Analysis of the human Ig isotype response to lactoferrin binding protein A from Neisseria meningitidis

Abstract

An effective vaccine for serogroup B meningococci has yet to be developed and attention has turned to subcapsular antigens of the meningococcus as possible vaccine candidates. Iron binding proteins are being studied, with most interest focused on the transferrin binding proteins (TbpA and TbpB) and the ferric binding protein (FbpA). This study describes the purification of lactoferrin binding protein A (LbpA) from two meningococcal strains and assesses the human isotype-specific serum antibody response to these proteins in patients with proven meningococcal disease due to a range of phenotypes. Overall, fewer than 50% of sera contained IgG that recognised LbpA isolated from either strain and this antibody response was not uniform between the two proteins. There was some evidence that the antibody response varied between meningococcal phenotypes. This study demonstrates that LbpA does not induce a highly cross-reactive antibody response, indicating that it is unlikely to be an effective vaccine antigen.

Lactoferrin binding protein, Neisseria meningitidis, Immune response, Meningococcal vaccine

1 Introduction

It is well understood that humoral antibody production is vital in protection against meningococci disease and it has been shown that the presence of serum bactericidal antibodies correlates with immunity to meningococcal disease [1]. Meningococcal disease is rare in neonates who are protected by maternal antibody [2]. The incidence of disease is high in young children and declines in adulthood as individuals acquire protective antibody. This acquired antibody protects against homologous and heterologous serogroups, serotypes and serosubtypes of Neisseria meningitidis, and it can therefore be inferred that this antibody is produced in response to antigens other than those detected in conventional serogrouping or serotyping schemes.

Capsular polysaccharide vaccines are available for meningococcal serogroups A, C, Y and W135 but only induce short-term immunity and have poor immunogenicity in children younger than 2 years, although polysaccharide-protein conjugate vaccines for serogroups A and C are proving safe and immunogenic in young infants [3]. However, the serogroup B capsule is poorly immunogenic which may partly be due to antigenic similarities with human neural cell adhesion molecules [4]. Thus, there is no polysaccharide vaccine available for serogroup B disease, which is the major disease-causing serogroup in western European countries [5]. As a result, subcapsular antigens are now under consideration as potential vaccine antigens. The PorA outer membrane protein (OMP) is a promising vaccine candidate, since antibodies against this protein were highly protective against bacterial challenge in an animal model [6,7]. However, there is variation in the level of expression of the PorA OMP [8] and a single nucleotide substitution in the PorA gene allows the organism to evade killing by previously specific bactericidal antibody [9].

Other potential vaccine antigens which may provide cross-protection against different serogroups of meningococci include the iron-regulated proteins. Iron is essential for bacterial survival and these proteins are surface-exposed and necessary for the uptake of iron from host carrier proteins such as transferrin and lactoferrin [10]. Most research has so far concentrated on the transferrin binding proteins TbpA and TbpB [11–14] and the ferric binding protein FbpA [15]. However, Neisseria spp. also express lactoferrin binding proteins, LbpA and the more recently identified LbpB [16]. LbpA is highly conserved [17], is present in both pathogenic and commensal Neisseria[18] and is essential for iron uptake from lactoferrin [19]. Lactoferrin is the major human iron carrier protein present on mucosal surfaces and is presumably an important source of iron for N. meningitidis during its colonisation of the nasopharynx. Murine monoclonal antibodies (mAbs) against LbpA reacted with more than 50% of meningococcal isolates tested, although they were not bactericidal [20].

Previously, we have studied the human antibody response to the transferrin binding proteins and shown the response to be cross-reactive to Tbps isolated from different strains [11]. In this current study, we examined the human immune response to LbpA purified from meningococcal strains in 70 well-defined acute and convalescent sera from patients with meningococcal disease, to ascertain whether this antigen is worthy of inclusion in a potential serogroup B meningococcal vaccine.

2 Materials and methods

2.1 Bacterial strains and growth conditions

N. meningitidis strains SD (B:15:P1.7,16) and B16B6 (B:2a:P1.2) were grown under iron-limited conditions as previously described [11].

2.2 Purification of LbpA

Washed bacterial cells were resuspended (20% w/v) in phosphate buffered saline (PBS), pH 7.4, containing 2% (v/v) Elugent (Calbiochem), incubated for 10 min at room temperature and then centrifuged at 27 500×g for 45 min. Ethylenediaminetetraacetic acid and N-lauroyl sarcosine were added to the supernatant to final concentrations of 10 mM and 0.5% (w/v), respectively, and this was applied to a column of lactoferrin-Sepharose, which was prepared by coupling 85 mg human lactoferrin to 5 g CNBr-Sepharose (Pharmacia) according to the manufacturer's instructions. The lactoferrin-Sepharose column was washed with at least 10 column volumes of 10 mM MES/NaOH, 150 mM NaCl, pH 5.5, containing 2% (v/v) Elugent (MES is 2-[N-morpholino] ethanesulfonic acid). LbpA was then eluted with 10 mM MES/TRIS, 2 M NaCl, 10 mM CaCl₂ pH 9.5, also containing 2% (v/v) Elugent. The wash and elution buffers were adapted from those described by Hu et al. [21]. Following each use, the lactoferrin-Sepharose was saturated with iron by washing with 10 column volumes of 40 mM TRIS, 2 mM NaHCO₃, pH 7.4, containing 25 µM Na citrate and 0.025 µM FeSO₄·7H₂O, and then washed with PBS. Fractions containing LbpA were detected by dot blots with lactoferrin-HRP as described by Schryvers and Morris [22].

2.3 SDS-PAGE

Isolated proteins were analysed using 10% (w/v) polyacrylamide gels as described previously [23].

2.4 Sera

Sera referred to the Meningococcal Reference Unit (MRU), Manchester Public Health Laboratory from cases of meningococcal disease were examined. All sera were from culture-proven cases. Strains included serogroups B, C, Y, and X; serotypes 1, 2a, 2b, 4, 14, 15 and 16; and subtypes P1.2, P1.4, P1.5, P1.7, P1.9, P1.10, P1.15 and P1.16. The date of onset of infection and the date serum was taken were known for each individual. Sera were divided into acute (up to 5 days post onset); convalescent (5–28 days post onset); and post-convalescent (greater than 28 days post onset).

2.5 ELISA

Each isolated LbpA was diluted to 1 µg ml⁻¹ in PBS, pH 7.4, added (100 µl per well) to a 96-well ELISA plate (Maxisorp Nunc-immuno) and incubated for 1 h at 37°C. The plate was then washed five times with PBS containing 0.05% (v/v) Tween 20 (PBST). Non-specific antigen binding sites were blocked by the addition of 200 µl PBS containing 3% (w/v) bovine serum albumin to each well. The plate was then incubated for 1 h at 37°C and then rewashed as above. Sera were diluted 1/800 for the IgM isotype and 1/250 for IgG and IgA isotypes in PBST containing 5% (v/v) normal goat serum and 1% (w/v) skimmed milk powder and 100 µl added per well. After incubation for a further 2 h at 37°C, the plate was rewashed as above and 100 µl per well of anti-human, polyclonal, goat IgM, IgG or IgA peroxidase-labelled conjugate (Sigma-Aldrich Co. Ltd., Poole, Dorset, UK), each diluted 1/1000, was added separately. Plates were then incubated for 2 h at 37°C, washed and o-phenylenediamine dihydrochloride substrate was added (100 µl per well) before a further incubation period of 30 min at room temperature. The reaction was stopped with 5 N H₂SO₄ (50 µl per well) and the optical density read at 492 nm (Titertek Multiskan MCC Type 341). A positive and a negative control serum was used in each test. The positive control serum originated from a case of meningococcal disease referred to the MRU. The negative control serum was obtained from a healthy individual with no antibodies against meningococcal OMPs demonstrable using an OMP-ELISA [24] and serum bactericidal antibody assay.

2.6 Calculation of results

Sera giving absorbance values greater than the mean absorbance of the negative control serum plus two standard deviations were designated reactive with the Lbp. The differences in antibody absorbances between LbpA from strains B16B6 and SD were investigated using regression modelling to perform the equivalent of variance leading to t-tests of group differences [25]. The ELISA absorbance values, divided by the cut-off value (negative control mean+2 S.D.) to standardise them, had a rather skewed distribution so, to satisfy the necessary assumption of an underlying normal distribution, they were transformed by taking their logarithms. The estimated mean differences and confidence limits on the log scale produced by this analysis gave, when transformed back to the original scale, the means and confidence intervals for the proportional differences in response to be expected, on average, between the categories being compared. This analysis was done using Microsoft Excel and the statistical package GLIM4 [26].

3 Results and discussion

This study represents the first investigation into the antibody response in patients with meningococcal disease to the lactoferrin binding protein, LbpA, isolated from two meningococcal strains, SD and B16B6. The LbpA proteins in this study were of the same molecular mass — approximately 92 kDa (Fig. 1). This is slightly lower than the published molecular mass of LbpA of 98 kDa [20] and may reflect variation between strains.

Figure 1

SDS-PAGE of LbpA isolated from N. meningitidis strains SD (lane 1) and B16B6 (lane 2).

The antibody levels produced against the two LbpA proteins isolated from two separate meningococcal strains were not uniform (Table 1). More of the sera tested reacted with LbpA isolated from strain B16B6 than with LbpA from strain SD, and there were significantly higher ELISA IgG absorbance values (P<0.0001) produced with LbpA from strain B16B6 (data not shown) as the coating antigen. In addition, IgM and IgA ELISA absorbance values in response to LbpA from strain B16B6 were 26% and 28% higher than those produced in response to LbpA from strain SD but these differences were not significant. This suggests either that the immunogenicity of the latter LbpA may have been affected by purification, or that circulating strains of N. meningitidis possessed LbpA types that were more antigenically similar to that from strain B16B6 or there were differences in the avidity of antibody produced. The former is less likely as the purified LbpA retained its ability to bind lactoferrin. In all instances fewer than 50% of sera recognised either LbpA. The main antibody isotype detected in the reactive sera was IgG and fewer than 10% of sera were found to contain IgM or IgA to either antigen. This is similar to the antibody response to the PorA OMP where a detectable IgM or IgA response is not induced, and different from the PorB OMP response where both of these antibody isotypes are induced [27,28].

Table 1

Results for acute, convalescent and post-convalescent sera which were ELISA-reactive with LbpA from strains SD and B16B6

	Acute (n=24)			Convalescent (n=26)			Post-convalescent (n=20)
	IgM	IgG	IgA	IgM	IgG	IgA	IgM	IgG	IgA
LbpA SD	1 (4%)	1 (4%)	1 (4%)	1 (4%)	1 (4%)	0 (0%)	1 (5%)	4 (20%)	0 (0%)
LbpA B16B6	1 (4%)	8 (33%)	2 (8%)	2 (8%)	12 (46%)	0 (0%)	1 (5%)	9 (45%)	1 (5%)

The percentage of sera containing IgG recognising LbpA increased in the convalescent and post-convalescent periods, indicating seroconversion following LbpA challenge. The ELISA IgG absorbance values produced in response to LbpA increased by 39% between acute and convalescent sera but this did not achieve statistical significance.

Sera were compared from a small number of different phenotypes representing the most frequently isolated phenotypes from the cases examined (Table 2). The magnitude of ELISA absorbances varied in relation to the infecting phenotype (P<0.001). Sera from individuals with N. meningitidis B:2b:P1:10 disease produced ELISA absorbance values that were 14% higher than those with B:15:P1.16 disease and sera from individuals with N. meningitidis C:2a:P1.2 disease produced ELISA absorbance values that were 52% lower than those with B:15:P1.16 disease. This antigenic heterogeneity was also highlighted by Pettersson et al. [20] who demonstrated that two mAbs produced against the 98-kDa LbpA of strain 2996 (B:2b:P1.2) only reacted with 44 and 42 of 74 meningococcal strains respectively. However, all patients infected by strain B:2b:P1.10 produced IgG that recognised LbpA from strain B16B6 (Table 2) suggesting that this particular strain possesses LbpA which is antigenically similar to strain B16B6 LbpA. This is also consistent with the findings of Pettersson et al. [20] who showed that the majority of serosubtype P1.10 isolates reacted with mAbs against the 98-kDa LbpA protein.

Table 2

Comparison of antibody isotype reactivity to the individual LbpAs from strains SD and B16B6 in convalescent sera following meningococcal disease due to phenotypes B:15:P1.16, B:2b:P1.10 and C:2a:P1.2

Sera from cases of phenotype	SD			B16B6
	IgM	IgG	IgA	IgM	IgG	IgA
B:15:P1.16	2/7	1/7	0/7	3/7	2/7	0/7
B:2b:P1.10	0/5	1/5	0/5	0/5	5/5	1/5
C:2a:P1.2	0/3	0/3	0/3	0/3	1/3	0/3

It is important to demonstrate a human antibody response to LbpA in order to establish its potential as a vaccine candidate. In this study a minority of patients with meningococcal infection were found to produce antibody to LbpA and this antibody response was not uniform between the two strains examined, suggesting that LbpA is not a promising cross-reactive vaccine antigen. This contrasts with previous studies which demonstrate that a strong cross-reactive immune response to TbpA and B is induced by meningococcal disease [11,12]. Further research is required to examine the antigenic variation of meningococcal Lbps and to demonstrate protective functional activity of human antibodies to meningococcal LbpA. It would also be of interest to investigate the immune response to LbpB and the immune response to LbpA and LbpB following prolonged carriage of meningococci in the nasopharynx. In a recent paper, convalescent antisera from one patient infected with Moraxella catarrhalis reacted with a protein thought to be LbpB but not with LbpA [29].

Acknowledgements

The authors would like to thank Dr Tony Swan at the PHLS Statistics Unit, Communicable Disease Surveillance Centre, Colindale for statistical support. The work at CAMR was supported by the UK Department of Health.