Polymorphism in a lincRNA Associates with a Doubled Risk of Pneumococcal Bacteremia in Kenyan Children

Abstract

Bacteremia (bacterial bloodstream infection) is a major cause of illness and death in sub-Saharan Africa but little is known about the role of human genetics in susceptibility. We conducted a genome-wide association study of bacteremia susceptibility in more than 5,000 Kenyan children as part of the Wellcome Trust Case Control Consortium 2 (WTCCC2). Both the blood-culture-proven bacteremia case subjects and healthy infants as controls were recruited from Kilifi, on the east coast of Kenya. Streptococcus pneumoniae is the most common cause of bacteremia in Kilifi and was thus the focus of this study. We identified an association between polymorphisms in a long intergenic non-coding RNA (lincRNA) gene (AC011288.2) and pneumococcal bacteremia and replicated the results in the same population (p combined = 1.69 × 10(-9); OR = 2.47, 95% CI = 1.84-3.31). The susceptibility allele is African specific, derived rather than ancestral, and occurs at low frequency (2.7% in control subjects and 6.4% in case subjects). Our further studies showed AC011288.2 expression only in neutrophils, a cell type that is known to play a major role in pneumococcal clearance. Identification of this novel association will further focus research on the role of lincRNAs in human infectious disease.

Figure 1

Signal of Association around rs140817150 in the Discovery Analysis of Pneumococcal Bacteremia Imputed SNPs are shown as circles and directly genotyped SNPs as triangles with colors indicating the correlation (r² in 1000 Genomes data) with rs140817150. A set of SNPs that contains the causal SNP with greater than 95% probability is ringed with circles. Annotated genes (blue) and lincRNAs (red) are shown in the bottom panel along with the fine-scale recombination rate.

Figure 2

rs140817150 lincRNA Association with the Main Bacterial Infections (A) Log transformed combined odds ratios and 95% confidence intervals of directly genotyped discovery and replication samples. The dotted line represents the log OR of 0 (OR of 1; no difference between case and control subjects). The values of point estimates and standard errors (in parentheses) are also given. Bacterial infection abbreviations are as follows: PNEUM, Streptococcus pneumoniae (pneumococcus); ACINET, Acinetobacter species; HAEMOPH, Haemophilus influenzae; ECOLI, Escherichia coli; SALMON, Salmonella (non-typhoidal); STREPBH, Streptococcus beta hemolytic; SAUR, Staphylococcus aureus. (B) The posterior probabilities on the models of association: no effect in any subtype (NULL), same effect in all subtypes (SAME), related effects across subtypes (REL), or the same non-zero effect only in PNEUM, ACINET, and HAEMOPH (P+A+H), in PNEUM and ACINET (P+A), or in PNEUM (P). (See Subjects and Methods.) Models are a priori assumed to be equally likely. Bayes factors, which compare the evidence (marginal likelihood) between any pair of models, can be calculated as the ratio of the posterior probability assigned to each model as reported under each bar of the plot.

Figure 3

lincRNA AC011288.2 Expression Measured in Neutrophils, Monocytes, B Cells, and NK Cells (A) Quantitative PCR of AC011288.2-002 in primary leucocyte subsets. To conservatively estimate the highest level of expression beneath the detectable level, CT values greater than 38 were re-assigned to 38, and normalized to β-actin expression. p values denote the significance of the relative expression levels of AC011288.2-001 in neutrophils compared to other cell types (Mann-Whitney test). (B) Unadjusted cycle number of amplification is shown. Filled squares denote detection of a fluorescent product after cycle 38 (Ct = 38), the limit of confident detection being based on careful inspection of the melting curves. Similar results for AC011288.2-001 were obtained (data not shown).