Mannose-binding lectin is present in human semen and modulates cellular adhesion of Neisseria gonorrhoeae in vitro

Authors


  • This work was presented in part at the 16th International Pathogenic Neisseria Conference, 7–12 September, 2008, Rotterdam, the Netherlands, abstract number (P084).

Summary

Mannose-binding lectin (MBL) is an innate immune molecule present in blood and some mucosal tissues, which can influence microbial attachment and inflammatory responses of host cells during infection. In this study MBL was found to be present at a low concentration in semen samples in the range 1·2–24·9 ng/ml. Co-incubation of bacteria with semen resulted in the binding of MBL to the bacterial surface. Neisseria gonorrhoeae is a common cause of genitourinary infection. MBL bound to N. gonorrhoeae with strain-to-strain variation in the intensity of binding and nature of the bacterial receptor. Pretreatment with MBL concentrations similar to those found in human serum modulated the adhesion of N. gonorrhoeae strain FA1090 but not strain MS11 to epithelial cells. This effect was dose-dependent. This work demonstrates that MBL is present in human semen and modifies cellular responses to N. gonorrhoeae in a concentration-dependent manner.

Introduction

Mannose-binding lectin (MBL) is a innate immune molecule of the collectin family, which is able to bind to repeating sugars structures on a wide range of human pathogens via its carbohydrate recognition domain (CRD) [1–4]. MBL activates complement via MBL-associated serum proteases (MASPs), which cleave C2 and C4 to form a C3 convertase leading to higher levels of serum killing and opsonization [5,6]. However, MBL also has a number of complement-independent effects on bacterial adhesion, internalization [7–9] and the inflammatory response elicited by a number of pathogens [10,11]. MBL is an acute phase protein produced mainly by the liver, but gene transcription has been demonstrated recently to occur in extra-hepatic sites, including the testis and the prostate gland [12]. While there have been no previous reports of MBL in seminal fluid, MBL can be detected in cervicovaginal lavage [13,14].

Neisseria gonorrhoeae causes one of the most common sexually transmitted infections in the world [15]. During the infective process the first point of contact for the gonococcus is the epithelial layer. In uncomplicated gonorrhoea the bacterium will adhere to the epithelial layer to provoke an inflammatory response leading to urethritis in males and cervicitis in females [15]. Additionally, the bacterium may invade the epithelial layer to reach the subepithelial tissue and go on to cause complications such as salpingitis, epididymitis, prostatitis, pelvic inflammatory disease and disseminated gonococcal infection [16]. As a result any modulation in the adhesion, invasion and inflammatory response of epithelial cells is likely to be of importance to the progression of disease.

MBL has been demonstrated previously to bind to N. gonorrhoeae Opa proteins [17] and the porin protein of the closely related N. meningitidis[17]. Lipo-oligosaccharide (LOS) structure is also a major determinant of MBL binding [18,19]. In work with the closely related N. meningitidis, we demonstrated that MBL increases phagocytic killing [8] while modulating tumour necrosis factor (TNF)-α, interleukin (IL)-6 and IL-1β release by human monocytes [20].

In the knowledge that MBL binds to important gonococcal adhesion and invasion factors, we set out to determine whether MBL is present in human semen and whether it modifies interactions of N. gonorrhoeae with epithelial cells.

Materials and methods

Purification of MBL

MBL was purified from human plasma paste (kindly provided by C. Dash, Blood Products Laboratory, Elstree, UK) using a two-step mannan affinity process [2,20,21]. Plasma paste was mixed with mannan agarose and MBL eluted by washing with 10 mM ethylenediamine tetraacetic acid (EDTA) in Tris-buffered saline (TBS), injected onto a second mannan agarose column and eluted with 100 mM mannose in TBS + 5 mM Ca2+. MBL was then purified further by size exclusion chromatography on a superose-6 10 mm/30 cm gel filtration column (Amersham Biosciences, Chalfont St Giles, UK) [22,23]. Purity was assessed by reducing sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblotting [24].

MBL detection in semen

Samples were obtained from men attending the Andrology Laboratory (Jessop Wing, Royal Hallamshire Hospital, Sheffield, UK) as part of infertility investigations. Liquefied samples were selected from 14 spermic men who had given informed consent for semen remnants to be donated to research. This collection was approved by the South Sheffield Research Ethics Committee (06/Q2305/21). For MBL detection, semen was centrifuged at 2000 g for 5 min and the supernatant tested for MBL by enzyme-linked immunosorbent assay (ELISA.) (limit of detection 0·02 ng/ml; Antibodyshop, Gentofte, Denmark).

Binding of MBL in semen to bacteria

Staphylococcus aureus strain SH1000 was grown to mid-log phase in brain heart infusion (BHI) broth and then incubated in 50 µl centrifuged seminal plasma for 20 min at 37°C. The bacteria were washed and incubated in 79 µl phosphate-buffered saline (PBS) +1 µl fluorescein isothiocyanate (FITC)-conjugated anti-MBL (131-01; Antibodyshop). Cells were then fixed in 1% paraformaldehyde and run on a Becton Dickinson fluorescence activated cell sorter (FACS)calibur (BD Biosciences, San Jose, CA, USA). SH1000 was used as we have found previously that it is a high MBL binding strain and the MBL concentration in semen is low.

Epithelial cell culture

The human conjunctival cell line, Chang (Wong–Kilbourne derivative of Chang conjunctiva; ATCC CCL-20·2) was maintained in Medium 199 (Gibco, Paisley, UK) with 10% heat-inactivated fetal calf serum (HIFCS) in tissue culture flasks at 37°C, 5% CO2. Cells were seeded at a cell density of 1·5 × 105 and used after 24 h of incubation. Chang cells were used because of their track record of use in published adhesion and invasion studies with gonococci [25,26]. It is important to note that the Chang cell line is a HeLa cell hybrid, as established by the presence of HeLa marker chromosomes in Chang cells [27].

Human papilloma virus E6- or E7-transformed human urethral epithelial cells (THUEC) have been described previously by Harvey [28] and were a kind gift from Dr M. Apicella, University of Iowa. Cells were seeded at a cell density of 2·5 × 105 in prostate epithelial growth medium (PrEGM) with 5% HIFCS at 37°C, 5% CO2. Cells were used after 3 days of incubation.

Preparation of bacterial strains and pretreatment with MBL

Stocks of N. gonorrhoeae, strains FA1090 [29] and MS11 [30], and S. aureus strain SH1000 [31] were maintained at –80°C. Strain FA1090 was grown for 24 h on Columbia agar with chocolated horse blood (Oxoid, Cambridge, UK) at 37°C, 5% CO2, then resuspended in veronal-buffered saline supplemented with 5 mM CaCl2 and 10 mM glucose (VBS-Glc). In both cases multiple colony picks were selected by dissecting microscopy to include a mixture of Opa and pili-expressing and non-expressing colonies. Strain MS11 was grown for 24 h on gonococcus (GC) agar (Difco, Detroit, MI, USA) supplemented by Vitox (Oxoid) at 37°C, 5% CO2, then resuspended as described above. S. aureus strain SH1000 was grown on blood agar 37°C, 5% CO2. Strains were grown on different media as they demonstrated different preferential growth media. MBL binding was affected non-significantly by differences in media (data not shown). With the exception of controls, bacterial solutions were supplemented with MBL (0·02, 0·4, 1·6 or 5 µg/ml) and incubated for 10 min at 37°C, 5% CO2, before washing and final suspension in the appropriate culture medium. In certain experiments MBL was incubated for 10 min at 37°C in VBS-Glc supplemented with 100 mM mannose (an inhibitor of carbohydrate binding) or 100 mM galactose (a non-inhibitor of binding) [32] immediately prior to addition to the organisms. As an additional control, MBL was heat-treated at 100°C for 10 min before addition to the bacterial pellet.

Whole cell ELISA for measurement of bacterial binding of MBL

The whole cell ELISA method has been described previously [17,33]. Briefly, MS11 and FA1090 were diluted to an optical density (OD)540 nm of 0·1 in TBS + 5 mM CaCl2 (TBS+); 100 µl of bacterial suspension was placed in wells of a 96-well tissue culture plate and allowed to dry overnight at 37°C, 5% CO2. Wells were blocked with 10% HIFCS TBS and incubated with MBL at different concentrations (5, 2·5, 1·25 and 0·625 µg/ml), mannose-treated MBL (mtMBL) or MBL with 10% HIFCS. Primary anti-MBL antibody (131-01; Antibodyshop) and secondary antibody were diluted in 10% HIFCS-TBS.

Effect of MBL on epithelial cell binding

Cells were washed twice with PBS warmed to 37°C and GC was added at a multiplicity of infection (MOI) of 100 (MS11) or 10 (FA1090) in tissue culture wells in duplicate. A lower MOI was used for FA1090 in comparison to MS11 because FA1090 bound to cells with higher efficiency (data not shown). After 3 h incubation at 37°C 5% CO2 cells were fixed with 1% (w/v) paraformaldehyde. The cells were then washed before staining with 85 ng/ml 4′,6-aminido-2-phenylindole dihydrochloride (DAPI; Molecular Probes, Eugene, OR, USA). Wells were incubated at room temperature in the dark for 10 min before a final wash. Coverslips were then mounted and counted blind to sample identity using a Leica DMRB fluorescent microscope at 100× objective. The following parameters were recorded: organisms per positive cell (defined as a cell with one or more organism within its cytoplasmic boundary), proportion of cells positive for one or more organisms and organisms per 100 cells counted.

Effect of MBL on epithelial cell invasion

A gentamicin exclusion assay was used to assess the invasion of cells by GC. Organisms were inoculated onto Chang cells at MOIs of 200, and then incubated for 6 h at 37°C, 5% CO2. Cells were then washed twice to remove non-adherent organisms and pulsed for 30 min with 400 µg/ml gentamicin, washed and lysed with 1% (w/v) saponin, and the viable counts of the organisms determined using a serial dilution technique. In relevant controls, no antibiotics were added. In preliminary work, this method was shown to eliminate 100% of identical inocula incubated in wells that lack cells (data not shown).

Effect of MBL on epithelial cell cytokine release

Both MS11 and FA1090 were inoculated onto cells at MOIs of 100. Cells were then incubated for 4, 6, 8 or 24 h. Supernates were collected and cytokine accumulation was measured by ELISA; IL-6, IL-8 (BD Biosciences). Additionally, a cytokine array (Raybio human inflammation antibody array III; Raybiotech, Norcross, GA, USA) was used according to the manufacturer's instructions to assess Chang cell cytokine production at 6 h post-infection.

Statistical analysis

Except where noted, in all cases each parameter after preincubation of organisms with MBL was calculated as a percentage of the parameter with buffer alone, which was set to 0%. Data were examined for normality using spss version 14 software (SPSS Inc., Chicago, IL, USA). Normally distributed data were compared to the buffer alone value of 0% by one-sample t-test. Where data were not normally distributed a Wilcoxon signed-rank test was used to compare MBL with buffer alone. In all cases, reported n numbers reflect separate experiments.

Results

MBL concentration within human semen

MBL was detected in all 14 samples of human seminal plasma derived from asymptomatic males, in the range 1·2–24·9 ng/ml with a median level of 5·41 ng/ml (Fig. 1).

Figure 1.

Mannose-binding lectin (MBL) concentrations in human seminal plasma. Determined by enzyme-linked immunosorbent assay; n = 14. Line denotes median.

MBL binds to N. gonorrhoeae in a dose-dependent manner

MBL bound to both MS11 and FA1090 in a dose-dependent manner (Fig. 2a). Significantly more MBL bound to MS11 than FA1090 (P = 0·0001). MBL binding to MS11 and SH1000 was inhibited by preincubation with mannose (P = 0·002), suggesting inhibition of CRD-dependent binding, but not 10% HIFCS (Fig. 2b). Conversely, MBL binding to FA1090 was inhibited by 10% HIFCS (P = 0·014) but not mannose.

Figure 2.

Mannose-binding lectin (MBL) binding to bacteria by whole cell enzyme-linked immunosorbent assay. Microtitre wells were coated with whole Neisseria gonorrhoeae strains MS11 or FA1090 or Staphylococcus aureus strain SH1000 to which MBL was added at a range of doses or pretreated with either 100 mM mannose or 10% heat-inactivated fetal calf serum (HIFCS). (a) MBL binding to MS11, FA1090 and SH1000 at a range of doses. (b) Pretreated MBL (5 µg/ml) binding to MS11 relative to MBL (5 µg/ml) in buffer alone. (c) Pretreated MBL (5 µg/ml) binding to FA1090 relative to MBL (5 µg/ml) in buffer alone. Values in (b) and (c) are the percentage binding of pretreated MBL relative to MBL in buffer alone, which is set at 100%. (a) Area under the curve for MS11 and FA1090 was compared by paired t-test, P = 0·0001, n ≥ 4, mean ± standard error of the mean (s.e.m.). (b,c) One-sample t-test *P < 0·05, **P < 0·01, n ≥ 2, mean ± s.e.m.

MBL from human semen binds to bacteria in a dose-dependent manner

MBL binding from human semen from four donors was detectable on SH1000 (Fig. 3a) and was dose-dependent and mannose-inhibitable (Fig. 3b). Seminal MBL binding was not detected on strains FA1090 or MS11, which are relatively poor MBL-binding strains compared with S. aureus. (Fig. 2a).

Figure 3.

Semen mannose-binding lectin (MBL) binding to Staphylococcus aureus. SH1000 was grown to mid-log phase in BHI broth and then incubated in seminal plasma from four separate donors or phosphate-buffered saline (PBS) followed by staining with fluorescein isothiocyanate-conjugated anti-MBL antibody or isotype control. Cells were then fixed in 1% paraformadyhyde and run on a Becton Dickinson fluorescence activated cell sorter (FACS)calibur. (b) Semen was diluted in PBS or preincubated with mannose to inhibit MBL binding.

MBL modifies adhesion of N. gonorrhoeae strain FA1090 to Chang epithelial cells

MBL exhibited a concentration- and strain-dependent effect on the association of N. gonorrhoeae to the Chang epithelial cell line (Fig. 4). MBL treatment (5 µg/ml) caused a 12% decrease from an untreated control value of 8·3 organisms of strain FA1090 per positive cell (P = 0·014). This effect was ablated by both mannose treatment of MBL (mtMBL) and by heat treatment of MBL (htMBL). Galactose-treated MBL (gtMBL) had an effect similar to MBL alone, but was not significant (Fig. 4a). MBL treatment (5 µg/ml) also decreased the number of organisms per 100 cells by 17% from an untreated control value of 147 organisms per 100 cells (P = 0·0045), while gtMBL effected a 12% decrease from the control (P = 0·043). This effect was inhibited by mannose and heat treatment (Fig. 4b).

Figure 4.

Association of mannose-binding lectin (MBL)-treated Neisseria gonorrhoeae with Chang epithelial cells. N. gonorrhoeae strain FA1090 was preincubated with MBL (0·02, 0·4, 1·6 or 5 µg/ml), heat-treated MBL (5 µg/ml; htMBL), mannose-treated MBL (5 µg/ml; mtMBL), galactose-treated MBL (5 µg/ml; gtMBL) or buffer alone. Organisms were then incubated with Chang cells at a multiplicity of infection of 10 for 3 h before fixation and fluorescent staining. Samples with MBL were calculated as a percentage of samples with buffer alone, which was set at 0%. (a,b) Carbohydrate inhibition; (c,d) dose–response. n ≥ 6, mean ± standard error of the mean. *P < 0·05, **P < 0·01 one-sample t-test.

In dose–response experiments three lower concentrations of MBL were used (0·02, 0·4 and 1·6 µg/ml). An increase in organisms per infected cell was observed when bacteria were pretreated with 0·4 or 1·6 µg/ml (Fig. 4c). This increase was significant at 0·4 µg/ml with an 18% increase observed from the untreated control value of 8·3 (P = 0·017). A non-significant increase was also seen in organisms per 100 cells (Fig. 4d). MBL had no significant effect on the proportion of FA1090 internalized by Chang cells (data not shown). MBL (5 µg/ml) had no significant effect on the binding of strain MS11 to epithelial cells (data not shown).

MBL does not modify adhesion of N. gonorrhoeae strain FA1090 to transformed human urethral epithelial cells

There was no effect of MBL (5 µg/ml or 20 ng/ml), mtMBL or gtMBL on the association or internalization of FA1090 to THUECs (data not shown).

MBL did not modify inflammatory responses of epithelial cells to N. gonorrhoeae

Cytokine array demonstrated that Chang cells produced IL-6 and IL-8 when infected by N. gonorrhoeae strains FA1090 or MS11 (data not shown). No effect of MBL pretreatment of N. gonorrhoeae was seen on the production of IL-6 or IL-8 by Chang cells at 4, 6, 8 or 24 h in response to infection with either strain (data not shown).

Discussion

In this study we have demonstrated that MBL is present in semen, binds to the GC and modulates binding of the organism to one of the two epithelial cells types examined in vitro in a dose-dependent and sugar-inhibitable manner.

The finding that MBL is present in semen is consistent with the work of Seyfarth et al.[12], who found that MBL mRNA is transcribed actively in both testis and prostate tissue. In addition, we found this MBL to be capable of binding to bacteria. In this case S. aureus was used, as we have demonstrated that it binds to MBL with great efficiency. We were unable to detect binding of MBL from semen to N. gonorrhoeae, but it seems likely that this may have been due to the sensitivity of the assay as the strains used have relatively low affinity for MBL compared to S. aureus. MBL semen concentrations are similar to those found in cervicovaginal lavage [13,14]. Within vaginal fluid, MBL concentration is influenced directly by MBL genotype [13].

In the current study an MBL concentration similar to that found in semen (0·02 µg/ml) did not affect binding of GC to epithelial cells. However, in natural gonococcal infection, MBL levels within inflamed genital mucosa are likely to rise as a result of exudation and increased acute phase protein production due to elevated levels of IL-6.

The presence of MBL in semen may have consequences for host defence against other infections of the genital tract. In the case of females, small changes in vaginal MBL appear to be related to susceptibility of individuals to recurrent vulvovaginal candidiasis (median 15·9 ng/ml in controls compared to 7·3 ng/ml in patients) [13]. MBL binds to Chlamydia trachomatis, and at low MBL concentrations (0·08 µg/ml) can inhibit C. trachomatis binding to HeLa cells [34]. Because this effect is seen at low concentrations of MBL it is possible that the MBL levels seen in semen may also prove inhibitory. It has also been demonstrated that MBL binds to and inhibits infection by human immunodeficiency virus (HIV) [35]. It seems possible that due to its presence in semen MBL may play a role not only in host defence, but also in disease transmission to sexual contacts.

The effects of MBL on gonococcal adherence were strain-dependent, possibly as a result of a difference in the MBL binding target. It was observed that binding of MBL to FA1090 was inhibited by serum but not carbohydrate while MBL binding to MS11 was inhibited by carbohydrate but not serum. This suggests that MBL binds to a carbohydrate ligand on MS11 but to a non-carbohydrate ligand on FA1090. As gonococcal LOS structure is a major determinant of MBL binding [18], it is suggested that LOS may be an important ligand on MS11 but not FA1090. Because LOS is abundant on the surface of gonococci it is possible that MBL binding to LOS may block binding to other potential targets. The absence of LOS binding in FA1090 may allow MBL to bind to non-carbohydrate MBL targets such as Opa [17] and influence its role in bacterial adherence. We tested the hypothesis that MBL binding to FA1090 is a protein–protein interaction by saturating with HIFCS and found that MBL binding to this strain was inhibited completely by serum, whereas binding to strain MS11 was not reduced significantly by the presence of HIFCS. MBL caused a biphasic modulation of binding of FA1090 to Chang epithelial cells. A reduction in the binding was observed at 5 µg/ml. However, at a lower dose (0·4 µg/ml) an increase in binding was seen. This is consistent with previous observations that the effects of MBL on Neisseria vary with concentration [8,20].

The lack of any effect of MBL on gonococcal adhesion to THUECs demonstrates that this effect is not universal to epithelial cells and may be due to differences in the primary binding targets of the gonococcus on the epithelial layer. Gonococcal adhesion to urethral epithelia is mediated largely by LOS-asialogycoprotein receptor interactions [36], whereas adhesion to the Chang cell line is dependent upon Opa–heparin sulphate proteoglycan interactions [26]. While the observed effects on adhesion were relatively small, this does not preclude them from being of significance, particularly if they act in co-ordination with other host immune factors.

The dose-dependence of the effect of MBL is of particular interest, as MBL serum concentrations vary in the range of 0·002–5·4 µg/ml due to common promoter polymorphisms [37]. This suggests that MBL genotype may be an important factor in gonococcal infection. In addition, MBL is an acute-phase protein [4], suggesting that MBL may have different roles at different stages in gonococcal infection as its local concentration rises as a result of inflammatory exudation.

In epithelial cells the effect of MBL on binding was mannose-inhibited, suggesting that the CRD of MBL is central to the process. However, we found no effect of mannose in MBL binding to strain FA1090. We interpret these data as suggesting that the CRD of MBL is involved in functional interactions with the host cell. Indeed, it has been demonstrated recently that MBL binds to healthy human cells [38]. That galactose abrogated the influence of MBL on binding to MDMs suggests that some level of osmotic interference may also be a factor, as galactose is not a MBL ligand [32].

In conclusion, we have shown that MBL is present in human semen, binds to GC and modifies host cellular responses to the organism. This may have consequences for the pathogenesis and transmission of gonococcal infection.

Acknowledgements

We thank Anne Cook for technical support, Phil Whitfield (Sheffield Teaching Hospitals, NHS Foundation Trust) for assistance with Semen MBL measurements and also C. Dash (Blood Products Laboratory, Elstree) for providing plasma paste. We are grateful to Nigel Klein (Institute of Child Health) for his valuable advice and Dr Mike Apicella (University of Iowa) for providing THUECs used in experiments. This study was supported by the Sheffield Teaching Hospitals Charitable Trust.

Disclosure

Declarations of interest: None.

Ancillary