IL17A genetic variation is associated with altered susceptibility to Gram-positive infection and mortality of severe sepsis

Introduction Interleukin 17A (IL17A) plays a key role in host defense against microbial infection including Gram-positive bacteria. Genetic factors contribute to the host defense, but the role of IL17A single nucleotide polymorphisms (SNPs) has not yet been investigated in severe sepsis. Therefore, we hypothesized that SNPs in the IL17A gene alter susceptibility to infection and clinical outcome of severe sepsis. Methods We tested for the association of IL17A SNPs with susceptibility to infection and clinical outcome of severe sepsis using two cohorts of European ancestry (derivation cohort, St Paul's Hospital (SPH), n = 679; validation cohort, Vasopressin and Septic Shock Trial (VASST), n = 517). The primary outcome variable was susceptibility to Gram-positive bacterial infection. The secondary outcome variable was 28-day mortality. Results Of four tested IL17A tag SNPs (rs4711998, rs8193036, rs2275913, rs1974226), rs1974226 SNP was associated with altered susceptibility to Gram-positive infection in the derivation SPH cohort (corrected P = 0.014). Patients having the rs1974226 GG genotype were more susceptible to Gram-positive infection, compared to AG/AA genotype in the two cohorts of severe sepsis (SPH, P = 0.0036, odds ratio (OR) 2.19, 95% confidence interval (CI) 1.28-3.72; VASST, P = 0.011, OR 1.95, 95%CI 1.16-3.27) and in the subgroup having lung infection (P = 0.017, OR 1.90, 95%CI 1.12-3.21). Furthermore, the IL17A rs1974226 G allele was associated with increased 28-day mortality in two cohorts (SPH, adjusted OR 1.44, 95%CI 1.04-2.02, P = 0.029; VASST, adjusted OR 1.67, 95%CI 1.17-2.40, P = 0.0052). Conclusions IL17A genetic variation is associated with altered susceptibility to Gram-positive infection and 28-day mortality of severe sepsis.

Recently two autosomal mutations of IL17 pathway genes were identified in patients with chronic mucocutaneous candidiasis disease, which is characterized by recurrent or persistent mucocutaneous infections by Candida albicans and Staphylococcus aureus [16]. The mutations, single nucleotide substitutions in the coding region of IL17 genes, abolish the IL17 immune response leading to persistent infections [16]. IL17 deficiency appears to differentially increase susceptibility to infection including S. aureus [12] and C. albicans [13][14][15]. These discoveries highlight the importance of the IL17A gene on infection and, in particular, on Gram-positive and fungal infections. Severe sepsis is a leading cause of death in intensive care units (ICUs) [17]. Multiple studies have shown that single nucleotide polymorphisms (SNPs) in important immune response genes alter susceptibility to infection and/or outcome of severe sepsis [18][19][20][21], but the role of IL17A SNPs has not yet been investigated in severe sepsis. Gram-positive infections, including S. aureus, are very common pathogens isolated from severe sepsis patients [22,23], while fungal infections are somewhat less common. Thus, we tested the hypothesis that IL17A genetic polymorphisms alter susceptibility of patients to Gram-positive infection using isolated pathogens from patients in two large severe sepsis cohorts of European ancestry. We subsequently tested for association of IL17A genotype with survival from severe sepsis.

Patients St Paul's Hospital (SPH) Cohort
Severe sepsis was defined by the presence of two or more diagnostic criteria for the systemic inflammatory response syndrome [24], proven or suspected infection, and at least one new organ dysfunction by Brussels criteria [25]. Septic shock was defined by severe sepsis plus hypotension [25] despite adequate fluid resuscitation. All patients admitted to the ICU at St Paul's Hospital (SPH) in Vancouver, Canada between July 2000 and January 2004 were screened (n = 1, 626). Of these, 691 patients of European ancestry had severe sepsis, were extensively phenotyped [26], and had DNA available. Twelve patients, who were also enrolled in the Vasopressin and Septic Shock Trial (VASST) [25] were excluded from this cohort to avoid duplication. Thus, 679 severe sepsis patients, including 453 septic shock patients, were analyzed in this study. The Institutional Review Board at SPH and the University of British Columbia (UBC) approved the study. For this fully anonymized analysis the UBC and SPH Review Boards waived the need for informed patient consent.

Vasopressin and Septic Shock Trial (VASST) Cohort
VASST was a multicenter, randomized, double blind, and controlled trial evaluating the efficacy of vasopressin versus norepinephrine in a total of 778 septic shock patients [25]. Of these, 517 patients of European ancestry had DNA available and were included in the analysis. The research ethics boards of all participating institutions approved this trial and written informed consent was obtained from all patients or their authorized representatives. The research ethics board at the coordinating center (UBC) approved the genetic analysis.

Microbiology
We assessed all microbiological cultures taken at the time of admission to the ICU for patients in the SPH cohort. Results from cultures that were collected from 48 hours prior to enrollment to 48 hours after enrollment in the VASST cohort [25] were analyzed. We defined bacterial infection in this study as clinical evidence of infection associated with positive microbiological culture, excluding the cultures judged to be positive due to contamination or colonization by the attending physician. We categorized positive microbial cultures into Grampositive bacteria, Gram-negative bacteria or fungus since these are the three major pathogen categories indentified in previous studies of sepsis in ICUs [23,27], and included in the analysis. The source of a positive culture was categorized as lung (sputum), blood, abdomen (peritoneal fluid, abscess drainage, biliary tract), skin (soft tissues or wounds), genitourinary system (urine) or other.

Selection of SNPs and genotyping
Tag SNPs for IL17A gene were identified using a multimarker tagging algorithm of Tagger software [28]. To identify tag SNPs, we used the sequence of the IL17A gene plus 2, 000 bp of 5' upstream sequence and considered all SNPs with a minor-allele frequency (MAF) > 10% in the HapMap Phase 2+3 Utah residents with Northern and Western European ancestry from the Centre d'Etude du Polymorphisme Humain (CEPH) collection (CEU) data [29] composed of 174 individuals and chose an r 2 threshold of 0.5. This identified four tag SNPs (rs4711998, rs8193036, rs2275913, rs1974226), which were genotyped in the SPH derivation cohort. IL17A rs1974226 was genotyped in the VASST replication cohort. DNA was extracted from the buffy coat of discarded blood samples using a QIAamp DNA maxi kit (Qiagen, Mississauga, ON, Canada) and genotyped using the Illumina Golden Gate assay (Illumina, San Diego, CA). We performed quality control of the genotyping using 5% repeat and HapMap Coriell DNA controls. For 5% repeat genotyping and Coriell controls, the concordance rate was > 99%. We did not discard any of the IL17A SNPs based on the quality control metrics.

Statistical analysis
The primary outcome variable was susceptibility to Gram-positive infection. To screen the four SNPs of the IL17A gene in the derivation cohort we used an Armitage's trend test followed by Bonferroni correction for multiple comparisons. Power for the genetic association of susceptibility to infection by Gram-positive bacteria was calculated using the Genetic Power Calculator [30]. We tested a joint analysis across the SPH and VASST cohorts on genetic susceptibility to Gram-positive, Gram-negative infection and Gram-positive infection by site of infection using a logistic regression controlling for the cohort. For the secondary analysis, we chose logistic regression to test for the genetic effect on 28-day mortality to allow for correction of potential confounding factors, including age, gender, surgical versus medical primary diagnosis and septic shock as covariates.
We tested for differences in baseline characteristics by genotype using a chi-square test for categorical data and a Kruskal Wallis test for continuous data. We tested for Hardy-Weinberg equilibrium using a chi-square test (threshold P-value < 0.05 was considered significant). We tested for differences in survival curves using a logrank test for trend. R2 measure of linkage disequilibrium was calculated using HapMap Phase 3 genotyping data (Han Chinese in Beijing (HCB), Japanese in Tokyo (JPT), CEU) in the genome variation server [31]. Differences were considered significant using a two-tailed P < 0.05. Analyses were performed using R (version 2.8.1) [32] and SPSS (SPSS, version 16, Chicago, IL) statistical software packages.

Results
A total of 679 severe sepsis patients of European ancestry in the derivation SPH cohort were successfully genotyped for four tag SNPs of IL17A. Patients with a positive microbiological culture (n = 301) had similar baseline characteristics to the entire population in the derivation cohort (Table 1) and had similar allele frequencies to those of HapMap European ancestry data ( Table 2). IL17A is reported to be essential for host defense against Gram-positive bacterial infection [1]. We first tested for the association of four SNPs with susceptibility to Gram-positive infection using the Armitage trend test for the additive model in the derivation cohort (Table 2). Of the four SNPs, the major G allele of IL17A rs1974226 G/A was significantly associated with increased susceptibility to infection by Gram-positive bacteria (percentages of positive culture: AA 38.5%, AG 46.4%, GG 64.2%, uncorrected P = 0.0037, Bonferroni corrected P = 0.014) ( Table 2).
Whether the observation that IL17A rs1974226 genotype altered susceptibility to infection, is specific to Gram-positive bacteria is unknown. Therefore, we next tested for genetic association of altered susceptibility to infection by three pathogen categories including Grampositive bacteria, Gram-negative bacteria and fungus in the SPH derivation cohort. Due to the small sample size of the minor homozygote AA genotype patients who had a positive culture (SPH AA genotype; Gram-positive n = 5, Gram-negative n = 8, fungus n = 0, mixed n = 1, total n = 14), we compared AA+AG genotype versus GG genotype in this analysis. In the derivation SPH cohort, the GG genotype patients had increased susceptibility to Gram-positive bacterial infection (AA+AG vs. GG, Gram-positive, P = 0.0036, odds ratio (OR) 2.19, 95% confidence interval (CI) 1.28-3.72) and decreased susceptibility to Gram-negative bacterial infection (AA +AG vs. GG, Gram-negative, P = 0.0086, OR 0.49 95% CI 0.29-0.84) (Figure 1). We subsequently tested for replication using the validation VASST cohort of European ancestry, which was successfully genotyped for IL17A rs1974226. VASST patients with a positive microbiological culture (n = 282) had a similar allele frequency (MAF = 0.195) compared to HapMap and SPH (Table 1) and were in Hardy-Weinberg equilibrium (P = 0.63). In accord with the observation in the derivation We further tested for the altered genetic susceptibility to Gram-positive bacteria by site of infection (Table 1). We combined two cohorts due to smaller subset sample size and analyzed using a logistic regression controlling for the cohort. The GG genotype patients had significantly increased Gram-positive bacterial infections compared to the AA/AG genotype in the lung (P = 0.017, OR  1.90, 95%CI 1.12-3.21), and non-significant trends in the same direction were observed in the other sites (blood, P = 0.53; abdomen, P = 0.91; skin and soft tissue (SST), P = 0.22; genitourinary system (GU), P = 0.25) (Figure 2). We next tested whether the IL17A rs1974226 altered the 28-day mortality of septic shock in the two cohorts. Because VASST only included septic shock patients, we limited our analysis of the SPH cohort to septic shock in order to examine cohorts of similar overall severity of illness. There was no baseline difference by the genotype of rs1974226 in age, gender, Acute Physiology and Chronic Health Evaluation (APACHE) II, surgical versus medical, pre-existing conditions, physiological or laboratory variables in the SPH cohort (Table 3). In the VASST cohort, patients who had rs1974226 GG or AG genotype had a decreased Pa O2 /F IO2 ratio compared to AA genotype in parallel with an increased rate of GG or AG genotype patients of chronic lung disease ( Table 3). The GG genotype patients had increased mortality over 28 days compared to the AG or AA genotype in the two cohorts of septic shock (Figure 3, SPH, P = 0.029; VASST, P = 0.010, log-rank test for trend: Table 4, G

Discussion
We found that patients who had the GG genotype of IL17A rs1974226 G/A SNP had increased susceptibility to Gram-positive bacterial infection in the derivation cohort and this result was replicated in the validation cohort. In addition, this genotype was significantly associated with adverse clinical outcome of severe sepsis. We found that the G allele of the IL17A SNP was associated with increased 28-day mortality in both cohorts of severe sepsis/septic shock patients. IL17A is an essential cytokine for host defense against bacteria [1,9], which is produced by a variety of cells [2,3] including T cells (T helper 17, γδ T, NKT cell), lymphoid-tissue inducer-like cells [33], neutrophils [34] and intestinal Paneth cell [35]. Cells triggered by microbes secrete IL17A, which is then recognized by an IL17 receptor [2][3][4]. Subsequent cellular signaling activates inflammatory pathways including NF-kB and MAPK/AP-1, which invoke production of pro-inflammatory cytokines, chemokines and antimicrobial peptides, which induce inflammation required for host defense [2][3][4]. Human IL17A-producing cells increase IL17A production against bacteria or their toxins [6,7]. Serum IL17A levels were elevated in bacterial sepsis patients with acute respiratory distress syndrome (ARDS) [8] and in animal models of abdominal bacterial infection [9,10]. IL17 deficiency in mice impairs microbial clearance, delays recovery from infection [1,11] and increases susceptibility to infection including S. aureus [12] and C. albicans [13][14][15]. In accord with the observations in IL17 deficient mice, low birth weight infants who had blood stream infections had deceased blood IL17 levels compared to those without blood stream infections [36]. These results are in accord with our current observations that severe sepsis patients who had the IL17A rs1974226 GG genotype had increased susceptibility to Gram-positive infection.
The human IL17A gene is composed of three exons (two introns) and located on chromosome 6p12 covering 4, 252 bases of genomic DNA. The transcript (1, 859 bp) has a relatively long 3'UTR region (1, 345 bp), where the rs1974226 SNP is located. Since the 3'UTR is involved in regulation of gene expression, such as mRNA stability and/or degradation as well as translation efficiency [37,38], a potential mechanism of the rs1974226 effect is alteration of gene regulation. Alternatively, other SNPs in high linkage disequilibrium with this SNP may have a biological function. In the Caucasian population the at-risk allele (G) is the major allele of rs1974226 G/A, potentially marking a haplotype that includes many rare functional SNPs that increase risk, or the rarer A allele may actually be protective. Of the four IL17A SNPs screened in this study, two IL17A Figure 3 Survival curves over 28 days by IL17A rs1974226 genotype in two cohorts of septic shock. Patients with IL17A rs1974226 GG genotype had increased mortality in the SPH and VASST cohort of septic shock compared to the AG or AA genotype patients (SPH P = 0.029, VASST, P = 0.010). SPH, St Paul's Hospital; VASST, Vasopressin and Septic Shock Trial. P values were calculated using a log-rank test for trend.
The GG genotype was associated with increased susceptibility to Gram-positive bacterial infection in two cohorts of severe sepsis patients, whereas a trend towards decreased susceptibility to Gram-negative bacteria was also observed. Genetic associations of pathogen specific susceptibility have been documented in pattern recognition receptors (PRRs) such as Toll-like receptors (TLRs) [42]. The potential mechanism of the different pathogen-specific susceptibility may be difference of PRRs.
IL17 induces neutrophil recruitment in the airways [43], which is an important mechanism of host defense for the lung. In accordance with this, we found that the rs1974226 GG genotype patients had significantly increased susceptibility to Gram-positive infection in the lung. While this association was significant for the lung, associations were not significant for other sites. This might be due to the small sample sizes of infections found in other sites. Other limitations of this study are, first, that we did not investigate mechanisms of action such as altered IL17A gene expression or alternative splicing by the IL17A rs1974226 SNP. Thus, further investigation regarding the genetic effect on IL17A mRNA structure and mRNA expression levels and protein levels would strengthen the results of this study. Second, the GG genotype had increased susceptibility to Gram-positive infection and 28-day mortality, but the causal link is not proven in this study. Third, this study finding of an association of the IL17A rs174226 GG genotype with increased susceptibility to a category of Gram-positive bacteria would be strengthened by testing for an association with a specific Gram-positive bacterial pathogen, such as S. aureus, in further larger studies.
Decreased susceptibility to Gram-negative bacteria observed in the GG genotype patients in the derivation cohort did not replicate; this needs further replication tests in larger cohorts.

Conclusions
To conclude, the IL17A rs1974226 GG genotype is associated with increased numbers of Gram-positive infections and increased 28-day mortality in severe sepsis patients.

Key messages
• Patients of European ancestry having the IL17A rs1974226 GG genotype were more susceptible to Gram-positive infection, compared to those having the AG/AA genotype in the derivation and validation cohorts of severe sepsis.
• The IL17A rs1974226 GG genotype patients had significantly increased Gram-positive bacterial infection compared to the AA/AG genotype in the subgroup having lung infection.
• The IL17A rs1974226 G allele was associated with increased 28-day mortality in both cohorts of severe sepsis/septic shock patients.