Serum Antibody Responses in Ethiopian Meningitis Patients Infected with Neisseria meningitidis Serogroup A Sequence Type 7

Abstract

To elucidate critical components of protective immune responses induced during the natural course of serogroup A meningococcal disease, we studied acute-, early-convalescent-, and late-convalescent-phase sera from Ethiopian patients during outbreaks in 2002 to 2003. Sera were obtained from laboratory-confirmed patients positive for serogroup A sequence type 7 (ST-7) meningococci (A:4/21:P1.20,9) (n = 71) and from Ethiopian controls (n = 113). The sera were analyzed using an enzyme-linked immunosorbent assay to measure levels of immunoglobulin G (IgG) against serogroup A polysaccharide (APS) and outer membrane vesicles (OMVs) and for serum bactericidal activity (SBA) using both rabbit and human complement sources. Despite relatively high SBA titers and high levels of IgG against APS and OMVs in acute-phase patient sera, significant increases were seen in the early convalescent phase. Antibody concentrations returned to acute-phase levels in the late convalescent phase. Considering all patients' sera, a significant but low correlation (r = 0.46) was observed between SBA with rabbit complement (rSBA) using an ST-5 reference strain and SBA with human complement (hSBA) using an ST-7 strain from Ethiopia. While rSBA demonstrated a significant linear relation with IgG against APS, hSBA demonstrated significant linear relationships with IgG against both APS and OMV. This study indicates that antibodies against both outer membrane proteins and APS may be important in providing the protection induced during disease, as measured by hSBA. Therefore, outer membrane proteins could also have a role as components of future meningococcal vaccines for the African meningitis belt.

Neisseria meningitidis is responsible for recurring epidemics of bacterial meningitis in the region of sub-Saharan Africa designated the meningitis belt (36). Most of the cases in this region are caused by serogroup A meningococci (MenA), although serogroups W135, C, and X are also involved (42). Molecular epidemiological studies of meningococcal strains have shown that a few complexes of related hypervirulent clones are responsible for the major part of the cases in the meningitis belt (42). Most recent serogroup A epidemics have been caused by a clonal group introduced to the meningitis belt from Mecca, Saudi Arabia, in 1987: the subgroup III/sequence type 5 (ST-5) clonal complex (67). Strains belonging to this complex have been very homogenous; nearly all expressed the same PorA (serosubtype P1.20,9) and PorB (serotype 4/21) (59, 62). Between 1988 and 1999, ST-5 reached all the countries of the meningitis belt, where it was responsible for severe epidemics (42). In the mid-1990s, bacteria of the closely related ST-7 emerged in Africa and progressively replaced the ST-5 strains (42). This replacement reflects a significant genetic and epidemiological change (41, 45).

Ethiopia is part of the meningitis belt, and many epidemics have been reported in the country since 1901 (45). In Ethiopia, the replacement of ST-5 by ST-7 strains occurred between 1995 and 2000. We recently showed that Ethiopian ST-5 and ST-7 strains differed in several loci associated with outer membrane antigens (45). These changes could be relevant for explaining the clonal replacement.

An effective polysaccharide (PS) vaccine that could prevent MenA disease in Africa has been available for nearly 3 decades (18, 46). However, the WHO does not recommend this vaccine for routine immunization, and mass vaccination is performed only in response to outbreaks, although there has been considerable dispute regarding this decision (1, 52). The new MenA conjugate vaccines (27) will hopefully provide long-term protection even when given to children below the age of 2 years. However, the immunoglobulin G (IgG) response and serum bactericidal activity (SBA) following immunization with MenA conjugate do not differ much from those observed with the MenA PS (APS) vaccine (14). Dissecting the humoral immune response following disease may contribute to a better understanding of the parameters that are important in relation to the development of improved vaccines.

A number of studies have determined the antibody response induced by MenA disease (2, 9, 10, 13, 17, 23, 31, 35, 50, 56, 58). In general, high levels of antibody against APS are normally observed in the population (40), but such levels may be observed in acute-phase sera from MenA patients as well (31, 58). Whereas anti-APS IgG antibodies can confer protection against MenA disease, high levels of anti-APS IgA antibodies in predisease or acute-phase sera may have a blocking effect on SBA and may be related to increased risk of MenA infection (20, 31). Despite preexistent high levels, IgG and IgM against APS increase significantly during early convalescent phase in a large proportion of MenA patients to levels of the same magnitude as those obtained following vaccination with APS vaccine (9, 10, 31, 56, 58). The antibodies induced by disease can be directed against APS, outer membrane proteins, lipooligosaccharide (LOS), and secreted proteins such as, e.g., IgA1 protease (9, 10, 13, 58), but these are not necessarily bactericidal (2, 23). Very few studies have characterized the noncapsular antibody response and the SBA mounted during disease, and the number of patients and the duration of follow-up were usually limited.

The objective of this study was to characterize the humoral response following MenA disease caused by subgroup III ST-7 meningococci in Ethiopia in 2002 and 2003. The antibody responses against APS and outer membrane vesicles (OMVs) were quantified by an enzyme-linked immunosorbent assay (ELISA), and the functional activity was measured by an SBA assay using two different complement sources (human and rabbit serum) with two different serogroup A strains.

(Part of this work was presented at the 15th International Pathogenic Neisseria Conference in Cairns, Australia, September 2006.)

MATERIALS AND METHODS

Patients.

Ninety-five meningitis patients presenting at two study areas in North Gondar Zone, Amhara Region, and Sidama and Gedio Zones, Southern Nations, Nationalities, and Peoples' Region (SNNPR), Ethiopia, were included in the study spanning April 2002 to June 2003, as described previously (45). Patients were recruited consecutively in the first meningitis season, until at least eight patients were included in each of the following intended age groups: infants (≥6 months to ≤2 years), young children (>2 to ≤6 years), older children and teenagers (>6 to ≤15 years), and adults (>15 years). During the second season, recruitment focused on including children below the age of 2 years, because too few had been included in the first season. Of 95 patients, 71 were laboratory confirmed with meningococcal meningitis (45) (Table 1). An attempt was made to obtain blood samples in the acute phase (0 to 7 days after the onset of disease), the early convalescent phase (8 days to 6 weeks), and the late convalescent phase (>72 days after the onset of disease). The date of onset of disease was defined as the date of the first severe symptom related to the meningitis disease episode reported by the patient. Late-convalescent-phase blood samples (range, 70 to 610 days after the onset of disease) were obtained through house visits by study teams seeking the patients' dwellings. Unless otherwise stated, the patients' sera described are those from patients confirmed positive for N. meningitidis by culture or PCR (Table 1). Eleven confirmed meningococcal disease patients were reported to have been vaccinated with APS vaccine previously; of these, two reported that they had been vaccinated 3 months prior to admission, seven reported that they had been vaccinated 8 to 12 months prior to admission, and two did not report the time of vaccination (45). Sera from patients not confirmed with N. meningitidis by culture or PCR (Table 1) (45) were included in the analysis for the purpose of later serological diagnostic confirmation.

TABLE 1.

Number of individuals, sera available for analysis, and demographic characteristics for Ethiopian patients with systemic MenA disease, nonconfirmed meningococcal disease patients, and controls

Patient category	No. of patients	Total no. of sera	No. of sera from the following age group (yr):				Median age (yr) (range)	Gendera (no. of females/males)
			<2	≥2-6	≥6-15	≥15
MDb	71	136	5	22	57	52	14 (0.7-50)	32/38
Nonconfirmed MD	24	34	5	3	5	21	14 (0.5-45)	7/14
Control	113	113	20	18	21	54	14 (0.8-68)	49/64

^aValues reflect the numbers of patients for whom data were available.

^bMD, meningococcal disease.

Controls.

Control blood samples were obtained from apparently healthy individuals and from patients with nonmeningococcal disease in Ethiopia. Controls were recruited from hospitals or health centers in Gondar in 2003 (n = 35; median age, 5.6 years; range, 0.8 to 55 years) and in SNNPR in 2002 (n = 39; median age, 28 years; range, 10 to 68 years) and 2003 (n = 39; median age, 12 years; range, 1 to 25 years) (Table 1). Twenty-one individuals from SNNPR had previously received the MenA/C PS vaccine. Of these, 4 had been vaccinated 5 to 11 months prior to sampling, 16 had been vaccinated 1 or more years previously, and the time of vaccination was not known for 1 individual. Forty-four sera from Norwegian high school students (median age, 13.5 years; range, 12.7 to 13.9 years) were also included; these were prevaccination sera from a clinical trial in Norway in 1992 (15). Unless otherwise stated, the control sera described are those from Ethiopia, for which data are given in Table 1.

Preparation of APS and methylated human serum albumin.

APS was prepared from strain Mk 83/94 (A:4/21:P1.20,9) (44) by phenol extraction as described elsewhere (19, 65). The purification yielded a product with negligible amounts of protein, LOS, and DNA, as assessed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis with Coomassie blue and silver staining and by high-performance liquid chromatography analysis (Perkin-Elmer, Wellesley, MA) using a TSK-GEL G5000 PWXL column with UV detection at a λ of 200 nm (Biotech Support Group, East Brunswick, NJ). Human serum albumin was methylated (37), and 1:1 complexes of APS and methylated HSA were prepared as previously described (11).

APS ELISA.

The APS ELISA was performed as described elsewhere (11). Binding of IgG and IgA antibodies to APS was detected using alkaline phosphatase-conjugated goat anti-human IgG (Sigma-Aldrich, St. Louis, MO) (11). The CDC1992 serum (National Institute of Biologicals and Standards, Potters Bar, United Kingdom) was used as an internal reference (24). In addition, two other internal control sera were included.

Preparation of OMVs.

OMVs were prepared by extraction with deoxycholate (dOMVs) from the MenA ST-7 strain Mk 686/02 and have been characterized previously (43). The outer membrane proteins PorA, PorB, RmpM, Opa, OpcA, NspA, NadA, Omp85, and TdfH were present, along with approximately 5% LOS L11 relative to protein (43).

OMV ELISA.

Binding of IgG to dOMVs was determined as described previously (43). An internal serum standard for anti-dOMV IgG response, defined as 1,000 arbitrary units (AU)/ml, was made by pooling early-convalescent-phase sera from five Ethiopian patients with confirmed MenA disease.

SBA assays.

SBA assays were performed using the “tilt method” as previously described (7, 43), but with some modifications. To avoid possible interference from antibiotics, sera were pretreated by mixing them with similar amounts of penicillinase at 100 international units/ml (Sigma-Aldrich) in Hanks buffered salt solution. Sera were incubated at room temperature for 10 min and heat inactivated at 56°C for 30 min. Twofold dilutions of control sera, penicillinase-treated sera, or monoclonal antibodies were tested with 10 μl of the inoculum per well in the presence of 25% complement in U-well polystyrene microplates (Greiner Bio-One, Frickenhausen, Germany). The complement used was either a single lot of serum from baby rabbits (PelFreez Biologicals, Clinical Systems, Brown Deer, WI) or plasma from a single human donor with no intrinsic bactericidal activity. The ST-5 strain F8238, isolated from a patient in Kenya in 1989, was used as a target strain (38) with rabbit complement (rSBA assay). The ST-7 strain Mk 686/02 (43, 45), isolated from an Ethiopian patient in 2002, was used as a target strain with human complement (hSBA assay). Both strains were serologically characterized as A:4/21:P1.20,9:L11 OpcA+ and had similar expression of PorA and NadA but different Opa protein repertoires (43, 45). The reaction mixture was incubated for 60 min at 37°C in air.

SBA titers were expressed as the reciprocal of the final serum dilution giving a ≥50% reduction in CFU at 60 min. Sera with titers below the SBA assay detection limit, <32 for the rSBA assay and <4 for the hSBA assay, were assigned a titer of 2. An rSBA titer of ≥128 was defined here as putatively protective against MenA disease, in order not to overestimate protection (5, 64). With the hSBA assay, a titer of ≥4 was defined as putatively protective (16). A seroresponder was defined as an individual demonstrating a ≥4-fold rise in the SBA titer. The seroconversion rate was defined as the proportion of individuals demonstrating an SBA titer below the detection level in the acute phase who subsequently achieved a ≥4-fold increase in the SBA titer in the early convalescent phase. The CDC1992 serum (24), two positive-control sera (sera from Norwegian teenagers after vaccination with APS plus serogroup C PS [15]), one negative-control serum (Norwegian prevaccination serum), and the murine anti-P1.9 PorA monoclonal antibody (MN5-A10F) (49) were used as internal assay references.

Data analysis.

Statistical analyses were performed using SPSS, version 13.0.1 for Windows (SPSS Inc., Chicago, IL). The data in the serological assays were considered to follow a nonnormal distribution and were therefore compared using nonparametric tests. For nonpaired data, group-to-group comparisons using Mann-Whitney t tests were first performed after an analysis-of-variance test (Kruskal-Wallis) had verified a significant overall difference between all groups. Paired data, such as groups of sera collected at different time points from the same patients, were compared using the Wilcoxon signed-rank sum test. Differences in proportions of responders above the predefined threshold values were compared using the chi-square test. Following log₁₀ transformation of data, the relation between the rSBA and hSBA assay results was analyzed using the Pearson correlation test, and the relation between antibody specificity and SBA results was determined using univariate linear regression.

Ethical clearance.

The investigators obtained ethical clearance from the AHRI/All Africa Leprosy TB & Rehabilitation Training Center (ALERT) Ethical Clearance Committee, the National Ethical Review Committee (Ethiopian Science and Technology Agency), and the Norwegian Regional Committee for Medical Research Ethics in Western Norway (REK III). Written informed consent was obtained from study participants or their parents/guardians (for those below the age of 18 years or those with disturbed consciousness) before enrollment.

RESULTS

Demographic characteristics of patients and controls.

Characteristics of the study participants are given in Table 1, together with the numbers of sera available for analysis. The Ethiopian meningococcal meningitis patients and controls had similar age and gender distributions (Table 1). However, in some assays, only selected sera were analyzed, and these were not always matched for age and gender.

IgG response against APS.

In sera from patients with meningococcal disease, the geometric mean concentrations (GMCs) of IgG against APS were 10.8 μg/ml in the acute phase, increasing to 33.4 μg/ml in the early convalescent phase and declining to 13.7 μg/ml in the late convalescent phase (Table 2; Fig. 1a). The proportions of patient sera with a GMC of IgG against APS of ≥2 μg/ml were 98% in the acute phase and 100% in the early and late convalescent phases. Limiting the statistical analysis to paired sera only, we found a significantly higher IgG level in the early-convalescent-phase sera than in the acute-phase sera (P < 0.001), while the difference in IgG levels between acute- and late-convalescent-phase sera proved not to be significant (P = 0.26) (Tables 2 and 3). The proportion of patients with ≥4-fold increases in anti-APS IgG levels from the acute to the early convalescent phase was 35%, while the proportion with ≥2-fold increases was 59% (Table 3; Fig. 2a). The anti-APS IgG levels in the acute- and late-convalescent-phase patient sera were not significantly different from those observed for controls, compared within each of the age groups. In contrast, anti-APS IgG levels in early-convalescent-phase sera were significantly higher than those for controls in all age groups, except for infants, where few sera were available. In acute-phase sera, we found no statistically significant difference in anti-APS IgG levels between age groups (P = 0.77) (Table 2). On the other hand, a significant difference in anti-APS IgG levels between age groups was seen among controls (P = 0.003), and the proportions of sera with anti-APS IgG levels above “thresholds” increased with age (Fig. 3a). The proportions of sera from patients with anti-APS IgG levels equal to or above the GMC for Ethiopian controls (9.4 μg/ml) were 49%, 93%, and 72% for acute-, early-convalescent-, and late-convalescent-phase sera, respectively. However, the proportion of early-convalescent-phase sera from patients with anti-APS IgG levels above the 99th percentile for the control sera was only 7.1%. Anti-APS IgG levels were significantly higher for teenage Ethiopian controls (≥13 to ≤15 years) than for teenage Norwegian controls (GMC, 1.9 μg anti-APS IgG/ml; range, 0.3 to 11.4 μg IgG/ml; n = 44) (P < 0.001).

TABLE 2.

GMCs of anti-APS IgG and anti-OMV IgG measured by ELISA in sera from Ethiopian patients with systemic MenA disease and controls

IgG	Age group (yr)	Analysis for:
		Patients						Controls
		Acute phase		Early convalescent phase		Late convalescent phase
		GMC (range)	na	GMC (range)	n	GMC (range)	n	GMC (range)	n
APS (μg/ml)	All	10.8 (1.3-88.3)	63	33.4 (4.0-174.6)	42	13.7 (2.1-54.2)	32	9.4 (0.3-125.0)	113
	0.5 to ≤2	3.8 (1.3-11.1)	2	7.7	1	5.6 (2.1-14.8)	2	4.5 (0.3-51.6)	20
	>2 to ≤6	13.7 (3.0-76.1)	10	51.9 (20.0-154.0)	8	9.3 (4.8-21.2)	4	6.3 (0.3-68.2)	18
	>6 to ≤15	11.4 (2.8-88.3)	26	27.2 (4.0-80.6)	17	14.6 (4.9-46.6)	14	11.3 (3.7-57.0)	21
	>15	9.9 (4.5-32.4)	24	36.5 (11.2-174.6)	16	16.9 (5.5-54.2)	12	13.0 (1.9-125.0)	54
OMV (AU/ml)	All	92 (10-777)	64	331 (32-3,000)	42	118 (10-2,452)	34	48 (10-202)	36
	<0.5 to ≤2	40 (10-156)	2	52	1	220 (54-895)	2	37 (10-202)	12
	>2 to ≤6	76 (10-314)	10	370 (64-1,400)	8	102 (67-159)	4	50 (30-69)	3
	>6 to ≤15	101 (20-777)	26	249 (34-2,222)	17	88 (33-544)	15	56 (24-95)	10
	>15 yr	95 (10-452)	25	473 (68-3,000)	16	155 (10-2,452)	13	54 (10-110)	11

^an, number of sera analyzed.

FIG. 1.

Development of antibody responses against MenA bacteria in sera from MenA disease patients over time after the onset of disease, as measured by ELISAs for IgG against APS (a) and IgG against OMVs (b). In the same sera, rSBA against strain F8238 (c) and hSBA against strain Mk 686/02 (d) were measured. conv., convalescent.

TABLE 3.

GM fold increases in antibody levels in paired sera from Ethiopian patients with systemic MenA disease

Antibody	Acute vs early conv.a phase					Acute vs late conv. phase
	GM fold increase (range)b	nc	P value of difference	% of patients with the following fold increase:		GM fold increase (range)	n	P value of difference
				≥2	≥4
Anti-APS IgG	3.0 (0.3-41.7)	37	<0.001	59	35	1.3 (0.1-9.3)	30	0.262
Anti-APS IgA	6.7 (0.9-150.4)	7	0.046	40	40	2.5 (0.5-17.8)	6	0.345
Anti-OMV IgG	3.3 (0.2-155.4)	38	<0.001	61	32	1.2 (0.09-111.3)	33	0.950
rSBA F8238	33 (0.06-32,768)	35	<0.001	60	57	2.5 (0-4,096)	29	0.248
hSBA Mk 686/02	6.3 (0.02-2,048)	33	<0.002	64	52	0.6 (0-512)	27	0.293

^aConv., convalescent.

^bCalculated by dividing the value for the individual serum sample from the early or late convalescent phase by the value for the corresponding acute-phase serum sample.

^cn, number of sera analyzed.

FIG. 2.

Antibody levels in paired acute- versus early-convalescent-phase sera from MenA disease patients. (a) Anti-APS IgG; (b) anti-OMV IgG; (c) rSBA; (d) hSBA.

FIG. 3.

Age-related differences in antibody levels among Ethiopian control sera, expressed as percentages of sera above thresholds (either GMCs of these control sera or 50% of the GM). (a) rSBA and anti-APS IgG; (b) IgA against APS; (c) anti-OMV IgG.

IgA response against APS.

In the patients' acute-phase sera, the GMC of IgA against APS was 1.1 μg/ml (n = 25). This level increased significantly, to 6.2 μg/ml, in early-convalescent-phase sera, while in late-convalescent-phase sera it decreased to 3.2 μg/ml, a level not significantly different from the acute-phase level (Table 3). No differences between age groups in the GMC of IgA against APS were observed among the acute-phase patient sera (P = 0.77). A higher level of anti-APS IgA was observed in control sera than in acute-phase patient sera for the >6- to 15-year age group (P = 0.006). As was the case for IgG, a significant difference in anti-APS IgA levels (P = 0.003) between age groups was observed for controls (Fig. 3b).

IgG response against OMVs.

The GMC of IgG against dOMVs of strain Mk 686/02 increased significantly from acute- to early-convalescent-phase sera, while it declined in late-convalescent-phase sera to the levels seen in acute-phase sera (Tables 2 and 3; Fig. 1b). The proportion of patients with ≥4-fold increases in anti-OMV IgG levels from acute to early convalescent phase was 32%, while the proportion with ≥2-fold increases was 61% (Table 3; Fig. 2b). We found no significant differences in IgG levels between the various age groups (P = 0.88). Higher levels of anti-OMV IgG were observed in acute-phase patient sera than in control sera for children >6 to ≤15 years old (P = 0.020) and for adults (P = 0.017) (Table 2). In early-convalescent-phase sera, anti-OMV IgG levels were significantly higher than in control sera for all age groups except infants. In late-convalescent-phase sera, anti-OMV IgG levels were still significantly higher than those in control sera among adults (P = 0.008). The proportions of control sera with anti-OMV IgG levels above two different thresholds (GMC and 50% of GMC for controls, respectively) showed only a slight tendency to increase with age (Fig. 3c). In control sera from Ethiopian individuals aged ≥13 to ≤15 years, the anti-OMV IgG GMC was almost twice as high (P = 0.001) as that in control sera from Norwegian teenagers (GMC, 29 U/ml; range, 21 to 40 U/ml, n = 10).

SBA.

The geometric mean titer (GMT) in acute-phase patient sera was 47.3 by the rSBA method, increasing to 3,104 in early-convalescent-phase sera and decreasing to 110 in late-convalescent-phase sera; the latter level was not significantly different from that found in acute-phase sera (Tables 3 and 4; Fig. 1c). Correspondingly, the proportions of patient sera with protective rSBA titers was 44% in the acute phase, 95% in the early convalescent phase, and 56% in the late convalescent phase (Table 4). The proportion of patients who were rSBA seroresponders (i.e., who demonstrated a ≥4-fold increase) was 57% (20/35) (Table 3), while the rSBA seroconversion rate was 100% (15/15). The rSBA GMT was significantly lower in acute-phase patient sera than in controls among adults (Table 4) (P < 0.001). This was also the case when acute-phase sera sampled at days 5 to 7 after onset were excluded from analyses.

TABLE 4.

rSBA against target strain F8238 and hSBA against Mk 686/02 in sera from Ethiopian systemic MenA disease patients and controls

Group	rSBA				hSBA
	GMT (range)	na	No. (%) of patients with titer of:		GMT (range)	na	No. (%) of patients with titer of:
			≥32	≥128			≥4	≥16
Patients
Acute phase
All	47.3 (2-65,536)	62b	29 (47)	27 (44)	11.7 (2-4,096)	47	25 (53)	19 (40)
0.5 to ≤2 yr	64 (2-2,048)	2			64	1
>2 to ≤6 yr	1,024 (2-65,536)	10			58.7 (2-4,096)	8
>6 to ≤15 yr	58.9 (2-32,768)	25			11.3 (2-512)	22
>15 yr	11.7 (2-4,096)	24			5.0 (2-128)	16
Early convalescent phase
All	3,104 (2-65,536)	40	38 (95)	38 (95)	81.2 (2-8,192)	35	27 (77)	24 (69)
0.5 to ≤2 yr		0				0
>2 to ≤6 yr	4,096 (128-65,536)	8			173 (2-2,048)	6
>6 to ≤15 yr	2,724 (2-32,768)	17			32 (2-2,048)	15
>15 yr	3,104 (2-65,536)	15			156 (2-8,192)	14
Late convalescent phase
All	110 (2-32,768)	32	21 (66)	18 (56)	8.0 (2-1,024)	28	13 (46)	2 (29)
0.5 to ≤2 yr	2 (2-2)	2			256	1
>2 to ≤6 yr	45.3 (2-2,048)	4			2.8 (2-8)	4
>6 to ≤15 yr	232 (2-32,768)	14			9.9 (2-1,024)	13
>15 yr	121 (2-4,096)	12			6.5 (2-256)	10
Controls
All	166 (2-131,072)	70	54 (77)	44 (64)	3.3 (2-256)	16	3 (19)	2 (13)
0.5 to ≤2 yr	23.1 (2-2,048)	17			3.1 (2-16)	8
>2 to ≤6 yr	192 (2-131,072)	17			6.7 (2-256)	4
>6 to ≤15 yr	192 (2-4,096)	17				0
>15 yr	737 (2-16,384)	19			2 (2-2)	4

^an, number of sera analyzed.

^bIncludes one patient for whom age was not reported.

In acute-phase patient sera, rSBA titers for children aged >2 to ≤6 years were significantly higher than those for individuals aged >6 to 15 years (P = 0.045) or >15 years (P = 0.001). rSBA titers were not significantly different between the latter two groups (P = 0.088). In control sera, rSBA titers for the <2-year age group were significantly lower than those for the >2- to 6-year, >6- to 15-year, and >15-year age groups (P < 0.036 for all), while GMTs for the three latter age groups were not significantly different. However, the proportion of sera with putative protective rSBA titers increased with age, from 40% for those aged <2 years to 82% for those aged ≥15 years (Fig. 3a). rSBA titers in control sera from Ethiopian teenagers were significantly higher than those in control sera from Norwegian teenagers (P < 0.001); the GMT of the Norwegian control sera was below the assay cutoff, and only 3/15 (20%) had putative protective rSBA titers.

By the hSBA assay, the GMT was 11.7 in acute-phase patient sera; it increased significantly, to 81.2, in early-convalescent-phase sera and then decreased to 8.0 in late-convalescent-phase sera (Tables 3 and 4). The hSBA titers in the late-convalescent-phase patient sera were not significantly different from those in the acute-phase sera (Tables 3 and 4). Correspondingly, among patients, the proportion of sera with putative protective hSBA titers increased significantly, from 53% in the acute phase to 77% in the early convalescent phase, before dropping to 46% in the late convalescent phase (Table 4). The proportion of patients who were hSBA seroresponders was 52%, while the hSBA seroconversion rate was 53% (8/15). Among adults, the hSBA GMT was higher in early-convalescent-phase patient sera than in control sera (P = 0.009). This was the case also among patients >2 to ≤6 years of age, although it was not statistically significant. No other significant differences in hSBA GMTs were observed between control sera and patient sera from any phase within the age groups (Table 4).

Vaccination status and serological results.

Self-reported data on whether the patients or controls enrolled in this study were vaccinated with APS in the last 3 years prior to enrollment were available from case record forms for 39/113 controls and 60/71 patients with meningococcal meningitis. Among the 60 patients reporting vaccination data, we observed no significant difference in anti-APS IgG levels between individuals who reported that they had been vaccinated (n = 11) and those who reported that they had not (n = 49). Likewise, for the 39 controls reporting vaccination data, we found no significant difference in anti-APS IgG levels between those who said they had been vaccinated (n = 21) and those who said they had not (n = 18).

Correlations between assays.

SBA is the primary surrogate assay for protection against meningococcal disease (7), but the complement source in the assay may be important for evaluation of the results. We therefore analyzed whether results from our two different SBA methods (rSBA and hSBA) showed covariation. Only a moderate but significant correlation between rSBA and hSBA titers was observed for acute-phase sera (r = 0.52; n = 46) and late-convalescent-phase sera (r = 0.41; n = 46) from patients and for controls (r = 0.51; n = 16), while no significant correlation was detected for early-convalescent-phase sera. Combining data from analyses of sera from all patient phases (acute, early convalescent, and late convalescent), we observed a moderate but significant correlation between titers obtained by the rSBA and hSBA assays (r = 0.46) (Fig. 4). The proportions with putatively protective titers by both the rSBA and the hSBA assay ranged from 13% to 74% (Table 5). While as much as 15% of sera (acute phase) showed rSBA titers of <128 and hSBA titers of ≥4, as much as 50% of sera (Ethiopian controls) showed hSBA titers of <4 and rSBA titers of ≥128 (Table 5). The proportions of sera without putative protective rSBA or hSBA titers were about one-third for acute- and late-convalescent-phase patient sera and for controls (Table 5). Some of the patients demonstrated a significant fall in anti-APS or anti-OMV IgG levels from the acute phase to the early convalescent phase (Fig. 2), but none of these were among those showing SBA seroconversion (i.e., having an SBA titer below detection level in the acute phase and subsequently demonstrating a ≥4-fold titer rise).

FIG. 4.

Correlation between rSBA titers and hSBA titers in all sera from Ethiopian MenA patients (r = 0.46; P < 0.001; n = 109). Dotted lines indicate suggested MenA SBA titer thresholds for protection (≥1:128 for rSBA and ≥1:4 for hSBA).

TABLE 5.

Proportions of sera with SBA titers above putative protective thresholds^a for the subset of sera for which both rSBA and hSBA were analyzed

Serum category	nb	No. (%) of sera with the following titer(s):
		rSBA, ≥1:128	hSBA, ≥1:4	rSBA, ≥1:128; hSBA, ≥1:4	rSBA, ≥1:128; hSBA, <1:4	rSBA, <1:128; hSBA, ≥1:4	rSBA, <1:128; hSBA, <1:4
Acute phase	46	22 (48)	25 (54)	18 (39)	5 (11)	7 (15)	16 (35)
Early convalescent phase	35	33 (94)	27 (77)	26 (74)	7 (20)	1 (3)	1 (3)
Late convalescent phase	28	15 (54)	13 (46)	8 (29)	6 (21)	4 (14)	10 (36)
Ethiopian controls	16	10 (63)	3 (19)	2 (13)	8 (50)	1 (6)	5 (31)

^aPutative protective SBA titer thresholds were ≥1:128 for rSBA and ≥1:4 for hSBA.

^bn, number of sera analyzed.

However, analysis of whether those patients achieving a putative protective hSBA titer were the same as those classified as seroresponders by the rSBA assay (i.e., a ≥4-fold rise from acute to early convalescent phase) showed no significant correlation between the results (P = 0.82; n = 33). Considering all patient sera, both the rSBA and hSBA titers correlated significantly with both anti-APS and anti-OMV IgG levels; however, there was low correlation between anti-OMV IgG levels and rSBA titers (Fig. 5). This pattern was confirmed by linear regression analysis, where we detected a significant linear relation between anti-APS IgG levels and rSBA titers (Fig. 5a) but not between anti-OMV IgG levels and rSBA titers (Fig. 5b). In contrast, significant linear relationships were detected both between anti-APS IgG levels and hSBA titers (Fig. 5c) and between anti-OMV IgG levels and hSBA titers (Fig. 5d). For Ethiopian controls, rSBA titers correlated significantly with levels of IgG against APS (r = 0.43).

FIG. 5.

Scatter plots showing correlation between data on hSBA, rSBA, anti-OMV IgG, and anti-APS IgG in all sera from Ethiopian MenA disease patients, along with the Pearson correlation coefficient r (P < 0.001 for all) and the least-squares linear regression equation, where significant. (a) Anti-APS IgG versus rSBA; (b) anti-OMV IgG versus rSBA; (c) anti-APS IgG versus hSBA; (d) anti-OMV IgG versus hSBA.

rSBA titers correlated significantly with anti-APS IgA levels when all patient sera or all controls were analyzed (r = 0.50 for both). In contrast, hSBA titers did not correlate with anti-APS IgA levels when either all patient sera or all controls were analyzed. We further analyzed anti-APS IgA levels in a subset of control sera with high anti-APS IgG levels (above 4.7 μg/ml, the concentration equivalent to 50% of the GMC of anti-APS IgG in control sera) from individuals aged >2 to ≤15 years in order to obtain similar age distributions when comparing sera with and without protective rSBA titers. For this subset of control sera, we then observed that anti-APS IgA levels were higher in sera with rSBA titers of ≥128 (GMC, 4.2 μg/ml; n = 6) than in sera with rSBA titers of <1:128 (GMC, 2.6 μg/ml; n = 15) (P = 0.036). In contrast, anti-APS IgA levels for acute-phase patient sera with and without protective rSBA titers were not significantly different.

Serological evaluation of patients not confirmed to have suffered meningococcal disease.

Following confirmation by culture and/or PCR of 71/95 patients having N. meningitidis in their cerebrospinal fluid, we serologically evaluated whether any of the remaining 24 meningitis patients could have suffered meningococcal meningitis as well (Table 1). Pairs of acute- and early-convalescent-phase sera were available from 10 of these patients. Based on the results observed in our study for patients with confirmed meningococcal meningitis (Table 3), a significant antibody increase indicating that the patient had gone through an episode of meningococcal meningitis was defined as a twofold rise in either the anti-APS or the anti-OMV IgG level. Significant rises in anti-APS IgG levels were observed for 6 of the 10 previously nonconfirmed patients. However, one of these had been immunized with APS vaccine 1 day after the onset of the disease; thus, the anti-APS IgG response could have been due to vaccination. Two of the five patients with significant anti-APS IgG increases also showed significant increases in anti-OMV IgG levels. Of the remaining four patients not confirmed with meningococcal disease, one showed a significant increase in anti-OMV IgG levels while demonstrating a significant decrease in anti-APS IgG levels. In summary, there were serological indications of MenA infection for at least 6 of the 10 patients, thus increasing the number of confirmed and probable meningococcal disease patients in this study to 77/95 (81.1%).

DISCUSSION

In this study we describe the antibody responses in patients from Ethiopia with meningococcal disease caused by N. meningitidis serogroup A ST-7 organisms. Our main findings are as follows: (i) most acute-phase patient sera and control sera show levels of IgG antibodies against APS substantially higher than those previously assumed to be sufficient for protection (31); (ii) MenA infections stimulate significant increases in levels of IgG against APS and OMVs and increase SBA levels in these patients during the first month after the onset of disease; (iii) approximately 8 months after the onset of MenA disease, the mean antibody levels decrease to levels found during the acute phase of the disease; (iv) the observed rSBA titer variation can apparently be explained by variations in anti-APS IgG levels but only to a minor degree by variations in anti-OMV IgG levels. In contrast, a major part of the observed variation in hSBA titers apparently can be explained by variations in levels of IgG against both APS and OMVs.

Sera and data collection for the study relied on the existing health infrastructure and the doctors' motivation for contributing to research amid a resource-challenged practice. The vaccination status was self-reported, since meningococcal vaccination usually happens as mass vaccination in response to epidemics and is seldom recorded in a personal vaccination card.

Results from comparisons of sera from acute-phase patients with sera from controls are very much dependent on the reported onset-to-admission time and the time it takes to mount a significant immune response after infection. The date of the first symptoms reported by patients may be an important uncertainty factor because of recall bias and possible differences in the understanding of what may constitute a “serious symptom.” We chose to include sera collected up to 7 days after the onset of the first symptoms as acute-phase sera. Thus, in some instances an immune response could already have been initiated in the period defined as acute. The anti-PS IgG response following PS vaccination usually starts to increase between 4 to 10 days after exposure to the antigen (8), and 12/62 acute-phase patient sera were collected ≥5 days after the reported onset of disease in our study. However, the antibody levels in the acute-phase sera collected at different days after the onset of disease did not show any increasing trend according to the number of days after the onset of disease (Fig. 1). We compared acute-phase sera obtained 0 to 4 days after the onset of disease with acute-phase sera obtained 5 to 7 days after onset but found no significant differences between groups in GM antibody levels or SBA titers (data not shown). Furthermore, excluding acute-phase sera collected at days 5 to 7 gave results similar to those obtained when all acute-phase sera were included. The anti-APS IgG levels in acute-phase sera were similar to those in control sera as well; thus, they were likely representative of the levels prior to response to the MenA infection. However, further studies should be performed comparing MenA SBA in acute-phase sera with that in control sera and also aiming at obtaining serial acute-phase sera closer to the onset of infection.

Antibody response against APS.

The anti-APS IgG levels in acute-phase patient sera were relatively high (∼10 μg/ml) but not significantly different from those in control sera (Table 2). This is in line with the findings of previous studies of patients in the meningitis belt (9, 20), and levels were similar to those among the adult population in Burkina Faso (40). In the meningitis belt, high levels of anti-APS antibodies are commonly found in healthy individuals (4, 9, 20, 21, 40). However, the levels we observed were higher than those reported for acute-phase sera from Finnish and Egyptian MenA disease patients as measured by a total-Ig radioimmunoassay (31, 58) and for acute-phase sera from Gambian MenA patients (9). Higher anti-APS IgG levels were seen in Ethiopian teenage controls than in Norwegian teenagers, confirming previous studies (4, 54). Apart from previous vaccination with APS vaccine or previous episodes of MenA disease, the main reason for the high levels is probably natural immunization, induced either by carriage of meningococci in the nasopharynx (9) or by enteric stimulation of cross-reactive PS structures (21, 51). Elevated levels of anti-APS IgG may also be due to nonspecific stimulation of APS-specific plasma cells, as observed for Finnish non-MenA bacterial meningitis patients (30). Since populations in developing countries are prone to undergo multiple infections, e.g., by bacteria or protozoa, this may be an important cause of the high anti-APS IgG levels.

The antibody level suggested to confer protection against MenA disease is ≥1 to 2 μg/ml of total Ig against APS (31, 47). One would therefore consider nearly all of our Ethiopian patients and controls to be protected. However, the suggested putative protective level was assigned from a phase III protection trial with APS vaccine in Finland in the 1970s using an radioimmunoassay and a single-donor reference serum (47). In using results from the standardized APS ELISA (11), one ought to apply this criterion carefully: for example, the well-characterized human reference serum CDC1992, used as a control in our study, was obtained by pooling sera from 14 U.S. individuals vaccinated with APS vaccine (24). It is well suited for analyzing APS vaccine-induced antibodies, but the avidities of these antibodies may differ from those of the antibodies induced by MenA disease or carriage in the African population. A Finnish study suggested that the cross-reactive anti-APS antibodies in normal sera were of lower avidity than those obtained by immunization with APS vaccine (30). Borrow and colleagues (28, 29) found that the avidity of anti-APS IgG correlated positively with MenA SBA in sera from MenA conjugate vaccinees and that an APS vaccine induced anti-APS IgG of lower avidity than the MenA conjugate vaccine. The level of anti-APS correlating with protection against MenA disease may need to be specifically established for the meningitis belt area. However, the protective level of antibodies is determined not only by the quantity but also by the quality of these antibodies; thus, the Ig isotype, IgG subclass, and avidity toward APS are probably all of importance, since these affect the SBA.

From the acute to the early convalescent phase of disease, we found that the IgG against APS increased significantly (Table 3), to levels equivalent to or higher than those observed following vaccination with APS vaccine (3, 25, 33). However, only 35% of the patients experienced a ≥4-fold increase in anti-APS IgG levels, while in Finnish MenA patients the corresponding proportion was 62% (31). The already high acute-phase levels for the Ethiopian patients compared to the Finnish patients may explain the difference observed. The high antibody levels, as well as the effect of MenA carriage acquisition (9), also complicate the use of a specific anti-APS IgG n-fold increase for serological confirmation of MenA disease in the meningitis belt.

In the late convalescent phase, the mean anti-APS IgG levels had returned to levels similar to those found in acute-phase sera (Table 2). As far as we know, this is the first study on long-term antibody levels following MenA disease. Similar reductions to prevaccination levels after ∼8 months have been observed in clinical trials with APS and conjugate-PS vaccines (14, 60), although others have seen elevated anti-APS IgG levels for a longer period after APS vaccination (66). In Ethiopian control sera, as well as in acute-phase patient sera, we found an age-related increase in the proportions of sera with anti-APS IgG levels above 4.7 μg/ml, the concentration equivalent to 50% of the GMC in controls (Fig. 3a). Such an age-related increase is in line with the findings of previous studies (31, 40).

Several studies have suggested that IgA antibodies against APS may render the population susceptible to meningococcal disease (4, 20, 21). Anti-PS IgA is reported not to activate complement via the classical pathway and is suggested to block the bactericidal effect of anti-PS IgG antibodies by competitive binding (21, 26). Provided that there is a constant IgG concentration, there should thus be an inverse relation between the IgA concentration and the SBA titer (21). In our study, anti-APS IgA levels correlated well with rSBA titers. Also, among Ethiopian controls aged >2 to ≤15 years who had high anti-APS IgG levels, the anti-APS IgA levels were significantly higher for those with protective rSBA titers than for those without. This contradicts the findings of a previous study of sera of the Sudanese population (4). The possible role of serum IgA in mediating the lysis of meningococci via the alternative complement activation pathway seems minor (61). The complement source may, however, be relevant for the potential ability of IgA to mediate SBA, but further studies are required to test this hypothesis.

Antibody responses against OMVs.

In line with the findings of a previous study (9), IgG levels against outer membrane antigens were significantly higher in acute-phase patient sera than in control sera when individuals above the age of 6 years were compared (Table 2). This contrasts with a similar comparison of anti-APS levels that showed no significant differences between acute-phase patient sera and control sera, and with a similar comparison of rSBA titers, which were lower in acute-phase patient sera than in control sera. This disparity may be explained by the exposure of a high proportion of patients to MenA carriage prior to disease, since carriage of MenA organisms has been shown to induce IgG against both capsular and noncapsular antigens (10). Significantly lower levels of anti-OMV IgG were observed for Norwegian teenagers, a finding that also may be explained by different exposure to MenA carriage. Acute-phase sera from patients with systemic serogroup B and C meningococcal disease in Norway demonstrated anti-OMV IgG levels similar to those for healthy population controls, while for septicemia patients, significantly lower anti-OMV IgG levels were demonstrated at admission (22). Our findings for Ethiopian patients contrast with these results, but further studies of septicemia patients from the meningitis belt would be interesting.

Meningococcal meningitis was also found to induce significant rises in levels of IgG antibodies against OMVs, in line with previous findings for MenA (9, 58) and MenB (22, 48, 53, 57) patients. However, only 32% showed a ≥4-fold increase in levels of IgG against OMVs, which can be explained by the already high levels in acute-phase sera. Considering that a ≥2-fold increase in anti-OMV IgG levels was demonstrated for 61% of confirmed MenA disease patients, approximately the same as the proportion demonstrating a ≥4-fold increase in SBA titers (Table 3), a ≥2-fold increase may be indicative of MenA disease in the meningitis belt.

Functional activity of antibodies as analyzed by SBA.

We found a significantly lower rSBA GMT in acute-phase patient sera than in control sera among adults, but not in the younger age groups. However, few sera were available for comparison in these age groups. A subcritical level of SBA was demonstrated to correlate with an increased risk of serogroup C meningococcal infection for U.S. military recruits, as measured with human complement against the prevailing serogroup C strain (16). However, this could not be confirmed by Greenwood et al. in a study of Gambian MenA patients, as measured by using baby rabbit complement and a local MenA epidemic strain (20). For both acute-phase patient sera and control sera, we observed high proportions considered protected by the rSBA assay. This is in agreement with results from studies of prevaccination sera in clinical vaccine trials (12). While the proportion of controls with protective rSBA titers increased with age, as was observed previously (3, 40), in acute-phase patient sera the titers peaked in children aged >2 to ≤6 years. Among Ethiopian controls, the proportion with hSBA titers of ≥4 was 19%. Using a different assay, Goldschneider et al. found a similar proportion with hSBA titers of ≥4 against a MenA strain among baseline sera (predisease sera collected at enrollment) from meningococcal patients, while the proportion was 72% for baseline sera from individuals who did not contract meningococcal disease (16). Achtman et al. have suggested that “disease might reflect the inability of the colonized individual to quickly mount a protective antibody response” (2); protection may thus consist of both a minimum level of circulating functional antibodies and a sufficient pool of memory B cells specific for functional antibodies (32). Establishing SBA assay conditions correlating with protection and assigning a protective MenA SBA titer seem pivotal for the evaluation of vaccines against MenA disease for the meningitis belt.

Significant titer rises were demonstrated for MenA patients from the acute phase to the early convalescent phase, and the proportion of sera with protective hSBA titers increased from 53% to 77%. This is higher than that reported for MenA meningitis patients infected with clone IV-1 or subgroup III strains from Gambia and Finland, respectively (2); in this study very few of the patients demonstrated the development of bactericidal antibodies following disease. The high percentages found in our study may be due to differences in the SBA assays, although in sera from MenB patients aged 10 to 17 years, 75 to 100% demonstrated the achievement of hSBA titers of ≥4 during the convalescent phase (48). The proportion of meningococcal disease patient sera attaining protective rSBA titers increased from 40% in the acute phase to 95% in the early convalescent phase, a proportion similar to that achieved during vaccination with MenA vaccines (14). Since the rSBA titers were already high in the acute-phase sera, it seems reasonable that seroresponse rates were only 57% for rSBA (Fig. 2c). The seroconversion rates for rSBA and hSBA were 100% and 53%, respectively. In comparison, vaccination of individuals 11 to 18 years old with APS conjugate vaccines resulted in seroresponse rates of 88% (rSBA) and 94% (hSBA) and seroconversion rates of 100% (rSBA) and 92% (hSBA) (34). Thus, in the short term, suffering from MenA disease seems to provide immunological protection in a similar proportion of individuals as immunization with APS and MenA conjugate vaccines.

A central question emerging from our findings is whether survivors of MenA disease in the meningitis belt are protected against reinfection with a similar strain, i.e., what the duration of the achieved protection is. In this study, the general picture was that although the proportion of individuals considered not protected by rSBA or hSBA was only 3% in early convalescent phase, this decreased in the late convalescent phase to a proportion similar to that in the acute phase (Table 5). The time between the collection of early- and late-convalescent-phase sera varied greatly; thus, reporting the median response alone hides the gradual decay in IgG levels and SBA titers.

The returning waves of MenA disease in the meningitis belt could be explained by the introduction of new clones in a population rendered susceptible due to waning immunity to the prevalent clone and lack of immunity to the newly introduced clone (39). The introduction of new serogroup A N. meningitidis clones into the meningitis belt has been well documented (2, 41, 42, 67), while the probable waning protective immunity following disease has been the subject of only a few studies (2, 9). Vaccination with APS vaccine has been shown to induce protection suggested to last as long as 3 years for adults (46). However, the dosing and presentation of antigen are very different in vaccination and infection, and thus, the magnitude of the response and the duration of protection conferred by infection are not necessarily indicative of an antigen's ability to induce long-term protection when used as a vaccine (63).

Correlations.

To evaluate the level of protection against MenA disease, we analyzed sera by two different SBA assays (rSBA and hSBA). A major part of the patients apparently achieved immunological protection following MenA disease, since 74 to 94% of the patients mounted a protective antibody response, by having putative protective titers of both rSBA and hSBA in early convalescent phase (Table 5). However, the individual results of our rSBA and hSBA assays correlated only modestly, and the individuals classified as protected by the two assays were not the same. This is in agreement with previous observations (34) and raises the questions of whether one of the methods may incorrectly categorize individuals as protected and whether the SBA test can be used only for evaluation of susceptibility to disease at a population level and not on an individual basis.

Our study showed that the choice of method (parameters) is critical in the evaluation of the ability of antibodies to confer SBA. In general, the titers obtained by the hSBA assay were considerably lower than those obtained by the rSBA assay (Fig. 4). This could be due to either the complement source or the target strain used. Using a limited number of sera, we compared the SBA titers against either strain F8238 or strain Mk 686/02 obtained with human complement with those obtained with rabbit complement with the same strain. We then found very low hSBA activity by using the F8238 target strain (data not shown), as also observed previously (64). On the other hand, the ST-7 strain Mk 686/02 was more sensitive to rabbit complement (data not shown). The different titers obtained by the rSBA and hSBA assays are primarily due to differences in the complement source and target strains. Human anti-meningococcal IgG is known to mediate higher MenA SBA titers in the presence of rabbit complement than in the presence of human complement (64). Also, IgM against meningococcal PS is suggested to have a higher ability to mediate SBA with rabbit complement than with human complement (55). Part of the MenA rSBA response observed in the acute phase in our study may thus have been caused by anti-MenA IgM. Although both target strains in the SBA assays were characterized as A:4/21:P1.20,9:FetA 3-1:L11 (43, 45), subtle differences in outer membrane antigens between the ST-5 and ST-7 target strains used in the SBA assay (6, 45), as well as in possibly variable amounts of capsule expressed, may have had an impact on the low correlation observed between hSBA and rSBA.

By analyzing the specificity of the SBA response mounted during disease, we showed that the variability in anti-APS IgG levels explained a substantial proportion of the observed variation in rSBA titers, whereas IgG against OMV apparently was less important for the rSBA variation. This was also reflected in the low correlation between rSBA titers and anti-OMV IgG levels. Again, minor outer membrane antigen differences between the rSBA target strain and the strain from which the dOMVs used in the ELISA were derived may have contributed to the observed correlations. In addition, some proportion of the rSBA variability may just as well be attributed to antibodies directed against, e.g., LOS or minor proteins not present in the dOMVs but expressed on live bacteria. In contrast to rSBA, the variability in hSBA could well be explained by both IgG against APS and IgG against OMVs, as reflected by correlation analyses (Fig. 5). It seems from the sera studied here that both IgG against APS and IgG against OMVs are important in inducing protection when the hSBA assay is used to assess the immune response. SBA assays performed with human complement may more closely reflect in vivo protection against meningococcal disease than SBA assays performed with rabbit complement. Thus, induction of antibodies toward noncapsular antigens may represent a possibility for the improvement of current vaccines against MenA disease in the meningitis belt (43).

Further studies are needed to determine the specificities of the noncapsular antibody response during MenA disease and their relative importance for providing protection against MenA disease. A notable proportion of sera had high concentrations of IgG against APS and OMVs without conferring bactericidal activity (Fig. 5), irrespective of the SBA method. Any protection possibly conferred by antibody-mediated opsonic activity following disease should thus be investigated with these MenA disease patient sera. However, a first priority should be to standardize an SBA assay using a recent representative serogroup A strain with human complement and then to establish whether the initial Goldschneider studies, showing a correlation between SBA and protection (16), can be paralleled in the meningitis belt.