sequence analysis indicated a deviation from the published sequences consistent with a potential disease-causing mutation in each case, as well as in the one family (H) in which heteroduplex analysis was not informative. The six infants who did not have the 121ins2 mutation on either allele were all homozygous for the novel mutation identified, as confirmed by restriction analysis or direct sequencing; a family history of consanguinity was obtained in each case. The only other variation from the published SP-B coding sequence observed in any patient was a polymorphism in codon 131 involving either a C or T, such that the codon encoded either threonine or isoleucine. Both sequence variants have been recognized previously (8, 9, 13). In some cDNA clones, there was a 12 bp deletion corresponding to the beginning of exon 8; this splice variant has been observed previously (10). The locations and nature of the mutations identified are summarized in Table 3, and included three nonsense and two frame-shift mutations, six missense mutations, one in-frame insertion, one in-frame deletion, and three mutations located at or near splice junctions. One mutation (1043ins3) was present in two families who were unrelated by history. A subsequent child born to Family J also developed fatal neonatal respiratory disease, and analysis of DNA from this child demonstrated that both affected siblings had the same SP-B genotypes. SP-B was undetectable in lung fluid from all deficient infants from whom such samples were obtained, including 13 homozygous for the 121ins2 mutation.

DNA from 30 control subjects without lung disease was examined for each of the mutations identified by either restriction analysis using enzymes listed in Table 3 or by heteroduplex analysis. Restriction or heteroduplex products indicative of a given mutation were not observed in any exons of the control patients. Restriction analysis was also utilized to confirm the presence of the mutation in each patient and in their parents when DNA was available (data not shown). For the infants who were heterozygous for the 121ins2 mutation and another mutation, the second mutation was inherited from the parent who did not carry the 121ins2 mutation, indicating that the mutations were on different alleles. Although DNA was not available from parents of the infant heterozygous for the 121ins2 and the C248X mutations, sequence analysis of 6 cDNA clones revealed that a given transcript had only one or the other mutation, indicating that they were derived from different alleles.

Analysis of SP-B-specific mRNA by RT-PCR demonstrated that the predominant SP-B transcripts from two infants with potential splicing mutations were smaller than those generated from other infants and control subjects (Figure 1). Sequence analysis demonstrated that the transcripts lacked the sequences corresponding to exons (2 and 4) whose splice junctions were altered. Exon 2 contains 128 bp, and its elimination from the transcript causes a frame-shift and introduction of a signal for the termination of translation at codon 24. Exon 4 contains 126 bp, and its skipping would thus result in an in-frame deletion. An RT-PCR product corresponding to the normal-sized transcript was not detected in association with the 209+4A G mutation; a weak RT-PCR product corresponding to the 121ins2 allele (as indicated by digestion with SfuI) was sometimes observed in association with the 121ins2/ 282-2delA genotype when additional cycles in the PCR reaction were used (data not shown). Lung tissue was not obtained for RNA preparation from either sibling with the 696G A mutation to confirm aberrant splicing, which could be predicted to occur as the last nucleotide in an exon is an adenine only 10% of the time, and a guanine is found 78% of the time in primates (23).

View larger version (47K): [in this window] [in a new window]   Figure 1.   RT-PCR of SP-B specific cDNA from SP-B-deficient infants. Total lung RNA (5 µg) was reverse-transcribed and amplified with PCR using primers that span the entire translated SP-B sequence. Genotypes of the SP-B-deficient patients are listed; WT/ WT was amplified from RNA from normal lung tissue. SP-B cDNA was smaller in the patients with the 282-2delA and the 209+4A G mutations, consistent with deletions of exons 4 and 2, respectively, compared with SP-B cDNA of the expected size from normal lung tissue, from an SP-B-deficient infant homozygous for the 121ins2 mutation, and from infants with the 121ins2/ L13P and 121ins2/C235R genotypes.

Immunohistochemistry

To evaluate surfactant protein production associated with the different SP-B gene mutations, immunostaining was performed on lung tissue from infants, as noted in Table 2, and from four infants homozygous for the 121ins2 mutation. In all SP-B-deficient infants, strong extracellular staining for SP-A was observed by immunohistochemistry (data not shown). In addition, intense staining with antisera directed against proSP-C was observed in the extracellular spaces and in type II epithelial cells in all of the patients for whom SP-B gene mutations were identified. Staining for proSP-C was so intense in SP-B deficient infants that positive results were often achieved when the antisera was used at dilutions of 1:32,000 to 1:64,000, suggesting that high concentrations of proSP-C peptides were present in the lung tissue from these patients. Extracellular proSP-C staining was never observed in control tissues, nor was cellular staining detected in control tissues at antibody dilutions greater than 1:4,000.

Staining for proSP-B and mature SP-B varied depending upon the mutation and the use of antigen retrieval methods. Staining for both mature SP-B and proSP-B was absent in patients with nonsense or frame-shift mutations, even with antigen retrieval methods. In contrast, staining for proSP-B without the use of antigen retrieval was observed in association with the C100G, C235R, R252C, and 1043ins3 mutations. Antigen retrieval, however, was required for detection of proSP-B staining in association with the L13P, 282-2delA, 1048del12, and 209+4A G mutations. ProSP-B staining was generally confined to alveolar type II epithelial cells, although with antigen retrieval methods staining of the proteinaceous material was observed in association with the C235R, R252C, and 1043ins3 mutations. Similar results were obtained using both antisera directed against full-length proSP-B and against the carboxy-terminal portion of proSP-B.

Staining for mature SP-B was observed in both alveolar type II cells and in extracellular material in association with the 1043ins3 mutation. After antigen retrieval, staining for mature SP-B was also detected in alveolar type II cells in association with the C100G, C235R, and R252C mutations, and in extracellular material with the C100G and R252C mutations. Immunohistochemical analyses of lung tissue from patients with the W60X, 209+4G A, C100G, and1043ins3 mutations are depicted in Figure 2.

View larger version (162K): [in this window] [in a new window]   Figure 2.   Immunohistochemical analysis of lung tissue for surfactant proteins. Formalin-fixed, parrafin-embedded sections were cut and stained for the surfactant proteins as described in Methods. Rows 1 ( panels A, F, K, P) and 2 ( panels B, G, L, O) represent staining for mature SP-B, Rows 3 ( panels C, H, M, R) and 4 ( panels D, I, N, S) staining for proSP-B, and Row 5 ( panels E, J, O, T ) staining for proSP-C; antigen retrieval was employed in Rows 2 and 4. Column 1 ( panels A-E ) shows lung tissue from the infant homozygous for W60X. Similar staining patterns were observed for infants with the 121ins2/W39X, 121ins2/134insAA, and 121ins2/121ins2 genotypes. Column 2 ( panels F-J ) shows lung tissue from the infant homozygous for 209+4G A. Similar staining patterns were observed for infants with the121ins2/282-2delA, 121ins2/ L13P, and 121ins2/1048del12 genotypes. Column 3 ( panels K-O) shows lung tissue from the infant heterozygous for 121ins2 and C100G; similar staining patterns were observed for patients with the 121ins2/C325R and R252C/R252C genotypes. Column 4 ( panels P-T ) is from the infant homozygous for 1043ins3. Extracellular staining for proSP-C is present in all SP-B-deficient infants. Original magnification: ×300.

Western Blot Analysis

In order to determine whether the surfactant proteins from patients with SP-B gene mutations migrated at the expected molecular weights, protein blotting was performed on tissue from patients in whom frozen tissue was obtained (Table 2). ProSP-B of M_r = 40,000 daltons was detected in control infants and SP-B deficient infants with the C235R, R252C, and C100G mutations, but not in lung from SP-B-deficient infants with the L13P, C248X, W60X, or 209+4A G mutations (Figure 3, top panel). ProSP-B was detected in lung from the infant with the 282-2delA mutation, but migrating at a lower molecular weight (M_r = 35,000 daltons), the size difference consistent with that predicted by the deletion of exon 4 sequence from the mRNA. ProSP-B of M_r = 26,000 daltons, representing the first intermediate in the cleavage of proSP-B to mature SP-B, was detected in control patients, and in SP-B-deficient patients with the C235R, R252C, C100G, and 282-2delA mutations. Similar results were obtained using the antisera directed against full-length proSP-B, including the detection of a protein corresponding to the amino terminal portion of proSP-B, which migrated at a lower molecular weight in the infant with the 282-2delA mutation compared with that in control patients and patients with missense mutations (not shown). Mature SP-B was readily detected in control tissues, but it was not detected in the lungs of SP-B-deficient infants with the exception of the infant homozygous for the R252C mutation, in whom the amount of mature SP-B detected was markedly reduced compared with that in control subjects (Figure 3, middle panel). Aberrantly processed SP-C peptides of M_r 9-12,000 daltons were detected in lung tissue from all SP-B-deficient infants (Figure 3, bottom panel).

View larger version (54K): [in this window] [in a new window]   Figure 3.   Protein blots for surfactant proteins in lung tissue and lung fluid. Lung homogenates containing 50 µg (left panel  ) or 100 µg (right panel )  of protein were separated on 10% Tricine gels, and immunoblotted with antisera to proSP-B (top), mature SP-B (middle), or the amino-terminal portion of proSP-C (bottom). 121 refers to the 121ins2 mutation. ProSP-B was absent in patients with the 121ins2/121ins2, 121ins2/C248X, W60X/W60X, 209+4G A/209+4G A, and 121ins2/L13P genotypes, but proSP-B bands were present in patients with 121ins2/ 282-2delA, 121ins2/C235R, R252C/R252C, and 121ins2/ C100G genotypes. The higher Mr proSP-B band in the patient with the 121ins2/282-2delA genotype was smaller than in control patients, consistent with the deletion of exon 4 in the mRNA. Mature SP-B was detectable in control tissues, but undetectable in tissues from infants with SP-B mutations, with the exception of the infant homozygous for the R252C mutation. Aberrantly processed SP-C peptides (Mr 9-12,000 daltons) were present in tissue of all infants with SP-B gene mutations. Control 1: normal lung, Control 2: pulmonary hypertension; Control 3: bronchopulmonary dysplasia.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The inability to produce surfactant protein B causes respiratory failure and death in the neonatal period (2). A frame-shift mutation, 121ins2, is the most frequently observed cause of hereditary SP-B deficiency and accounted for 42 of 64 (65%) mutant alleles identified in this study. However, an additional 15 mutations associated with SP-B deficiency were identified. Thus, there is considerably more allelic variation in hereditary SP-B deficiency than previously recognized. None of the mutations was found in control patients, indicating that they are not likely to be common polymorphisms, and each mutation identified was the only deviation from the published SP-B sequence observed on a given allele, supporting the hypothesis that it was responsible for disease. The 1043ins3 mutation was the only newly identified mutation observed in more than one unrelated family, who were both of Pakistani descent. Whether the 1043ins3 is a relatively common mutation in this population will require further study. Although the presence of common mutations may be useful in diagnosis, the sensitivity of genetic testing is limited by the extent of allelic variability in this disorder (6). The locations of disease-causing mutations identified in the SP-B gene to date are shown in Figure 4.

View larger version (12K): [in this window] [in a new window]   Figure 4.   Location of mutations within the SP-B gene. The locations of all known mutations in the SP-B gene are indicated by symbols corresponding to the type of mutation. Exons are represented by boxes; introns by lines. The nucleotide corresponding to the location of the mutations is given above the symbol.

The possibility of a mutation hotspot in exon 4 was suggested in a recent report of another frame-shift mutation in close proximity to the site of the 121ins2 mutation (13). Although we confirmed the presence of this mutation in a patient from the same kindred, as well as two other new mutations located in or near exon 4, we also identified mutations scattered throughout the SP-B gene. Mutations in relatively close proximity to one another were also located in exon 2 (4 mutations within 83 bp), exon 7 (4 mutations within 52 bp), and exon 9 (2 mutations within 5 bp). The region of exon 9 contained consensus sequences (CCCTG and TG A/G A/G G/T A/C) that have been associated with small insertions and deletions (13). The total number of SP-B gene mutations identified to date is relatively small, and the apparent clustering of mutations may reflect the likelihood of a mutation in that region of the gene to cause disease and, therefore, to be identified. We did not observe any examples of apparent de novo mutations, as mutations in affected infants were inherited from their parents in all cases examined. Whether there are regions of the SP-B gene that are more prone to mutation than others will require further study.

Mutations that alter the amino acid sequence of proSP-B may provide insights into the roles of different proSP-B domains in lung function or proSP-B processing. The first 23 amino acids of preproSP-B are thought to constitute a signal peptide that is cotranslationally cleaved (3, 24). A proline for leucine substitution in this region was associated with markedly reduced proSP-B expression. Disruption of the hydrophobic domain of the signal peptide may have impaired translocation into the endoplasmic reticulum and thus prevented cotranslational processing of preproSP-B. An analagous substitution in the hydrophobic core of the signal peptide of antithrombin prevented translocation into microsomes and post-translational glycosylation and resulted in antithrombin deficiency (25).

The repeating periodicity of the cysteines and hydrophobic residues in proSP-B suggest that it is organized into three tandem domains that have structural homology to the saposins, lysosomal proteins that bind lipids and activate lysosomal hydrolases (26). The middle domain corresponds to the mature 79 amino acid SP-B peptide stored in lamellar bodies and secreted into the air spaces; the role(s) of the amino- and carboxy-terminal flanking domains of proSP-B are uncertain, but may be important in transport of SP-B through the cell, or have more direct biologic function (24, 26). The amino-terminal domain is proteolytically cleaved first in a step that is not thought to be cell-specific (27), and the subsequent removal of the carboxy-terminal domain occurs at a site distal to the transgolgi during transport from multivesicular to lamellar bodies, and is type II cell-specific (28, 29).

Three mutations were identified in the amino-terminal domain of proSP-B, which was essential for targeting of proSP-B to secretory granules in an in vitro system (20). ProSP-B derived from the 282-2delA and C100G mutations was processed to the 26,000-dalton intermediate, although the altered protein sequences resulting from the mutations would have been removed in this proteolytic step. The failure to further process the intermediate to the active SP-B peptide may have resulted from the inability to route proSP-B to the proper cellular compartment needed for complete proteolytic processing. The 282-2delA mutation results in deletion of 42 amino acids from the amino-terminal domain, and the C49R and C100G mutations eliminate cysteines, which are likely important in formation of intramolecular disulfide bridges important for the proprotein tertiary structure (26, 30). Impaired trafficking caused by misfolding of proteins is an increasingly recognized mechanism by which mutations lead to disease, as is the case for the cystic fibrosis gene product (31). The C235R and R252C mutations also apparently interfered with processing of proSP-B to mature SP-B as proSP-B expression was comparable to that of control subjects, but mature SP-B expression was markedly reduced. As these two mutations were in the region encoding mature SP-B, they could have also interfered with SP-B function, either by altering the normal disulfide bonding of mature SP-B (26) or by altering the pattern of positively charged residues, which may be important for SP-B's interaction with negatively charged phospholipids (3, 32). However, a synthetic peptide containing a similar substitution in the mature SP-B region (R236C) was surface active in vitro and in vivo (33).

A role for the carboxy-terminal domain of proSP-B in the processing of proSP-C was suggested by the finding that genetically engineered mice expressing an SP-B gene construct that lacked this domain had aberrantly processed SP-C peptides in their lung tissue (34). Two mutations (1043ins3, 1048 del12) located in close proximity to one another were identified in the region encoding this domain, and were associated with aberrant processing of proSP-C. However, the infants with these mutations developed fatal lung disease in contrast to the genetically engineered mice that survived (34). We have not identified patients with frame-shift or nonsense mutations in the region of the SP-B gene encoding the carboxy-terminal domain of proSP-B, and it is possible that such mutations could be associated with production of functional SP-B and prolonged survival. The mechanism for the disruption of SP-C processing associated with SP-B deficiency remains unclear. One hypothesis is that mature SP-B and/or properly folded proSP-B are necessary for the formation of the intracellular organelles (lamellar bodies) needed for the final processing steps of SP-B and SP-C (24, 26). The findings that the number of lamellar bodies are markedly reduced in SP-B deficient mice and infants with hereditary SP-B deficiency supports this concept (7, 35).

The development of lung disease in infants who made reduced amounts of mature SP-B supports the notion that a critical level of SP-B production is needed to prevent lung disease, and it suggests that individuals with even one SP-B mutation limiting SP-B production may be at risk for lung disease if other factors further decrease SP-B production. This may account for the findings in a full-term infant with transient SP-B deficiency who had a missense mutation on one allele (12). In support of this hypothesis, genetically engineered mice heterozygous for one null SP-B allele had half-normal levels of SP-B mRNA and protein, exhibited abnormal lung compliance, and were more susceptible to hyperoxic lung injury when compared with homozygous wild-type litter mates (36, 37).

Although SP-B and proSP-B expression varied among distinct SP-B gene mutations, marked extracellular staining of lung tissue with antisera to proSP-C, and/or accumulation of 9-12 kD proSP-C 9-12 peptides by protein blotting, were observed in all patients in whom SP-B gene mutations were identified on both alleles. Extracellular staining with proSP-C antisera was never observed in any control tissue during this or previous studies (6). Extracellular staining for proSP-C or detection of aberrantly processed proSP-C peptides may thus be more reliable markers for hereditary SP-B deficiency than the lack of SP-B.

In summary, we have characterized 13 novel mutations in the SP-B gene responsible for hereditary SP-B deficiency, some of which allowed for production of reduced amounts of mature SP-B as well as proSP-B that was not processed to the functional, mature SP-B peptide, indicating a greater degree of allelic and biochemical heterogeneity in this disorder than previously appreciated. Identification of mutations in distinct domains of the proSP-B molecule may provide insights into their functions and further define their roles in surfactant function and processing of proSP-B and proSP-C.