Variation in Toll-like receptor 4 and susceptibility to group a meningococcal meningitis in Gambian children

Meningitis is an important cause of childhood mortality in sub-Saharan Africa. A case-control study investigating the role of the Toll-like receptor 4 gene in host susceptibility to Group A Neisseria meningitidis meningitis was conducted in 524 Gambian children. There was no association between a functional Toll-like receptor 4 gene polymorphism and meningitis, suggesting that this variant is not a major determinant of disease in this population. Toll-like receptors (TLRs) are an integral component of the innate immune system required for early host defense against a range of invading pathogens. TLR4, interacting with CD14 and MD-2, recognizes lipopolysaccharide (LPS), a cell wall component of Gram-negative bacteria, including Neisseria meningitidis. Signaling occurs via the nuclear factor kappa B pathway to initiate an inflammatory response cascade characterized by the up-regulation of costimulatory molecules and proinflammatory cytokines and chemokines. Uncontrolled activation of this pathway leads to the clinical manifestation of septic shock, which can be induced by LPS injection in volunteers, 1 whereas localized central nervous system inflammation leads to meningitis. 2 Bacterial meningitis is an important cause of childhood mortality and morbidity in sub-Saharan Africa. Epidemics of Group A N. meningitidis infection occur regularly within the “meningitis belt” of sub-Saharan Africa, extending from Sudan in the East across to Senegal and The Gambia in the West. It is increasingly recognized that host genetic factors influence the immune response to pathogens, including meningococci. Rare familial defects in the terminal components of the complement pathway predispose to meningococcal infection and variation in a number of genes, such as mannose-binding lectin, has been associated with sporadic meningococcal disease in outbred European populations. 3 Epidemics provide an excellent opportunity to investigate the genetics of susceptibility to infection given that exposure rates within the affected communities are high and all individuals are exposed to the same strain of pathogen. Arbour et al. 4 identified two missense functional mutations in the extracellular domain of TLR4 (Asp229Gly and Thr399Ile) that are associated with hyporesponsiveness to LPS in alveolar macrophages and epithelial cells. We have investigated the role of the Asp229Gly TLR4 polymorphism in susceptibility to Group A N. meningitidis meningitis in sub-Saharan Africa. The Thr399Ile polymorphism described by Arbour et al. 4 was absent in a different Gambian population sample (MJN, unpublished observations) and therefore not included in this study. Methods. Study population. This study had the approval of the Medical Research Council (MRC)/Gambia Government Ethics Committee. During 1997 an epidemic of Group A N. meningitidis meningitis occurred in the Upper River Division of The Gambia. We conducted a retrospective case-control study in which cases were identified with a database set up at the MRC Laboratories in Basse (the regional town center) as part of the epidemic surveillance and traced 6 months after the end of the epidemic. Case definition was based on a clinical history of meningitis (fever, headache, photophobia, vomiting, neck stiffness), which was confirmed on the child’s health record, and was treated with parenteral chloramphenicol. Group A N. meningitidis was confirmed by bacterial culture in >200 clinical cases of meningitis presenting to Basse Health Centre during the first month of the epidemic. The number of cases soon exceeded the laboratory capacity, and bacteriologic confirmation was not possible for all cases. Therefore infants younger than the age of 18 months were excluded because of the nonspecific presentation of meningitis in this age group and the possibility of other pathogens as the cause of a clinical meningitis syndrome. The upper age limit was 18 years. Because cases were traced 6 months after the epidemic, children who died as a result of their disease were not included in this study. For each case a healthy unaffected control, matched for age, sex and ethnic group, was identified at the case’s compound. A compound in The Gambia comprises a compound head, his wives (usually two to four) and children and extended family members, visitors and their families. Where possible a healthy unrelated child who lived at the same compound was selected to control for similar exposure to N. meningitidis. If such a child was not identified, an unrelated child from a neighboring compound with prolonged contact with the case (e.g. close playmate) was selected. Occasionally when an unrelated child was not identifiable, a related child (maximum relatedness was first cousin) from the same compound was selected as a control, to ensure similar exposure to N. meningitidis. We calculated that a sample size of 250 in each group had 80% power to detect an odds ratio of 1.4 (alpha = 0.05). Genotyping. Venous blood (2 to 5 ml) was collected from cases and controls into EDTA. DNA was extracted using the Nucleon kit (Tepnel Life Sciences, Manchester, UK). Genotypes for the TLR4 Asp299Gly polymorphism, which results from an A to G substitution at nucleotide position 896, were determined with the amplification refractory mutation system as described elsewhere. 5 The allele-specific primers 5′-AGACTACTACCTCGATGA for the A allele and 5′-AGACTACTACCTCGATGG for the G allele together with the consensus primer 5′-GCATTCCCACCTTTGTTGG amplify a 218-bp fragment. Allele and genotype frequencies in the cases and controls were compared with chi square and Stata 7.0 (Stata Corp.) software. Deviation from Hardy-Weinberg equilibrium was assessed with a Monte Carlo permutation procedure (HWSim software available at http://krunch.med.yale.edu/hwsim/). Results. We evaluated 262 meningitis cases (mean age, 7.9 years; range, 1.5 to 18 years) of whom 122 (46.5%) were female, and Group A N. meningitidis infection was confirmed in 120 (45.8%) cases by positive cerebrospinal fluid culture. Cerebrospinal fluid had not been not collected from the remaining 142 cases because of clinical and laboratory overload during epidemic conditions. We identified 262 controls matched for age (mean age, 7.61 years; range, 1.5 to 18 years), sex [122 female (46.5%)] and ethnic group. Allele and genotype frequencies for the Asp299Gly polymorphism are shown in Table 1. No differences were observed in either allele or genotype frequency between the two groups. Both groups were in Hardy-Weinberg equilibrium.TABLE 1: Genotype and allele frequencies for the Asp299Gly TLR4 polymorphism in Gambian meningitis cases and controlsDiscussion. We have found no association between the Asp299Gly TLR4 polymorphism and Group A N. meningitidis meningitis, suggesting that there is no major role for this polymorphism in susceptibility to meningitis in Gambian children. TLR4 is a credible candidate gene for susceptibility to N. meningitidis infection. TLR4 has been identified as the major LPS sensor, and the polymorphism we studied has been shown to influence LPS responses in humans. 4 Mutations of the murine Tlr4 gene selectively impede LPS signal transduction in C3H/HeJ and C57BL/10ScCr mice, rendering them resistant to LPS yet highly susceptible to Gram-negative infection. 6TLR4 is expressed in the leptomeninges and choroid plexus in rat brain, and LPS modulates this expression. 7 The data presented here do not exclude the presence of other functional TLR4 polymorphisms that could influence susceptibility to N. meningitidis disease. However, others have shown that N. meningitidis can induce proinflammatory cytokine responses via pathways independent of CD14 and TLR4, indicating that LPS is not the sole mediator of N. meningitidis induced inflammation. 8 Studies with an LPS-deficient N. meningitidis mutant have shown that even in the absence of LPS, the Gram-negative cell wall remains a potent inflammatory stimulant, utilizing signaling pathways (e.g. TLR2) independent of those involved in LPS signaling. 9 These studies, together with the data we report, suggest that host response genes other than TLR4 are more important in the pathogenesis of N. meningitidis meningitis. We have investigated meningitis, the more common clinical phenotype during N. meningitidis epidemics in the sub-Saharan African meningitis belt. The study was retrospective (cases were traced via clinic records after the epidemic) and therefore could not include children who died as a result of severe disease. We also excluded infants because microbiologic proof of infection was not available for every case and the presentation of meningitis in this age group is nonspecific. We have therefore not excluded the possibility that TLR4 modulates disease severity or phenotype (i.e. sepsis rather than meningitis) or plays a role in disease in the very young. Others have investigated variation in TLR4 and severity of meningococcal infection. Read et al. 10 also failed to identify an association between TLR4 and meningococcal disease. Although no differentiation is made in this study between sepsis and meningitis, 9% of cases died as a result of infection, suggesting a more severe phenotype. There are two other important differences between the two studies. First Read et al. 10 studied a genetically different population. This is reflected in the Asp299Gly allele frequencies; the rare allele had a frequency of 11% in our Gambian population, compared with 5.9% in white English blood donors. Secondly the epidemiology of meningococcal infection differs geographically, with sporadic cases caused by to Group B and Group C serotypes being more common in Europe, compared with epidemic Group A disease in Africa. Acknowledgments. We thank all the children and their guardians for their cooperation with this study; Bakari Dibba, Barjen Drammeh, Galo Touray, M. Lamin Dukureh, Samuel Mendy and Pa Amodu Suso for expert assistance in the field; Dr. Sarah Dunstan for provision of ARMS PCR primer sequences; and Michael Levin and Keith McAdam for their support. This work was supported by the Medical Research Council (UK). AAA was supported by the Sir Halley Stewart Trust