Homozygous, Intragenic Tandem Duplication of SFTPB Causes Neonatal Respiratory Failure

To the Editor:

SP-B (surfactant protein-B) deficiency (OMIM #265120) results from biallelic loss-of-function variants in SFTPB (surfactant protein-B gene) (NM_000542.5/GRCh38), causes progressive neonatal respiratory failure among term infants, and is lethal without lung transplantation (1, 2). SP-B is expressed by alveolar epithelial type II cells (AEC2s) and assembled with SP-C and phospholipids in lamellar bodies to produce surfactant. SFTPB comprises 11 exons, the last of which is untranslated, and SP-B is synthesized as a 381–amino acid proprotein (3). The most common pathogenic variant, c.361 delCinsGAA;p.Pro121Glufs*95, is identified in approximately two-thirds of infants with SP-B deficiency (1, 2). More than 20 SFTPB pathogenic variants, including single-nucleotide variants, small insertions and deletions, and a large (∼3 kb) deletion that included exons 7 and 8, have been identified (1, 2, 4–9). We identified term infant male (proband) and female full siblings who underwent lung transplantation at 4 months for or died at 8 weeks of progressive neonatal respiratory failure, respectively (see clinical details Figure E1 in the data supplement). Next-generation sequencing (NGS) gene panel (SFTPB, SFTPC, ABCA3, NKX2-1, FOXF1), conventional/low-resolution chromosomal microarray analysis (CMA), and trio exome sequencing identified a 9.09-kb triplication (gain of two copies) of exons 1–10 of SFTPB in the proband. However, the zygosity, phase, or inheritance of the copy number gain were unable to be determined.

This study was approved by the Washington University Human Research Protection Office, and we obtained parental informed consent. We performed long-read sequencing (PacBio) on high–molecular weight DNA from the proband and identified a novel 9,244-bp homozygous, intragenic tandem duplication of SFTPB that included exons 1–10, with a breakpoint between exons 10 and 11 (Figure E2). Using an exon-targeted custom Agilent 400k CMA + SNP array (Baylor Genetics), we confirmed both infants were homozygous for the tandem duplication of SFTPB, and both parents are heterozygous carriers (Figures E3 and E4). The duplication was flanked on both sides by AluSx sequences, a major subfamily of Alu elements, repetitive DNA sequences that are frequently located in introns, contain random variants, and may alter splicing. In addition to the duplication, three large regions (chromosomes 2, 5, and 11; 42.9 Mb in total) of homozygosity were identified in the siblings, suggesting a shared parental ancestor.

To confirm pathogenicity of this novel SFTPB duplication, we performed functional studies on proband explant lung tissue (see data supplement for detailed methods). Explant histology revealed diffuse AEC2 hyperplasia, extensive alveolar septal widening, patchy mixed inflammation, proteinosis-like material and macrophages in the alveolar spaces, and early interstitial fibrosis (Figure 1A) (2). Electron microscopy revealed hyperplastic AEC2s, many containing cytoplasmic membrane-bound inclusions, with some resembling composite bodies with small vesicles and loosely concentric to irregular membrane whorls (Figure 1B). A few maturing lamellar bodies were also present. Immunoblotting results demonstrated absence of mature SP-B and presence of aberrantly processed proSP-C, similar to the explant from an infant homozygous for the most common SFTPB loss-of-function variant (p.Pro121Glufs*95) and in contrast to an infant with pulmonary hypertension without SFTPB variants (mature SP-B present, aberrant proSP-C absent; Figures 1C, E5, and E6). Immunofluorescence staining with SP-B antibody, which recognizes the proprotein and mature forms of SP-B, was absent in the proband AEC2s, and immunostaining for proSP-C was increased (Figure 1E), similar to infants with biallelic SFTPB loss-of-function variants (2, 10). ABCA3 staining was present in the proband AEC2s and similar to lung tissue from an infant with SP-B deficiency homozygous for p.Pro121Glufs*95 (Figure E7). SFTPB transcript abundance was markedly reduced in the proband explant (Figure E8), consistent with nonsense-mediated decay or an unstable RNA transcript.

Figure 1.

(A) Light microscopy of proband explant lung tissue demonstrates diffuse alveolar epithelial type II cell (AEC2) hyperplasia and alveolar septal widening with patchy mixed inflammation. Alveolar spaces contain macrophages and proteinosis-like material (asterisks); hematoxylin and eosin. Scale bar, 100 μm. (B) Electron microscopy of focal AEC2 with composite bodies (arrowheads), maturing lamellar bodies (arrows), and heterogeneous electron dense inclusions (asterisks), some resembling residual bodies. Tubular myelin (star) is present in the alveolar space. Scale bar, 1 μm. (C) Comparison of immunoblotting of explant lung tissue from the proband, an infant homozygous for the SFTPB c.361 delCinsGAA;p.Pro121Glufs*95 variant, and an infant with pulmonary hypertension without SFTPB variants demonstrates absence of mature SP-B (surfactant protein-B) and aberrant proSP-C in the proband explant lung tissue. (D and E) Enhanced resolution immunofluorescence images for surfactant proteins proSP-C and SP-B. Images from autopsy lung tissue from a 6-month-old previously healthy infant (control donor) (D) and from proband explant lung tissue (E). Mature SP-B staining was not detected in the proband lung tissue, and intense proSP-C accumulation (aberrant proSP-C) was seen in the hyperplastic AEC2 cells in the proband. Scale bars, 5 μm.

Although SP-B deficiency is an extremely rare disease, affected infants have provided important insights into surfactant biology in the human neonatal lung (1, 2). Genetic surfactant dysfunction disorders should be considered in term/late-preterm infants with persistent respiratory failure after a week of age and after exclusion of infectious, anatomic, and cardiac abnormalities. Given the substantial morbidity associated with lung biopsy in critically ill infants, genetic testing is increasingly pursued as a first-tier diagnostic tool in the neonatal intensive care unit (11). NGS gene panels and exome sequencing target coding regions that contain the majority of disease-causing variants. NGS reliably identifies single nucleotide variant, small insertions and deletions, and intronic variants that border exon–intron junctions (generally ±5 bp). However, current NGS technologies use short-read (∼150–300 bp) lengths, which can miss structural variants, including large deletions, duplications, inversions, and translocations, repetitive regions, and noncoding region variants and has limited ability to phase variants (12). Although exon-targeted CMA + SNP testing detected the homozygous SFTPB duplication in this case, many clinical CMA testing platforms are not exon targeted, and copy number gains of this size (∼9.2 kb) may not be consistently detected. In contrast, long-read sequencing uses read length fragments of ∼10–25 kb that span large genomic regions and improve alignment and resolution of copy number variants, complex structural variants, and repetitive regions. Long-read sequencing permits phasing of variants to determine zygosity. Although a large deletion including exons 7–8 in SFTPB has been reported (4, 9), our report represents the first example of a homozygous, pathogenic partial-gene copy number gain in SFTPB. Long-read sequencing is currently available on a research basis (13). However, clinical use in the near future may permit detection of pathogenic structural variants or intronic variants not detected by short-read sequencing.

The SFTPB duplication identified in this family was flanked by Alu elements, ∼300-bp sequences that arose in primates ∼65 million years ago and are the most abundant repetitive elements in the human genome (14, 15). Alu elements can mediate nonallelic chromosomal rearrangements that give rise to copy number variants in the intervening region (14, 15). Although Alu elements can alter gene transcription and translation, they are frequently found at endpoints of segmental duplications and breakpoints of genomic rearrangements (16). They can also disrupt gene expression when inserted into exonic regions or cause alternative splicing when inserted into intronic regions (16). Alu element insertion has been implicated as a pathogenic mechanism in several diseases, including alveolar capillary dysplasia with misalignment of the pulmonary veins, neurofibromatosis, neurodegenerative diseases, and cancers (17–19). Although tandem gene duplications may be associated with an increase in gene expression and may contribute to evolution of new genomic functions, loss-of-function tandem gains are rare (20, 21).