Pulmonary surfactant is a complex mixture of phospholipids and proteins that is secreted into the alveolar space, reduces surface tension, prevents end expiratory alveolar collapse, and is required for gas exchange. Surfactant is synthesized in lamellar bodies, specialized intracellular organelles derived from lysosomes in alveolar epithelial type II cells (AEC2, aka ATII or AT2 cells). Phospholipids are transported into the lamellar bodies by ABCA3 and assemble with surfactant proteins B and C to form surfactant. The lamellar bodies are released into the alveolar lumen via exocytosis. Pathogenic variants in genes that encode key components of pulmonary surfactant include surfactant protein B (SP-B, SFTPB gene), surfactant protein C (SP-C, SFTPC gene), and the ATP-binding cassette transporter A3 (ABCA3, ABCA3 gene) and are leading inherited causes of neonatal respiratory distress syndrome (RDS) and childhood interstitial lung disease (chILD) (Garmany et al., 2008). Treatments for these monogenic pulmonary diseases are limited and non-specific, and for many patients, lung transplant may be the only option for survival beyond the first months of life (Wambach et al., 2014). Recent advances offer renewed opportunities to develop gene therapies for genetic disorders of surfactant dysfunction resulting from pathogenic variants in SFTPB, SFTPC, and ABCA3. There is no “one size fits all” approach for developing gene therapy vectors for SP-B, SP-C, and ABCA3 deficiencies. Non-viral vectors (i.e., plasmid DNA, nanoparticles, and or in vitro transcribed mRNA) and Epstein-Barr based plasmid DNA each have important qualities and have advanced the lung gene transfer field (reviewed in (Stribling et al., 1992; Tu et al., 2000)). This review will focus on adeno-associated virus (AAV), lentiviral, or adenoviral (Ad)-based delivery vehicles for a gene addition approach or delivery of gene editing tools to complement or repair disorders of surfactant dysfunction, as well as appropriate models to assess gene transfer efficacy. Our goal for this review is to introduce potential viral vector-mediated gene therapy options for surfactant diseases and provide enough details about each deficiency to highlight why gene therapy may be particularly challenging for this disease class.

Newborns with SP-B deficiency typically present with RDS shortly after birth and die of progressive respiratory failure in the first few months of life without a lung transplant. A few children with biallelic SFTPB variants and chronic respiratory insufficiency have been reported (Dunbar et al., 2000). SP-B deficiency is an autosomal recessive disease and the most common pathogenic variant p. Pro133Glnfs*95 (previously known as ‘121ins2’) results in a frameshift and nonsense-mediated decay of the mRNA transcript. Treatment with exogenous surfactant enriched in SP-B protein is ineffective (Thompson, 2001) and lung transplant remains infrequent due to inavailability of suitable neonatal donor lungs (Eldridge et al., 2017). Gene addition (Barnett et al., 2017; Leibel et al., 2019; Kang et al., 2020) and gene editing approaches (Mahiny et al., 2015; Jacob et al., 2017) have demonstrated that complementing SP-B expression in AEC2s restores the phenotypic defect in vitro and in vivo. Although gene editing approaches may be successful, the diversity of pathogenic variants suggests that a gene addition approach may be a more near term goal for SP-B deficiency.

SP-B plays a major role in the assembly and function of pulmonary surfactant and is required for proper lamellar body biogenesis in AEC2s (Figure 1). It commonly exists as a homodimer and is a small, 79 amino acid hydrophobic protein that permeabilizes, cross-links, mixes, and fuses cell membranes. Mature SP-B results from proteolytic modifications involving the 381 amino acid preproSP-B peptide followed by additional glycosylation and proteolytic events of proSP-B (Guttentag, 2008) (Figure 2A). Pathogenic SFTPB variants that disrupt glycosylation or proteolytic cleavage sites have been identified (Lin et al., 2000). While both viral and nonviral delivery strategies offer promise, here we will focus on viral vectors.

SP-C is a 35-amino acid hydrophobic protein that organizes with SP-B to lower the surface tension of surfactant (Figure 1). The SFTPC locus spans ∼3,500 bp and is expressed from six exons on chromosome 8. A 197-amino acid protein is first produced with an N-terminal propeptide domain and C-terminal BRICHOS domain (pro-SPC). As the mature protein is processed, the N-terminal domain is cleaved from the proSPC form to SP-C. The mature protein has a shortened N-terminal domain, linker, and C-terminal BRICHOS domain (Figure 2B).

Pathogenic variants in SFTPC can arise de novo (∼50% of cases) or are inherited (∼50% of cases) (Nogee et al., 2002). A pathogenic splicing variant in SFTPC within the first base of intron 4 (c.460+1G > A) was first reported in 2001 in an infant and mother with interstitial pneumonitis (Nogee et al., 2001). Subsequently, additional pathogenic SFTPC variants have been identified, the most frequent of which is p. I73T (Cameron et al., 2005). Individuals with pathogenic SFTPC variants most commonly present with sporadic or familial interstitial lung disease as infants, children, or adults, and less frequently with neonatal RDS (Nogee et al., 2002). The disease course is highly variable with some infants presenting with severe respiratory failure requiring lung transplant, others exhibiting chronic disease managed with long-term mechanical ventilation (Liptzin et al., 2015), and others remaining relatively asymptomatic (Thomas et al., 2002). Genotype alone is not predictive of disease presentation, severity, and or course. Pathogenic SFTPC variants located in the C-terminal BRICHOS domain, which acts as a chaperone to promote protein stability, result in increased endoplasmic reticulum stress, inflammation, and spontaneous pulmonary fibrosis in a murine model (Katzen et al., 2019). Non-BRICHOS variants, which include p. I73T, result in defective AEC2 macroautophagy, inflammation, remodeling, and fibrosis (Nureki et al., 2018).

SP-C associated interstitial lung disease is a gain-of-toxic-function disease inherited in an autosomal dominant pattern with variable penetrance (Thomas et al., 2002). Expression of a mutant SFTPC allele is responsible for the surfactant dysfunction phenotype. The resultant gain-of-toxic-function requires silencing of the mutant allele. Therefore, a gene addition approach is not an option. An allele specific gene knockout or knockdown approach can be used to disrupt and inactivate or reduce the abundance of the mutant SFTPC product. CRISPR/Cas9 nuclease-mediated gene knockout has been validated in mice in utero by targeting the p.173T allele, resulting in improved lung morphology, and increased survival of offspring (Alapati et al., 2019). RNAi and antisense RNA strategies can also be envisioned.

ABCA3 deficiency, an autosomal recessive disorder, and is caused by pathogenic variants in ABCA3. Biallelic variants in ABCA3 cause severe neonatal RDS, chILD, and adult pulmonary fibrosis (Shulenin et al., 2004; Klay et al., 2020; Tomer et al., 2021). Infants with biallelic ABCA3 frameshift or nonsense variants present with neonatal RDS at birth and die within the first year of life without lung transplant (Wambach et al., 2014). The respiratory phenotypes and disease courses for individuals with ABCA3 missense variants are variable and include neonatal RDS and interstitial lung disease.

ABCA3 is a phospholipid transporter located at the lysosomal-derived lamellar body limiting membrane and plays a critical role in surfactant assembly and lamellar body formation (Figure 1). The ABCA3 cDNA encodes a 1,704 amino acid polypeptide. Mature ABCA3 protein is folded in the endoplasmic reticulum and undergoes glycosylation in the Golgi (Beers and Mulugeta, 2017). A second post-translational modification involves the N-terminal proteolytic cleavage of the 190 kD protein, shortening the protein to 150 kD (Engelbrecht et al., 2010) (Figure 2C). The lamellar bodies from infants and children with ABCA3 deficiency are small with dense bodies and are described as having a “fried-egg appearance” (Wert et al., 2009). Two mechanistic classes of ABCA3 missense variants have been identified: disruption of intracellular trafficking and impaired phospholipid transport (Matsumura et al., 2006; Denman et al., 2018; Wambach et al., 2020). The most common pathogenic variant is p. E292V, a missense variant that impairs phospholipid transport (Wambach et al., 2016). This variant is commonly associated with chILD (Wambach et al., 2014) and is present in 0.4% of individuals in Genome Aggregation database (gnomad.broadinstitute.org, accessed September 2021).

The overall goal for gene therapy to treat loss-of-function diseases is to efficiently and persistently (long-term) express ABCA3 at physiological levels in AEC2s. A gene addition strategy involves complementing the loss of function with a full length ABCA3 cDNA. Because most ABCA3 variants are rare and private, this poses a challenge for developing variant-specific gene editing or targeted drug approaches. The size constraints of some viral vectors present challenges for a transgene as large as ABCA3.

Immortalized pulmonary epithelial cell lines with characteristics of AEC2s such as A549 cells are a first line model system because they are easy to passage and maintain, form lamellar body-like structures which can be visualized by electron microscopy, and are readily transduced by viral vectors. A549 ABCA3 ^−/− cells are established and lack well-developed lamellar body-like structures (Wambach et al., 2020). However, a major limitation for tumor-derived cell lines such as A549 is that they do not produce surfactant. Alveolar cells need to differentiate in order to produce surfactant and express SP-C, a common marker used to identify AEC2s.

Alveolar epithelial cell organoids derived from human embryonic or pluripotent stem cells generate alveolospheres in culture which display functional properties of AEC2s. These cells can proliferate, differentiate, and be modified to encompass surfactant dysfunction phenotypes (Jacob et al., 2017; Jacob et al., 2019). iPSC organoids derived from a patient with SP-B deficiency mirrored the disease phenotype including decreased surfactant production. This defect was restored when transduced with a lentiviral vector carrying SFTPB, including proper lamellar body formation (Leibel et al., 2019). In total, AEC2 organoids are a useful model to assess strategies to restore function of surfactant components.

Measuring restoration of surfactant function is challenging. Surfactant proteins B and C are typically quantified to confirm the presence of surfactant proteins in cell lysates or secretions. The measurement of phospholipid transport using dipalmitylphosphatidylcholine (DPPC) release is one method to assess surfactant production (Jacob et al., 2017). Instruments such as the pulsating bubble surfactometer can be used to measure the surface tension lowering properties of cell secretions.

Mouse models of each of these surfactant-based genetic diseases have been generated. SP-B knockout mice develop severe respiratory distress and die within hours of birth and exhibit dense, abnormal lamellar body-like organelles (Clark et al., 1995; Stahlman et al., 2000). Restoring SP-B using an AAV vector improved survival in conditional SP-B knockout mice (Kang et al., 2020). Alternatively, SP-C knockout mice are viable and can survive to adulthood. They produce lamellar bodies and surfactant proteins A, B, and D. Phenotypic responses include decreased stability of surfactant at low volumes, pneumonitis and emphysema (Glasser et al., 2001; Whitsett and Weaver, 2002). ABCA3 knockout mice have a similar outcome as SP-B knockout mice and do not survive (Ban et al., 2007; Fitzgerald et al., 2007). Conditionally null ABCA3 mice have also been generated (Besnard et al., 2010).

Gene based therapies for disorders of surfactant dysfunction resulting from pathogenic variants in SFTPB, SFTPC, and ABCA3 are now realizable. Generating a delivery vector for either a gene addition or gene editing approach requires the careful consideration of vector design, transgene expression cassette, preclinical models used, and assays to assess the correction of surfactant production and function.