As a way of background, pulmonary surfactant, a lipoprotein complex, is essential for life. It prevents alveolar lung collapse by lowering the surface tension at the air-liquid interface of the lung alveolus. Lung alveoli are the distal airspaces in the lung, lined by epithelial Type I and Type II cells. Under normal conditions macrophages are the only immune cells present in the alveolar space. The alveolus is covered by a thin liquid layer, called hypophase. Surfactant is found at the surface of the hypophase, i.e., at the air-liquid interface, as well as in the hypophase as a surfactant reservoir. Through its ability to maintain lung alveolar stability, surfactant enables the lung to carry out its key function of O₂/CO₂ exchange. Deficiency of surfactant in prematurely born infants and dysfunction of surfactant in adults can potentially lead to serious breathing problems including death.

Pulmonary surfactant is composed of about 90% of lipids, primarily phospholipids and four non-serum proteins, the surfactant protein A (SP-A), SP-B, SP-C, and SP-D. SP-B and SP-C are hydrophobic proteins and are involved in activities that primarily affect the surfactant function. For example, surfactant is found in the form of a monomolecular surface film at the air-liquid interface of the alveolus, which is responsible for the reduction of surface tension and thus prevention of lung collapse, and it is also found as surfactant reservoir in the hypophase. The two surfactant compartments are interconnected. During a breath there is reorganization of surfactant layers and the hydrophobic proteins are key for surfactant multilayer connection and bringing lipids from the hypophase to the air-liquid interface (1, 2). SP-A and SP-D are hydrophilic proteins and both belong to the collectin family of proteins. These are primarily involved in innate immunity, regulation of inflammatory processes and may serve as a link to adaptive immunity. In addition, SP-A contributes to various aspects of surfactant structure such as in the formation of tubular myelin (an extracellular structural form of surfactant), and the reorganization of surfactant in the hypophase. SP-D may play a role in surfactant homeostasis (1).

SP-A was the only known surfactant protein at the early times of clinical surfactant replacement trials (3, 4). The success of a human clinical study in 1980 (5), where pulmonary surfactant derived from cow lung (a natural source) was used successfully to treat prematurely born babies at risk for respiratory problems, and the failure of previous studies where off the shelf lipids were used (6), raised interest in the study of SP-A and led to the discovery of the other surfactant proteins (7–15). Although the surfactant proteins in the early years were known by various names, a nomenclature, used today, was agreed upon soon after their discovery (16). SP-A, the focus in this review, in addition to its surfactant-related functions, plays a role in the lung innate immune response and regulation of inflammatory processes (1).

Unlike rodents that have a single gene, humans (17, 18) and primates (19) have two genes, the result of gene duplication about 26.5 million years ago (19). The human SFTPA locus has been mapped at q22-q23 of chromosome 10 (20–22). This locus consists of two functional genes, SFTPA1 and SFTPA2, in opposite transcriptional orientation with a pseudogene, SFTPA3P (23), in reverse orientation relative to SFTPA1 at about 15kb away from the 5′ region of SFTPA1 (21). The two genes are in linkage disequilibrium (24). The SFTPA locus also includes another pseudogene, the MBL3P (mannose binding lectin family member A3 pseudogene) ( Figure 1A ) (25). Although one study found the SFTPA locus close to the MBP locus (26), another study found that the MBP locus is located at a large distance, at about 25,000-35,000kb, from the SFTPA locus (21). Radiation hybrid mapping has placed SFTPA2 and SFTPD (another surfactant protein gene) on the 5′region of SFTPA1 at about 40 and 120 kb, respectively. Their orientation relative to the centromere is SFTPD-SFTPA2-SFTPA3P-SFTPA1-telomere. Although the evolutionary advantage of the SFTPA gene duplication has not been studied, one can only speculate. The SP-A protein is shown to be relatively conserved through the major vertebrate groups and it has been proposed that the surfactant system is an evolutionary prerequisite for air-breathing species (27). We postulate that the original role of SP-A was surfactant-related and at some time in evolution was “co-opted” to serve in host defense. Perhaps with the dual role, the SFTPA gene was subject to evolutionary selection that led to gene duplication. The available literature, as discussed elsewhere (1) indicates that for the most part both gene products carry similar functions but one seems to do a better job in host defense activities and the other in surfactant-related activities. Nonetheless, the functional complementation of the two protein products may mean that the gene duplication in the primate lineage was followed by subfunctionalization via selective pressure that keeps both genes functional (28). This is in contrast with another host defense gene, the mannose binding lectin (MBL), which underwent pseudogenization and lost its second functional gene in humans (29). At present, this hypothesis is simply speculative.

The SFTPA1 and SFTPA2 genes encode proteins that contain both collagenous and carbohydrate regions (22, 30) that places them in the family collagenous C-type lectins or collectins. SP-D, another surfactant protein encoded by the SFTPD gene, also contains both collagenous and carbohydrate regions and it is placed in the same family of proteins (31), along with the mannose binding protein (22). Collectins are soluble pattern recognition receptors and are part of the innate immune system (32–35). They bind various molecules on the surface of microorganisms, such as carbohydrate containing structures and lipids and may eliminate microorganisms employing various mechanisms. Collectins may also modulate regulation of inflammatory or allergic processes and the adaptive immune system. In addition, they play a role in the clearance of apoptotic cells (36).

The human surfactant protein A genes have been identified with extensive genetic and epigenetic variability in coding and non-coding regions and this variability has been associated with many pulmonary diseases (37–43). The impact of the coding genetic variability on function has been reviewed elsewhere (1). Briefly, the most frequently observed coding variants in the general population for each SFTPA gene are six (1A, 1A⁰, 1A¹, 1A², 1A³, 1A⁵) for SP-A2 and four (6A, 6A², 6A³, 6A⁴) for SP-A1 (30, 44). However, the focus in the present review is primarily on the role of untranslated and flanking regions on the regulation of human SFTPA genes.

The 5′ flanking regions of the SFTPA1 and SFTPA2 genes share some conserved cis-acting regulatory elements with differing degrees of sequence conservation. Such elements have been shown to play a role in the differential regulation of the two genes, via certain transcription factors, such as NKX2.1/TTF-1 and NF-kB, or transcription coactivators such as the CBP/p300 factors. These may explain the differential regulation under different stimuli, such as in the presence of dexamethasone or cAMP analogs (45–47). In addition, epigenetic regulation can partially explain the differential regulation of the two genes. DNA methylation sites have been identified upstream of the transcription start site (TSS) of both SFTPA1 and SFTPA2 (48, 49), while histone modifications have been implicated in the regulation under certain stimuli (50).

The human surfactant protein is extensively co- and post translationally modified resulting in a large number of isoforms as shown by two-dimensional gel electrophoresis. Following the use of metabolic inhibitors and enzymes, this complex group of isoforms is reduced to a small number of isoforms that coincide with the isoforms of primary translation products of human lung RNA (51–53). The cloning of the genomic SFTPA1 and SFTPA2 sequences (17, 18) and of their cDNAs (52), enabled comparison of the two groups of sequences and this comparison revealed 5′-untranslated region (UTR) splice variability and 3′-UTR sequence variability in the SFTPA1 and SFTPA2 transcripts, as discussed below, as well as sequence variability within the coding region, but this has been reviewed elsewhere (1).

In humans, multiple transcripts have been identified that are due to 5′-UTR splice variability (54). Although in other species a single transcript (55) or more than one transcript (19, 27) have been identified, it is not known whether these are due to differences in splicing or polyadenylation. Karinch et al. (54), employing primer extension in an attempt to map the transcription start site of each human SFTPA gene and 5′RACE (rapid amplification of cDNA ends) to fully characterize their transcripts, using RNA from two unrelated individuals, discovered an extensive 5′-UTR splice variability as shown in Figure 1B (upper panel). Exon C′, not shown in Figure 1 , and a similar 5′-UTR splice variability was also described by McCormick et al. with a different terminology (56). In this review, we use the Karinch terminology because this has been largely used in subsequent publications. SFTPA1 transcripts appear to use A, A′ and A′′ transcription start sites of “exon A” with equal frequency but the SFTPA2 transcripts use only the A site of “exon A”. The significance of this (if any) has not been addressed further (54).

The major 5′-UTR variants for SFTPA2, ABD and ABD′, are distinguished from the major SFTPA1 5′-UTR variant by the inclusion of exon B (eB). The difference between D and D′ is 3 nucleotides with the D′ having the additional 3 nucleotides, the result of a splice site favorability between D and D′ due a single nucleotide change (57). Two minor 5′-UTR splice variants were observed for SFTPA1 (ACD′ and AB′D′) plus some rare (not shown in Figure 1B ) variants (54). The transcripts from each gene carrying a different major or minor 5′-UTR splice variant are translated both in vitro (except for the AB′D′) (54) and in vivo as shown by polysome bound RNA (58). However, differences in both relative translatability and relative levels of the splice variants were observed among individuals (58).

Extensive sequence variability including small deletions/insertions was also observed in 3′-UTR (54, 56, 59, 60). This variability included an 11-nucleotide (11-nt) insertion/deletion ( Figure 1C ) that was initially described for an SP-A1 variant named 6A¹ (59). This 6A¹ variant was identical to the most frequently found SFTPA1 variant, 6A², except at a single nucleotide (54). However, since the 6A¹ variation was not present in subsequent sequencing data, it was presumed that this was a sequencing error and the 6A¹ is referred to as 6A² thereafter. This 11nt sequence provides potential binding sites for miRNA (61), shown to regulate expression (62). In addition to the 11-nt, other elements have been identified, by sequence comparison, in the 3′-UTR of SFTPA1 and SFTPA2. These include the minimum AU-rich element motif UUAUUUAUU shown elsewhere to mediate mRNA degradation (63). This is present in SFTPA2 variants at position 926-935 but not in the SFTPA1 variants (61).

Under baseline conditions, SFTPA mRNA levels vary significantly among individuals (64) with a sixfold difference between high and low expression among individuals (57). The lack of correlation between total SFTPA mRNA levels and the SFTPA1/SFTPA2 transcript ratio indicated that the levels in an individual may vary as a function of the SFTPA genotype, where the level of transcription and/or stability of mRNA may differ among variants (57). A variability in SP-A protein levels in bronchoalveolar lavage is observed among individuals (65–68) and during development (69). Together these indicate that there may be mechanisms that differentially affect the expression of the SP-A variants. In fact, in response to a variety of stimulatory or inhibitory regimens in fetal lung explants or in the human adenocarcinoma H441 cell line, the levels of human SP-A protein and/or mRNA change significantly (70–73). Also, SP-A levels may change in certain disease states (65–67, 74–76). Furthermore, inhibitory or stimulatory substances were shown to differentially affect the regulation of SFTPA1 and SFTPA2 mRNA in fetal lung explants or cell lines (46, 77–79). Although differences in protein and mRNA levels in health and in various disease states have been observed, to the best of our knowledge no comprehensive study has been done to correlate mRNA levels of each human SFTPA gene/variant and protein levels in samples from patients or healthy individuals or experimental systems where each transcript in its entirety is studied. We speculate that given the opportunity for complex regulation, as discussed in this review, it is probably unrealistic to think that a direct correlation of the overall mRNA and protein levels could exist without providing additional specific information. Such information may include reference to the specific genetic variant, the splice variant, the specific conditions, i.e., exposure to various insults including environmental stressors, and other. Experimental models of 5′-UTR or 3′-UTR regions of the SFTPA1/A2 variants are shown to exhibit differences in response to various insults and thus the mRNA/protein levels of the two genes and/or of their variants may differ. It would probably require additional reagents and approaches, such as gene- and variant-specific antibodies, and direct RNA sequencing to make such a correlation meaningful.

Alterations of SFTPA in lung cancer have been observed in various experimental settings (133–139). In a few cases, SFTPA was shown to be useful as a marker of differential diagnosis of metastatic lung cancer and mesotheliomas (138, 140–143). A high-throughput approach identified aberrant methylation of CpG sites of several genes to associate with lung cancer (144) indicating that this may be a contributing process in lung cancer. A subsequent DNA CpG methylation profiling of the SFTPA1 gene promoter identified two CpG SFTPA1 sites (SFTPA1_370 and SFTPA1_1080, Figure 4C ) to be hypomethylated in lung cancer (adenocarcinoma and squamous cell carcinoma). In normal lung tissue the level of methylation of another SFTPA1_1468 CpG site (not shown to significantly differ in its methylation content in cancer lung tissue) was associated with the level of SFTPA1 transcripts. This CpG is located 160 nucleotides upstream of the TATAA box. The high level of unmethylated CpG_1468 was correlated with a high level of SFTPA1 transcripts indicating that the methylation status of this site may play a role in SP-A1 expression. This site is absent from SFTPA2. Of relevance, rare SFTPA1 transcripts, including a more frequently found SFTPA1 transcript coding for the 6A⁴ protein variant, were shown to associate with risk for lung cancer (145). A CpG DNA site methylation difference was also observed between normal and cancer lung tissue for the SFTPA2 gene promoter (49). This CpG site is located -2215 upstream of the transcription start site ( Figure 4B ) and exhibited a higher level of methylation in lung cancer especially in adenocarcinoma. Moreover, the level of SFTPA2 mRNA and protein were reduced in lung cancer whereas the mRNA level of DNA methyltransferases (DNMT1 and DNMT2) was increased. An in-silico analysis revealed a number of potential binding sites of transcription factors around this CpG methylation site indicating that apart from its potential as a marker, this CpG site modification may interfere with the binding of regulatory factors affecting SFTPA2 expression.

DNA methylation, which is affected by environmental factors, air pollution, smoking, diet among others (146–150), may be a regulatory mechanism for SFTPA gene expression. The differential regulation of the methylation status of specific CpGs in SFTPA1 and SFTPA2 maybe one of the epigenetic processes that does not only apply to lung cancer but also to other health states (151). Furthermore, epigenetic phenomena that modulate changes in gene function without changing the nucleotide sequence include several processes, such as DNA methylation, histone modifications, miRNAs, and splice variants. The role of miRNAs, splice variants and DNA methylation has been discussed above. Histone acetylation and methylation have been shown to affect SFTPA expression in the lung during development and hypoxia (50, 116, 152). The developmental timing of SFTPA1 expression is associated with enhanced acetylation and decreased methylation of histone H3 at the SFTPA promoter. Histone methyltransferases Suv39H1 and Suv39H2 are bound to the TBE prior to induction of SFTPA by cAMP (152). Their transcript levels are inversely correlated with the developmental pattern of SP-A expression (152). Increased O₂ tension facilitates the induction of histone H3 acetylation on lysines 9 and 14 at the TBE, while hypoxia induces dimethylation of lysine 9 of H3 (50) and recruitment of Suv39H1 and Suv39H3 to the TBE (152). Dexamethasone treatment increased nuclear levels of the histone deacetylases HDAC-1 and HDAC-2, but not the total levels of histone H3, in human fetal type II cells (116). ChIP analysis indicated that dexamethasone also increases the occupancy of the TBE specifically by HDAC-1 and HDAC-2. The cAMP analogue Bt₂cAMP increased the levels of acetylated and phosphorylated histone H3; this effect was antagonized by dexamethasone, which promoted demethylation of H3K9 globally and locally, in the TBE region of the SFTPA promoter (116).

Although differential allele expression has not been studied for the human SFTPA genes either in the lung or extrapulmonary tissues, it is worth noting that this mechanism may be operative under certain conditions, as it is shown for the rat SFTPA under unperturbed conditions (153). SFTPA was shown under these conditions to exhibit a balanced biallelic expression in the lung but in colon was both balanced and imbalanced and family studies indicated that inheritable factor(s) may contribute to the regulation of differential allele expression (153).

Understanding differences between human SFTPA1 and SFTPA2 genes and their corresponding variants is both useful and important as such information could be implemented in personalized regimens. For example, in a pilot study where precision cut lung slices from human donor lungs were treated with varying concentrations of methylprednisolone, a pharmacologic relationship was observed between treatment and SFTPA genotype (154). As lung donors and recipients are treated with immunosuppressive regimens, one may in future pharmacogenetic studies of lung transplants further investigate as well as consider immunosuppressive regimes tailored according to the donor’s genetic background. Recent developments in sequencing technologies, particularly the advent of long-read sequencing, could contribute to our understanding of variant frequencies and their potential regulation. Furthermore, identification of key regulators, either cis or trans, can be useful in modulating specific SFTPA1 or SFTPA2 gene expression as it may be appropriate i.e., in the case of the prematurely born infants where levels of SP-A are low and/or in the course of infection. Of interest, the SFTPA2 transcript coding for the 1A⁰ protein variant has been shown to exhibit a protective effect in terms of survival in both animal models after infection and/or other insults (155, 156) and in lung transplant patients (39). Thus, in such cases it could be advantageous to modulate expression of SP-A1 and/or SP-A2 proteins.

The work presented in this review indicates that the human SFTPA genes are under extensive and complex regulatory control that merits further experimentation. Our current knowledge on this topic is based for the most part on studies where different regions of a given gene was studied i.e., flanking region, 5′-UTR, 3′-UTR. So at the present time it is difficult to know exactly how a given variant will respond to a stimulus or under a certain environmental condition in its entirety. For example, regulatory mechanisms of one region (i.e., 5′-UTR) may attenuate or enhance i.e., an up or down regulation imparted by mechanisms operative in another region (i.e., 3′-UTR, flanking region) or even nullify effects. A major advance to our understanding could be achieved if the entire SFTPA locus is studied as a unit in order to better assess the commonalities and differences between the two genes as well as the potential role of the SFTPA3P on the regulation of the functional genes, as pseudogenes may contribute to the regulation of their functional counterpart (157–159). It is currently unknown whether SFTPA3P is expressed or not, and if so, under what conditions and at what developmental stage. If it is expressed, it could act as a lincRNA, thus regulating the neighboring parental genes in a number of possible ways (160). Given that the two functional genes are in opposite transcriptional orientation and may share regulatory cis elements that may or may not work in a coordinated regulation, examining the entire locus as a whole would be a sensible approach, albeit one with substantial challenges. Identifying an appropriate study system would be a start. Given the size of the locus, generation of humanized mice would be technically challenging, so studies in non-human primates might be more appropriate.

JF had the idea for the article. JF and NT performed the literature search and wrote the manuscript. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.