Gene Therapeutics for Surfactant Dysfunction Disorders: Targeting the Alveolar Type 2 Epithelial Cell

Abstract

Genetic disorders of surfactant dysfunction result in significant morbidity and mortality, among infants, children, and adults. Available medical interventions are limited, nonspecific, and generally ineffective. As such, the need for effective therapies remains. Pathogenic variants in the SFTPB, SFTPC, and ABCA3 genes, each of which encode proteins essential for proper pulmonary surfactant production and function, result in interstitial lung disease in infants, children, and adults, and lead to morbidity and early mortality. Expression of these genes is predominantly limited to the alveolar type 2 (AT2) epithelial cells present in the distal airspaces of the lungs, thus providing an unequivocal cellular origin of disease pathogenesis. While several treatment strategies are under development, a gene-based therapeutic holds great promise as a definitive therapy. Importantly for clinical translation, the genes associated with surfactant dysfunction are both well characterized and amenable to a gene-therapeutic-based strategy. This review focuses on the pathophysiology associated with these genetic disorders of surfactant dysfunction, and also provides an overview of the current state of gene-based therapeutics designed to target and transduce the AT2 cells.

INTRODUCTION

The alveolar region of the human lungs is comprised of an extensively complex epithelial surface that mediates the exchange of O₂ and CO₂ required for respiration at birth. The alveolar surface is lined by alveolar type 1 (AT1) and type 2 (AT2) epithelial cells, which are in close apposition with endothelial cells.¹ In humans, the maintenance of pulmonary function is dependent on the proper function of AT2 cells, which synthesize and secrete surfactant, and serve as facultative progenitors of the distal lung. Pulmonary surfactant is composed of phospholipids (such as phosphatidylcholine and phosphatidylglycerol) and proteins (surfactant protein [SP]-A, SP-B, SP-C, and SP-D; produced intracellularly) that in combination reduce the surface tension and prevent alveolar collapse at end expiration (the composition and function of surfactant are reviewed in detail by Parra and Perez-Gil²).

The SFTPB, SFTPC, and ABCA3 genes encode for key components required for proper production and function of pulmonary surfactant, and are predominantly expressed in the AT2 cells. Pathogenic variants in these genes result in diffuse parenchymal lung disease, which presents as severe neonatal respiratory distress syndrome (RDS), childhood interstitial lung disease (chILD), or adult pulmonary fibrosis, all of which are associated with substantial morbidity and mortality.^3–6

Available medical therapies for genetic disorders of surfactant dysfunction are limited, not specific to the underlying genetic defect, and generally ineffective.^7,8 Lung transplantation, which is associated with substantial morbidity and mortality, has been performed for progressive respiratory failure associated with these disorders.⁹ The genes central to the proper composition and function of pulmonary surfactant are well characterized, allowing for the development of a gene-based therapeutic strategy to target AT2 cells.

Herein, we provide an overview of the pathophysiology associated with each of these disorders that originate in the AT2 cells, the in vivo and in vitro model systems that recapitulate specific aspects of the disease phenotype and discuss the current state of gene therapeutics designed to target AT2 cells.

GENETICS

RDS typically occurs in preterm infants due to a developmentally regulated deficiency of pulmonary surfactant production and presents shortly after birth with clinical symptoms that include tachypnea, subcostal and intercostal retractions, nasal flaring, grunting, hypoxemia, and impaired ventilation.¹⁰ Reports of late preterm and full-term infants with severe, progressive RDS prompted the investigation of a genetically linked disruption of pulmonary surfactant production.⁵ In 1993, Nogee et al. first reported on a full-term infant with congenital alveolar proteinosis and lethal neonatal RDS, who was later found to be homozygous for a loss-of-function variant in the SFTPB.³ This report illustrated the importance of genetic testing full-term infants with severe and progressively worsening RDS.

In the United States, genetic testing for surfactant dysfunction disorders is now widely offered (https://www.ncbi.nlm.nih.gov/gtr/tests). The genetics and pulmonary pathophysiology associated with rare variants in the SFTPB, SFTPC, and ABCA3 genes that cause surfactant dysfunction disorders are discussed in detail below. While not discussed here, other genes associated with surfactant dysfunction and diffuse lung diseases are reviewed in detail by Nogee¹¹ and Nogee and Ryan.¹²

SP-B DEFICIENCY

The mature SP-B is a highly hydrophobic 79-amino acid (aa) protein derived from proteolytic processing of a 393-aa proprotein. SP-B is expressed primarily by AT2 cells, and is required for the maintenance of surface tension and prevention of alveolar collapse at end expiration^13,14 (https://gtexportal.org, accessed July 2022). SP-B deficiency is an autosomal recessive disorder that results from biallelic loss of function variants in SFTPB.³ SP-B deficiency is considered to be the rarest genetic surfactant dysfunction disorder and affects ∼1 in 5 to 10 million births¹⁵ (https://gnomad.broadinstitute.org, accessed July 2022).

The most common pathogenic SFTPB variant, c.397delCinsGAA; p.Pro133GlnfsTer95 (also noted as “121ins2” in earlier publications^16–18) has a carrier frequency of ∼1 in 2,300 individuals, with an increased frequency of ∼1 in 1,200 among European, non-Finnish-descent individuals (Genome Aggregation Database, gnomAD; https://gnomad.broadinstitute.org/, v2.1.1, accessed July 2022), likely due to a founder effect. Further, this SFTPB variant has been identified in approximately two-thirds of infants with SP-B deficiency.¹⁷

These infants present with severe RDS shortly after birth and in the absence of lung transplantation will succumb to progressive respiratory failure within their first 3 months of life.⁹ Rarely, SP-B deficiency is associated with delayed disease onset during infancy or survival past the first year of life owing to SFTPB variants that likely retain partial residual protein function.¹⁹ Notably, parents of affected infants and adult carriers of SFTPB c.397delCinsGAA loss-of-function variant are typically healthy,¹⁷ substantiating that ∼50% of SP-B levels are sufficient for normal lung function.

SP-C DYSFUNCTION

Pathogenic variants in SFTPC have been identified among individuals with diverse lung diseases, including neonatal RDS, chILD, and adult pulmonary fibrosis.^4,20,21 Lung disease resulting from SP-C dysfunction is rare, and its prevalence is difficult to estimate accurately. While information on the exact disease phenotype is not readily available, several pathogenic SFTPC variants are listed in gnomAD, suggesting a prevalence of ∼1 in 20,000 individuals. Pathogenic SFTPC variants are autosomal dominant with variable penetrance, even within families carrying the same SFTPC variant.^4,20–22 Approximately half of SFTPC variants are inherited, and half arise de novo.^23,24 The most common pathogenic SFTPC variant, c.218T>C (p.Ile73Thr, p.I73T), has been identified in multiple unrelated individuals of diverse ethnic backgrounds, and is present in ∼25% of individuals with SP-C–related lung disease.²¹

Individuals with pathogenic SFTPC variants typically present with chILD⁴ or early-onset pulmonary fibrosis. Unlike surfactant proteins A, B, or D, SP-C is produced and expressed in the lungs exclusively by the AT2 cells.²⁵ Mature SP-C is a 35-aa, extremely hydrophobic protein that is derived from the proteolytic processing of an 191 or an 197 aa proprotein (proSP-C).²⁴ The terminal ∼100 aa of proSP-C have a high-level homology to the BRICHOS proteins that are misfolded in some familial (i.e., British and Danish) dementias.²⁶ Proper folding of the BRICHOS domain is necessary for the processing and trafficking of proSP-C to the lamellar bodies (LBs) in AT2 cells,²⁷ allowing for the incorporation of mature SP-C into pulmonary surfactant.

SFTPC variants result in a toxic gain of function, and histological examination of explanted lungs from symptomatic individuals revealed an abundant expression of proSP-C.²⁸ Most pathogenic SFTPC variants are located within the BRICHOS domain, with rare published reports of pathogenic variants within the mature SP-C.²⁹ Cell-based model systems^30–34 and mouse models^35–38 have been utilized to demonstrate that BRICHOS domain variants result in the accumulation of misfolded proteins leading to endoplasmic reticulum (ER) stress. This subsequently activates the unfolded protein response (UPR), and results in apoptosis, inflammation, and spontaneous lung fibrosis.

Non-BRICHOS SFTPC pathogenic variants, including the p.I73T variant, result in defective macroautophagy, inflammation, and spontaneous lung fibrosis.^28,39,40 It is difficult to predict the disease course of patients with pathogenic SFTPC variants, and to date a genotype–phenotype correlation has not been established.⁷ While some patients with SFTPC variants have progressively worsening respiratory failure necessitating lung transplantation,⁹ others stabilize with supportive care (i.e., supplemental O₂, chronic mechanical ventilation, and/or anti-inflammatory medications^7,41), and a subset remains relatively asymptomatic.²²

ATP-BINDING CASSETTE TRANSPORTER A3 DEFICIENCY

ATP-binding cassette transporter A3 (ABCA3) deficiency is an autosomal recessive disorder and is the most common genetic disorder of surfactant dysfunction.⁵ The typical clinical presentation of ABCA3 deficiency includes neonatal RDS and chILD during infancy or early childhood.⁴² While the incidence is rare, adults with biallelic ABCA3 variants that present with early-onset pulmonary fibrosis have been reported.^8,42–45 ABCA3 is a 1,704 aa protein that shares considerable structural homology with another well-studied ATP-binding cassette transporter protein, the cystic fibrosis transmembrane conductance regulator (CFTR); similar to CFTR, ABCA3 has two membrane-spanning domains and two nucleotide-binding domains.⁴⁶ ABCA3 is developmentally regulated and is highly expressed in AT2 cells where it localizes to the limiting membranes of LBs. ABCA3 transports phospholipids, specifically phosphatidylcholine, from the cytoplasm to the LBs, where the phospholipids combine with SP-B and SP-C to produce pulmonary surfactant.^47,48 ABCA3 is also expressed at low levels in the brain, kidney, thyroid, and other tissues.⁴⁷

To date, >300 disease-associated ABCA3 variants have been reported, with the vast majority being either missense or private (i.e., unique to individuals and families).⁴² The most common pathogenic ABCA3 variant is c.875A>T; p.Glu292Val (p.E292V), and is present in ∼1 in 220 individuals, with higher carrier frequencies reported among individuals of non-Finnish European descent (gnomAD). Although the p.E292V variant is commonly associated with chILD, there have been reports of individuals homozygous for this variant presenting with severe neonatal RDS or early-onset pulmonary fibrosis.^42,45,49

Using the gnomAD frequencies of the ABCA3 loss-of-function variants and the previously identified disease-associated variants (including p.E292V), we estimate the incidence of ABCA3 deficiency to be ∼1 in 90,000 live births. However, this incidence is more than likely an underestimate, as most pathogenic ABCA3 variants are either missense variants without available functional data or are private,^8,42 and are therefore not always captured in the large databases containing exome and genome data from adults.

There are two mechanistic classes of ABCA3 missense variants: those that disrupt intracellular trafficking of ABCA3 to the LBs (i.e., p.L101P), and those that impair phospholipid transport into the LBs (i.e., p.E292V).^50–52 Infants with biallelic loss-of-function ABCA3 variants present with severe neonatal RDS and typically succumb to their disease within their first year of life.^8,42 The timing of disease onset, severity, and progression for infants and children with missense ABCA3 variants are variable, and to date, survival outcomes are difficult to predict.^8,42

CURRENT TREATMENTS FOR GENETIC DISORDERS OF SURFACTANT DYSFUNCTION

Available medical therapies are limited, not gene specific, and generally ineffective.^7,8 The most common, noninvasive therapeutic strategy for infants with neonatal RDS, due to SP-B or ABCA3 deficiency, is exogenous surfactant replacement that leads to transient improvement in O₂ requirement lasting a few hours, although the response diminishes over time.²⁴ Anti-inflammatory medications, which include corticosteroids, hydroxychloroquine, and azithromycin, have also been used as treatments for both infants and children with ABCA3 deficiency, although with limited efficacy.^8,53,54 Children with pathogenic SFTPC variants may benefit from a regimen of steroid and hydroxychloroquine treatments in addition to chronic mechanical ventilation.^7,41

Despite its obvious limitations, lung transplantation remains the definitive treatment option for progressive respiratory failure among infants and children but is only available at a few quaternary centers in the United States. The 5-year survival for infant and pediatric lung transplantation for genetic disorders of surfactant dysfunction is ∼60%, and is similar to that of other transplant indications.⁹ Transplantation-associated morbidities are frequent, and include hypertension, renal insufficiency, and post-transplant lymphoproliferative disorder.⁹

Furthermore, infants who undergo lung transplantation are at an elevated risk of speech and motor delays, hearing loss, and frequently require gastrostomy tube placement for nutritional support.⁹ Bronchiolitis obliterans remains the most common cause of chronic lung transplant failure and develops in 30–50% of recipients within 5 years of transplantation.⁹ While infant and pediatric lung transplantations for genetic disorders of surfactant dysfunction have been performed since 1993,^9,55 morbidity and mortality remain significant, thus highlighting the need for new and advanced therapies that include molecular therapeutics.

PRECLINICAL MODELS OF GENETIC DISORDERS OF SURFACTANT DYSFUNCTION

To assess the translatability of any type of gene therapeutic, it is important to utilize sophisticated in vivo and in vitro model systems that mimic key features of the diseased alveoli. These model systems are discussed in detail below.

MOUSE MODELS OF SFTPB DEFICIENCY

In humans, SP-B deficiency presents as RDS within hours of birth. This RDS phenotype was successfully recapitulated in a mouse model generated by the targeted disruption of the Sftpb locus in Swiss black mice, resulting in neonatal lethality in the affected offspring.⁵⁶ The Sftpb knockout (KO) mouse, a model of SP-B deficiency that develops lethal respiratory failure shortly after birth, is characterized by an absence of LBs in AT2 cells, a lack of tubular myelin in the airspaces, and an aberrant processing of SP-C proprotein.⁵⁶

Although Sftpb KO mice closely phenocopied SP-B–deficient infants, it is interesting to note that upon histological examination of their lungs at E17.5, there were no signs of proteinosis or destruction of the respiratory epithelium, in stark contrast to the phenotypic changes observed in the lungs of SP-B–deficient infants.⁵⁶ This suggests that SP-B deficiency alone is not the cause for the complex repertoire of histological abnormalities observed in the lungs of SP-B–deficient infants.

Transgenic SP-CrtTA:teto₇:SP-B mice on the FVB/N background in which expression of Sftpb is doxycycline regulated also recapitulate the disease phenotype of SP-B deficiency.⁵⁷ In this model, when SP-B expression in AT2 cells was decreased by ∼75%, mice presented with respiratory failure, indicating that SP-B must be expressed at a minimum 25% of normal levels for proper surfactant homeostasis and respiratory function.⁵⁷ Lin et al. demonstrated that restoration of SP-B expression in AT2 cells, but not in club cells, was effective at rescuing the neonatal lethal phenotype of Sftpb KO mice.⁵⁸ The authors also showed that expression of SP-B in some, but not all, AT2 cells allowed for the postnatal survival of Sftpb KO mice.⁵⁸ In a subsequent study, Nesslein et al. demonstrated that in lung regions in which AT2 cells did not express SP-B, there was an obvious development of focal air space enlargement.⁵⁹

MOUSE MODELS OF SFTPC DYSFUNCTION

Several transgenic and knock-in (KI) mice with specific pathogenic SFTPC variants have been engineered to model the human pathophysiology associated with SP-C dysfunction.^35–38,40 Pathogenic variants in the BRICHOS domain result in a misfolded SP-C proprotein that is retained in the ER, activates ER stress, and subsequently undergoes proteasome-mediated degradation.³² An animal model of SP-C dysfunction, the Sftpc^Δexon4 mouse, which expresses the BRICHOS domain variant Δexon4 under control of the 13 kb mouse Sftpc promoter, is characterized by severe defects in lung morphogenesis that lead to perinatal lethality as early as E18.³⁵ The severity of lung disease in the Sftpc^Δexon4 mouse model was reported to be directly proportional to the level of expression of the mutant SP-C proprotein.³⁵

The mutant SP-C proprotein is unstable, and its transient expression in AT2 cells results in ER stress that induces apoptosis.³⁵ The p.L188Q SFTPC variant was identified in a large kindred and is inherited in an autosomal dominant manner with incomplete penetrance.²² In the C57Bl6/J L188Q transgenic mouse model with intact WT Sftpc alleles, doxycycline-induced expression of the human p.L188Q BRICHOS domain variant²² did not result in overt structural abnormalities in the lung, likely due to the low levels of p.L188Q expression.³⁶ Intratracheal challenge of C57Bl6/J L188Q transgenic mice with bleomycin resulted in lung fibrosis 3 weeks postchallenge, and was associated with activation of the ER stress response and an increased rate of apoptosis of AT2 cells.³⁶

The p.L184Q KI mice on the C57Bl6/n background express the murine version of the human p.L188Q variant,²² and present with oxidative and ER stress during postnatal AT2 cell expansion³⁸; adult KI L184Q mice challenged with bleomycin, instilled intratracheally, developed persistent lung disease.³⁸ C57Bl6 mice harboring the human p.C121G BRICHOS variant³⁷ demonstrate disruptions in the disulfide bonds of the SP-C proprotein, as seen in human patients, and develop spontaneous lung fibrosis.³⁷ Induction of p.C121G allele expression in adult mutant mice activates the UPR and spontaneous fibrotic remodeling of the lung parenchyma after a phase of acute alveolitis.³⁷ Collectively, the findings of the studies noted above emphasize the role of ER stress-mediated initiation or progression of fibrotic lung disease associated with the expression of Sftpc BRICHOS domain variants.

The SP-C linker domain variant p.I73T is the most commonly identified pathogenic variant in SFTPC, which results in the formation of a misfolded proprotein that escapes protein quality control and is misrouted to the plasma membrane.³⁹ Similar to the Δexon4 study described above, both heterozygous and homozygous p.I73T C57Bl6 mouse embryos present with arrested lung development as early as E18.5, which is associated with an acute increase in SP-C p.I73T mutant proprotein expression.⁴⁰ Tamoxifen-mediated induction of p.I73T expression in 8- to 12-week-old mice resulted in acute alveolar remodeling in the lungs, and tamoxifen-dose–dependent lethality as early as 5 days after induction; expression of p.I73T in adult mice also resulted in a phase of acute alveolitis followed by fibrotic lung remodeling.⁴⁰ Taken together, the Sftpc ^I73T and Sftpc^Δexon4 mouse models strongly suggest that the cytotoxic aggregation of mutant SP-C proprotein is associated with lethal lung disease.

MOUSE MODELS OF ABCA3 DEFICIENCY

Targeted disruption of the Abca3 locus in mice results in neonatal lethality and is associated with a profound loss of LBs in AT2 cells.^60–63 Lungs of Abca3-deficient mice are characterized by a significant reduction in surfactant lipids, highlighting the essential role of ABCA3 in surfactant lipid metabolism and LB biogenesis.^60–63 In studies by Besnard et al., conditional deletion of Abca3 from AT2 cells of the SP-C-rtTA:TetO-Cre:Abca3^f/f mouse model (mixed strain background) resulted in surfactant deficiency and respiratory distress; 67% of the Abca3^Δ/Δ pups died shortly after birth with surviving mice developing emphysema by 9 months of age.⁶⁴

Notably, the surviving Abca3^Δ/Δ mice present with two distinct AT2 cell populations: those that expressed ABCA3 and those that did not, suggesting that surviving ABCA3-expressing cells in the Abca3^Δ/Δ mouse may compensate for ABCA3 deficiency and maintain adequate surfactant production. Consistent with this hypothesis, nontargeted ABCA3-expressing cells restored pulmonary homeostasis in the Abca3-conditional KO mouse model, in which Abca3 expression was conditionally ablated through tamoxifen treatment for 4 days.⁶⁵

Twelve weeks after tamoxifen treatment, expression of Abca3 messenger RNA (mRNA) in AT2 cells was completely restored, concurrent with the proliferation of ABCA3-expressing AT2 cells that restored normal lung structure and pulmonary function.⁶⁵ The remarkable selective advantage of a small population of ABCA3-expressing cells (∼23% of proSP-C⁺ AT2 cells⁶⁵) that was adequate for the restoration of alveolar homeostasis substantiates the therapeutic potential of a gene-based intervention for ABCA3 deficiency.

IN VITRO MODELS OF GENETIC DISORDERS OF SURFACTANT DYSFUNCTION

In addition to animal models of genetic disorders of surfactant dysfunction, there is a need for in vitro model systems that recapitulate human disease pathobiology, which can be harnessed to assess variant pathogenicity, inform clinical decision making, and allow for the evaluation and development of effective therapeutic strategies. Overexpression of specific SFTPC^{30–34,39,66–69} and ABCA3^{50–52,61,70–72} variants by either plasmid transfection or adenovirus-based vector-mediated transduction of HEK293T and A549 cells has been utilized to study the downstream effects of mutant protein expression.

Recently, an ABCA3^−/− A549 parent cell line with a single recombination target site or “landing pad” for genomic integration has been used for stable expression and functional characterization of individual ABCA3 missense variants.⁷² Others have attempted to ablate SFTPB in A549 and H441 cell lines through gene editing,⁷³ or have utilized patient-specific induced pluripotent stem cells (iPSCs) expressing mutant alleles from the endogenous locus that were differentiated to AT2-like cells.^18,28,74

To model SP-B deficiency in vitro, Munis et al⁷³ used CRISPR/Cas9 gene editing of H441 cells grown at an air–liquid interface to generate SFTPB KO cells. In this model system, SFTPB KO cells had reduced transepithelial electrical resistance (TEER), which suggested a compromised epithelial barrier. Transduction of SFTPB KO cells with an F/HN-pseudotyped simian immunodeficiency virus (SIV)-based lentivirus (LV) vector encoding SFTPB resulted in measurable SP-B expression and restoration of functional epithelial barrier properties.⁷³

In A549 or HEK293 cells transfected with various SFTPC BRICHOS variants (SFTPC^Δexon4, SFTPC^C121G, SFTPC^C186G, or SFTPC^L188Q), retention of proSP-C in the ER and subsequent intracellular aggregation of the misfolded proprotein has been reported, resulting in activation of the ER stress and UPR pathways leading to apoptotic cell death.^{30,31,33,35,66,68} In vitro studies evaluating the consequences of overexpression of SFTPC^I73T in A549 or HEK293 cells revealed a toxic gain of function due to the altered intracellular trafficking of proSP-C^I73T to the plasma membrane.^34,39,67,69 Subsequent accumulation of misprocessed isoforms within the endosomes resulted in proteostasis perturbations characterized by a late block in autophagy and impaired mitophagy.³⁹

While A549 and H441 (lung adenocarcinoma epithelial) or HEK293 (human embryonic kidney) cells are typically used to study variants of surfactant genes, these lines do not possess the necessary AT2 machinery for surfactant production.^75,76 Three-dimensional (3D) model systems in combination with the development of methods for deriving human AT2-like cells from patient-derived iPSCs (iAT2s) provide a physiologically relevant disease model to study the biology of human distal lungs and aid in the development of novel molecular therapies for genetic disorders of surfactant dysfunction: iAT2s can produce surfactant, and are transcriptionally and ultrastructurally similar to primary AT2 cells.^18,77

Specifically, iAT2s have been used to study the processing of SP-B and SP-C proproteins harboring pathogenic variants using patient-specific mutant iAT2s and their gene-edited progeny.^18,28 Jacob et al. generated iAT2s from an infant carrying the most common pathogenic SFTPB variant, p.Pro133GlnfsTer95, and reported that these iAT2s exhibited reduced SFTPB mRNA expression as a result of the nonsense-mediated decay of the mutant SFTPB transcript.¹⁸

Furthermore, no detectable SP-B protein or LBs were identified, as confirmed by transmission electron microscopy, recapitulating the patient disease phenotype. Intriguingly, “footprint-free” CRISPR/Cas9-based gene editing corrected the SFTPB variant, leading to increased SFTPB mRNA that restored LB ultrastructure and surfactant processing.¹⁸ Another study evaluated the therapeutic benefit of a LV vector expressing hSFTPB in patient-derived iPSCs homozygous for the p.Pro133GlnfsTer95 SFTPB variant.⁷⁴ Differentiation of gene-corrected iPSCs into 3D cultures restored both LB formation and processing of surfactant proteins and phospholipids.⁷⁴

Recently, patient-specific iPSCs heterozygous for the most common pathogenic SFTPC variant, SFTPC^I73T, were utilized to study the intrinsic epithelial dysfunction at the inception of lung disease.²⁸ When comparing the syngeneic mutant iAT2s to the corrected iAT2s, the mutant iAT2s misprocessed and mistrafficked proSP-C, in a manner similar to that observed in AT2 cells in vivo.²⁸ The accumulation of misprocessed proSP-C extended the observations made through the forced overexpression of mutant protein in heterologous cell lines,^34,39,67,78 patient bronchoalveolar lavage samples,⁶⁷ and the Sftpc^I73T mouse model.⁴⁰

Downstream consequences of mistrafficked proSP-C in the iAT2 model included proteostasis perturbations, mitochondrial dysfunction, metabolic reprogramming, and inflammatory activation.²⁸ Notably, gene editing to ablate the expression of mutant SFTPC restored normal AT2 cell-specific function in the patient-derived iAT2s in vitro.⁷⁹ Taken together, these studies suggest that patient-specific iAT2s can indeed be used as a reliable preclinical in vitro model.

Despite the advantages of mouse models,^37,38,40 there are significant differences in the degree and type of surfactant dysfunction-associated lung disease that develops in these models, which limit their effective use for human disease modeling and evaluation of therapeutics. The ideal human AT2 cell dysfunction model should allow for the study of disease pathogenesis by utilizing patient-derived cells that would reveal the cascade of mechanistic events associated with the inception, progression, and resolution of lung disease. However, the study of pathobiology in primary human AT2 cells is hindered by the difficulty of accessing patient-derived samples, especially from patients at the early stages of the disease. Although iPSC-based models provide a unique opportunity for disease modeling using human cells, they also have several limitations.

The main limitation is that most of the established protocols generate cells that are less mature when compared with their in vivo counterparts, due in part to their more proliferative state compared with the typically quiescent AT2 cells.⁸⁰ Importantly, primary AT2 cells in culture are more proliferative and lose some degree of maturation compared with their quiescent in vivo parent cells.⁸⁰ Further, primary AT2 cells derived from patients with pathogenic variants in the surfactant protein genes have yet to be stably maintained in vitro, limiting comparative studies between patient-specific iAT2s and primary AT2 cells cultured from these patients.

GENE THERAPY FOR GENETIC DISORDERS OF SURFACTANT DYSFUNCTION

Gene therapy remains the most promising therapeutic strategy for genetic disorders of surfactant dysfunction. In addition to the availability of several candidate gene therapeutics, the associated genes, SFTPB, SFTPC, and ABCA3, are extensively characterized,^81,82 and several animal models and sophisticated cell culture systems are available for preclinical testing. The two most important challenges of gene therapy for these disorders are delivery to target cell and the cell-specific expression of the therapeutic transgene.

Several strategies can now be employed for AT2 cell-restricted transgene expression, including, but not limited to, cell-specific promoters and/or enhancer elements. Various virus and nonvirus-based vector systems have been utilized for effective gene transfer to the respiratory epithelium.⁸³ Furthermore, there are technologies now available and also under development that allow for precise deposition of gene therapeutics to specific regions of the neonatal lungs. These technologies are reviewed by Katz et al. and McCarron et al.^84,85 In the next section, we review the viral-based and nonviral-based gene therapeutics that have a demonstrated transduction of primary AT2 cells (also summarized in Table 1).

Table 1.

Viral-based vectors used for transduction of cells of the alveolar epithelium in vivo and alveolar type 2 cells in vitro

Vector	Serotype or Envelope	Transgene	AT2 Cell-Specific Ex Vivo or In Vivo Model System(s)	Vector Dose	References
AdV-based vector	Ad5	β-Galactosidase (LacZ)	Human fetal lung AT2 cells and fibroblast cocultures	1.0 × 103 p/cell	86
			Human fetal lung explants (22 weeks gestation)	1.0 × 1011 p/explant
			Mouse lungs	2.5 × 1011 p/mouse, IT
		SP-C1–197, SP-CΔexon4	Mouse AT2 cells	50 MOI	35
		GFP, Bcl-xL, Bcl-2	Mouse lungs	2.0 × 1010 optical particle unit/mouse, IT	87
			Mouse AT2 cells	2.0 × 107–2.0 × 109 PFU/well
		FoxM1B, FoxM1B-TA, LacZ	Mouse AT2 cells	20 MOI	88
			Mouse lungs	5.0 × 108 PFU, IT
		ABCA3 variants	Isolated human precursor lung epithelial cells	3.0 × 103 VP/cell	72
AAV-based vector	AAV1	hALP	Mouse lungs	6.0 × 1010–3.0 × 1011 GC/mouse, IN	98
	AAV2	LacZ	Mouse lungs	1.0 × 1010 p/mouse, IN	97
		GFP	Human embryonic stem cell-derived lung bud organoids	3.5 × 108–1.0 × 109 GC	103
	AAV4	GFP	Pig lungs	1.0 × 1012 VG/pig (bronchoscope-guided delivery to left lobe)	100
	AAV5	hALP	Mouse lungs	6.0 × 1010–3.0 × 1011 GC/mouse, IN	98
		LacZ	Mouse lungs	1.0 × 1010 p/mouse, IN	97
	AAV6	ffLuc, hALP	Mouse lungs	1.0 × 1011 VG/mouse, IN	99
		GFP	Human embryonic stem cell-derived lung bud organoid	3.5 × 108–1.0 × 109 GC	103
	AAV6.2	GFP	Human embryonic stem cell-derived lung bud organoid	3.5 × 108–1.0 × 109 GC	103
	AAV6.2FF	ffLuc, hALP	Mouse lungs	1.0 × 1011 VG/mouse, IN	99
		GFP	Human embryonic stem cell-derived lung bud organoid	3.5 × 108–1.0 × 109 GC	103
		mSftpb, hSFTPB, GFP, mCherry, ffLuc	Mouse lungs	1.0 × 1011 VG/mouse, IT	104
			Precision cut lung slices (human)	1.0 × 108–1.0 × 1010 VG
		mCherry, Cre, Sftpc-miR	Mouse lungs	5.0 × 1011–4.0 × 1012 VG/mouse, IN	105
	AAV8	ACE2, GFP	Mouse lungs	4.0 × 1011 VP/mouse, IT	101
	AAV9	LacZ, α1-antitrypsin	Mouse lungs	1.0 × 1011 GC/mouse, IN	102
LV-based vector	VSV-G	EGFP	Rat AT2 cells	0.1–50 MOI	108
		EGFP	Mouse lungs	1.0 × 108 international units, IT	110
		CDK4, hTERT, SV40	Human AT2 cells	Not available	109
		GFP	iPSC-derived AT2 cells	50 MOI	111
		GFP	Mouse lungs	0.5–3.0 × 108 TU/mouse, IT	116
	SARS-CoV2 Spike	GFP	iPSC-derived AT2 cells	30 MOI	111
	Sendai Virus—F and HN proteins	ffLuc	Mouse lungs	5.0 × 108 TU/mouse, IN	117
	Mokola virus glycoprotein	GFP	Mouse lungs	0.5–3.0 × 108 TU/mouse, IT	116
	Lymphocytic choriomeningitis virus glycoprotein	GFP	Mouse lungs	0.5–3.0 × 108 TU/mouse, IT	116
	Ebola virus glycoprotein	GFP	Mouse lungs	0.5–3.0 × 108 TU/mouse, IT	116

AAV, adeno-associated virus; ABCA3, ATP-binding cassette transporter A3; AdV, adenovirus; AT2, alveolar type 2; ffLuc, firefly luciferase; GC, genome copies; hALP, human placental alkaline phosphatase; IN, intranasal; iPSC, induced pluripotent stem cell; IT, intratracheal; LV, lentivirus; MOI, multiplicity of infection; p, particles; PFU, plaque-forming unit; SP-C, surfactant protein C; VG, vector genomes; VP, viral particles; VSV-G, vesicular stomatitis virus glycoprotein.

ADENOVIRAL VECTORS

Recombinant adenoviral (rAdV) vectors have broad tropism in various species in vivo, and rAdV-mediated gene transfer to AT2 cells both in vivo and ex vivo has been studied extensively.^{35,72,86–89} In a study by Strayer et al,⁸⁶ a replication-deficient AdV vector expressing LacZ was evaluated in adult mice, ex vivo in normal human lung explants, and in vitro in A549 and also H441 cells. As evidenced by β-galactosidase, the rAdV-mediated SP-B–driven LacZ expression was restricted to AT2 cells and club cells.⁸⁶ Another AdV vector, AdV5F35++, a chimeric rAdV5 vector with the AdV35 fiber with two mutations (D207G, T245A), resulted in extensive transduction of AT2 cells in the lungs of nonhuman primates.⁸⁹

In studies by Alapati et al., in utero rAdV-mediated gene editing of the p.I73T Sftpc variant in heterozygous mouse fetuses at E16 increased the survival of newborn mice beyond 24 h, concurrent with a marked improvement in alveolarization at postnatal day 7,⁷⁹ thus providing important proof of concept for therapeutic virus vector-based gene editing for toxic gain of function associated with SFTPC variants.

While the large packaging capacity of AdV vectors can certainly accommodate the coding sequence of all the surfactant-associated genes (SFTPB [1.2 kb], SFTPC [0.6 kb], and ABCA3 [5.1 kb]), AdV vectors trigger host cellular immune responses,^90,91 which may lead to detrimental cytopathic effects when delivered into the distal lungs. Despite approaches to engineer AdV vectors for improved targeting and to prevent in vivo immunological responses,⁹² AdV vectors result in poor and transient (i.e., <14 days⁹³) expression in vivo, which necessitates AdV vector readministration to maintain the therapeutic benefit conferred by the expression of the transgene. To date, effective readministration of AdV vectors in lungs in vivo is hampered by pre-existing and/or acquired vector-specific antibodies.

ADENO-ASSOCIATED VIRAL VECTORS

Compared with rAdV vectors, recombinant adeno-associated viral (AAV) vectors have a favorable safety profile,^94–96 and lead to sustained expression for the lifetime of the transduced cell. The small (∼4.7 kb) packaging capacity of rAAV limits their use to the SFTPB and SFTPC genes. Several AAV capsids (i.e., AAV5, AAV6, AAV8, AAV9) were shown to transduce AT2 cells in vivo^97–102 and in ex vivo organoid models.¹⁰³ rAAV6.2FF, containing the p.F129L, p.Y445F, and p.Y731F mutations in the AAV6 capsid, was used by Kang et al¹⁰⁴ to express the SFTPB complementary DNA (cDNA) in the SP-CrtTA:(teto)₇:SP-B–conditional KO mouse model. Restoration of SP-B proprotein expression led to the survival of neonatal and adult mice coupled with phenotypic changes in the diseased alveoli after gene therapy; LB structure was maintained but most importantly, respiratory failure was rescued.

Notably, the authors showed that intratracheal delivery of rAAV6.2FF did not elicit any local or systemic immune responses.¹⁰⁴ In a subsequent study, Rindler et al¹⁰⁵ used the rAAV6.2FF vector to deliver microRNAs (miRNAs) in the lungs of C57Bl6 mice to target AT2 cells and inhibit Sftpc expression, resulting in a >90% reduction in expression, thus providing a rationale for a gene silencing strategy for pathogenic SFTPC variants. Since AAV vectors remain episomal and do not integrate, vector readministration is necessary for maintenance of therapeutic transgene expression.

The effectiveness of readministration can be impacted by the level of serum circulating AAV vector-specific neutralizing antibodies. Although readministration of either AAV2/9¹⁰² or AAV2/6¹⁰⁶ in mouse lungs was shown to be feasible, the rapid increase in AAV vector-specific neutralizing antibodies precluded the successful intranasal or intratracheal readministration and subsequent transduction with AAV2,¹⁰⁶ AAV2/5,¹⁰² or AAV5/5.¹⁰⁷ Collectively, these studies suggest the use of an immunosuppressive regimen or alternate AAV serotypes for effective readministration.

LV VECTORS

LV vectors have a packaging capacity of ∼7.5 kb and can thus accommodate large transgenes, rendering them a highly suitable vector platform for the in vivo delivery of ABCA3. In addition, LV vectors can be pseudotyped with various envelope glycoproteins to either enhance or confer tissue- or cell-specific tropism. Vesicular stomatitis virus glycoprotein (VSV-G)–pseudotyped LV vectors have been previously shown to transduce AT2 cells and iAT2s.^108–111 In combination with the SP-C promoter (3.7 kb, 5′ flanking sequence), transgene expression was restricted to the AT2 cells in vivo.¹¹⁰

Notably, the 3.7 kb SP-C promoter was shown to regulate expression in the bronchiolar epithelium,^112,113 whereas the shorter 2 kb SP-C promoter improved targeting to the AT2 cells.¹¹⁴ It should be noted that while the 2 kb promoter drove relatively weak transgene expression in AT2-like cell lines (i.e., MLE12, MLE15, and A549), vector-mediated gene expression was improved by the addition of a CMV enhancer element.¹¹⁵ VSV-G–pseudotyped LV vector–mediated delivery of SFTPB to iPSC-derived lung organoids harboring the p.Pro133GlnfsTer95 SFTPB variant highlights the potential of the therapeutic application of LV vectors for the treatment of genetic disorders of surfactant dysfunction. HIV-based LV vectors pseudotyped with glycoproteins derived from the Ebola Zaire virus, the lymphocytic choriomeningitis virus, the Mokola virus,¹¹⁶ and SIV-based vector pseudotyped with glycoproteins derived from the Sendai virus have been reported to transduce AT2 cells in the distal lungs in vivo.¹¹⁷

These findings warrant further research into enhanced AT2 cell targeting through the use of alternate glycoproteins to pseudotype LV-based gene therapeutics, particularly given that VSV-G–pseudotyped LV vectors have relatively poor lung epithelial cell tropism in vivo.¹¹⁸ To sustain long-lived expression in the airways, LV vectors should be able to be readministered. Sinn et al. demonstrated in mice successful intranasal readministration of GP64-pseudotyped feline immunodeficiency virus-based vector.¹¹⁹

Further, Griesenbach et al. demonstrated the feasibility of daily or monthly readministration of the F/HN-pseudotyped SIV vector into mouse airways.¹¹⁷ Collectively, these studies provide important proof of concept for sustaining and/or enhancing transgene expression in the lungs. While marginal, the theoretical risk for insertional mutagenesis induced by the integration of LV vectors remains. Current LV vectors appear to target terminally differentiated cells, adding an additional layer of safety should this adverse event occur.

NONVIRAL VECTORS

Compared with viral-based gene therapeutics, nonviral-based gene therapeutics have a favorable in vivo immunogenicity profile.^120,121 Although nonviral vectors (such as lipid nanoparticles) have been used to target AT2 cells ex vivo,¹²² although with moderate success, there is an apparent lack of reported in vivo alveoli-directed gene therapeutic applications, suggesting the limitation of this specific platform technology.

CONCLUSION

A gene therapeutic for the genetic disorders of surfactant dysfunction holds great promise. Current gene therapy strategies require improvement of AT2 cell-specific expression and the incorporation of silencing strategies to control off-target effects, particularly in cells of the bronchiolar epithelium. Molecular studies conducted in mouse models of surfactant dysfunction disorders suggest that targeting 20–30% of AT2 cells is sufficient to reverse the phenotypic changes associated with specific surfactant gene variants (i.e., SFTPB and ABCA3 variants). The translation, however, of these studies to the target patient population, particularly the neonates with actively developing lungs, is unknown.

Furthermore, the application of a gene therapeutic in these patients is complicated by the limitations of timely identification of candidate patients—infants are often identified at different centers, and very few centers in the United States offer lung transplantation. For these disorders that are characterized by progressively worsening lung disease, gene therapy may provide a bridge to lung transplantation by prolonging the viability of the diseased lungs.

Gene therapy for ultrarare diseases is actively being pursued, and its importance is emphasized by the newly formed Bespoke Gene Therapy Consortium, a partnership between the National Institutes of Health, the U.S. Food and Drug Administration, and multiple public and private organizations that aim to develop customized gene therapeutics for these diseases. In conclusion, the application of gene therapy in neonates, infants, and children with genetic disorders of surfactant dysfunction is feasible, and warrants further investigation for future clinical development.

AUTHORs' CONTRIBUTIONS

S.S. contributed to conceptualization, writing—original draft, review and editing. K.-D.A. and J.A.W. performed writing—original draft, review and editing. M.P.L. provided conceptualization, supervision, writing—review and editing.

AUTHOR DISCLOSURE

S.S. and M.P.L. are employees of Spirovant Sciences, Inc. All other authors have no competing financial interests.

FUNDING INFORMATION

This work was supported in part by an I.M. Rosenzweig Junior Investigator Award from The Pulmonary Fibrosis Foundation and an Integrated Pilot Grant Award through Boston University Clinical & Translational Science Institute (1UL1TR001430) to K.-D.A., NIH NHLBI (R01HL149853 and U01HL13475) and Children's Discovery Institute grants to J.A.W.