Abstract
Immunoreactive surfactant protein-D (SP-D) was assessed in human fetal, newborn, and adult tissues. In the fetal lung, SP-D was detected on airway surfaces by 10 weeks' gestation, staining increasing in the distal airways, decreasing in the proximal conducting airways with advancing gestation. In lungs from near-term infants and adults, SP-D was detected in Type II cells, serous cells of tracheobronchial glands, and subsets of cells lining peripheral airways. Immunostaining was decreased or absent in areas of lungs of neonates after injury to Type II cells, infection, or hemorrhage and was decreased in collapsed or unseptated airways from older infants with bronchopulmonary dysplasia. SP-D was also detected in many organs at all ages. SP-D was readily detected in epithelial cells and luminal material in lacrimal glands, salivary glands, pancreas, bile ducts, renal tubules, esophageal muscle and glands, parietal cells of the stomach, crypts of Lieberkuhn, sebaceous and eccrine sweat glands, Von Ebner's glands, endocervical glands, seminal vesicles, adrenal cortex, myocardium, and anterior pituitary gland. SP-D is a widely distributed member of the “collectin” family of polypeptides secreted onto luminal surfaces by epithelial cells lining ducts of many organs, where it likely plays a role in innate host defense.
SP-D, also termed collectin-7 (col-7), is a 43-kD member of the C-type lectin family of polypeptides, sharing considerable homology with other mammalian lectins, including SP-A, conglutinin, and mannose-binding protein (McCormack and Whitsett 1996; see Wright 1997 for review; Crouch and Wright 2001). SP-D binds carbohydrates and lipids in a calcium-dependent manner and probably plays a role in innate host defense against various bacterial, fungal, and viral pathogens. SP-D interacts with a number of microorganisms, including influenza A virus (Hartshorn et al. 1998), Cryptococcus neoformans, Aspergillus, and various bacteria, including Mycobacterium tuberculosis, Salmonella, Klebsiella pneumoniae, Pseudomonas aeruginosa, and Escherichia coli (Sastry and Ezekowitz 1993; Crouch et al. 1995; Pikaar et al. 1995; McCormack and Whitsett 1996; Wright 1997; Ferguson et al. 1999; Restrepo et al. 1999). SP-D binds to the core antigens in LPS in a manner distinct from SP-A binding. Targeted deletion of murine SP-D (Botas et al. 1998; Korfhagen et al. 1998), demonstrated its important role in lipid homeostasis in the lung. SP-D-null mice developed alveolar lipidosis associated with activation of alveolar macrophages, increased oxidant production, and metalloproteinase activation (Wert et al. 2000). SP-D-deficient mice developed severe emphysema and lung fibrosis, demonstrating a critical role of SP-D in the regulation of pulmonary inflammation and remodeling (Wert et al. 2000). SP-D-deficient mice were also highly susceptible to infection by influenza A virus in vivo (LeVine et al. 2001). Although most studies regarding the function of SP-D have focused on its potential role in the lung, SP-D mRNA was detected in various organs, including the stomach, kidney, and mesentery (Fisher and Mason 1995; Motwani et al. 1995; Chailley-Heu et al. 1997). Recent studies demonstrated widespread distribution of SP-D and SP-D mRNA in adult human tissue. (Madsen et al. 2000). Cellular sites of SP-D proteins in a variety of human tissues at various times in development have not been described. The present study utilized a highly specific SP-D antiserum to localize SP-D in human tissues.
Materials and Methods
Tissue Preparation
This was a retrospective study that was approved by the Committee for the Protection of Human Subjects, Vanderbilt University Medical Center. Fetal tissues were obtained immediately after hysterotomy, hysterectomy, or spontaneous abortion. Infant samples were obtained at surgical biopsy, or by postmortem sampling within 2 hr of death. Child and adult samples were obtained postmortem. Tissues were fixed in 10% phosphate-buffered formalin, dehydrated through graded ethanols, and embedded in paraffin. Four-μm-thick serial sections were cut and mounted separately on Superfrost Plus (Fisher; Atlanta, GA) glass sides.
Purification of Mouse SP-D and Production of Antiserum
Lung lavage fluid from granulocyte-macrophage colony-stimulating factor (GM-CSF) and SP-A double-null mutant mice was used as a source of SP-D (Dranoff et al. 1994; Korfhagen et al. 1996). A modification of the methods of Strong et al. (1998) was used to prepare the purified protein. Briefly, the pH of the mouse lung lavage fluid was adjusted with Tris-HCl (pH 7.4) and made to 10 mM EDTA, stirred for an hour at room temperature (RT), and centrifuged at 10,000 × g to separate the lipid-rich pellet from the supernatant containing most of the SP-D. The supernatant was made 20 mM with respect to CaCl2, readjusted to pH 7.4, and passed through a maltosyl-agarose column. The column was washed to background absorbance and then washed with 1.0 M NaCl in Tris-HCl (pH 7.4), 10 mM EDTA. The SP-D was eluted off the column by the addition of 50 mM MnCl2 and the SP-D-containing fractions were identified by SDS-PAGE. The fractions containing SP-D were dialyzed and the protein concentration was determined. New Zealand White rabbits were injected three times at 2-week intervals with 250 μg of mouse SP-D in incomplete Freund's adjuvant.
Purification of Antiserum
Lungs were removed from SP-D-null mutant mice (Korfhagen et al. 1998) and homogenized in PBS. The homogenate was centrifuged to remove the large particles and the supernatant was covalently linked to Affi-gel (BioRad Laboratories; Hercules, CA) after dialysis with 50 mM HEPES buffer, pH 7.8. Lung homogenate crosslinked to Affi-gel beads was mixed with the SP-D antiserum overnight at 4C. The adsorbed antiserum was separated from the beads by centrifugation. The adsorbed rabbit anti-SP-D antiserum detected SP-D from mouse, sheep, and human lavage fluid as assessed by Western blotting (Figure 1A and not shown). The antiserum did not react with tissue or lung lavage from the SP-D —/— mouse as assessed by immunochemistry or Western blotting analysis, respectively. The SP-D antiserum stained mouse lung and pancreas from wild-type mice but did not stain these tissues for the SP-D —/— mouse.
Specificity of SP-D ELISA
Microtiter plates were coated with 0.2 μg material from alveolar proteinosis of mouse SP-D overnight. Human SP-A or mouse SP-D (0.001–1 μg/ml) was incubated overnight with rabbit anti-mouse SP-D (1:20,000). Antibody-antigen mixtures were incubated in the SP-D-coated microtiter plate for 2 hr. After washing, goat anti-rabbit IgG-peroxidase conjugate (1:1000) was added to the microtiter plate for 1 hr. After washing, the color was visualized using ortho-phenylenediamine in the presence of hydrogen peroxide and the absorbance was determined at 490 nm. The absorbance was graphed vs the concentration of SP-A or SP-D (Figure 1B).
Immunolocalization
Slides were deparaffinized by placing in three xylene baths, 10 min each, and then rehydrated through two baths of ethanol (5 min each in 100%). Endogenous peroxide was quenched for 20 minutes in a bath of 0.3% H2O2 in methanol followed by 95% and then 80% ethanol 5 min each and finally placed in PBS, pH 7.2. Nonspecific staining was blocked by exposing slides for 30 minutes to prediluted normal goat serum (BioGenex catalog HK112–9K; San Ramon, CA). The appropriately diluted primary antibody (1:5000 for SPD) was then applied and allowed to incubate overnight at 4C. Slides were washed three times in PBS, pH 7.2, 5 minutes each and then a prediluted biotinylated goat anti-rabbit immunoglobulin reagent was applied to the slides for 30 min (BioGenex SS Kit; catalog AP500–5R).
(
After washing again, a prediluted peroxidase-conjugated streptavidin (BioGenex kit) was applied for 30 min. After more washes, the peroxidase activity was then localized by reaction with 0.5% 3,3′-diaminobenzidine-0.01% H2O2 and counterstained with hematoxylin.
Results
Characterization of the SP-D Antiserum
To minimize crossreactivity with SP-A, the initial SP-D antiserum was prepared against mouse SP-D purified from a double transgenic mouse lacking both SP-A (Korfhagen et al. 1996) and GM-CSF (Dranoff et al. 1994). SP-D accumulates to high concentrations in these mice and the surfactant does not contain SP-A. The initial antiserum reacted against monomeric SP-D (Mr 43,000) but also reacted with a larger ogliomeric form of SP-D (Mr 90,000) and an unknown protein of Mr 60,000. Adsorption of the antiserum against lung homogenates from the SP-D —/— mouse removed all reactivity with the Mr 60,000 protein but left strong reactivity against SP-D (Figure 1A). An ELISA was established demonstrating specificity of the antiserum for human SP-D and negligible reactivity with human SP-A isolated from alveolar proteinosis fluid (Figure 1B).
Distribution of SP-D in Adult and Term Newborn Lung
SP-D immunoreactivity was initially assessed in adult human lung, demonstrating diffuse intracellular staining in Type II epithelial cells. Characteristic extracellular luminal staining was observed along the surfaces of the airspaces (Figure 2). It was also demonstrated around lamellar bodies in Type II cells of near-term infants, as well as rimming their terminal air spaces (Figure 3).
SP-D in Non-pulmonary Tissues
SP-D was readily detected in a variety of non-pulmonary tissues from fetuses, children, and adults. The presence of SP-D staining in non-pulmonary tissues was distinct from our previous findings demonstrating the lack of staining for SP-A in those tissues. Although SP-D reacted strongly in eccrine sweat glands and gastric fundus, no staining was seen in the same samples with anti SP-A antiserum (data not shown). In initial studies, intense SP-D staining was noted in the intercalated ducts of the exocrine pancreas and in the lumen of larger ducts (Figure 4). To further assess the specificity of reactivity in non-pulmonary tissue, reactivity was assessed in the presence of excess purified mouse SP-D. Co-incubation with exogenous SP-D ablated immunoreactivity against pancreatic tissue (Figure 5).
SP-D was detected in epithelial cells and in luminal material of the ducts of many adult tissues, including lacrimal glands, salivary glands, intercalated ducts of the pancreas, hepatocytes and intra- and extrahepatic bile ducts, esophageal glands, breast, sebaceous and eccrine sweat glands of the skin, and Von Ebner's glands of the tongue. Parietal cells of the stomach, proximal and distal renal tubules, and podocytes of the glomeruli were immunostained. In the reproductive tract, seminal vesicles and endocervical glands stained. SP-D was also observed in endocrine tissues, including adrenal cortex, and in follicular stellate cells of the anterior pituitary gland. Myocardial cells from the right atrium also stained intensely for SP-D (Figures 6–14). Consistent with widespread distribution of staining, SP-D was detected in human amniotic fluid and saliva but was not detected by ELISA in seminal fluid, urine, breast milk, or tears.
SP-D in Fetal Tissue and Neonatal Lung
SP-D was detected in lungs of all fetuses and newborns. In 10–20-week fetuses, all open airways were rimmed with SP-D (Figures 15 and 16). Thereafter, bronchioles were lightly rimmed or unstained. With advancing gestation there was a distal-proximal gradient of SP-D staining, terminal airways being most heavily stained. Staining was less intense in larger bronchioles. Tracheas were rimmed with SP-D up to 15 weeks, and some term fetal tracheas showed staining of basal and intermediate cells (Figure 19). Serous cells of tracheal glands were stained from 18 weeks onward (Figure 17). In the near-term infants, SP-D was detected in the cytoplasm of Type II cells (Figure 18), and airways were rimmed with stain (Figure 3). In lungs from infants with hyaline membrane disease (HMD) and bronchopulmonary dysplasia (BPD), only open terminal airways were rimmed with SP-D. Injured areas lined with hyaline membranes, or alveoli filled with hemorrhage, infection, or edema fluid, were lightly stained or unstained (Figures 20 and 21) SP-D was not detected in bronchioles and bronchi from these infants. However, serous cells of bronchial and tracheal glands were consistently well stained, particularly in infants with lung inflammation. In infants, both layers of striated muscle surrounding the upper esophagus were immunostained but staining was not uniform in distribution (Figure 14). Lungs from older infants with BPD were lightly stained around dilated, unseptated open airways, as were lungs of older infants without a history of HMD whose lungs remained unseptated (Figures 22 and 23).
As seen in adult tissues, pancreas, stomach, and duodenum were stained in fetuses and newborns, as were distal convoluted kidney tubules, seminal vesicles and anterior pituitary gland (Table 1).
A number of tissues tested remained unstained (Table 2), and all tissues showed some variability in the degree of staining, not related to age but most probably to the patients' illness and its inflammatory response in various tissues.
Discussion
SP-D was readily detected in fetal and postnatal human lung as intracellular staining in Type II epithelial cells, and in serous cells of tracheobronchial glands. Staining was prominent along luminal surfaces of terminal airspaces. Intense and specific SP-D staining was also observed in duct cells and in open ducts in various organs and glands in fetuses, infants, and adults. In infants, SP-D was decreased in pulmonary tissues in areas of acute and chronic injury or collapse of alveoli. The abundance of SP-D in luminal material in various organs, including lung, liver, pancreas, lacrimal and salivary glands, tongue, and cervix, supports its proposed role in innate defense at potential sites of invasion by pathogens. The expression of SP-D at unexpected sites, including myocardium, and in the striated muscle surrounding the esophagus supports a potential anti-inflammatory or host defense role for SP-D in non-epithelial organs.
SP-D in the Developing Lung
Early in gestation, SP-D was detected in open airways throughout the lung. As gestation progressed, SP-D staining increased in the lung periphery, with terminal airways being stained most heavily near term and thereafter. Tracheobronchial glands were stained from 18 weeks' gestation and thereafter. In late gestation, fetal lung, and adult lung, SP-D was detected primarily in Type II epithelial cells, consistent with previous studies in rodents in which SP-D mRNA and protein were detected in alveolar cells in both rat and mouse (Fisher and Mason 1995). In the present study, SP-D staining was also prominent on luminal surfaces in alveolar airspaces. SP-D was not detected in cells of conducting airways of the adult human lung, although luminal staining was present on ciliated surfaces. Basal and intermediate cells of the trachea were occasionally stained in infants, but the findings in humans contrast with those in rodents, in which SP-A and SP-D staining was detected in non-ciliated cells of conducting airways (Crouch et al. 1992; Voorhout et al. 1992; Wong et al. 1996).
In the human lung, serous cells and luminal contents of tracheobronchial glands were immunostained for SP-D, suggesting that the primary source of SP-D in the conducting regions of the human lung may be derived from secretions of tracheobronchial glands. This finding may provide a basis for the recently demonstrated lack of SP-D in lung washings from patients with cystic fibrosis (Postle et al. 1999). If tracheobronchial gland secretions are the primary source of SP-D in the conducting airways and these glands are destroyed or impacted with mucus, SP-D may have limited access to airway surfaces in cystic fibrosis. The finding that SP-D was present on luminal surfaces of the lung supports its proposed role in innate defense. SP-D binds to bacterial lipopolysaccharide as well as to carbohydrate-rich surfaces of many bacteria, viruses, fungi, and allergens. Therefore, SP-D is constitutively present on airway surfaces and may play a role in the initial recognition of respired particles and pathogens.
Normal adult lung immunostained for SP-D, showing staining of Type II cells and light rimming of alveolar walls. Peroxidase-anti-peroxidase and hematoxylin. Bar = 6.4 μm.
Lungs of a near-term infant showing staining of Type II cells around lamellar bodies (arrow) and rimming of alveolar surfaces with stain. Peroxidase-anti-peroxidase and hematoxylin. Bar = 6.4 μm.
Adult pancreas immunostained for SP-D, showing intense staining of the cells emptying into the intercalated ducts of the pancreatic acini and staining of the luminal contents. Islets are unstained. Peroxidase-anti-peroxidase and hematoxylin. Bar = 16 μm.
Adult pancreas immunostained for SP-D, showing ablation of stains when tissue was incubated with SP-D. Perioxdase-anti-peroxidase and hematoxylin. Bar = 32 μm.
Adult submandibular gland immunostained for SP-D, showing heavy staining of intra-acinar ducts and of material secreted into the lumen of these ducts. Peroxidase-anti-peroxidase and hematoxylin. Bar = 16 μm.
Cross-section of the fundal glands of the adult stomach immunostaining for SP-D, showing intense staining of parietal cells. Peroxidase-anti-peroxidase and hematoxylin. Bar = 16 μm.
Immunostaining for SP-D in adult kidney, showing staining of lining cells of proximal and distal convoluted tubules. Peroxidase-anti-peroxidase and hematoxylin. Bar = 32 μm.
Immunostaining of the podocytes of a glomerulus from a 4-month-old infant. Peroxidase-anti-peroxidase and hematoxylin. Bar = 16 μm.
Pulmonary findings in SP-D —/— null mice support a critical and unexpected role for SP-D in the regulation of oxidant generation, metalloproteinase activation, and cytokine responses in alveolar macrophages, effects that are independent of exposure to pathogens (Wert et al. 2000). SP-D —/— mice develop severe postnatal emphysema, associated with foamy macrophages, and increased metalloproteinase activity, likely related to findings associated with the intrinsic activation of alveolar macrophages by oxidants in the absence of SP-D. Recent studies demonstrated the spontaneous activations of NF-κB in alveolar macrophages from SP-D-deficient mice (Yoshida et al. 2001). These findings support the concept that SP-D plays an important modulatory role in oxidant production and inflammation in the lung (Wert et al. 2000). Recent studies also support the concept that SP-D has a direct antioxidant role in the lung (Bridges et al. 2000). Although similar pathological findings were not observed in other tissues of SP-D-null mice, it remains possible that SP-D plays a similar role in control of innate defense or regulation of inflammation in other tissues that immunostain for SP-D.
The distribution of SP-D immunostaining seen in various tissues is consistent with that of SP-D mRNA seen in other species (Fisher and Mason 1995; Chailley-Heu et al. 1997; Madsen et al. 2000) and with recent immunohistochemical and RNA studies in human tissues (Madsen et al. 2000). The observed widespread distribution of SP-D is distinct from that of SP-A, which is confined to pulmonary tissue in the human. The antibody used in the present study was generated against surfactant from SP-A-null mice and was adsorbed extensively to lung tissue from SP-D gene-targeted mice to remove activity against crossreacting antigens, including SP-A. The ELISA established with the antibody is highly selective for human SP-D compared to SP-A. The distinct manner of staining between SP-A and SP-D in human tissue, the specificity of the ELISA and inhibition of non-pulmonary SP-D staining with blocking experiments with SP-D, and a lack of reactivity of the antiserum in tissues from the SP-D of the mouse supports the specificity of the immunoreactivity at physiological concentrations. It remains possible that the antiserum recognizes other related epitopes or other lectins, including SP-A, at higher concentrations. The distribution of staining with the anti-SP-D antibody is also consistent with recent studies regarding the distribution of SP-D mRNA and immunostaining for SP-D seen in various adult tissues (Madsen et al. 2000).
The abundance of SP-D on luminal surfaces is consistent with its role in innate defense against pathogens at sites of entry into various organs. The finding that SP-D was expressed at high concentrations in secretory cells and in luminal contents of secretory ducts of liver, pancreas, salivary glands, glands of the skin, and other tissues would provide concentrations of this collectin at sites of potential invasion by microorganisms. The paucity of SP-D staining of lungs of infants with acute HMD and in advancing BPD in areas with alveolar hemorrhage, infection, edema, or alveolar collapse, where Type II cell function may be compromised, may render the lung susceptible to secondary infection and inflammation. The abundant expression of SP-D in tissues such as myocardium, anterior pituitary, and esophageal muscle however, is more difficult to explain on the basis of innate host defense function for SP-D. The finding that SP-D regulates NF-κB translocation, oxidant production, metalloproteinase activation, and cytokine responses in the lung may indicate that similar roles are played by SP-D in other tissues, in which SP-D may regulate inflammation and remodeling involved in the pathogenesis of various acute and chronic disorders.
Adult sebaceous gland, showing widespread immunostaining for SP-D around the oil droplets. Dermis is unstained. Peroxidase-anti-peroxidase and hematoxylin. Bar = 32 μm.
Cervical glands of the adult uterus immunostained for SP-D, showing rimming of the lumens of glands with stain. Peroxidase-anti-peroxidase and hematoxylin. Bar = 64 μm.
Fetal anterior pituitary gland, showing intense immunostaining for SP-D of the follicular stellate cells and of the acinar contents. Bar = 16 μm.
Adult cardiac muscle of the right atrium immunostained for SP-D, showing intense staining at focal sites in the muscle (arrow), with adjacent more lightly stained areas. Peroxidase-anti-peroxidase and hematoxylin. Bar = 32 μm.
Cross-section of upper esophageal striated muscle from a 23-week fetus, showing the vesicular appearance of the cells intensely immunostained for SP-D. Peroxidase-anti-peroxidase and hematoxylin. Bar = 6.4 μm.
Lung of a 10-week fetus immunostained for SP-D, showing conducting and terminal airways rimmed by staining. Only the terminal buds that have no lumen are unstained. Peroxidase-anti-peroxidase and hematoxylin. Bar = 6.4 μm.
Lung of a 23-week fetus immunostained for SP-D, showing rimming of all terminal airways but little or no staining of a conducting airway. Peroxidase-anti-peroxidase and hematoxylin. Bar = 32 μm.
Bronchial glands of a 23-week fetus immunostained for SP-D, showing heavy staining of serous cells of glands. Peroxidase-anti-peroxidase and hematoxylin. Bar = 32 μm.
Lung of a normal 36-week infant immunostained for SP-D, showing widespread rimming of open alveolar luminal surfaces. Per-oxidase-anti-peroxidase and hematoxylin. Bar = 32 μm.
Trachea of an infected term infant, showing staining of basal and intermediate cells and rimming of gland lumens. Peroxidase-anti-peroxidase and hematoxylin. Bar = 32 μm.
Lung of a 25-week infant who survived 4 hr with HMD and infection. SP-D was detected in a few cuboidal cell-lined terminal airways that were rimmed with stain. Proximal airways with sloughed cell membranes and polymorphonuclear leukocytes in the lumen were weakly stained or unstained. Peroxidase-anti-peroxidase and hematoxylin. Bar = 16 μm.
Lung of a 36-week infant with severe HMD who died at 15 days. Little or no SP-D staining was detected in terminal airways filled with red blood cells (arrowhead), but open airways were rimmed with stain (arrow). Peroxidase-anti-peroxidase and hematoxylin. Bar = 32 μm.
Lung of a 27-week infant who survived for 28 days with HMD and BPD. Terminal airways are collapsed and more proximal airways are dilated, unseptated, and very lightly rimmed with SP-D staining. Peroxidase-anti-peroxidase and hematoxylin. Bar = 64 μm.
Lung from a 28-week gestation infant who survived 42 days without HMD. The terminal airways are large, open, unseptated, and faintly rimmed with SP-D immunostaining. Peroxidase-anti-peroxidase and hematoxylin. Bar = 64 μm.
Tissues positively stained for SP-D a
aFigures represent number of tissues reviewed.
In the present study, considerable variability in SP-D staining was seen in pathological samples from the lungs and other tissues obtained postmortem. It is unlikely that this variability relates to sample collection or fixation because tissues were obtained and fixed rapidly after death, especially in fetuses and newborns. The variability in SP-D staining is likely related to the biological differences in this heterogeneous population of infants. The heterogeneity of SP-D staining is probably influenced by developmental, spatial, and inflammatory stimuli that influence SP-D gene expression or stability. In vivo studies support the concept that SP-D gene expression is strongly increased by endotracheal endotoxin (McIntosh et al. 1996), hyperoxia (Aderibigbe et al. 1999), and fibroblast growth factor-7 (Xu et al. 1998). Regional differences and variability among individual infants in the intensity of SP-D staining likely reflect the roles of oxygen therapy, injury, and/or infection that have influenced the clinical course of these infants.
Tissues unstained with SP-D a
aFigures represent the number of tissues reviewed.
Footnotes
Acknowledgements
Supported by HL56387 and HL61646.
We wish to thank Ms Ann Maher, Ms Robin Roller, and Mr Jeffrey Phillips for secretarial help, Mr Brent Weedman for photography, Mr Terry Johnson for preparation of illustrations, and Ms Sandra Olson for immunohistochemistry.