The presence of SNO-SP-D in BAL from clinically relevant samples has recently been demonstrated. SNO-SP-D was indirectly identified in BAL from patients with Hermansky–Pudlak syndrome type 1 (122) and directly detected in samples from asthmatic patients after segmental challenge with allergen (242). Hence, a role for SNO-SP-D as a disease marker is conceivable, as it is enriched in BAL from patients with pulmonary disease (122, 242). An ELISA-based method to measure circulating SNO-SP-D in COPD patients has been developed using anti-SP-D capture antibodies and antibodies reacting with S-nitrosylated groups for detection; however, this approach did not identify an association with disease severity and found only a weak correlation with radiologist score of emphysema, while possible associations with disease activity have not been verified (286). Nevertheless, detection of circulating SNO-SP-D in additional pulmonary diseases is warranted.

A similar ELISA-based approach has been applied for measurement of fucosylated SP-D, demonstrating the presence of a core-fucose in N-glycans groups on serum SP-D, and an association of circulating levels of fucosylated SP-D with COPD outcomes. This modification represents a promising circulating disease-associated SP-D biomarker candidate, although in initial experiments measured levels of fucosylated SP-D could not adequately separate never-smokers, COPD smokers, and COPD (241).

An alternative approach to simple SDS-PAGE-based measurement of SP-D degradation products in BAL or serum is permeation chromatography-based size separation; however, both techniques may be biased by genetic size variants (encoded by rs721917) and by oxidative crosslinking. Moreover, enzymatic neutrophil elastase degradation of SP-D and consecutive production of monoclonal antibodies against proteolysis products did not appear to result in immunological recognition of disease-specific SP-D breakdown products (286). Although various proteases are known to degrade SP-D, this approach has not yet been developed sufficiently to enable detection of disease-induced SP-D neoepitopes.

Fetuses carrying the SFTPD rs721917 Met11 allele have been associated with spontaneous preterm birth (235), and an initial genetic association of SP-D with pulmonary outcomes in premature infants was suggested by the observations that 2-marker SP-D/SP-A haplotypes including the Met11 allelic SP-D variant were protective against the development of RDS (306) and harmful in BPD (232), respectively. Subsequently, the rs1923537 polymorphism, located downstream of SFTPD, was demonstrated as associated with RDS and the requirement for oxygen supplementation at day 28 (a proxy of mild BPD) in very early preterm birth infants (234), whereas other SFTPD SNPs, including rs721917 (Met11Thr), were not associated with either RDS or BPD in that study or a subsequent investigation (307). Similar tendencies were observed in another study, although the genetic Met11 SP-D variant appeared to be protective for BPD defined radiologically and with requirement for supplemental oxygen at a gestational age of 36 weeks (295). In the most recent study, several SNPs in SFTPD, including the Met11 variant, were found to be positively associated with circulating SP-D levels and harmful in respiratory distress, the requirement for oxygen supplementation at day 28, and respiratory support (233).

Although they generated conflicting data, the above studies support a contribution of SP-D genetic variation to pulmonary outcomes in prematurity, and the discrepancies between studies may be explained by differences in statistical power, different mean gestational ages and the associated variation in requirement for oxygen supplementation at day 28/36, or the investigated SNPs or haplotypes may reflect variation at other, causative variants. No association between SFTPD alleles and neonatal mortality has been identified, and this may explain the persistence of high-frequency SNPs associated with respiratory outcomes in prematurity. No association was identified between SP-D SNP variation and diffuse lung disease enriched for genetic surfactant dysfunction (308).

Surfactant protein D expression in the fetal distal airways increases with advancing gestation and is evident in 10-week-old fetuses; however, in lungs from infants with RDS and BPD, only open terminal airways were bordered with SP-D expression. Injured areas lined with hyaline membranes, or alveoli filled with hemorrhage, infection, or edema fluid, were lightly stained or unstained, whereas serous cells of bronchial and tracheal glands were consistently stained, particularly in infants with lung inflammation (10).

The percentage of multimeric SP-D in neonatal BAL, which is capable of binding microbial compounds, appeared to be lower in preterm than term infants (309), and alveolar SP-D may essentially be absent in the presence of RDS, increasing after surfactant treatment (310, 311). However, an increase in preterm BAL SP-D the first day (day 1) after birth has been demonstrated (309, 312), in parallel with a transient increase in capillary SP-D levels. This transient, circulatory increase was predominantly observed in infants homozygous for the rs721917 Met11 allele and was more apparent in infants with respiratory distress or receiving respiratory support. Such observations suggest that genetic SP-D variation determines the magnitude of induction of respiratory support-dependent systemic SP-D levels. Serum SP-D levels, which were not significantly affected by antenatal steroids, were positively associated with gestational age, mode of delivery, risk of later septicemia, and risk of respiratory distress (65, 233, 313). A recent study correlated serum SP-D measured at later time points (day 3 and day 7 after birth) in preterm infants and found no relation with the requirement for mechanical ventilation or oxygen, or with the development of BPD using that approach (314).

Mouse fetal pulmonary SP-D expression increases with advancing gestation, and levels predominantly elevate shortly before birth, stimulated by vascular endothelial growth factor signaling and glucocorticoid treatment (54, 66–69, 71, 315, 316). Pulmonary SP-D expression is therefore very low in experimental prematurity (317); however, various studies of the Sftpd^−/− lung phenotype do not support a role for SP-D in normal fetal or postnatal lung development, although the vast accumulation of phospholipids in the Sftpd^−/− lung have effects on surfactant homeostasis. Rather, these studies support a role for SP-D in emphysema and fibrosis development during airway remodeling processes later in life, as recently reviewed by Bersani et al. (318).

An early study of baboons showed that both the expression and protein secretion of pulmonary SP-D precede that of SP-A in normal gestational development in the baboon, and they are comparable to, or exceed, adult levels during advancing gestational age. In line with observations of the induction of circulating SP-D with respiratory support, data from premature baboons receiving 100% oxygen for 10 days to produce chronic lung injury indicated that lavage concentrations of SP-A reached a low percentage of that of normal adults, while those of SP-D equaled the amounts present in normal adults. The combined lavage SP-A/SP-D pool reached a low percentage of that of normal adults and the authors concluded that the combined decreased concentration of surfactant host-defense proteins may augment proclivity to infection and worsening injury (319). This suggestion was partly supported by a subsequent experiment demonstrating an increased tendency for lung infection in the same model; however, the major experimental effects were related to massive depletion of SP-A (320). Surprisingly, Sftpd^−/− mice are resistant to hyperoxia, which may be partly explained by phospholipid and SP-B-mediated induction of surfactant resistance to inactivation (321).

Consistent with clinical lung maturation, the major effects of experimental chorioamnionitis are fetal lung inflammation, increased airway surfactants, and increased lung volumes (322–324), while the pro-inflammatory stimulus of chorioamnionitis is also commonly associated with preterm delivery and subsequent RDS (325, 326). A clinical study did not support an association of increased SP-D in amniotic fluid with intra-amniotic infection (327). In contrast, studies of mice and lambs have demonstrated a clear induction of SP-D expression in the fetal lung after LPS treatment (328, 329), although an earlier study indicated greater fluctuations in SP-D expression (330). Furthermore, studies using transgenic mice overexpressing rat SP-D under the SP-C promoter demonstrated that SP-D enhances cytokine production in the fetal and maternal compartments on maternal LPS exposure. Moreover, a significantly higher proportion of the pups born to dams overexpressing SP-D were stillborn after LPS treatment compared with those from wild-type mice (331). Moreover, mice that are doubly deficient in both SP-D and SP-A had delayed parturition and decreased expression of inflammatory and contractile genes (332). A recent study extended these findings, demonstrating that Sftpd^−/− female mice have fertility defects, evidenced by smaller litter size, increased pre-implantation embryo loss, and elevated uterine inflammation, when mated with wild-type males; however, in support of previous findings, maternal LPS administration did not result in increased embryo loss or pro-inflammatory responses in Sftpd^−/− females (333).

In contrast to the surprising association between high intra-amniotic SP-D and preterm birth, addition of recombinant human SP-D to commercial surfactant containing SP-B and SP-C alone improved surfactant function by protecting the premature lung from ventilation-induced inflammation and by increasing its resistance to protein inhibition of surfactant function and changing its biophysical properties and structure when tested in lambs shortly after birth (334). Although pulmonary inflammation was not blocked by SP-D, exogenous SP-D was effective in a model of endotoxin shock in newborn preterm lambs, where intratracheal administration prevented systemic inflammation and decreased cytokine expression in the spleen and liver (169).

Although structural SP-D polymorphisms are not associated with allergic bronchial asthma (220), an association with decreased atopy was identified in black subjects (221).

Several studies have documented induction of BAL or systemic SP-D levels in asthma, which may be attributable to a combination of induced SP-D synthesis in airway epithelia (335) and increased air–blood barrier integrity, as described ealier.

Levels of SP-D are increased in BAL samples from allergic asthma patients (336), may further increase after segmental allergen challenge, and be correlated with those of BAL eosinophils, which in turn are correlated with NO content in BAL and oxidized SP-D species (242). Baseline SP-D levels are elevated in serum from patients with allergic asthma, and further elevated after allergen challenge, which is predictive for the late asthmatic response and for eosinophil cationic protein concentrations post-challenge (337).

Recent studies have supported these initial observations and demonstrated that sputum SP-D is increased in severe asthma or severity of exacerbation (338, 339), that serum SP-D increases stepwise in mild to moderate and severe disease, and correlates inversely with lung function and directly with small airway resistance (273). Mackay et al. (19) demonstrated that serum SP-D was increased in severe asthma with mixed eosinophilic and neutrophilic inflammation and enriched for SP-D breakdown products. The latter observation was supported by the recent findings of Fakih et al. (217) of a significantly decreased HMW/LMW serum SP-D ratio in asthmatic patients. In the studies of Mackay et al., the BAL/serum SP-D ratio was reported to be decreased, implicating depletion of SP-D from BAL due to leakage of degraded SP-D to the serum. Furthermore, serum SP-D levels were inversely associated with alveolar neutrophil infiltration and alveolar endotoxin levels (19).

Investigations that did not support a relationship between asthma and SP-D variation include one of the earliest studies of serum SP-D (340), and a recent study of serum SP-D by Akiki et al. (341). These discrepancies may be explained by differences in the distributions of mild and severe asthma between investigations. Serum SP-D enrichment is observed in pulmonary allergies (342–346) and allergic rhinitis (219), which is considered to be a manifestation of the same underlying disease processes as asthma (347), and basic research has demonstrated a link between SP-D and T helper 2 cell (Th2)-mediated inflammation. Thus, differences in clinical observations may also be due to a lack of stratification for different distributions of allergic asthma, or asthma characterized by a Th2-high profile, which accounts for approximately 50% of patients with steroid naïve asthma (348), and the impact of neutrophilic inflammation or pathogenic load, as implicated by the observations of Mackay et al. (19).

Allergic asthma models have provided insight into the physiological influence of SP-D in the development of Th2 type allergic asthma. The use of Sftpd^−/− mice as allergic asthma models has provided both clear and subtle allergic asthma phenotypes; however, uniform data have been obtained from studies using exogenous administration of SP-D (110, 197, 349). A large body of studies have used rfhSP-D in treatment protocols and, although rfhSP-D did not efficiently aggregate and opsonize pollen grains relative to native SP-D (92), both SP-D and rfhSP-D appear to dampen the majority of aspects of the allergic phenotype. Intranasal administration of SP-D/rfhSP-D in murine models of pulmonary hypersensitivity induced by diverse allergens and antigens suppresses specific IgE levels in serum, reduces peripheral and pulmonary eosinophilia, and causes T helper 1 cell polarization from the allergic Th2-mediated inflammation to varying degrees (110, 123, 144, 203, 204, 206, 221, 349–353). In one model, beneficial effects of exogenous SP-D were observed when it was administered 6 h after, but not 24 h before, allergen challenge. A single application of rfhSP-D to allergen-sensitized mice led to a dampening of the allergic airway response equilibrium, similar to the effect of budesonide (352).

Moreover, allergen exposure induced SP-D protein levels in an IL-4/IL-13-dependent manner, resulting in increased murine alveolar SP-D levels (353, 354), and this negative feedback loop appears to protect the airways from inflammatory damage after allergen inhalation, as described in a recent review (355). SP-D in lavage and tissue is derived from AT-II cells and Club cells; however, it is also synthesized in hyperplastic goblet cells of inflamed lungs (356).

Additional beneficial SP-D effects in asthma may be anticipated in virus-induced asthma, because of the anti-viral effects of SP-D; however, deleterious effects may also be expected due to disease-induced formation of SNO-SP-D. Disease-related proteolysis is suspected to render a fraction of SP-D malfunctioning or deleterious (260), nevertheless, the overall effects of SP-D appear to be beneficial in the context of allergic asthma. The main reported individual steps leading to SP-D-dependent experimental phenotypes are illustrated in Figure 3.

Some of the effects illustrated in Figure 3 remain unclear. SP-D can increase pollen starch granule (PSG)-positive cells in vitro and accelerate PSG binding/uptake in vivo; however, studies by Winkler et al. (93) demonstrated that it did not affect total clearance of PSGs from the mouse lung nor enhance T-cell proliferation induced by PSG-positive dendritic cells. Hence, the different results obtained using human cell cultures or clinically isolated cells, compared with those from mice models, require further investigation.

A wider role for SP-D in Th2 immunity was recently proposed in a study of infection with the helminth Nippostrongylus brasiliensis. Elevated SP-D production in Th2 immunity is partly driven by IL-4 and IL-13. In turn, SP-D can exert negative feedback control of Th2 responses (90, 353); however, whereas allergic asthma studies identified increased type 2 immunity in Sftpd^−/− mice, similar effects were not observed in N. brasiliensis infections. The results of these investigations demonstrate that elevated SP-D can enhance type 2 immunity and suggest that SP-D is an important modulator of protective IL-13 producing type 2 innate lymphoid cell-mediated and alveolar macrophage responses against N. brasiliensis (90).

Associations of polymorphisms or haplotypes in, or flanking, SFTPD with COPD, emphysema, and COPD survival have been identified in various populations and by GWA studies (215, 216, 224–226, 360–362). As reviewed by Lock-Johansson et al. (363), specific SNP associations have not been validated in all investigated populations (363). The SNPs associated with COPD and circulating SP-D levels differed within some investigations, suggesting distinct genetic influences on COPD susceptibility and SP-D levels. In addition, two coding SNPs in SFTPD were associated with expiratory lung function in a study of preclinical smoke-induced lung injury (212), suggesting that structural variants of SP-D affect the susceptibility of COPD development in smokers. Recently, Mendelian randomization analyses were used to analyze serum SP-D-associated genetic variants and their association to COPD in the largest study executed until today. Variants were tested for association with COPD risk in 11,157 cases and 36,699 controls and with 11 years decline of lung function in 4,061 individuals. This study concluded that variants associated with increased serum SP-D levels decreased the risk of COPD and slowed the lung function decline (364).

The majority of studies of smokers have reported reduced alveolar levels of SP-D (252, 269, 278, 365), and alveolar epithelial injury after LPS instillation was more severe in smokers than non-smokers, with increased circulatory SP-D and decreased BAL SP-D (277). These clinical observations are supported by in vitro observations that nicotine can cause reduced levels of SP-D in human airway epithelial cells (366). Additional studies have demonstrated increased circulating SP-D with tobacco smoking or other types of noxious exposure, including data from a twin study showing that smoking monozygotic twins had markedly increased serum SP-D relative to their non-smoking twins, despite the very high heritability of serum SP-D levels (208, 252, 280, 340, 367, 368); however, smoke-induced serum levels may not be clearly evident in mixed populations of respiratory patients, where SP-D levels may be influenced by other processes (369). As described earlier, cigarette smoke disrupts the quaternary structure of SP-D molecules (252). However, an initial attempt to construct an immunoassay for proteolytic SP-D breakdown products in serum did not provide additional information regarding COPD (286).

Whether SP-D levels can differentiate COPD from smoking or other respiratory diseases is uncertain. Some studies have noted significant decreases in SP-D in BAL samples from COPD patients compared with current smokers (252, 279), while others have not (269, 278). In a multicenter study, “Evaluation of COPD Longitudinally to Identify Predictive Surrogate Endpoints” (ECLIPSE), of approximately 2,000 individuals with COPD, higher serum SP-D levels were reported among COPD patients relative to current and former smokers without respiratory obstruction. As the largest difference in serum SP-D levels occurred between non-smokers and current/former smokers, it was concluded that SP-D is a powerful biomarker for smoking; however, there was no difference in serum SP-D levels in individuals with COPD, or in smoker controls with chronic bronchitis, compared with those who did not have this symptom (280). In addition, there are examples of smaller studies that have not shown significant differences between COPD smokers and control smokers (226, 370, 371).

Chronic obstructive pulmonary disease is a heterogeneous disease with several features that overlap with asthma, which has important therapeutic implications for some patients. Data from a randomized control trial of inhaled glucocorticoid therapy in COPD (Groningen Leiden universities chronic obstructive pulmonary disease) confirmed that alterations in airway gene expression may coexist in asthma and COPD, and suggested that Th2 inflammation is important in a subset of patients with COPD who have no history of asthma (372). Despite the vast body of experimental evidence for a role of SP-D in eosinophil-mediated allergy/asthmatic disease, there is no reported evidence linking changes in SP-D expression/function and eosinophil activity in COPD; however, independent studies have provided evidence that supports elevation of circulating SP-D in COPD compared with asthma patients; thus, SP-D may differentiate these two diseases. Yet, these studies may not have controlled sufficiently for the effects of smoking or the severity of asthma (340, 341). Size separation of HMW and LMW serum SP-D appeared to provide a clear distinction between controls and COPD subjects, with the HMW/LMW ratio significantly decreased in COPD after controlling for smoking and additional confounders; however, a similar association was evident for asthmatic patients (217). Additional studies have suggested that smoking-related SP-D variation is prognostic in lung cancer (301, 302).

Semi-quantitative studies on aging Sftpd^−/− mice (395) have indicated that they develop progressive emphysema, increasing septal wall thickness (395–397), and subpleural fibrosis (194), which may indicate a phenotype that combines emphysematous and fibrotic changes in the lungs. Highly detailed stereological investigations of the long-term progression of lung parenchymal remodeling, destruction, and septal wall thickening in aging Sftpd^−/− mice were recently undertaken by Schneider et al. (398). These investigations concluded that the unchallenged aging Sftpd^−/− pulmonary phenotype was characterized by airspace enlargement only, without septal wall thickening (fibrosis) or destruction (emphysema). The age-induced airway enlargement was suggested to result from unexplained altered airway mechanics. The authors further suggested that the discrepancies in observations with previous studies might partly result from the different genetic backgrounds of the mouse strains used and ongoing alveolarization in C57BL6 mice (398).

Although the role of SP-D in spontaneous airway remodeling was recently questioned, several studies of Sftpd^−/− mice challenged with noxious stimuli have provided clear evidence for a role of SP-D in pulmonary inflammation and emphysema development, which are essential components of COPD pathophysiology. These results support a role for SP-D in modulating alveolar macrophage activation, oxidant production, and MMP activity, leading to emphysema-like and fibrotic changes of the lung, and histological and morphometric studies have demonstrated airspace enlargement and/or emphysema development associated with Sftpd deficiency. Inflammatory and structural changes may be reversed or prevented by local treatment with rfhSP-D or full-length SP-D (200, 202, 399, 400). These changes appear to predominantly affect the alveolar compartment, and there is no apparent SP-D-dependent effect on vascular physiology after cigarette smoke exposure (401).

In contrast to findings in humans with COPD, in Sftpd^−/− mice neutrophil infiltration was not associated with lung remodeling during emphysema development (175). Moreover, alveolar levels of SP-D in mice and rats may increase after pulmonary exposure to noxious stimuli (276, 402–405), in contrast to the majority of findings in humans; however, reductions of SP-D in rat BAL after exposure to noxious stimuli have been reported (406). Viral and bacterial respiratory tract infections amplify the chronic inflammation in COPD by triggering pathogen-recognition receptors, as recently reviewed by Brusselle et al. (407). Additional beneficial SP-D effects in COPD development or progression may be anticipated due to the anti-viral and anti-bacterial effects of SP-D. The main individual steps contributing to the SP-D-depressed experimental phenotypes relevant to COPD are illustrated in Figure 4.

The phenotype of mice deficient for the proposed SP-D receptor, GPR116, is highly similar to that reported for Sftpd^−/− mice, including alveolar enlargement, enhanced ROS production, accumulation of foamy macrophages, and enhanced expression of Mmp12. The surfactant lipid changed induced by SP-D may be attributable to its effects on alveolar type II cell activity (177, 193, 414). Although GPR116 deficiency also causes an emphysema-like phenotype that is associated with alveolar macrophage activation (177, 416), the phenotypes of neither GPR116 nor Sftpd-deficient mice fully reflect those of human COPD. Mouse alveolar macrophages become foamy with lipid-laden phagosomes due to extensive uptake of surfactant lipids (177, 417, 418); however, in human pathology, foamy macrophages are found in diverse disorders, including diffuse pan-bronchiolitis or bronchiolitis associated with bronchiecstasis (419), diseases of surfactant homeostasis (420), extrinsic allergic alveolitis (421), and tuberculosis, rather than in COPD. Moreover, the alveolar accumulation of surfactant lipids in the two types of gene deficiency appears to reflect aspects of pulmonary alveolar lipoproteinosis, derived from defective granulocyte macrophage colony-stimulating factor signaling (124, 396, 422), or other types of macrophage exhaustion. Thus, the appearance of foamy macrophages in the Sftpd^−/− lung may not be specific to COPD-related processes, although induction of MMP-12-synthesizing foamy macrophages results after exposure of mice to cigarette smoke (276), and foamy macrophages provide a source of ROS and inflammatory signaling in mouse models as a result of Sftpd deficiency (125).

A recent study of constitutional circulating SP-D measurements in an elderly twin population, where subclinical disease was likely to be present, demonstrated a relationship between the highest SP-D levels and increased all-cause mortality. Female twins with the highest SP-D levels had a significantly increased risk of dying before their co-twin during the study follow-up period. Adjustment of this analysis for intrapair differences in smoking pack-years did not affect the association, indicating that SP-D is not merely a proxy for smoking in reflecting mortality (444). A study of dementia in the elderly reached a similar conclusion about the association with all-cause mortality, without a gender-bias (445). Hence, it is possible that SP-D is associated with CVD, which is the number one cause of death globally.¹

The sources of circulating SP-D are not entirely clear, and it is possible that spill-over of SP-D from the diseased arterial wall into the circulation may affect total serum levels, in addition to the lung, as a main contributor. COPD and coronary artery disease (CAD) frequently coexist, COPD is an independent risk factor for CAD and increased risk for cardiovascular mortality, and patients with diagnosed and treated COPD are at increased risk for hospitalization and death due to CVD (432, 446–451). These correlations suggest that the association of SP-D with all-cause mortality may be rather complex. Some recent studies have demonstrated that SP-D levels correlate with alveolar leakage in heart failure (433), and with the presence of submassive pulmonary embolism (434), supporting the hypothesis that variation in circulatory SP-D results from disease-mediated lung damage in some types of CVD.

Recently obtained data indicated nominal associations between SFTPD coding variants, including the Met11Thr polymorphism, and subclinical atherosclerosis, and demonstrated that the direction of the effects was dependent on tobacco smoking, but independent of circulating SP-D levels (214). An additional study identified a borderline significant association of the Met11Thr polymorphism with advanced atherosclerosis, and this variation was observed to significantly contribute to the risk of arterial stenosis when included in a complex model with other inflammatory gene variants (222).

As suggested above, innate immune factors, including SP-D, may provide a link between respiratory and vascular morbidity and mortality. In this context, circulating SP-D was positively associated with both all-cause mortality and cardiovascular mortality in a large study of the Vancouver Coronary Angiography Cohort by Hill et al. (429), and inclusion of SP-D into a multi-marker analysis improved risk prediction across all levels of risk. There was also a significant positive association of SP-D with triple vessel disease and a borderline association with total cholesterol; however, no associations with other serum lipid fractions or diabetes mellitus were observed in that study. SP-D was also associated with all-cause mortality in the ECLIPSE COPD cohort (287), as described above, and in a study of the relationships among COPD, coronary artery calcification, and mortality using the same cohort, circulatory SP-D was correlated with Agatston score (432). The relationship between SP-D and CVD is further supported by the observation of Hu et al. (452) of positive associations between circulatory SP-D levels and carotid artery intima-media thickness and severe coronary artery calcification, which persisted after multivariate adjustment. Hill et al. (429) and Hu et al. (452) did not report BAL SP-D or other means of monitoring the presence of lung disease and, therefore, did not address the question of the organ source of circulatory SP-D.

Surfactant protein D expression, as well as expression of the SP-D-receptor, OSCAR, which induces secretion of TNF-α by CCR2 + inflammatory monocytes exposed to SP-D, was localized to the tunica intima and tunica media of clinical atherosclerotic plaques (48, 133). Arterial SP-D expression is documented both in endothelial cells (48) and coronary smooth muscle cells (47), where the pattern of expression regulation implicates SP-D in modulation of inflammation, and TNF-α induces SP-D expression in both cell types (47, 63). Direct interaction between LMW, but not HMW, SP-D and HDL, LDL, and oxLDL was demonstrated using ELISA-based assays (36). These observations suggest that the dominant effects of SP-D in the vasculature may not be the multivalent interactions with microbes, but rather may include specialized LMW SP-D-mediated effects on lipid metabolism.

The SP-D receptor, OSCAR, is expressed by cell types relevant for atherogenesis, including monocytic cells (133), neutrophils (453), dendritic cells (454, 455), and endothelial cells (456). Sftpd deficiency is shown to be protective in a diet-induced atherosclerosis model in C57BL/6 mice. Endogenous SP-D can influence systemic plasma lipids and inflammation, reducing HDL cholesterol and triglyceride levels while increasing those of TNF-α in mice maintained long term on a cholate-rich diet. Intravenous treatment with rfhSP-D for 5 days resulted in reduction in HDL cholesterol, but also, in parallel, in LDL cholesterol and total cholesterol, while there was no effect on triglycerides. The role of SP-D-mediated modulation of plasma lipids in atherogenesis is thus uncertain. However, the long-term effects of these events were not studied (48). A subsequent study using double-deficient apolipoprotein E (Apoe)^−/−Sftpd^−/− mice showed similar quantitative effects in decreasing lesion size. Moreover, Sftpd deficiency resulted in less macrophage accumulation, and more smooth muscle cell coverage of lesions (457). In that study, endogenous SP-D reduced several plasma lipid fractions, including total cholesterol and induced circulatory IL-6, rather than TNF-α. In contrast to the two previous studies, a study using double-deficient LDL-receptor (Ldlr)^−/−Sftpd^−/− mice did not provide clear evidence for a role of SP-D in atherogenesis (Sorensen, unpublished data).

Bridges et al. (410) reported that SP-D was a potent inhibitor of oxidation of LDL to oxLDL in in vitro assays, but later ascribed the majority of the effects to residual EDTA in the SP-D vehicle in a published erratum. However, the consequences of Sftpd deficiency in atherogenesis may also result from the antioxidant effects of SP-D and the effects on the formation of ROS and in NO-mediated signaling, although these aspects have not been investigated in detail in the vascular system.

A potential link between low SP-D levels and the sequela of atherosclerosis was recently demonstrated using cerebrospinal fluid samples from patients with cerebral infarction (41). Almost simultaneously, a mouse model of cerebral ischemia was analyzed, but showed no effect of Sftpd deficiency; however, that model was based on permanent occlusion of the middle cerebral artery and did not involve atherosclerosis (49).

The main reported individual steps leading to SP-D-enhanced experimental outcomes relevant for atherosclerosis are illustrated in Figure 5.

In addition to the association of SP-D with CVD, possible causes of the association between circulating SP-D and all-cause mortality may include the role of SP-D in metabolic homeostasis. As described earlier, the brain, pancreas, gut, and endothelium express SP-D, and these locations suggest potential effects in systemic metabolism. Diabetes and metabolic syndrome are frequent comorbidities in COPD patients (459–462). COPD, diabetes, and the metabolic syndrome are all related to systemic inflammation, which may explain their cooccurrence (463). Moreover, pulmonary function impairment is well recognized in diabetes and may be attributable to increased pulmonary bacterial colonization due to hyperglycemia and exacerbation of COPD, among the additional causes recently discussed by Kinney et al. (464). By contrast, a reduced ability to eradicate pathogens may cause frequent respiratory tract infections, and chronic inflammation resulting in insulin resistance and type 2 diabetes, as reviewed by Fernández-Real and Pickup (465). Moreover, cigarette smoking is an independent predictor of type 2 diabetes, although the link between lower vital capacity and diabetes risk is independent of cigarette exposure and stronger in never-smokers (466).

Obesity is the most frequent cause of insulin resistance (467) and, although impaired lung function is present before the development of metabolic syndrome or diabetes, the associations between pulmonary and metabolic impairment may, at least partly, result from the effects of obesity, as recently shown in the longitudinal Strong Heart Study (468). Multiple cellular and molecular mechanisms of the innate immune system may contribute to this relationship, as reviewed by Lackey and Olefsky (469). Several studies have independently identified positive associations between lung function impairment and full metabolic syndrome or obesity (470–473), and metabolic biomarkers are predictors of impaired lung function in response to particulate inhalation (474).

Overall, the close relationship between metabolic and respiratory impairment suggests that lung innate immunity, as inferred from genetic or phenotypic SP-D variation, could reside at the intersect of inflammation, obesity, and insulin resistance.

Multivariate genetic analysis was used to investigate genetic associations of serum SP-D with factors involved in metabolic syndrome in a twin study (GEMINAKAR). The study found that genetic correlations between traits differed, but were close to zero for SP-D versus body mass index (BMI), which may indicate that the common genetic contribution to the regulation of the traits (i.e., pleiotropy) is significant but limited. There was no significant genetic correlation between serum SP-D and HDL cholesterol or total cholesterol, and the study did not investigate possible genetic associations between SP-D and diabetes (208). A subsequent study by Ortega et al. (51) demonstrated that a common SNP variant in the SFTPD promotor, which is positively associated with circulating SP-D levels, was enriched in individuals with lower BMI and decreased percentage fat mass. A study by Pueyo et al. (229) showed that the presence of the variant SP-D allele, Met11, increased the odds of developing insulin resistance and type 2 diabetes, especially among women. Moreover, this SNP, and others in the SFTPD genomic region, were significantly associated with glucose homeostasis, insulin sensitivity, and type 2 diabetes, according to GWA datasets, and these associations were independent of circulating SP-D concentrations (229). A borderline significant inverse association was observed between free LPS concentration in plasma and circulating SP-D in subjects homozygous for Thr11, but not those homozygous for Met11, in a small subpopulation of the study. This observation indicates that the different genotypes may predispose to differing buffering capacity in metabolic endotoxemia (229), consistent with observations by Leth-Larsen et al. (37), showing that Thr11 SP-D bound preferentially to LPS compared with Met11 SP-D (37).

Fernández-Real et al. (427) evaluated serum SP-D in four different cohorts and found that SP-D concentrations were significantly decreased in non-obese populations with type 2 diabetes, and that they were correlated with several metabolic variables, including BMI, and fasting and post-load glucose levels. Obese subjects showed significantly decreased serum SP-D concentrations, and weight loss led to further significantly decreased serum SP-D concentrations, while serum SP-D concentrations were not significantly associated with lung function parameters in these studies. When multiple linear regressions were performed using data from subjects with altered glucose tolerance, fasting glucose and fasting triglycerides were identified as independently contributing to 33% of SP-D variance (427). However, a recent study by López-Cano et al. (475) demonstrated increased serum SP-D in diabetes in obese patients with type 2 diabetes. Moreover, serum SP-D in this population was independently and inversely associated with FEV₁ in the obese diabetes patients and the authors concluded serum SP-D might be a useful biomarker for detecting lung impairment in obese patients with type 2 diabetes.

The findings of Ortega et al. (51) supported previous observations that SFTPD was expressed in human adipose tissue and adipocytes, and this expression was decreased in both omental and subcutaneous adipose tissue from obese subjects with and without type 2 diabetes, compared with the control group. Omental SP-D expression was inversely associated with several metabolic variables, including BMI, percentage fat mass, waist circumference, and fasting glucose; and positively associated with the expression of insulin receptor substrate 1 and fatty acid synthase, suggesting that SP-D may effect glucose tolerance in fat tissue or vice versa.

At least two previous studies demonstrated decreased constitutional circulating SP-D in obese or overweight subjects after correction for gender, smoking, and age, which may support an initial influence of the processes leading to accumulation of fat tissues on the correlations between SP-D and metabolic impairment. As a part of the Danish population-based twin study on metabolic syndrome, GEMINAKAR, serum SP-D was identified as significantly and inversely associated with weight and waist circumference in men, and with BMI in both sexes (476). Similar findings were obtained from a Chinese population-based cohort (271). Moreover, the relationship was also observed in patients who were obese and had ALI (477).

Three independent studies have demonstrated a link between Sftpd deficiency and obesity (457, 476, 478). According to a study by Stidsen et al. (478), body mass development in Sftpd-deficient mice was significantly increased compared with that of corresponding wild-type mice receiving both ad libitum and fixed energy intake diets. The study also identified a significant redistribution of body fat in parallel to obesity-related insulin resistance, as well as increased metabolic endotoxemia associated with the Sftpd deficiency in this model (478). Relative insulin resistance in Sftpd-deficient mice was later validated by Hirano et al. (457). Additional links between SP-D and body fat mass include the observation of decreased SP-D expression in human adipose tissue with increasing BMI (51), and direct SP-D interactions with fatty acids (479).

Furthermore, SP-D was linked to metabolic homeostasis in two independent studies seeking to identify genes involved in β cell adaptation during mouse pregnancy or in rat perinatal regenerating pancreas. The studies demonstrated that Sftpd mRNA and protein levels in β cells exhibited large but transient perinatal increases (480, 481). TNF-α- and IL-1β-induced expression of SP-D in islets was demonstrated (480), suggesting that TNF-α induction of SP-D is a general effect for many cell types.

No direct role of SP-D in experimental diabetes development has been demonstrated; however, the development of diabetes in mother rats is accompanied by delayed pulmonary histogenesis and decreased SP-D expression in the pups (482). Moreover, viral neutralization by SP-D in vitro was abolished in the presence of glucose at levels commonly found in diabetic mice (483), which may contribute to reduced anti-viral protection in diabetes.

Important information has been obtained about SP-D variation in premature infants, including the normal range of SP-D in term and preterm newborns and the influence of perinatal factors; however, there is a paucity of independent studies validating the observed association between the highest serum SP-D levels measured day 1 after birth and the most severe outcomes in premature infants, which warrants further validation.

Recent research findings support increased circulating SP-D in moderate or severe asthma and in addition to allergic rhinitis. Yet, SP-D breakdown products were also enriched in the blood system in severe asthma, indicating that such fragments may provide a more specific measure of disease activity.

Although SP-D levels are highly increased in response to smoking, the highest circulating SP-D levels appear to be not only proxies for smoking but also related to disease outcomes after adjustment for smoking, hence providing a robust marker for treatment effects. Association with mortality has been identified in CVD in cohorts selected, and not selected, for COPD. The presence of CVD is thus anticipated to be a significant contributor to induction of circulating SP-D levels in COPD. The authors of the ECLIPSE study stressed that division of the COPD population into more homogenous subgroups with similar prognoses provides the potential to test interventions. That study further underscored that monitoring of biomarker arrays including SP-D is useful as an indicator of disease prognosis, or for prediction and monitoring of clinical responses to an intervention; however, this strategy is yet to be validated in randomized clinical trials.

In general, there appears to be an untapped potential for monitoring circulating LMW SP-D or SP-D neoepitopes, with the aim of separating disease cases from controls, as well as for use in disease prognosis. Specific proteases are known to degrade SP-D, and this aspect has not yet been pursued sufficiently to identify disease-induced SP-D neoepitopes. More elaborate immunoassay-based methods, which have been previously used to detect circulating extracellular matrix epitopes, are warranted for studies of SP-D breakdown product variation in respiratory, as well as non-respiratory, diseases and experimental data support the presence of such circulating breakdown products. This approach has the potential to provide measures of disease and organ specific SP-D breakdown products.

It remains to be determined whether those subjects presenting with the highest circulating SP-D levels are prone to the development of clinical disease due to the presence of preclinical lung injury or CVD and if monitoring of SP-D may facilitate early intervention.

Circulating SP-D is confounded by adiposity and diabetes, and experimental and genetic links have been established to these disorders; however, these associations have not been explained mechanistically. The correlation between circulating SP-D and adiposity is inverse and may, therefore, counteract phenotypic associations with both respiratory disease and CVD.