Abstract
Objective. Surfactant protein B deficiency is a lethal cause of respiratory distress in infancy that results most commonly from a homozygous frameshift mutation (121ins2). Using independent clinical ascertainment and molecular methods in different populations, we sought to determine allele frequency.
Study Design. Using clinical characteristics of the phenotype of affected infants, we screened the Missouri linked birth–death database (n = 1 052 544) to ascertain potentially affected infants. We used molecular amplification and restriction enzyme digestion of DNA samples from a metropolitan New York birth cohort (n = 6599) to estimate allele frequency.
Results. The point estimate and 95% confidence interval of the 121ins2 allele frequency in the Missouri cohort are 1/1000 individuals (.03–5.6/1000) and in the New York cohort are .15/1000 (.08–.25/1000). These estimates are not statistically different.
Conclusions. The close approximation of these independent estimates suggests accurate gene frequency (approximately one 121ins2 mutation per 1000–3000 individuals) despite its rare occurrence and that this mutation does not account for the majority of full-term infants with lethal respiratory distress.
- CI =
- confidence interval •
- ICD-9 =
- International Classification of Diseases, 9th Revision
To estimate frequencies of rare recessive genes, multiplication from small population segments has been used but may exaggerate or underestimate frequency estimates due to ethnic stratification, environmental selection, or genotype–phenotype heterogeneity.1–5 More reliable frequency estimates have used large clinical datasets to ascertain affected individuals by clinical phenotype or molecular methods for identification of rare recessive gene carriers.6–9 However, either method may be affected by genotype–phenotype heterogeneity or lack of detectable phenotype in heterozygous individuals.10 To determine the contribution of a rare disease to the frequency of a common diagnosis, we have used both clinical ascertainment and molecular identification.
Surfactant protein B deficiency is a rare inherited cause of lethal respiratory distress in infancy.11 The most frequent mutation results in a frameshift at codon 121 of the surfactant protein B gene (121ins2).12 ,13 The pulmonary surfactant isolated from affected infants fails to lower surface tension.14Although other mutations have been identified in individual families, the 121ins2 mutation is present on both alleles in ∼60% of affected infants and on 1 allele in an additional 25%.15 Estimates of allele frequency have been difficult because of the low frequency of the disease, the difficulty in identifying clinically affected infants, and the lack of a clinical phenotype in heterozygous infants.16 All homozygous infants reported to date have exhibited a consistent clinical phenotype of early onset, severe respiratory distress followed by progressive respiratory failure and death within the first year of life.13 No survivors without lung transplantation have been reported.15 ,17Because of the continuing contribution of respiratory failure among full-term infants to neonatal morbidity and mortality,18we wished to estimate the frequency of the 121ins2 allele in the general population.
Using population-based, independent clinical ascertainment and molecular methods in different populations, we find that point estimates (∼1 per 1000–3000 individuals) and confidence intervals (CIs) of 121ins2 allele frequency are low and are not statistically different. The close approximation of these independent estimates suggests accurate gene frequency despite its rare occurrence and that this mutation does not account for the majority of full-term infants with lethal respiratory distress.
METHODS
Linked Birth–Death Database
The Missouri Department of Health linked birth–death certificate database includes all infants born between 1979 and 1992 with birth weights greater than 500 g (n = 1 052 554).19 The ethnic composition of this population is 83.5% white, 15% black, and 1.5% Hispanic. Using Statistical Analysis System software (SAS Institute, Cary, NC), we identified all infants in this interval with characteristics consistent with surfactant protein B deficiency based on our previous experience (Table 1).13 ,15 ,20 Using this strategy, both of the previously reported surfactant protein B-deficient infants from Missouri were identified in the proteinosis code (International Classification of Diseases, 9th Revision[ICD-9] 516).11 ,14 To examine the possibility of novel clinical phenotypes in this cohort, one of us (D.E.D.) studied lung tissue from all 14 available autopsies from these infants by light microscopy and immunohistochemical staining, as previously described.20 ,21
Ascertainment of Surfactant Protein B-Deficient Infants by Clinical Phenotype in Missouri Linked Birth–Death Database, 1979–1992
DNA Samples, Molecular Amplification, Restriction Enzyme Analysis, and DNA Sequencing
DNA samples (n = 6599) were isolated from cord blood obtained by the New York Blood Center in New York City.22 The ethnic composition of this population is 55% white, 17% black, 23% Hispanic, and 5% Oriental. To detect the 121ins2 mutation, a 312-bp fragment that spans the restriction site forSfuI present in the 121ins2 mutation was amplified with polymerase chain reaction from each genomic DNA sample. The primers used were (5′) GAA CTC CAG CAC CCT GGG GGA (3′)(sense) and (5′) GCT CCC CCA TGG GTG GGC ACA (3′)(antisense). Each reaction mixture contained 6 ng of genomic DNA in a 10-μL reaction which also contained: .8 μmol each primer, .2 mM each dNTP, 2.0 mM MgCl2, 60 mM Tris-HCl (pH 9.5), 15 mM (NH4)2SO4, and .75 U Taq DNA polymerase (Sigma, St Louis, MO). After denaturation (95oC, 1 cycle for 3.5 minutes), 30 cycles of denaturation at 95°C for 45 seconds, annealing at 65°C (30 cycles for 1 minute), elongation at 72°C (30 cycles for 1 minute), and a second elongation at 72°C (1 cycle for 2 minutes) were performed in a 96-well plate using a MJ Research thermocycler (MJ Research, Watertown, MA). Each plate also contained normal, heterozygous, and homozygous DNA samples. Amplified fragments were treated with 2.5 U SfuI (Sigma), and the restriction enzyme products were subjected to agarose gel electrophoresis and ethidium bromide staining (Fig 1). Initial amplification failed in 8% of samples that required repeat amplification. All agarose gels were photographed and scanned into a computer database. When SfuI restriction fragments (210-bp and 104-bp) were detected, presence of the 121ins2 mutation was confirmed by direct sequencing of the amplification products using dye terminator cycle sequencing in an ABI 373A sequencer. To exclude the possibility that amplified 121ins2 allele originated from transplacental maternal cellular traffic,22 maternal DNA samples from infants heterozygous for the 121ins2 mutation were similarly analyzed. Mixing experiments using selected samples and known homozygous or heterozygous DNA consistently demonstrated theSfuI restriction fragments (data not shown). These studies were approved by the Human Subjects Committee of the Washington University School of Medicine.
Molecular detection of 121ins2 allele by restriction enzyme digestion with SfuI. The right hand seven lanes each contain DNA obtained from the New York Blood Center. The labeled sample is heterozygous for the 121ins2 mutation. Lanes 1 to 4 contain molecular mass markers, DNA from a known normal individual, DNA from a known 121ins2 heterozygote, and DNA from a known 121ins2 homozygote, respectively.
Statistical Analysis
CIs were derived from estimates of the Poisson distribution.23 χ2 analysis was performed using Statview for Macintosh computers (Version 4.5, Abacus Concepts, Berkeley, CA).
RESULTS
Clinical Estimate
Among the 91 infants whose clinical database information and ICD-9 diagnosis codes were suggestive of surfactant protein B deficiency, 89 infants had clinical information in chart reviews or autopsy reports that was not consistent with the known surfactant protein B clinical phenotype (Table 1). The majority of these infants had no evidence of respiratory distress at birth, were discharged from the hospital shortly after birth, and returned with respiratory problems. To examine the possibility that some of these infants represented novel clinical phenotypes of surfactant protein B deficiency, we recovered lung tissue from all 14 available autopsies from these 91 infants. Detailed review of hematoxylin and eosin stained lung sections from these 14 patients revealed 2 patients with proteinosis whose pathologic characteristics were consistent with surfactant protein B deficiency. Both patients were siblings who had been previously recognized and reported.11 Among the remaining 4 patients with proteinosis, 2 did not have the pathologic changes characteristic of proteinosis in the sections examined, and 2 had extensive necrotizing tracheitis and bronchiolitis. To determine whether other pathologic phenotypes in this population resulted from deficiency of surfactant protein B, one of us (D.E.D.) used immunohistochemical staining to evaluate the expression of surfactant proteins A, B, and proSP-C in lung tissue from all 14 available autopsies. Lung tissue from the 2 siblings with proteinosis lacked detectable surfactant protein B, had increased staining for pro-surfactant protein C, and sparse epithelial-cell staining for surfactant protein A.11 The distribution and intensity of immunostaining for surfactant proteins A, B, and proSP-C in lung tissue from the remaining 12 autopsies were similar to that observed in unaffected infants.11 The most common diagnosis among these 14 patients was sudden infant death syndrome (n = 6). To identify additional families who might have had multiple affected infants, we searched the database for infants with ICD-9 codes 514, 516, 518, and 519 whose mothers had the same maiden name or social security number. No additional infants were identified.
Because both affected infants were siblings, we estimate allele frequency to be 1/1000 individuals (95% CI: .03–5.6/1000) in the Missouri cohort. Because this estimate is based on ascertainment by clinical phenotype and heterozygous infants are asymptomatic, it must be taken as an approximation of the lower bound of the allele distribution. If sibship frequency in the Missouri cohort lies between 1% and 10%, no change is observed in the point estimate or CI.
Molecular Estimate
Molecular amplification and SfuI restriction enzyme digestion revealed 2 unrelated infants of 6599 examined (13 198 chromosomes) who carried the 121ins2 mutation on 1 allele, and the presence of the mutation in each sample was confirmed by direct sequencing. No infants homozygous for the 121ins2 mutation were detected. These data place the allele frequency at .15/1000 individuals (95% CI: .08–.25/1000) in the New York cohort. Neither maternal DNA sample contained the 121ins2 mutation. The estimates of allele frequency by phenotype or genotype are not significantly different (P = .37).
DISCUSSION
To estimate the frequency of the 121ins2 allele, we used independent ascertainment and molecular methods in distinct populations. The frequencies of the allele in the 2 cohorts (∼1/1000–3000 individuals) are not statistically different. Methodologic and statistical differences in the clinical and molecular methods may contribute to the range in estimates. For example, no correction for the sibship of the 2 identified patients in the Missouri cohort was made in the denominator of the estimate. Also, because clinical phenotype was the ascertainment tool used to screen the Missouri cohort, heterozygous infants would have only been detected incidentally in the context of another allele that would have led to the disease phenotype.
Our estimate of 121ins2 frequency suggests that this mutation is not a principal contributor to lethal respiratory failure among full-term infants. However, our estimate does suggest that sufficient donor availability and respiratory support strategies are available to make lung transplantation a feasible therapy for management of infants homozygous for this mutation.17 ,24 ,25 The consistent clinical phenotype observed in patients homozygous for the 121ins2 mutation, the possibility of antenatal confirmation of the diagnosis with molecular methods, accessibility via the tracheobronchial tree of type II pneumocytes for repeated transfection with a surfactant protein B-containing construct that reconstitutes the pulmonary surfactant production, and the inevitable progression to death within the first 6 months of life of affected infants suggest that infants with surfactant protein B deficiency might be candidates for gene replacement therapy.26–28 Our estimate also suggests that this allele occurs with sufficient frequency to test this strategy.
Infants homozygous for the 121ins2 allele reported to date are of predominantly Western European origin (L. M. Nogee, unpublished data). Studies of haplotypes on chromosome 2, band 2p12 to p11.2 that contain this mutation may provide informative polymorphisms to trace the origin of this mutation.29
Although developmental delay of phospholipid incorporation into the pulmonary surfactant is the most commonly recognized cause of respiratory distress among newborn infants, disturbances in genetic regulation of surfactant protein B may disrupt neonatal lung function pathologically, physiologically, and biochemically.30–36Studies of different ethnic groups, gender, surfactant proteins A and B, and targeted gene ablation in murine lineages have strongly suggested that genetic mechanisms contribute significantly to lethal and nonlethal respiratory distress in infancy.37–44 The incidence of nonlethal respiratory distress in infancy caused by disorders of genetic regulation of surfactant protein B may be considerably greater than our allele estimates suggest. Intragenic polymorphisms may provide markers for genotype–phenotype correlation with less profound but clinically significant disturbances in surfactant protein B regulation.
ACKNOWLEDGMENTS
This work was supported in part by National Institutes of Health HL/HD Grant 54187 to Dr Cole.
We thank Garland Land and Wayne Schramm for assistance with the Missouri Department of Health infant birth–death database, Dr Michael DeBaun for advice concerning statistical analysis, Megin Wehmueller and Norman Mollen for technical assistance, and Dawn Rouse for secretarial support.
Footnotes
- Received April 13, 1999.
- Accepted June 11, 1999.
Reprint requests to (F.S.C.) Division of Newborn Medicine, St Louis Children's Hospital, 1 Children's Pl, St Louis, MO 63110. E-mail: cole{at}kids.wustl.edu
REFERENCES
- Copyright © 2000 American Academy of Pediatrics
