Genetic control of the variable innate immune response to asymptomatic bacteriuria

Abstract

The severity of urinary tract infection (UTI) reflects the quality and magnitude of the host response. While strong local and systemic innate immune activation occurs in patients with acute pyelonephritis, the response to asymptomatic bacteriuria (ABU) is low. The immune response repertoire in ABU has not been characterized, due to the inherent problem to distinguish bacterial differences from host-determined variation. In this study, we investigated the host response to ABU and genetic variants affecting innate immune signaling and UTI susceptibility. Patients were subjected to therapeutic urinary tract inoculation with E. coli 83972 to ensure that they were exposed to the same E. coli strain. The innate immune response repertoire was characterized in urine samples, collected from each patient before and after inoculation with bacteria or PBS, if during the placebo arm of the study. Long-term E. coli 83972 ABU was established in 23 participants, who were followed for up to twelve months and the innate immune response was quantified in 233 urine samples. Neutrophil numbers increased in all but two patients and in an extended urine cytokine/chemokine analysis (31 proteins), the chemoattractants IL-8 and GRO-α, RANTES, Eotaxin-1 and MCP-1, the T cell chemoattractant and antibacterial peptide IP-10, inflammatory regulators IL-1-α and sIL-1RA and the T lymphocyte/dendritic cell product sIL-2Rα were detected and variably increased, compared to sterile samples. IL-6, which is associated with symptomatic UTI, remained low and numerous specific immune mediators were not detected. The patients were also genotyped for UTI-associated IRF3 and TLR4 promoter polymorphisms. Patients with ABU associated TLR4 polymorphisms had low neutrophil numbers, IL-6, IP-10, MCP-1 and sIL-2Rα concentrations. Patients with the ABU-associated IRF3 genotype had lower neutrophils, IL-6 and MCP-1 responses than the remaining group. The results suggest that the host-specific, low immune response to ABU mainly includes innate immune mediators and that host genetics directly influence the magnitude of this response.

Figure 1. Patients and samples.

Samples from patients participating in a clinical trial of induced E. coli 83972 ABU were analyzed. All collected urine samples were subjected to PMN, IL-6 and IL-8 quantification, and blood samples from eleven patients were collected for genotyping of promoter polymorphisms in TLR4 and IRF3. Blood and urine samples from these eleven patients were also selected for an extended urine protein analysis.

Figure 2. Host response to E. coli 83972 bacteriuria.

E. coli 83972 ABU triggered an increase in PMN numbers and IL-8 concentrations (p<0.0001) but IL-6 levels were unchanged (n.s., Mann-Whitney test). Group-wise comparison of monthly urine samples collected during E. coli 83927 ABU or after PBS inoculations. Coded patient IDs are noted on the x-axis. A. Means + SEs of neutrophil numbers, IL-8 and IL-6 concentrations during E. coli 83972 bacteriuria (pink) or sterile conditions (blue). B. Intra-individual comparison of samples obtained during E. coli 83972 bacteriuria (pink) and sterile intervals (blue). C. Box-plot of intra-individual host response variation during E. coli 83972 ABU.

Figure 3. Consistency of the individual host response to E. coli 83972 inoculation.

A. The host response in urine samples from the first (blue) and second inoculations (grey) were compared (Geometric means + SEs) in six patients that had received repeated inoculations. B. Kinetics of the host response during the first and second ABU episode in one high and one low responder.

Figure 4. Urine cytokine/chemokine proteome response to E. coli 83972 ABU.

Cytokine/chemokine levels in individual patients during ABU (grey, medians of all samples in each individual). A pool of 20 sterile urine samples from 9 patients (blue, median of all samples) is used for comparison. Coded patient IDs are noted on the x-axis, in the same order as in Figure 1. Significant differences between sterile and ABU samples from individual patients (Kruskal-Wallis, Dunn's post test) are shown (*p<0.05, **<0.01, ***<0.001). A. Granulocyte (IL-8, GRO-α) or monocyte (MCP-1) chemotaxis. B. Cytokines involved in IL-1 signaling and inhibition. C. Cytokines involved in T lymphocyte chemotaxis and regulation. D. Cytokines involved in eosinophil chemotaxis. E. Scatter diagrams illustrate significant correlations between neutrophil counts and neutrophil chemoattractants IL-8 and GRO-α, and the monocyte chemoattractant MCP-1. For the remaining proteins, no correlation with neutrophil numbers was found. For sample numbers, group means and statistical analysis see Table 2.

Figure 5. Promoter polymorphisms and the host response to ABU.

A. Three of five TLR4 genotypes associated with primary ABU were detected (blue hexagons) in five of the eleven patients. These patients had significantly lower neutrophil numbers (p<0.002) and IL-6 (p<0.0001), MCP-1 (p<0.01), IP-10 (p<0.0001), and sIL-2Rα (p<0.0001) concentrations than the patients with non-ABU associated TLR4 genotypes XIX, IV, XX and IX (red squares). Each column represents one patient and each hexagon or square one monthly urine sample. B. The heterozygous IRF3 promoter genotype associated with ABU (A/G-C/T, blue hexagon) was detected in four of the eleven patients, who had significantly lower neutrophil numbers (p = 0.01) and IL-6 (p<0.001) and MCP-1 (p = 0.0001) concentrations than patients with the homozygous, pyelonephritis-associated genotype (A/A-C/C, red square). Each column represents one patient and each hexagon or square one monthly urine sample.