Scholarly article on topic 'A broadly-protective vaccine against meningococcal disease in sub-Saharan Africa based on Generalized Modules for Membrane Antigens (GMMA)'

A broadly-protective vaccine against meningococcal disease in sub-Saharan Africa based on Generalized Modules for Membrane Antigens (GMMA) Academic research paper on "Biological sciences"

CC BY-NC-ND
0
0
Share paper
Academic journal
Vaccine
OECD Field of science
Keywords
{" Neisseria meningitidis " / Meningococcus / Meningitis / Vaccine / "Outer membrane vesicles" / "Factor H binding protein" / GMMA}

Abstract of research paper on Biological sciences, author of scientific article — Oliver Koeberling, Emma Ispasanie, Julia Hauser, Omar Rossi, Gerd Pluschke, et al.

Abstract Introduction Neisseria meningitidis causes epidemics of meningitis in sub-Saharan Africa. These have mainly been caused by capsular group A strains, but W and X strains are increasingly contributing to the burden of disease. Therefore, an affordable vaccine that provides broad protection against meningococcal disease in sub-Saharan Africa is required. Methods We prepared Generalized Modules for Membrane Antigens (GMMA) from a recombinant serogroup W strain expressing PorA P1.5,2, which is predominant among African W isolates. The strain was engineered with deleted capsule locus genes, lpxL1 and gna33 genes and over-expressed fHbp variant 1, which is expressed by the majority of serogroup A and X isolates. Results We screened nine W strains with deleted capsule locus and gna33 for high-level GMMA release. A mutant with five-fold increased GMMA release compared with the wild type was further engineered with a lpxL1 deletion and over-expression of fHbp. GMMA from the production strain had 50-fold lower ability to stimulate IL-6 release from human PBMC and caused 1000-fold lower TLR-4 activation in Human Embryonic Kidney cells than non-detoxified GMMA. In mice, the GMMA vaccine induced bactericidal antibody responses against African W strains expressing homologous PorA and fHbp v.1 or v.2 (geometric mean titres [GMT]=80,000–200,000), and invasive African A and X strains expressing a heterologous PorA and fHbp variant 1 (GMT=20–2500 and 18–5500, respectively). Sera from mice immunised with GMMA without over-expressed fHbp v.1 were unable to kill the A and X strains, indicating that bactericidal antibodies against these strains are directed against fHbp. Conclusion A GMMA vaccine produced from a recombinant African N. meningitidis W strain with deleted capsule locus, lpxL1, gna33 and overexpressed fHbp v.1 has potential as an affordable vaccine with broad coverage against strains from all main serogroups currently causing meningococcal meningitis in sub-Saharan Africa.

Academic research paper on topic "A broadly-protective vaccine against meningococcal disease in sub-Saharan Africa based on Generalized Modules for Membrane Antigens (GMMA)"

Contents lists available at ScienceDirect

Vaccine

journal homepage www.elsevier.com/locate/vaccine

A broadly-protective vaccine against meningococcal disease in sub-Saharan Africa based on Generalized Modules for Membrane Antigens (GMMA)

Oliver Koeberlinga, Emma Ispasaniebc, Julia Hauserb,c, Omar Rossia, Gerd Pluschke Dominique A. Caugantd, Allan Saula, Calman A. MacLennan

a Novartis Vaccines Institute for Global Health, Siena, Italy b Swiss Tropical and Public Health Institute, Basel, Switzerland c University of Basel, Switzerland d Norwegian Institute of Public Health, Oslo, Norway e University of Birmingham, Birmingham, United Kingdom

ARTICLE INFO

Article history:

Received 6 February 2014

Received in revised form 19 March 2014

Accepted 20 March 2014

Available online 3 April 2014

Keywords:

Neisseria meningitidis Meningococcus Meningitis Vaccine

Outer membrane vesicles Factor H binding protein GMMA

ABSTRACT

Introduction: Neisseria meningitidis causes epidemics of meningitis in sub-Saharan Africa. These have mainly been caused by capsular group A strains, but W and X strains are increasingly contributing to the burden of disease. Therefore, an affordable vaccine that provides broad protection against meningococcal disease in sub-Saharan Africa is required.

Methods: We prepared Generalized Modules for Membrane Antigens (GMMA) from a recombinant serogroup W strain expressing PorA P1.5,2, which is predominant among African W isolates. The strain was engineered with deleted capsule locus genes, lpxL1 and gna33 genes and over-expressed fHbp variant 1, which is expressed by the majority of serogroup A and X isolates.

Results: We screened nine W strains with deleted capsule locus and gna33 for high-level GMMA release. A mutant with five-fold increased GMMA release compared with the wild type was further engineered with a lpxL1 deletion and over-expression of fHbp. GMMA from the production strain had 50-fold lower ability to stimulate IL-6 release from human PBMC and caused 1000-fold lower TLR-4 activation in Human Embryonic Kidney cells than non-detoxified GMMA. In mice, the GMMA vaccine induced bactericidal antibody responses against African W strains expressing homologous PorA and fHbp v.1 or v.2 (geometric mean titres [GMT] = 80,000-200,000), and invasive African Aand X strains expressing a heterologous PorA and fHbp variant 1 (GMT = 20-2500 and 18-5500, respectively). Sera from mice immunised with GMMA without over-expressed fHbp v.1 were unable to kill the A and X strains, indicating that bactericidal antibodies against these strains are directed against fHbp.

Conclusion: A GMMA vaccine produced from a recombinant African N. meningitidis W strain with deleted capsule locus, lpxL1, gna33 and overexpressed fHbp v.1 has potential as an affordable vaccine with broad coverage against strains from all main serogroups currently causing meningococcal meningitis in sub-Saharan Africa.

© 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Neisseria meningitidis is a major cause of epidemics in sub-Saharan Africa [1]. These were mainly caused by strains belonging

* Corresponding author at: Novartis Vaccines Institute for Global Health, Via Fiorentina 1, 53100 Siena, Italy. Tel.: +39 0577 539240; fax: +39 0577 243352.

E-mail addresses: calman.maclennan@novartis.com, c.maclennan@bham.ac.uk (C.A. MacLennan).

to capsular group A, but there has been an increasing contribution of serogroups W and X strains with epidemic potential in the last two decades [2-5]. A serogroup A polysaccharide conjugate vaccine (MenAfriVac) has been developed for preventive mass immunization in the African meningitis belt [6]. The vaccine is highly effective at prevention of serogroup A invasive disease and carriage [7-9], but group W and X strains remain a persistent problem. This underlines the need for an affordable vaccine that provides protection against the main serogroups causing meningitis in Africa and potentially against serogroups that may emerge in the region in the future.

http://dx.doi.Org/10.1016/j.vaccine.2014.03.068 0264-410X/© 2014 Elsevier Ltd. All rights reserved.

GMMA generated from strains engineered to over-express immunogenic antigens that are present across all serogroups, constitute an attractive approach to vaccination. The term GMMA (Generalised Modules for Membrane Antigens) provides a clear distinction from conventional detergent-extracted outer membrane vesicles (dOMV), and native outer membrane vesicle (NOMV), which are released spontaneously from Gram-negative bacteria. GMMA differ in two crucial aspects from NOMV. First, to induce GMMA formation, the membrane structure has been modified by the deletion of genes encoding key structural components, including gna33 (meningococcus) and tolR (Shigella and Salmonella [10]). Second, as a consequence of the genetic modification, large quantities of outer membrane bud off (the Italian word for bud is 'gemma') to provide a practical source of membrane material for vaccine production, leading to potential cost reduction. While NOMV have been used for immunogenicity studies, the yields are too low for practical vaccines.

The most promising candidate protein vaccine antigen discovered for meningococcus is factor H binding protein fHbp. The extraction process required to make dOMV removes lipoproteins, including fHbp, and increases the cost of production of dOMV relative to GMMA. The fHbp gene is present in most invasive meningococcal isolates independent of the serogroup. fHbp can be divided into three antigenic variants (v. 1, 2 or 3) [11] or into at least nine modular groups based on the combination of five variable a and P fHbp segments [12,13]. Individual peptides within each variant are identified by a unique peptide ID. The outer membrane protein, PorA, is highly immunogenic but antibodies tend to provide subtype-specific protection [14]. African meningococcal isolates are relatively conserved in relation to fHbp variant and PorA subtype [15,16]. Invasive serogroup A and X strains predominantly express fHbp v.1. PorA subtype P1.5,2 is shared by most serogroup W strains and P1.20,9 is expressed by the majority of A strains [15]. This epidemiological pattern makes a protein-based vaccine both a possible and attractive approach for sub-Saharan Africa.

A vaccine for the meningitis belt needs to be affordable and large-scale low-cost production of a GMMA vaccine has to be feasible. Deletions of gna33 or rmpM, that augment the release of these outer membrane particles can reduce costs [17-21]. In this study, we selected a vaccine strain based on a panel of African W strain capsule and gna33 double knock-out mutants. The isolate with the highest GMMA production was then further engineered for the deletion of lpxLl and over-expression of fHbp v.1 (lD1).This genetic approach may form the basis for a broadly-protective, safe and economic vaccine for sub-Saharan Africa.

2. Materials and methods

2.1. N. meningitidis strains

Three African serogroup W, seven A and seven X strains were the target strains for serum bactericidal assays. Nine African serogroup W strains were screened as potential vaccine production strains (Table 1). Carrier strain 1630 (ST-11) expressing PorA subvariant P1.5,2 and fHbp v.2 (1D23) was chosen for GMMA production [22]. To abolish capsule production, a fragment of the bacterial chromosome containing synX, ctrA and the promoter controlling their expression, was replaced with a spectinomycin-resistance gene. First, the recombination sites were amplified with primers ctrAf_Xma:CCCCCCGGGCAGGAAAGCGCTGCATAG and ctrAr_XbaCGTCTAGAGGTTCAACGGCAAATGTGC; SynLKpnCGG-GGTACCCGTGGAATGTTTCTGCTCAA and Synr.SpeGGACTAGTCCA-TTAGGCCTAAATGCCTG from genomic DNA from strain 1630. The fragments were inserted into plasmid pComPtac [23] upstream and downstream ofthe chloramphenicol resistance gene.

Subsequently the chloramphenicol resistance gene was replaced with a spectinomycin resistance cassette. The lpxLl gene was deleted by replacement with a kanamycin resistance gene [24], and the gna33 gene with an erythromycin resistance cassette [25]. fHbp expression was up-regulated using multicopy plasmid encoding fHbp v.1 (lD1) [26].

2.2. GMMA preparation

Bacteria were grown at 37 °C, 5% CO2 in 50 mL of a modified version of a meningococcus defined medium described previously [27] at 180 rpm until early stationary phase. Cells were harvested (2200 g, 30min, 4 °C) and the culture supernatant containing the GMMA was filtered through a 0.22 |im pore-size membrane (Mil-lipore, Billerica, MA, USA). To collect GMMA, the supernatant was ultracentrifuged (142,000 x g, 2h, 4 °C). The membrane pellet was washed with phosphate buffered saline (PBS), resuspended in PBS and sterile filtered. GMMA concentration was measured according to protein content by Lowry assay (Sigma-Aldrich, St. Louis, MO, USA). For protein and lipooligosaccharide analysis, GMMA were separated by SDS-PAGE using a 12% gel and MOPS or MES buffer (Invitrogen, Carlsbad, CA, USA). Total proteins were stained with Coomassie Blue stain. The amount of PorA was determined by den-sitometric quantification of the PorA protein in relation to total measurable protein. Lipooligosaccharide was visualized by treatment of the gel with periodic acid and staining with silver nitrate. The gel was developed with a solution containing 50 mg/L citric acid and 0.05% formaldehyde. fHbp was detected by Western blot using a polyclonal antibody raised in mice against recombinant fHbp lD1.

2.3. IL-6 release by human peripheral blood mononuclear cells (PBMC) stimulated with GMMA

PBMC were separated from whole blood using Ficoll-Paque Plus density gradient (Amersham Pharmacia Biotec), washed with PBS and resuspended in 10% heat-inactivated fetal bovine serum (FBS)/10% Dimethyl sulfoxide and stored in liquid nitrogen until use. For stimulation, PBMCs were thawed, washed with PBS/2.5 mM EDTA and 20 |g/mL DNAse (Sigma-Aldrich, St. Louis, MO, USA) and resuspended in RPMl-1640 complete (with 25 mM HEPES, glu-tamine, 10% FBS +1% Antibiotics Pen-Strep). 2 x 105 cells/well were stimulated with GMMA (1-10-6 |g/mL final concentration) for 4 h at 37 °C. Cells were removed by centrifugation and IL-6 in the super-natants was measured by ELISA using 0.1 |g of an anti-human IL-6 antibody (eBioscience, San Diego, CA, USA). A Biotin-labelled antihuman lL-6 antibody was used for detection (e-Bioscience).

2.4. Measurement ofTLR-4 stimulation by NF-kB luciferase reporter assay

Human Embryonic Kidney 293 (HEK293) cells expressing luciferase under control of the NF-kB promoter and stably trans-fected with human Toll-like receptor (TLR) 4, MD2 and CD14 were used. 25,000 cells/well were added to microclear luciferase plates (PBl International) and incubated for 24 h at 37 °C. GMMA (1-1.28 x 10-5 |g/mL final concentration) were added and incubated for 5 h. Cells were separated from the supernatant and lysed with passive lysis buffer (Promega, Madison, Wl, USA). Luciferase assay reagent (Promega) was added and fluorescence was detected using a luminometer LMaxll 384 (Molecular Devices).

2.5. Mouse immunization

Female CD-1 mice were obtained from Charles River Laboratories (Wilmington, MA, USA). Eight mice per group were immunised

Table 1

Characteristics of African N. meningitidis wild type strains used for screening of GMMA production strains and in serum bactericidal assays.

Strain Serogroup Year of isolation Country Source Sequence type fHbp variant3 fHbp peptide lDa (% identity to fHbp lD1) PorA subtype fHbp expression (%)b

N. meningitidis wild type strains used for screening of the GMMA production strain

1485 W 2003 Ghana Carrier 11 2 23 5,2

1630 W 2004 Ghana Carrier 11 2 23 5,2

1629 W 2004 Ghana Carrier 11 2 23 5,2

1681 W 2004 Ghana Case 11 2 23 5,2

1682 W 2004 Ghana Case 11 2 23 5,2

1846 W 2004 Ghana Case 11 2 23 5,2

1888 W 2004 Ghana Carrier 11 2 23 5,2

1973 W 2005 Ghana Carrier 11 2 23 5,2

2882 W 2007 Ghana Carrier 11 2 22 5,2

N. meningitidis wild type strains used in serum bactericidal assays

Mali 10/09 W 2009 Mali Case 11 2 23 (70) 5,2 70

BF2/11 W 2011 Burkina Faso Case 11 1 9 (94) 5,2 42

N1361 A 2002 Ghana Case ND 1 5 (96) 20,9 80

N2008 A 2005 Ghana Case ND 1 5 20,9 75

BF6/06 A 2006 Burkina Faso Case 2859 1 5 20,9 52

N2181 A 2006 Burkina Faso Case ND 1 5 20,9 75

Su14/07 A 2007 Sudan Case 7 1 5 20,9 40

N2602 A 2007 Burkina Faso Case ND 1 5 20,9 76

Mali21/10 A 2010 Mali Case 8639 1 5 20,9 33

BF2/97 X 1997 Burkina Faso Case 751 1 73 (93) 5-1,10-1 50

BF12/03 X 2003 Burkina Faso Case 751 1 73 5-1,10-1 53

MRS2006093 X 2006 Ghana Case 181 1 74 (93) 5-1,10-1 75

Ug9/06 X 2006 Uganda Case 5403 1 74 19,26 200

BF7/07 X 2007 Burkina Faso Case 181 1 74 5-1,10-1 160

Ug11/07 X 2007 Uganda Case 5403 1 74 19,26 4

BF16/10 X 2010 Burkina Faso Case 181 1 74 5-1,10-1 101

a Determined by sequencing of the fHbp gene and analysis of the protein sequence using the N. meningitidis database on http://pubmlst.org/neisseria/fHbp/. b fHbp expression was measured by Western blot of whole cell lysates previously described [1]. Percentage expression is in relation to group B strain H44/76, a relatively high expresser of fHbp v.1. Expression of fHbp v.2 in strain Mali 10/09 is expressed as percentage of fHbp expression in group B strain 8047, a relatively high expresser of fHbp v.2.

intraperitoneally three times with 2 weeks intervals. Serum samples were obtained 2 weeks after the third dose. GMMA from the serogroup W Triple KO (lpxLl, capsule, gna33 KO), OE fHbp strain were given at 0.2, 1 and 5 | g doses based on total protein. Two other groups of mice received 5 | g of GMMA from the Double KO (lpxLl, gna33 KO) OE fHbp mutant or 5 |g GMMA from the Triple KO mutant strain. Control mice were immunised with 5 | g recombinant fHbp lD1 or aluminium hydroxide only. All vaccines were adsorbed on 3 mg/mL Aluminium hydroxide in a 100 |L formulation containing 10 mM Histidine and 0.9 mg/mL NaCl. Sera were stored at -80 °C until use. All animal work was approved by the ltalian Animal Ethics Committee (AEC project number 14112011).

2.6. Serological analysis

Anti-fHbp lgG antibody titres were measured by ELlSA as previously described [28]. The coating antigen was 1 |g/mL non-lipidated recombinant hexa-Histidine-tagged fHbp lD1 [11]. Serial five-fold dilutions of the serum samples starting at 1:100 were analysed. Secondary antibody was a 1:2000 dilution of alkaline phosphatase-conjugated goat-anti mouse lgG (lnvitrogen, cat, no 62-6522, Lot 437983A). The titre was defined as the extrapolated dilution resulting in absorption of 1 at 405 nm after 30 min of incubation with 1 mg/mL 4-nitrophenyl phosphate disodium salt hexahydrate (Sigma-Aldrich) diluted in 1 M diethanolamine and 0.5 mM MgCl2, pH 9.8.

Serum bactericidal antibody (SBA) activities were measured as described before [28]. Bacteria were incubated at 37 °C, 5% CO2 in Mueller-Hinton broth containing 0.25% glucose and 0.02 mM Cytidine-5'-monophospho-N-acetylneuraminic acid sodium salt (Sigma-Aldrich). The cells were washed with Dulbecco's PBS buffer (Sigma-Aldrich) containing 1% BSA. Each reaction mixture contained approximately 400 colony-forming units, 20% human

complement screened for lack of bactericidal activity against the target strain and serial dilutions of the serum samples starting at 1:10. Bactericidal titres were defined as the reciprocal extrapolated dilution resulting in 50% killing of bacteria after 60 min incubation at 37 °C compared to the mean number of bacteria in five control reactions at time 0.

2.7. Statistical analysis

For statistical analysis, antibody titres were log 10 transformed. ELlSA titres <100 were assigned the value 50, SBA titres <10 were assigned the value 5. Mann-Whitney U test was used to compare pairs of values. A probability value of <0.05 was considered statistically significant. The analysis was performed with the Graph Pad Prism software 5.01.

3. Results

3.1. Selection of the serogroup W GMMA production strain and generation of mutants

Nine group W strains (six carrier and three case isolates) with PorA subtype P1.5,2, collected in Ghana between 2003 and 2007, were screened as candidate GMMA production strains. To identify the isolate with highest GMMA production, gna33 was deleted from all strains. ln some isolates, simultaneous deletion of the capsule decreased the GMMA release compared to the gna33 single knock-out (KO). Therefore, we generated gna33 and capsule double KO mutants of the nine W strains and compared GMMA production. These double-mutant strains released two to five-fold higher amounts of GMMA than a representative group W wild type strain (Fig. 1A). Strain 1630 (gna33 KO, capsule KO), which released the

Table 2

N. meningitidis vaccine strains and GMMA vaccines used in the study.

Vaccine strain characteristics

Fig. 1. GMMA release by engineered meningococcal W strains A. W strains engineered to have deleted capsule and gna33 KO were grown in small-scale shake-flasks. Strain 1630 was selected as the GMMA-production strain for further studies. WT= GMMA release by a representative wild type strain. Bars indicate the mean and standard error of three independent experiments. B. Upper panel: SDS-PAGE and Coomassie Blue stain of 5 |xg GMMA. Middle panel: Silver stain of

Designation of vaccine strain and GMMA used for immunization

Prototype vaccine candidate Serogroup W, strain 1630 Capsule KO, Ipxil KO, gna33 KO, over-expressed fHbp 1D1

Control vaccines Serogroup W, strain 1630 Capsule expressed, Ipxil KO, gna33 KO, over-expressed fHbp 1D1 Serogroup W, strain 1630 Capsule KO, Ipxil KO, gna33 KO

Triple KO, OE fHbp

Double KO, OE fHbp

Triple KO

highest quantity of GMMA, was selected for further genetic manipulation.

To generate the final vaccine strain, we deleted IpxLI and engineered the mutant to over-express fHbp v.1, designated 'Triple KO, OE fHbp'. We also prepared two isogenic group W control strains: one with deleted IpxLI and gna33, over-expressed fHbp v.1 with the capsule still expressed ('Double KO, OE fHbp'), and another with deleted IpxLI, capsule and gna33, but no fHbp overexpression ('Triple KO') (Table 2). SDS-PAGE and Coomassie Blue staining of the proteins revealed a similar protein pattern in the three GMMA preparations. Densitometry indicated that in all three GMMA preparations, the relative amount of PorA to total protein is 5%. By silver stain, the GMMA contained similar levels of lipooligosaccharide. By capture ELISA, with recombinant fHbp as standard, approximately 3% of the total protein in GMMA from the Triple KO, OE fHbp was fHbp, and by Western blot, the two GMMA over-expressing fHbp had similar fHbp levels.

3.2. IL-6 release by human PBMCand TLR-4 activation in HEK293 cells after stimulation with GMMA

To assess the endotoxic activity of the GMMA, we measured the release of IL-6 by human PBMC after stimulation with different concentrations of GMMA from the Triple KO, OE fHbp mutant and the parent serogroup W wild type strain (Fig. 1C). Approximately 50-fold higher concentrations of GMMA from the mutant strain were required to stimulate the release of 200 pg/mL IL-6, confirming the decrease in endotoxic activity. We measured the ability of the GMMA to stimulate human TLR-4 in transfected HEK293 cells (Fig. 1D). Low concentrations of GMMA from the wild type bacteria stimulated TLR-4, as measured by increased NF-kB expression. Approximately 1000-fold higher concentrations of GMMA from the Triple KO, OE fHbp mutant were required for equivalent TLR-4 stimulation. These results are consistent with a strongly decreased ability of the LOS in GMMA from the serogroup W mutant to

lipooligosaccharide in 0.5 |xg GMMA. Lower panel: detection of fHbp in GMMA by Western blot using a polyclonal anti-fHbp v.1 antibody. M = molecular weight marker. Lane 1, GMMA Triple KO, OE fHbp; lane 2, GMMA Double KO, OE fHbp; lane 3, GMMA Triple KO. C. IL-6 release by human PBMCs stimulated with different concentrations of GMMA with deleted capsule, gna33 and lpxLI and over-expressed fHbp v.1 for four hours. IL-6 release into culture supernatants was analysed by ELISA. D. Stimulation of TLR-4 in HEK293 cells transfected with human Toll-like receptor (TLR) 4, MD2 and CD14 and luciferase expressed under control of the NF-kB promoter. Cells were stimulated with GMMA for five hours, lysed and emitted light was quantitated with a luminometer. The readings were divided by the control cells stimulated with PBS. Mean results and standard deviations from two independent experiments were plotted. Black circles = GMMA from the group W wild type strain 1630 used to construct the mutants. White triangles = Triple KO, OE fHbp: GMMA from the group W mutant strain with deleted capsule, lpxL1 and gna33 and over-expressed fHbp ID1.

Fig. 2. IgG anti-fHbp antibody responses elicited in mice as measured by ELISA. Groups of eight mice were immunised with three doses of vaccine, 2 weeks apart. The serum samples analysed were obtained 2 weeks after the third dose. Each symbol represents an individual serum sample, the line indicates the geometric mean titre of each vaccine group. GMMA used for immunization: Triple KO, OE fHbp: capsule, IpxLI and gna33 KO with over-expressed fHbp ID1. Double KO, OE fHbp: IpxLI and gna33 KO with over-expressed fHbp ID1 and capsule expression. Triple KO: capsule, IpxLI and gna33 KO without over-expressed fHbp. rHis-fHbp: recombinant hexa-histidine tagged fHbp ID1; Alum: aluminium hydroxide. Numbers above the x-axis show the vaccine dose in |xg. Statistical analysis between pairs of groups was by Mann-Whitney U test.

activate TLR-4 compared with GMMA from the non-detoxified parent wild type strain.

3.3. Antibody responses elicited in mice immunised with GMMA

We measured anti-fHbp v.1 antibody responses in individual serum samples by ELISA. GMMA from all mutants with over-expressed fHbp elicited high anti-fHbp antibody responses, even at the lowest dose of 0.2 |g (Fig. 2). 5 |g Triple KO, OE fHbp GMMA induced significantly higher geometric mean titres than 5 |ig Double KO, OE fHbp GMMA (P =0.03) or 5 |g of recombinant fHbp v.1 (P<0.001). GMMA from the Triple KO mutant without fHbp overexpression induced no measurable anti-fHbp antibody responses.

Fig. 3. Serum bactericidal antibody responses of immunised mice against African meningococci group W (panel A), group A (panel B) and group X strains (panel C) measured with human complement. Group A and X strains were ordered based on their relative fHbp expression, increasing from the left to the right. Serum samples analysed were obtained 2 weeks after the third dose (see legend to Fig. 4). The bars show reciprocal geometric mean titres (±95% confidence interval) from four serum samples per GMMA vaccine group, containing sera from two mice each. For the strains labelled with an asterisk and the negative control group (Alum) two serum pools were analysed containing sera from four mice each and error bars indicate the standard error of the mean. GMMA vaccines: Triple KO, OE fHbp: capsule, lpxLI and gna33 KO with over-expressed fHbp ID1, hatched bars. Triple KO: capsule, lpxLI and gna33 KO without over-expressed fHbp, grey bars. rHis-fHbp: recombinant hexa-Histidine tagged fHbp ID1, white bars; Alum: aluminium hydroxide, black bars. Dotted lines indicate the lowest serum dilution tested (1:10).

3.4. SBA responses of mice immunised with GMMA from recombinant serogroup W strains

The three serogroup W test strains were isolated in Ghana, Mali and Burkina Faso and expressed PorA subtype P1.5,2, which is identical to that expressed by the GMMA vaccine strains. Strain BF2/11 expressed fHbp v.1 (1D9) and the two other strains expressed fHbp v.2 (1D23). The seven group A strains tested were collected in Ghana, Burkina Faso, Sudan and Mali. They expressed a heterologous PorA compared to that in the GMMA, and fHbp v.1 (1D5). fHbp expression in these test strains ranged from 33 to 80% of that of a reference group B strain H44/76 with relatively high fHbp expression [11]. The seven group X strains were isolated in Burkina Faso, Ghana and Uganda. Two strains from Burkina Faso expressed fHbp 1D73, the other isolates expressed 1D74. The strains from Burkina Faso were sequence type 751 and 181, respectively. The two strains from Uganda were ST5403 and expressed PorA subtype P1.19,26, while the other five group X strains were P1.5-1,10-1. The two strains from Uganda differed from each other by the level of fHbp

expression. Strain Ug11/07 had 4% and Ug9/06 has 200% of the fHbp expression level compared to the reference strain (Table 1).

GMMA with or without fHbp over-expression elicited high bactericidal titres that were not significantly different from each other against the three W strains expressing either fHbp v.1 or v.2 (Fig. 3A). This is consistent with previous observations that bactericidal activity against strains sharing the same PorA as the GMMA-production strain is predominantly mediated by anti-PorA antibodies [26].

GMMA from the Triple KO, OE fHbp strain induced antibodies that were able to kill six out of seven serogroup A strains (geometric mean titres [GMT] ranging from 20 to 2500) (Fig. 3B). The only isolate that was resistant to killing was readily killed by a mouse serum raised against group A polysaccharide conjugate vaccine. The antibodies induced by the GMMA from the Triple KO, OE fHbp strain were able to kill all serogroup X strains tested (GMT = 18-5500) (Fig. 3C). GMMA produced from the W strain which lacked fHbp v.1 over-expression (Triple KO), induced antibodies that were only able to kill one X strain (BF7/07), consistent with the majority of

Fig. 4. Serum bactericidal antibody responses induced by meningococcal W GMMA A. Dose-dependent serum bactericidal antibody responses induced by group W triple KO, OE fHbp GMMA (capsule, lpxLl and gna33 KO, over-expressed fHbp lD1) measured with human complement. Mice were immunised three times 2 weeks apart with 0.2|g (white bars), 1 |g (hatched bars) or 5 |g (squared bars). Serum samples were obtained 2 weeks after the third dose and four pools containing sera from two mice each were measured against African serogroup W strain 1630, serogroup A strain N2602 and serogroup X strain BF7/07. The dotted line indicates the lowest serum dilution tested (1:10). Spearman Rank test was used for the statistical analysis of the correlation between the dose and geometric mean titres of each strain. B. Serum bactericidal antibody responses elicited by GMMA from W with or without expression of capsule against one African serogroup W, two serogroup A and three serogroup X strains. GMMA vaccine groups: Triple KO, OE fHbp: capsule, lpxLl and gna33 KO with over-expressed fHbp lD1, checked bars. Double KO, OE fHbp: lpxLl and gna33 KO with over-expressed fHbp lD1 and capsule expression, white bars. Statistical analysis between pairs of groups was done by Mann-Whitney U test and asterisks indicate probability values of <0.05. The bars show reciprocal geometric mean titres (±standard error of the mean) from four serum samples per GMMA vaccine group, containing sera from two mice each.

bactericidal antibodies induced by the GMMA vaccine being directed against fHbp. Antibodies made against the recombinant fHbp lD1 were only bactericidal against serogroup X strain Ug9/06 with the highest fHbp expression.

We investigated the dose-dependent bactericidal antibody response against one W (1630), A (N2602) and X (BF7/07) isolate (Fig. 4A). Sera raised against GMMA with over-expressed fHbp were bactericidal against these strains in a dose-dependent manner (Spearman Rank P = 0.001 for group A and P< 0.0001 for group W and X) with killing occurring at all three doses (0.2,1 and 5 |g). GMMA from the triple KO, OE fHbp mutant was prepared from a mutant with deleted capsule expression in order to attenuate virulence of the vaccine strain and reduce serogroup-specific antibody production. To test the latter, we investigated whether maintaining capsule expression in the GMMA-producing strain affects the bactericidal antibody response. Sera from mice immunised with GMMA prepared from the Triple KO, OE fHbp vaccine strain had significantly higher SBA activity against three of five A and X strains tested than GMMA from the isogenic mutant that expressed the

capsule (Fig. 4B). These data are consistent with the hypothesis that deletion of capsule biosynthesis in the GMMA-production strain not only decreases virulence, but also increases antibody responses towards non-capsular antigens, such as fHbp.

4. Discussion

The group A polysaccharide conjugate vaccine, MenAfriVac, is highly effective at prevention of serogroup A invasive disease and carriage [7-9]. However, other serogroups, in particular W and more recently X, are increasingly contributing to the burden of meningococcal disease in sub-Saharan Africa [3,29-32]. Additionally, other meningococcal serogroups, e.g. group C, that, although not having caused outbreaks in recent years, may become a threat in the future. The challenge for future vaccine approaches for the meningitis belt is to develop a meningococcal vaccine that is not only affordable, but provides broad cross-serogroup protection against meningococcus, and complements the roll out pneumo-coccal vaccination to deal with the problem of pneumococcal meningitis in the region.

GMMA from recombinant meningococcal strains offer a promising option. They contain protein antigens (e.g. fHbp) which induce antibodies with serogroup independent cross protection. ln addition, a simple, economic and scalable procedure for their preparation has been developed with minimal downstream processing required, which enables large quantities of GMMA vaccine to be produced at low cost [10]. While strains containing deletions of lpxLl and capsule synthesis genes with up-regulated fHbp expression have been described [33,34], our approach incorporates the additional deletion of gna33 in order to enhance the level of GMMA production, and consequently the potential affordability of the vaccine for use in Africa. The mechanism of up-regulation of GMMA production is not fully understood. Our findings indicate that GMMA release by different gna33 KO strains is variable, indicating a requirement to screen multiple strains for high level GMMA release.

We tested bactericidal activity of sera from immunised mice against 17 group A, W and X strains. Five | g of the GMMA from the Triple KO, OE fHbp group W strain induced SBA responses against 16 (94%) of these isolates. Ability to kill the A and X strains was attributable to fHbp which comprises only about 3% of the total GMMA protein. ln comparison, 5 |g recombinant fHbp lD1 induced a detectable bactericidal antibody response only against one X strain which had the highest level of fHbp expression. This is consistent with previous studies with NOMV demonstrating that fHbp expressed in the native membrane environment induces antibodies with greater functional activity than vaccines containing recombinant fHbp [15,35,36].

Previous studies have demonstrated broad cross-protection of NOMV vaccines against a panel of diverse African strains [15,34,37]. We did not compare our GMMA vaccine directly with NOMV. Nevertheless, the strong bactericidal activity of the GMMA-induced antibodies against strains with homologous or heterologous PorA and different fHbp lD types (lD 5, 73 and 74), suggests that the new combination of mutations, including deletion of gna33, that all affect the outer surface, does not impair the immunogenicity of the main antigens, fHbp and PorA. lt has been shown that decreased SBA titres are induced when mice expressing human factor H are immunised with NOMV over-expressing wild type fHbp [38]. This can be overcome by introducing the R41S mutation into the fHbp gene of the vaccine-producing strain [38,39]. The aim of the current study was to serve as a first proof of concept in mice for a GMMA meningococcal candidate vaccine and the R41S mutation was not incorporated into our vaccine design. We are currently investigating the utility of this mutation in GMMA vaccines.

For safety and immunological reasons, we engineered the vaccine strain to have deleted lpxLI and be non-encapsulated which is associated with the inability to cause invasive disease [ 40]. As described for group B strains, deletion of lpxLI resulted in decreased ability of the group W GMMA to stimulate Il-6 release by human PBMC and activate TLR-4. These data indicate that genetic detoxification of meningococcal LOS by inactivation of lpxLI is a common mechanism among different serogroups.

Consistent with our hypothesis that removal of the capsule would enhance the level of bactericidal activity induced against non-W serogroups, GMMA produced by the non-encapsulated mutant W strain induced higher bactericidal titres against A and X strains, than the isogenic encapsulated control. The underlying mechanisms require further investigation. Capsular polysaccharide on GMMA may mask fHbp epitopes from the immune system, particularly from fHbp-specific B cells. An alternative explanation is that capsular polysaccharide on GMMA may serve as an anti-genic competitor, interfering and decreasing the immune response to common protein antigens such as fHbp, although addition of external group A polysaccharide conjugate did not impair antibody responses to protein antigens in a meningococcal NOMV vaccine [ 34]. Thermostability is also highly desirable for any new vaccine targeted at the African meningitis belt and we are currently investigating this quality in our GMMA vaccine.

In conclusion, the findings of this study provide support for a GMMA-based vaccine approach as an affordable and broadly-protective vaccine strategy against meningococcal meningitis for Africa.

Conflict of interest

OK, OR, AS and CAM are employees of the Novartis Vaccines Institute for Global Health. CAM is the recipient of a clinical research fellowship from GlaxoSmithKline.

Acknowledgements

We thank Dan Granoff, Children's Hospital Oakland Research Institute, Oakland, USA for providing plasmid pFP12-fHbp and Ugo DOro, Novartis Vaccines, Siena, Italy for providing TLR4-expressing HEK293 cells. This work was supported by a European Union FP7 Industry and Academia Partnerships and Pathways award, GENDRIVAX (Genome-driven vaccine development for bacterial infections). This is a collaboration between the Novartis Vaccines Institute for Global Health, Swiss Tropical and Public Health Institute, Kenyan Medical Research Institute and Wellcome Trust Sanger Institute and [ grant number 251522]. The funding source had no involvement in the study design; in the collection, analysis and interpretation of the data; in the writing of the report; or in the decision to submit the article for publication.

References

[ 1] Halperin SA, BettingerJA, Greenwood B, Harrison LH, Jelfs J, Ladhani SN, et al. The changing and dynamic epidemiology of meningococcal disease. Vaccine 2012;30(Suppl. 2):B26-36.

2] Laforce FM, Okwo-Bele JM. Eliminating epidemic Group A meningococ-cal meningitis in Africa through a new vaccine. Health Aff (Millwood) 2011;30:1049-57.

[ 3] Xie O, Pollard AJ, Mueller JE, Norheim G. Emergence of serogroup X meningococcal disease in Africa: need for a vaccine. Vaccine 2013;31:2852-61.

[ 4] Collard JM, Maman Z, Yacouba H, Djibo S, Nicolas P, Jusot JF, et al. Increase in Neisseria meningitidis serogroup W135, Niger, 2010. Emerg Infect Dis 2010;16:1496-8.

[ 5] Delrieu I, Yaro S, Tamekloe TA, Njanpop-Lafourcade BM, Tall H, Jaillard P, et al. Emergence of epidemic Neisseria meningitidis serogroup X meningitis in Togo and Burkina Faso. PLoS One 2011;6:e19513.

[ 6] Marc LF, Ravenscroft N, Djingarey M, Viviani S. Epidemic meningitis due to Group A Neisseria meningitidis in the African meningitis belt: a persistent problem with an imminent solution. Vaccine 2009;27(SuppI. 2):B13-9.

[ 7] Daugla DM, Gami JP, Gamougam K, Naibei N, Mbainadji L, Narbe M, et al. Effect of a serogroup A meningococcal conjugate vaccine (PsA-TT) on serogroup A meningococcal meningitis and carriage in Chad: a community study [ corrected]. Lancet 2014:383:40-7.

[8] Kristiansen PA, Ba AK, Sanou I, Ouedraogo AS, Ouedraogo R,Sangare L, et al. Phe-notypic and genotypic characterization of meningococcal carriage and disease isolates in Burkina Faso after mass vaccination with a serogroup a conjugate vaccine. BMC Infect Dis 2013:13:363.

[9] Kristiansen PA, Diomande F, Ba AK, Sanou I, Ouedraogo AS, Ouedraogo R, et al. Impact of the serogroup A meningococcal conjugate vaccine, MenAfriVac, on carriage and herd immunity. Clin Infect Dis 2013:56:354-63.

10] Berlanda SF, Colucci AM, Maggiore L, Sanzone S, Rossi O, Ferlenghi I, et al. High yield production process for Shigella outer membrane particles. PLoS One 2012;7:e35616.

11] Masignani V, Comanducci M, Giuliani MM, Bambini S, Adu-Bobie J, Arico B, et al. Vaccination against Neisseriameningitidis using three variants ofthe lipoprotein GNA1870. J Exp Med 2003:197:789-99.

12] Beernink PT, Granoff DM. The modular architecture of meningococcal factor H-binding protein. Microbiology 2009:155:2873-83.

13] Pajon R, Beernink PT, Harrison LH, Granoff DM. Frequency of factor H-binding protein modular groups and susceptibility to cross-reactive bactericidal activity in invasive meningococcal isolates Vaccine 2010:28:2122-9.

14] Tappero JW, Lagos R, Ballesteros AM, Plikaytis B, Williams D, Dykes J, et al. Immunogenicity of 2 serogroup B outer-membrane protein meningococcal vaccines: a randomized controlled trial in Chile. JAMA 1999:281:1520-7.

15] Pajon R, Fergus AM, KoeberlingO, Caugant DA, Granoff DM. Meningococcal fac-torH binding proteins in epidemic strains from Africa: implications for vaccine development. PLoS NeglTrop Dis 2011:5:e1302.

16] Huber CA, Pfluger V, Hamid AW, Forgor AA, Hodgson A, Sie A, et al. Lack of antigenic diversification of major outer membrane proteins during clonal waves of Neisseria meningitidis serogroup A colonization and disease. Pathog Dis 2013:67:4-10.

17] Ferrari G, Garaguso I, Adu-Bobie J, Doro F, Taddei AR, Biolchi A, et al. Outer membrane vesicles from group B Neisseria meningitidis delta gna33 mutant: proteomic and immunological comparison with detergent-derived outer membrane vesicles. Proteomics 2006:6:1856-66.

18] Berlanda SF, Doro F, Rodriguez-Ortega MJ, Stella M, Liberatori S, Taddei AR, et al. Proteomics characterization of outer membrane vesicles from the extraintestinal pathogenic Escherichia coli DeltatolR IHE3034 mutant. Mol Cell Proteomics 2008:7:473-85.

19] Henry T, Pommier S, Journet L, Bernadac A, Gorvel JP, Lloubes R. Improved methods for producing outer membrane vesicles in Gram-negativ bacteria. Res Microbiol 2004:155:437-46.

20] van de Waterbeemd B, Streefland M, van der Ley P, Zomer B, van DH, Martens D, et al. Improved OMV vaccine against Neisseria meningitidis using genetically engineered strains and a detergent-free purificatio process. Vaccine 2010:28:4810-6.

21] Grizot S, Buchanan SK. Structure of the OmpA-like domain of RmpM from Neisseriameningitidis. Mol Microbiol 2004:51:1027-37.

22] Forgor AA, Leimkugel J, Hodgson A, Bugri A, Dangy JP, Gagneux S, et al. Emergence of W135 meningococcal meningitis in Ghana. Trop Med Int Health 2005:10:1229-34.

23] Ieva R, Alaimo C, Delany I, Spohn G, Rappuoli R, Scarlato V. CrgA is an inducible LysR-type regulator of Neisseria meningitidis, acting both as a repressor and as an activator of gene transcription. J Bacteriol 2005:187:3421-30.

24] Koeberling O, Seubert A, Granoff DM. Bactericidal antibody responses elicited by a meningococcal outer membrane vesicle vaccine with overexpressed factor H-binding protein and genetically attenuated endotoxin. J Infect Dis 2008:198:262-70.

25] Adu-Bobie J, Lupetti P, Brunelli B, Granoff D, Norais N, Ferrari G, et al. GNA33 of Neisseria meningitidis is a lipoprotein required for cell separation, membrane architecture, and virulence. Infect Immun 2004:72:1914-9.

26] Hou VC, Koeberling O, Welsch JA, Granoff DM. Protective antibody responses elicited by a meningococcal outer membrane vesicle vaccine with over-expressed genome-derived neisserial antigen 1870. J Infect Dis 2005:192: 580-90.

27] Egen RC, Fortin LA, Sun WWQ.Animal component free meningococcal polysaccharide fermentation and seedbank development. Patent No. US 7,399,615 B2: 2008.

28] Beernink PT, ShaughnessyJ, Ram S, Granoff DM. Impaired immunogenicity of a meningococcal factor H-binding protein vaccine engineered to eliminate factor h binding. Clin Vaccine Immunol 2010:17:1074-8.

29] Massenet D, Inrombe J, Mevoula DE, Nicolas P. Serogroup W135 meningococcal meningitis, Northern Cameroon, 2007-2008. Emerg Infect Dis 2009:15:340-2.

30] Koumare B, Ouedraogo-Traore R, Sanou I, Yada AA, Sow I, Lusamba PS, et al. The first large epidemic of meningococcal disease caused by serogroup W135, Burkina Faso, 2002. Vaccine 2007:25(Suppl. 1):A37-41.

31] Boisier P, Nicolas P, Djibo S, Taha MK, Jeanne I, Mainassara HB, et al. Meningococcal meningitis: unprecedented incidence of serogroup X-related cases in 2006 in Niger. Clin Infect Dis 2007:44:657-63.

32] Gagneux SP, Hodgson A, Smith TA, Wirth T, Ehrhard I, Morelli G, et al. Prospective study of a serogroup X Neisseria meningitidis outbreak in northern Ghana. J Infect Dis 2002:185:618-26.

[33] Zollinger WD, Donets MA, Schmiel DH, Pinto VB, Labrie J, Moran EE. Design and evaluation in mice of a broadly protective meningococcal group B native outer membrane vesicle vaccine. Vaccine 2010;28:5057-67.

[34] Pajon R, Fergus AM, Granoff DM. Mutant native outer membrane vesicles combined with a serogroup A polysaccharide conjugate vaccine for prevention of meningococcal epidemics in Africa. PLoS One 2013;8: e66536.

[35] Koeberling O, Delany I, Granoff DM. A critical threshold of meningococcal factor H binding protein expression is required for increased breadth of protective antibodies elicited by native outer membrane vesicle vaccines. Clin Vaccine Immunol 2011;18:736-42.

[36] Koeberling O, Giuntini S, Seubert A, Granoff DM. Meningococcal outer membrane vesicle vaccines derived from mutant strains engineered to express factor H binding proteins from antigenic variant groups 1 and 2. Clin Vaccine Immunol 2009;16:156-62.

[37] Beernink PT, Caugant DA, Welsch JA, Koeberling O, Granoff DM. Meningococcal factor H-binding protein variants expressed by epidemic capsular group A, W-135, and X strains from Africa. J Infect Dis 2009;199:1360-8.

[38] Beernink PT, Shaughnessy J, Braga EM, Liu Q, Rice PA, Ram S, et al. A meningococcal factor H binding protein mutant that eliminates factor H binding enhances protective antibody responses to vaccination. J Immunol 2011;186:3606-14.

[39] Beernink PT, Shaughnessy J, Pajon R, Braga EM, Ram S, Granoff DM. The effect of human factor H on immunogenicity of meningococcal native outer membrane vesicle vaccines with over-expressed factor H binding protein. PLoS Pathog 2012;8:e1002688.

[40] Vogel U, Weinberger A, Frank R, Muller A, Kohl J, Atkinson JP, et al. Complement factor C3 deposition and serum resistance in isogenic capsule and lipooligosac-charide sialic acid mutants of serogroup B Neisseria meningitidis. Infect Immun 1997;65:4022-9.