Surfactant protein A, an innate immune factor, is expressed in the vaginal mucosa and is present in vaginal lavage fluid

Authors


Colin MacNeill, Department of Obstetrics and Gynecology, Pennsylvania State University College of Medicine, 500 University Drive, Hershey, PA 17033, USA. E-mail: cmacneill@psu.edu

Summary

Surfactant protein A (SP-A), first identified as a component of the lung surfactant system, is now recognized to be an important contributor to host defence mechanisms. SP-A can facilitate phagocytosis by opsonizing bacteria, fungi and viruses, stimulate the oxidative burst by phagocytes and modulate pro-inflammatory cytokine production by phagocytic cells. SP-A can also provide a link between innate and adaptive immune responses by promoting differentiation and chemotaxis of dendritic cells. Because of the obvious relevance of these mechanisms to the host defence and ‘gate keeping’ functions of the lower genital tract, we examined human vaginal mucosa for SP-A protein and transcripts and analysed vaginal lavage fluid for SP-A. By immunocytochemistry, SP-A was identified in two layers of the vaginal epithelium: the deep intermediate layer (the site of newly differentiated epithelial cells); and the superficial layer (comprising dead epithelial cells), where SP-A is probably extracellular and associated with a glycocalyx. Transcripts of SP-A were identified by Northern blot analysis in RNA isolated from vaginal wall and shown, by sequencing of reverse transcription–polymerase chain reaction products, to be derived from each of the two closely related SP-A genes, SP-A1 and SP-A2. SP-A was identified in vaginal lavage fluid by two-dimensional gel electrophoresis, and confirmed by mass spectrometry. This study provides evidence, for the first time, that SP-A is produced in a squamous epithelium, namely the vaginal mucosa, and has a localization that would allow it to contribute to both the innate and adaptive immune response. The findings support the hypothesis that in the vagina, as in lung, SP-A is an essential component of the host-defence system. A corollary hypothesis is that qualitative and quantitative alterations of normal SP-A may play a role in the pathogenesis of lower genital tract inflammatory conditions.

Abbreviations
BCIP

bromo-chloro-indolyl-phosphate

ICC

immunocytochemistry

MALDI-TOF

matrix-assisted laser desorption ionization–time of flight

MS

mass spectrometry

NBT

nitroblue tetrazolium

RT–PCR

reverse transcription–polymerase chain reaction

SP-A

surfactant protein A.

Introduction

Surfactant protein A (SP-A) is a member of the collectin (collagenous lectin) family of proteins that has important host defence functions.1,2 A common feature of these proteins is that they possess, in addition to a collagen-like domain, a carbohydrate-binding lectin domain.1,2 The latter recognizes structural patterns of cell-surface carbohydrates, such as lipopolysacharide and mannan, which are present on many classes of micro-organism.3 This property underlies some of the well-characterized host-defence functions of collectins (including those of SP-A), notably the ability to facilitate phagocytosis of micro-organisms by opsonizing bacteria, fungi and viruses.1,2,4 In addition, SP-A can stimulate chemotaxis and the cytolytic oxidative burst by phagocytes, can act as an antioxidant and is able to modulate pro-inflammatory cytokine production by peripheral blood mononuclear cells, splenocytes, macrophages and macrophage-like cell lines.1,2,4,5 Moreover, SP-A has been shown to modulate the differentiation and chemotaxis of cells of monocytic origin, including dendritic cells, thereby providing a link between innate and adaptive immune responses.6 The functional versatility of SP-A can be attributed to its complex, highly organized quaternary structure, with distinct binding domains for phospholipids, glycolypids, diverse carbohydrates, calcium and lipopolysacharide, in addition to the carbohydrate-binding lectin domain.7 Recent studies of mice that are homozygous null for SP-A attest to the importance of SP-A in modulating the inflammatory response. Although SP-A−/− mice have normal lung compliance, and survive and breed normally under germ-free conditions, under standard housing they display reduced alveolar macrophage binding and uptake of group B streptococci, and die of pneumonia.8,9

Although SP-A was first identified as a component of the pulmonary surfactant complex, there is a large and growing body of evidence that its principal function is probably host defence.1,2,4 Therefore, we reasoned that the vaginal epithelium, as one of the critical interfaces with the external environment, would be a site in which SP-A is probably expressed. The vaginal mucosa must tolerate a range of commensal organisms, yet be able to defend the host against potentially pathogenic organisms.10 Mechanisms regulating vaginal host defences, particularly those involving locally produced factors, are just starting to be identified.10–12 Understanding these mechanisms is essential for developing effective prevention and for treating the consequences of a failure of vaginal host defence. This might include invasion of the host by organisms responsible for conditions ranging from sexually transmitted diseases to spontaneous preterm labour.13,14

Here, we report that SP-A is expressed in the pre- and postmenopausal vaginal mucosa, and present in vaginal lavage fluid. To our knowledge, this is the first study to demostrate the expression of a member of this important class of innate immune factors in a stratified squamous epithelium, specifically the lower genital tract, at a critical interface between host and pathogens.

Materials and methods

Collection of vaginal tissue and fluid

Human vaginal tissue was obtained from 12 pre- and 12 postmenopausal patients undergoing hysterectomy for benign uterine disease or vaginal repair for low-grade cystocele or rectocele. In nine of the premenopausal patients, hysterectomy specimens were available for histological dating of the endometrium to determine the stage of the menstrual cycle. Based on the criteria of Noyes et al.,15 five of the nine premenopausal specimens were from women judged to be in the proliferative phase, two in the mid-cycle, and two in the secretory phase of the cycle. Of the 12 postmenopausal patients, two were prescribed oral conjugated equine oestrogens and two were prescribed topical conjugated equine oestrogens. The other eight did not receive any form of hormonal replacement. Full-thickness fresh surgical specimens, which included epithelium and the underlying lamina propria, were fixed for 12–24 hr in 10% neutral-buffered formalin for immunocytochemistry (ICC) or snap-frozen at −80° for the extraction of RNA.16 The vaginal mucosal fluid phase was sampled from a healthy premenopausal subject by rinsing the vaginal wall repeatedly with 2 ml of normal saline, and stored at −80° until use. The cervical os and coagula of cervical mucus were purposefully avoided in the sampling technique. Tissues were collected according to protocols approved by the Institutional Review Board of the College of Medicine of the Pennsylvania State University.

Analysis of vaginal tissue SP-A protein

Antibody A rabbit polyclonal anti-human SP-A immunoglobulin G (IgG), raised against SP-A from alveolar proteinosis material that was purified by isoelectric focusing, has been previously described.17

Immunocytochemistry Sections (4 µm) from formalin-fixed paraffin-embedded full-thickness vaginal epithelium with underlying lamina propria were processed using low-temperature antigen retrieval for ICC, as described previously.16 The primary antibody was used at dilutions of 1 : 500–1 : 1000. The signal from the biotinylated goat anti-rabbit secondary antibody was amplified using Vectastain ABC reagent (Vector, Burlingame, CA), with alkaline phosphatase as the reporter and Vector Red (Vector) or bromo-chloro-indolyl phosphate (BCIP) and nitroblue tetrazolium (NBT) (Sigma Chemical Co., St Louis, MO) as the substrate and chromogen, respectively. Endogenous alkaline phosphatase was inhibited by preincubating sections with 0·2 N HCl for 5 min. From each tissue block, two adjacent sections were processed for ICC. One was not counterstained, while the other was exposed briefly to haematoxylin, to stain the nuclei. Control sections were incubated with preimmune serum followed by secondary antibody, or with secondary antibody only. Sections were viewed by two of the authors (J.W. and C.M.), first independently and then together. Images were recorded using a Nikon DXM 1200 digital camera under Nomarski optics and using Nikon ACT1, version 2, software.

Analysis of vaginal tissue for SP-A mRNA

In order to confirm that SP-A is expressed in vaginal epithelium, we examined total RNA isolated from vaginal mucosa for the presence of transcripts of SP-A by Northern analysis. In addition, reverse transcription–polymerase chain reaction (RT–PCR) was carried out to determine whether both of the two linked genes encoding SP-A in humans –SP-A1 and SP-A2– are expressed in the human vaginal mucosa.18 While SP-A proteins transcribed from these two genes differ by fewer than 10 amino acid residues, these differences have been shown to affect the host-defence functions in vitro.17

Preparation of RNA Tissues (≈ 200 mg) stored at −80° were homogenized in 2 ml of RNAzol B (Tel.Test, Friendswood, TX) using a glass-Teflon Dounce homogenizer. After adding an equal volume of chloroform, the homogenate was shaken vigorously for 15 seconds, chilled on ice for 5 min and centrifuged (12 000 g, 4°) for 15 min. The aqueous phase was collected and total RNA precipitated with an equal volume of isopropanol.

Northern analysis Twenty micrograms of total RNA from vaginal mucosa or lung (positive control) was separated on a 1% agarose-formaldehyde gel, transferred onto a GeneScreen Plus membrane (NEN, Boston, MA) and immobilized by ultraviolet cross-linking. The membrane was hybridized with 32P-labelled anti-sense SP-A probe [106 counts per minute (c.p.m.)/ml] in ULTRAhyb solution (Ambion, Austin, TX) at 68° overnight, washed according to the manufacturer's instructions and exposed to Kodak XAR 5 film at −80°.

RT–PCR and sequencing Total RNA (100 ng) was incubated with 15 ng of oligo-dT primer at 70° for 10 min and then cooled to room temperature for 15 min. RT was carried out using Maloney murine leukaemia virus reverse transcriptase (Gibco BRL, Gaithersburg, MD), as described previously.19 One microlitre of the reaction product was used as a template in a 50-µl PCR. The primers, 5′-GACGTTTGTGTTGGAAGCCCTGG-3′ (sense) and 5′-GGTACCAGTTGGTGTAGTTCACAG-3′ (antisense), amplify a 578-bp segment spanning SP-A1 and SP-A2 exons one to four. This segment includes sequences that differentiate between transcripts from SP-A1 and SP-A2, the two linked genes that encode SP-A.18 The PCR buffer, primer, dNTP and Taq polymerase concentrations used were those recommended by the manufacturer (Applied Biosystems, Inc., Foster City, CA). Cycle parameters were as follows, for 35 cycles: denaturation at 94° for 2 min, then at 94° for 45 seconds; annealing at 58° for 45 seconds; extension at 72° for 45 seconds. Reaction products were separated by electrophoresis through a 2% agarose gel containing ethidium bromide. Gel bands (≈ 578 bp) from two subjects were excised and the DNA extracted and cloned into plasmid pCRII (Invitrogen, Carlsbad, CA). Eight clones from each of two subjects were sequenced in both directions by the Penn State College of Medicine Biomolecular Core Facility.

Analysis of vaginal lavage fluid for SP-A protein

To determine whether SP-A was secreted into the vaginal fluid phase, lavage fluid was analysed for SP-A by gel electrophoresis; the sequence of SP-A immunoreactive protein isolated from such gels was determined by mass spectrometry (MS).

Two-dimensional (2D) gel electrophoresis and immunoblotting To disrupt aggregates present in the vaginal fluid, Tris–HCl was added to the vaginal lavage to a final concentration of 1 mm, EDTA to 1 mm and Tween-20 to 0·1%. Resulting solutions were cleared of cellular debris by centrifugation at 150 g for 1 min. The supernatant was concentrated 20-fold using an Ultrafree-MC (10 000 molecular weight cut-off) microconcentrator (Millipore Corp., Bedford, MA) and resuspended in a mixture of 2D lysis buffer and sample buffer (3 : 1; Amersham Biosciences, Piscataway, NJ). Isoelectric focusing was performed in duplicate by loading a 50-µl aliquot onto each 11-cm gel strip containing an immobilized pH 3–10 gradient (Immobiline DryStrip; Amersham Biosciences) and run for 9 hr at 300 V, then for 8 hr at 2990 V (30870 V/hr). Proteins were separated in the second dimension on a horizontal 12·5% sodium dodecyl sulphate (SDS)–polyacrylamide ExcelGel (Amersham Biosciences). One of the two gels was transferred to a nitrocellulose membrane (Trans-Blot, 0·45 µm; Bio-Rad Laboratories Inc., Hercules, CA) using a NovaBlot transfer apparatus (Amersham Biosciences). The membrane was blocked with 1% bovine serum albumin (BSA) in phosphate-buffered saline (PBS) and immunostained using anti-human SP-A primary antibody, followed by horseradish peroxidase (HRP)-conjugated goat anti-rabbit secondary antibody. The signal was detected with Western Lightning enhanced chemiluminescence reagent (Perkin Elmer Life Sciences, Boston, MA). The other gel was silver-stained, using a glutaraldehyde-free method of sensitization, according to the manufacturer's instructions (SilverQuest; Invitrogen Corp.). The gel was then washed in water for 10 min, drained and photographed using a Polaroid camera. The silver-stained gel was overlaid on the radiograph of the immunostained gel, and silver-stained gel spots corresponding to anti-SP-A immunoreactive spots were excised for MS analysis (see below). A comparable size piece of gel from a protein-free area of the gel was excised as a negative control for MS.

Mass spectrometry In order to determine the nature of the immunoreactive protein identified by the 2D gel immunoblot, tryptic peptides from gel fragments were analysed by matrix-assisted laser desorption ionization–time of flight (MALDI-TOF) MS. Excised silver-stained gel fragments were destained and subjected to partial trypsin digestion by dehydrating the gel fragments in 100% methanol for 5 min, rehydration in 30% methanol in water, followed by three washes, for 10 min each, in 100 mm ammonium bicarbonate containing 30% acetonitrile. The gel pieces were crushed, rinsed in water and dried in a Speed Vac for 30 min. Gel pieces were resuspended in 50 mm ammonium bicarbonate buffer containing 7 ng/ml sequencing grade trypsin (Promega Corp., Madison, WI) and incubated overnight at 37°. The supernatant was collected following high-speed centrifugation for 1 min. The remaining peptides were extracted from the gel pieces using 20 µl of 50% acetonitrile containing 0·1% trifluoroacetic acid and then combined with the above supernatant. Solutions were concentrated to 5 µl in a Speed Vac and submitted to the Penn State College of Medicine Core Facility for MS.

Tryptic peptides were mixed with an equal volume of matrix (α-cyano-4-hydroxy trans cinnamic acid), passed through a Zip-Tip and spotted onto a stainless steel sample stage where the sample was allowed to evaporate at room temperature. Mass spectra were obtained using a Reflectron MALDI-TOF mass spectrophotometer (Voyager DE-PRO; PerSeptive Biosystems/Applied Biosystems). Peptide spectra were calibrated using several matrix ion peaks as internal standards. Peptide spectra were compared with human SP-A amino acid sequence (gi:13346506) from the National Center for Biotechnology Information (NCBI) database.

Results

Localization of SP-A in the vaginal mucosa

In all tissue sections from premenopausal donors, SP-A immunoreactive protein was present in the vaginal epithelium and localized in two distinct regions: in the cytoplasm of cells in the deep portion of the intermediate layer adjacent to the parabasal layer; and in the superficial layer (Fig. 1a, panels 1–3; Fig. 1b, panel 1; Fig. 1c, panel 1). The deep intermediate layer contains cells that accumulate glycogen and start the process of progressive loss of biosynthetic organelles as they undergo terminal differentiation, whereas the superficial layer contains essentially dead cells with pyknotic nuclei that are destined to be shed into the vaginal lumen.20 There were no obvious differences in localization or intensity of the immunoreactive protein as a function of the ovarian cycle, as determined by histological dating of endometrium obtained at the time of hysterectomy (data not shown). There was no immunostaining seen in cells in the upper intermediate layer, i.e. the layer that separates the strongly immunopositive deep intermediate and superficial layers (Fig. 1a, panels 1 and 2). This is in contrast to glycogen, which is present throughout the intermediate layer (Fig. 1c, panels 2 and 3).20 Pretreating sections with α-amylase did not affect immunostaining for SP-A, indicating that glycogen did not affect SP-A immunostaining (compare Fig. 1a, panel 1 with Fig. 1b, panel 1 and Fig. 1c, panel 1). Cells in the basal layer, where epithelial stem cells are localized, and cells in the parabasal layer, which separates the basal from the intermediate layer, were consistently immunonegative (Fig. 1a, panel 1).

Figure 1.

Topography of immunoreactive surfactant protein A (SP-A) in premenopausal vagina. (a) SP-A identified by immunocytochemistry (ICC), using Vector Red as the chromogen. (b) Preimmune and secondary antibody controls for ICC for SP-A using alkaline phosphatase reporter, and bromo-chloro-indolyl-phospahate (BCIP) and nitroblue tetrazolium (NBT) as the substrate and chromogen, respectively. (c) Correlation of localization of immunoreactive SP-A with glycogen. (a) Panel 1: at low magnification, SP-A (red reaction product) is seen in two distinct layers: in the basal portion of the intermediate layer, composed of rounded, metabolically active cells; and in the superficial layer, composed of flattened dead cells with pyknotic nuclei. Note the absence of immunostaining in the more superficial cells of the intermediate layer. The arrow points to one of the papillae comprising a core of subepithelial cells [shown in panels 2 and 3 of (a) by the letter ‘p’] surrounded by basal and parabasal cells. (a) Panels 2 and 3 show panel 1 at successively higher magnifications: panel 2 highlights the virtual absence of immunostaining in the more superficial portion of the intermediate layer, while panel 3 highlights the contrast between the strongly immunostaining cells in the intermediate layer and the minimal or no immunostaining of cells in the basal layer. The arrow points to the layer of cells demarcating the epithelium from the subepithelial layer (lamina propria). (b) Panel 1: immunoreactive SP-A (blue/black reaction product of BCIP/NBT) in a section from which glycogen had been removed by pretreatment with α-amylase. Note the similarity in localization of immunoreactive SP-A in the deep intermediate layer with that in sections containing glycogen (a, panel 1). (b) Panels 2 and 3: the absence of immunostaining in preimmune serum (panel 2) and secondary antibody (panel 3) controls. (c) Panels 1, 2 and 3: comparison of the topography of immunoreactive SP-A with that of glycogen. (c) Panel 1: section immunoreacted for SP-A after digestion of glycogen with α-amylase (blue/black reaction product of BCIP/NBT. (c) Panels 2 and 3: glycogen, stained magenta with pararosanaline, using periodic acid-Schiff (PAS) as the chromogenic dye, is seen throughout the intermediate layer. (c) Panel 4: elimination of staining for glycogen with PAS in a section pretreated with α-amylase. The stain for glycogen using pararosaniline in the PAS reaction, unlike basic fuchsin dye, does not stain the amylase-resistant intercellular material (see the Discussion). Sections were counterstained with haematoxylin in panels 1, 2 and 3 of (a), and in panels 2, 3 and 4 of (c). Images were recorded with a Nikon DXM 1200 digital camera using Nomarski optics and Nikon ACT1, version 2, software.

Immunostaining within the superficial layer in some specimens appeared continuous, while in others it was interrupted by weakly or minimally immunopositive regions. It should be noted, however, that because this study was carried out on surgical specimens, part of the superficial layer was probably removed during the preparation of the vaginal mucosa for surgery. In one study, wiping the vaginal mucosa with rough absorbent paper was reported to decrease the glycogen concentration of vaginal epithelium by ≈ 20%.21

Control sections preincubated with preimmune serum or with secondary antibody only were uniformly negative (Fig. 1b, panels 2 and 3).

As expected, in specimens from postmenopausal women there were marked differences in the thickness of the mucosa, the intermediate layer was narrowed and stratification was less evident.20 Nevertheless, SPA immunostaining was present in all specimens and was localized, as in the premenopausal epithelium, in cells adjacent to the parabasal layer and in the superficial layer (Fig. 2). The distribution of such immunopositive cells in the vaginal epithelium from postmenopausal subjects, like the structural organization of the epithelium, varied greatly (Fig. 2). Groups of cells with strong immunostaining were seen to be interposed between cells with minimal immunostaining. In four of the 12 specimens from postmenopausal subjects the epithelium was markedly thickened (acanthosis) and appeared to be covered by an immunonegative keratinized layer (Fig. 2a, panels 7 and 8). These changes are probably the result of trauma associated with prolapse of the vaginal vault, the reason for surgical repair of the vagina. Changes suggesting oestrogenization of the vagina were seen in only one of the patients prescribed equine oestrogens (Fig. 2b, panels 1 and 2). We have no information on whether the other three patients were taking the oestrogens as prescribed. Together, these findings suggest that SP-A is expressed in the vaginal epithelium constitutively, and is not dependent on ovarian hormones.

Figure 2.

Topography of immunoreactive surfactant protein A (SP-A) in the vaginal epithelium of nine postmenopausal women. Immunocytochemistry was carried out, as described in the Materials and methods, using Vector Red as the chromogenic substrate for the alkaline phosphatase reporter and haematoxylin to stain nuclei. (a) Panels 1–8: tissue sections from from four donors; panels 1, 3, 5 and 7 (upper panels) at lower magnification and panels 2, 4, 6 and 8 at higher magnification illustrate the large differences in thickness of vaginal epithelium of women not receiving oestrogen. Panels (a)7 and (a)8, as well as (f), show thickening of the epithelium (acanthosis) and keratinization of the superficial layer, a response to chronic trauma resulting from prolapse of the vagina. The most obvious structural difference between pre- and postmenopausal vagina is in the intermediate layer. It is much thinner and constitutes a much less distinct layer in the postmenopausal that in the oestrogenized premenopausal vaginal epithelium. Nevertheless, SP-A-immunopositive cells, although more patchy than in premenopausal vaginal mucosa, can be seen in the deep intermediate layer in all specimens and there are areas showing strong immunostaining for SP-A in the superficial layer. (b) Panels 1 and 2 show the more oestrogenized appearance of vaginal epithelium of one of the subjects prescribed equine oestrogens. It shows a better developed intermediate layer than in sections from the other 11 postmenopausal subjects and therefore a much clearer separation of the immunopositive cells in the deep intermediate layer and the immunopositive superficial layer. Panels (c)–(f) exemplify the marked variation in SP-A distribution in postmenopausal subjects.

Identification of SP-A transcripts in the vaginal mucosa

To determine the presence and size of SP-A transcripts in the vagina, total RNA isolated from the vaginal mucosa was subjected to Northern blot analysis, with RNA from lung serving as a positive control. SP-A transcripts from the vagina were found to be equal in length to those from the lung (Fig. 3). However, while the signal from hybridization to the transcripts from lung could be obtained after 10 min exposure of the membrane, obtaining a comparable signal from vaginal RNA required 36 hr of exposure. In interpreting this difference in relative abundance of SP-A transcripts in the two tissues it is important to bear in mind differences between them in the proportion of cells expressing SP-A. In lung, ≈ 16% of the tissue used for RNA extraction comprises cells that express SP-A, the type II pneumocytes.22 According to the immunocytochemical findings, SP-A expression is essentially limited to cells in the deep intermediate layer of the vaginal epithelium. These clearly constitute only a small fraction of the cells present in full-thickness tissue specimens (epithelium +lamina propria) from which the RNA was extracted. The superficial layer which also shows intense immunostaining comprises mostly dead cells that are an unlikely source of transcripts.

Figure 3.

Northern blot analysis of surfactant protein A (SP-A) transcripts in vaginal mucosa. Total RNA (20 µg) from lung (lane 1) and vagina (lane 2) was separated by a 1% agarose formaldehyde gel and transferred to a GeneScreen membrane. The membrane was hybridized with a 32P-labelled anti-sense SP-A RNA probe and exposed for 10 min (lung) or 36 hr (vagina) at − 80°.

Vaginal RT–PCR products from three individuals are shown in Fig. 4. The size of the amplicon from each of the vaginal samples was identical to that from the lung. To determine whether SP-A transcripts in the vagina are derived from one or both SP-A genes, clonally selected RT–PCR products from two of the tissue donors were sequenced. Sequence comparison indicates that the vaginal transcripts are identical to those characterized in the lung, and that both SP-A1 and SP-A2 transcripts are expressed (data not shown). Although the SP-A cDNA that was sequenced from the vagina was not full length, the Northern blot analysis indicated that lung and vaginal transcripts are identical in size. Therefore, parts of the vaginal SP-A cDNA that were not analysed at the sequence level, including the 5′- and 3′ untranslated regions (UTRs), would not be expected to be significantly different from those previously characterized from the lung. Confirmation of the full sequence, including the 3′- and 5′-UTRs, will be required to exclude the possibility of small tissue-specific differences that may affect gene expression.

Figure 4.

Vaginal surfactant protein A (SP-A) transcripts demonstrated by reverse transcription–polymerase chain reaction (RT–PCR). Total RNA (100 ng) from lung (lane 1) or vaginal mucosa (lanes 2–4) was reverse transcribed. PCR reactions were primed with oligonucleotides that span exons 1–4 and recognize sequence common to SP-A1 and SP-A2. The faint band marked by the small white arrow is identical in size to genomic SP-A.

SP-A is present in the vaginal lavage fluid

The presence of SP-A immunoreactive cells in the superficial layer of the vaginal mucosa led us to postulate that SP-A may be released into the fluid phase associated with the vaginal mucosal surface. Indeed, immunoreactive SP-A was identified in fluid collected by lavage from a healthy premenopausal donor and separated by 2D SDS–poluacrylamide gel electrophoresis (PAGE) (Fig. 5). A comparison of the 2D gels of human lung SP-A and vaginal fluid SP-A revealed that the latter has an isoelectric point which is slightly more basic, and an apparent molecular mass of 60 000–70 000, which is slightly higher, than that of the SP-A dimer from the lung (data not shown).23

Figure 5.

Two-dimensional gel electrophoresis immunoblot of vaginal fluid proteins. The blot was immunostained with rabbit anti-human surfactant protein A (SP-A) primary antibody and goat anti-rabbit secondary antibody conjugated with horseradish peroxidase, followed by detection with enhanced chemiluminescence. The arrow indicates the spot that was excised from a parallel silver stained gel and partially digested with trypsin. The resulting tryptic peptides were analysed by mass spectrometry.

To confirm the identity of the protein detected in the vaginal lavage, MALDI-TOF mass spectrometry was used to analyse tryptic peptides from the silver-stained gel spot that corresponded to the SP-A immunoreactive spot (Fig. 5). The peptides were found to cover 55% of the amino acid sequence of pulmonary SP-A (Fig. 6).24 This greatly exceeds the 15% coverage that is considered sufficient to establish the identity of a protein.25

Figure 6.

Matrix-assisted laser desorption ionization–time of flight (MALDI-TOF) identification of surfactant protein A (SP-A) tryptic peptides in vaginal fluid. A piece of gel corresponding to the spot shown in Fig. 5 was digested with trypsin and peptides were analysed by MALDI-TOF. Of the complete SP-A amino acid sequence depicted, the peptides shown in bold italics and underlined were identified in the trypsinized gel piece.

Together, the findings provide evidence for the presence of immunoreactive SP-A in the vaginal epithelium at two discrete sites, of SP-A protein in the fluid phase and that both SP-A1 and SP-A2 genes are expressed in the vaginal epithelium.

Discussion

Locally produced factors that modulate the host response to pathogens represent an early and critical aspect of host defence. The role of such innate immune factors in the vaginal host defence has only recently begun to receive attention.11,12 The data presented here indicate that one such factor, SP-A, is produced in the vaginal epithelium within a specific epithelial cell population in the intermediate layer, is concentrated in the superficial layer and can be recovered intact from vaginal lavage fluid. The finding of no obvious cyclic changes in SP-A immunoreactivity in the premenopausal vaginal epithelium, and, more importantly, its persistence in the postmenopausal vaginal epithelium, indicate that SP-A is expressed in this tissue constitutively, independently of ovarian hormones.

The discrete localization of SP-A immunoreactivity in the deep intermediate layer, adjacent to the basal layer of the vaginal mucosa, is consistent with the notion that SP-A could participate in the afferent arm of the immune response in this tissue.11,12 According to the current understanding of early responses of the vagina to pathogens, phagocytic cells in the dermis, and the basal and parabasal layers of the squamous epithelium traverse intercellular channels permeating the epithelium, sample pathogens then return to local subdermal lymphoid aggregates where antigen presentation occurs.10,26 Langerhans' cells – the resident dendritic cells in the vagina and one of the putative targets of SP-A – have also been identified in the deep intermediate layer.6,27,28 The localization of cells expressing SP-A in the same layer puts this collectin in place to contribute to the process of antigen presentation and activation of dendritic cells. Similarly, the concentration of SP-A in the superficial layer of the vaginal mucosa clearly places it at a strategic site for participating in the immediate, innate immune component of host defence. However, the fact that the immunopositive superficial layer is separated from those producing it in the deep intermediate layer by several strata of immunonegative cells raises some obvious questions. How does SP-A reach the superficial layer and what is responsible for anchoring SP-A to this layer?

One plausible answer to the question of the origin of SP-A in the superficial layer is provided by evidence for the presence of channels between cells throughout the intermediate layer.10,26,29 These channels could provide a route for transfer of SP-A from cells in the deep intermediate layer to the surface layer of the vaginal mucosa. The upper genital tract could also contribute to SP-A in the superficial layer of the vagina, as secretory products of the mucosa lining the uterus and cervix are known to reach the vagina. Therefore, it will be important to determine whether and where in the upper genital tract SP-A is expressed.

To address the question of what is responsible for anchoring SP-A to the superficial layer requires consideration of the composition of the material occupying the extracellular space. The presence of a significant extracellular space, occupied by material that stains with basic fuchsin Schiff reagent and is resistant to α-amylase, and that surrounds cells of the superficial layer of the vaginal epithelium, was first described in 1951.30 The staining has been attributed to polysaccharides and/or glycoproteins and to constitute an example of a prominent glycocalyx.31,32 While the composition of the glycocalyx in the superficial layer of the human vagina remains to be determined, it is reasonable to propose that it would include molecules with which SP-A could associate via its carbohydrate-recognition domain(s). Together, the findings direct attention to the need to consider the extracellular components of the vaginal epithelium as part of the host-defence system.

The demonstration that both SP-A1 and SP-A2 genes are expressed in the vaginal mucosa has functional implications. In the lung, two gene products have been described.18 Functional differences between SP-A derived from a single gene, and SP-A derived from both genes, have been reported.17 For example, cytokine production by macrophage-like THP-1 cells is greater when cells are treated with in vitro-expressed SP-A2 than with SP-A1, and is greater upon treatment with co-expressed SP-A1 and SP-A2 than with either SP-A1 or SP-A2 alone. Therefore, it will be of interest to determine whether quantitative differences in SP-A1 relative to SP-A2 in the vagina contribute to an individual's response to micro-organisms.

Finally, it should be noted that the electrophoretic characteristics of vaginal SP-A differ somewhat from those of lung. Electrophoretic characteristics of lung SP-A are largely caused by various post-translational modifications, including glycosylation, sialylation, acetylation and proline hydroxylation, modifications that may affect function.23,24 Therefore, it will be of interest to determine how SP-A is processed in the vagina and whether there are differences in the processing of SP-A among individuals and in health and disease.

In conclusion, SP-A is expressed in the vaginal mucosa, is localized in specific cellular layers and can be recovered from vaginal lavage fluid. Together, the findings suggest that SP-A is an integral component of vaginal host defence and provide the foundation for future investigation as to its role in innate and adaptive immunity in the lower genital tract. The vagina plays a critical role in preventing systemic infection resulting from pathogens breaching the vaginal mucosa or reaching the upper genital tract and abdominal cavity. Hence, knowledge of the basis for qualitative and quantitative alterations of vaginal SP-A has the potential to contribute to our understanding of the pathogenesis of both local and ascending inflammatory conditions of the female genital tract. These include conditions that have substantial impact on human health, notably, sexually transmitted diseases and preterm labour.13 Learning about SP-A and other innate host defence factors in the vagina may also provide insight into their role at other mucosal surfaces that are less readily accessible to investigation.

Acknowledgments

The authors are grateful to Dr Mary K. Howett for many helpful discussions and to Dr Gary Clawson, Director of the Jake Gittlin Cancer Institute, for use of the Institute's photomicrography equipment. This study was supported by grants NIH R21 DE-14041 to Colin MacNeill, PHS PO1 AI-37829 to Judith Weisz, NIH R21 DE-14041 to David S. Phelps, NIH 5R37HL-034788-16 to Joanna Floros and by a Penn State College of Medicine Dean's Feasibility Grant to Zhenwu Lin. Photomicrography equipment supplied by The Jake Gittlin Cancer Center was supported by NIH R01 CA-40145.

Ancillary