Background
Sepsis is a major health challenge. Despite improved treatment options, sepsis remains a leading cause of death in intensive care units[
1]. Lipopolysaccharide (LPS), or endotoxin, the major outer membrane component of gram-negative bacteria, is a potent inflammatory response stimulator[
2]. In addition, LPS triggers inflammation in gram-negative sepsis[
3]. Excessive amounts of gut-derived LPS released during intestinal hypo-perfusion have also been implicated in sepsis caused by gram-positive and fungal infections[
4,
5]. LPS signaling is initiated by the activation of the myeloid differentiation factor 2 and toll-like receptor 4 (TLR4) complexes on myeloid cells[
2,
6]. TLR4 has recently been shown to recognize endogenous danger-type, or ‘alarmin,’ factors, thereby implicating TLR4 as a tissue injury and microbial invasion sensor[
7]. Studies using mouse strains deficient in TLR4 signaling[
8,
9] or expression[
10‐
13] or those using TLR inhibitors in wild-type mice[
14,
15] confirmed that TLR4 contributes to bacterial clearance and the host inflammatory response in the infection setting[
16].
Two missense single nucleotide polymorphisms (SNPs) in the
TLR4 gene, Asp299Gly/Thr399Ile, have been reported to be associated with endotoxin hypo-responsiveness to inhaled LPS[
17]. This investigation was followed by a series of studies that explored the potential impact of these SNPs on the incidence and course of infectious diseases[
18], such as septic shock with gram-negative bacterial infection[
19]. Although some studies have shown a relevance of the Asp299Gly/Thr399Ile SNPs in gram-negative infections, others did not confirm this association[
20‐
22]. Furthermore, recent studies using primary cells isolated from individuals bearing these mutations have indicated that the Asp299Gly/Thr399Ile haplotype has little or no effect on LPS responsiveness[
23].
Recently, Sato et al. demonstrated the biological significance of a genetic variation of the
TLR4 gene called rs11536889. Functional analyses revealed that
TLR4 rs11536889 contributes to the translational regulation of TLR4 expression and has some influence on the response to LPS, possibly by binding to microRNAs, which act in post-transcriptional regulation[
24]. A large study that included prostate cancer patients and age-matched controls from Sweden revealed an association between
TLR4 rs11536889 and prostate cancer[
25]. Later, Hishida et al. observed that
TLR4 rs11536889 genotypes are associated with severe gastric atrophy in
helicobacter pylori-seropositive Japanese subjects[
26]. Zhou et al. found that the
TLR4 rs11536889 SNP is significantly associated with
hepatitis type B virus recurrence after liver transplantation[
27]. In addition, Miedema et al. found that this SNP is associated with an increased risk of chemotherapy-induced neutropenia in children with acute lymphoblastic leukemia[
28]. These observations suggest that the rs11536889 genetic variation of the TLR4 gene may influence human inflammatory and/or malignant diseases[
24].
This study aimed at exploring whether the putative regulatory TLR4 rs11536889 genotypes relate to organ failure severity in critically ill patients with sepsis during their time in the intensive care unit. The outcomes of wild-type GG were compared to those of GC/CC.
Discussion
The present study addressed the question of whether the putative regulatory TLR4 rs11536889 genotypes are related to organ failure in critically ill patients with sepsis.
The primary endpoint, organ failure, was quantified using SOFA scores as a specific clinical marker in patients with sepsis and was significantly higher in
TLR4 rs11536889 GG patients compared with those of
TLR4 rs11536889 GC and CC patients (Table
2).
The
TLR4 rs11536889 genotype distribution among the septic patients was similar to database entries regarding healthy Caucasians and also followed the Hardy-Weinberg equilibrium. The
TLR4 rs11536889 genotypes, however, were not associated with any of the recorded baseline characteristics. As assessed by scoring the sepsis type (sepsis and severe sepsis vs. septic shock) and SOFA and APACHE II scores, we found that the
TLR4 rs11536889 genotypes were also not related to the septic disease severity at sepsis onset (Table
1). We believe that the SOFA scores at sepsis onset did not differ between the GG versus GC and CC genotypes because of the phenotypic heterogeneity of the sepsis syndrome. This heterogeneity is influenced by many factors, including the pathogenic organism responsible for the infection, its location, and the amount of time passed since the onset of infection, as well as other individual parameters. To detect genotypic differences, a longitudinal observation involving SOFA scores quantified over the study period is much more promising (Table
2).
As shown by Sato et al., monocytes from
TLR4 rs11536889 CC subjects expressed significantly higher levels of TLR4 compared with those from TLR4 rs11536889 GG and GC subjects. When PBMCs were stimulated with LPS, a TLR4 ligand, the cells from the
TLR4 rs11536889 CC and GC subjects secreted significantly higher levels of the proinflammatory cytokine IL-8 compared to cells from the GG subjects[
24]. Accordingly, these previous investigations support the assumption that
TLR4 rs11536889 GG sepsis patients present severe organ dysfunction (as measured using SOFA scores) because of attenuated TLR4 proinflammatory signaling in response to LPS compared to C allele carriers.
These significant results, with respect to organ dysfunction, reveal severe morbidity among
TLR4 rs11536889 GG patients (according to SOFA scores) and together with the fact that GG patients are assumed to present attenuated TLR4 expression[
24], offer an explanation why synthetic TLR4 antagonists have failed to produce a clinical benefit in patients with severe sepsis[
14,
15]. These agents may alter the inflammatory response via TLR4 to pathogens, thereby contributing to organ dysfunction in these patients.
The SOFA-renal score was higher in
TLR4 rs11536889 GG patients, indicating severe renal dysfunction in this group. Because
TLR4 rs11536889 GG patients may exhibit decreased
TLR4 expression, our results are consistent with former observations indicating that decreased TLR4 expression in chronic kidney disease patients was associated with attenuated proinflammatory cytokine synthesis during infection[
34]. The observed severe hepatic dysfunction measured using the SOFA-hepatic score among
TLR4 rs11536889 GG subjects, indicating severe hyperbilirubinemia in this group, is in agreement with recent findings reported by Deng et al.[
16] that TLR4 signaling is essential for LPS clearance by hepatocytes during sepsis. The absence of an association between the rs11536889 genotypes and the SOFA respiratory score may be attributed to the fact that the SOFA respiratory score is somewhat weak because it only refers to the oxygenation index. This score depends on several factors, such as ventilator settings during mechanical ventilation and different ventilator settings that result in different oxygenation indices, which lead to score variation. We believe that there was no significant difference in the SOFA-CNS score between the genotypes mainly because sepsis patients are treated with sedating medication, which impacts the CNS and thus affects the SOFA-CNS score. The SOFA-Cardiovascular score most likely did not differ between GG patients and C allele carriers because this score is only based on the catecholamine therapy needed, which depends on volumetric status and cardiac function.
Analysis of the 28-day and 90-day mortality revealed no significant differences among the
TLR4 rs11536889 genotypes. The severe organ failure observed in G homozygous patients may not have contributed to higher mortality rates because the patients received sufficient intensive care treatment, which allowed their organ failures to be managed appropriately. The patients were treated according to current guidelines for the treatment of sepsis (Surviving Sepsis Campaign)[
35].
Additionally, there was a significantly higher gram-negative infection rate among the C allele carriers (81%) compared with the rate observed in the GG patients (62%; p = 0.0062) (Table
3). This result is in agreement with previous observations linking this polymorphism with an increased susceptibility to infection[
24,
26,
27]. This observation of higher susceptibility of C allele carriers to gram-negative infections should be thoroughly examined in future studies to detect any causality between the SNP and susceptibility to gram-negative infections.
A possible limitation of this study is the possibility that the studied TLR4 rs11536889 SNP associated with organ failure in patients with sepsis is in linkage disequilibrium with SNPs in another nearby gene and that these latter genes are responsible for the observed phenotypic effects.
To the best of our knowledge, this is the first investigation evaluating this putative regulatory polymorphism in this key innate immune receptor in adult Caucasian sepsis patients, showing a significant association between the TLR4 rs11536889 GG genotype and the severity of organ failure (renal, coagulation and hepatic). According to these results, it would be worthwhile to further assess the TLR4 rs11536889 polymorphism for its relevance to sepsis in larger and independent cohorts.
Competing interests
The authors declare no competing interests.
Authors’ contributions
All authors have contributed to the study design, the acquisition of data (clinical and experimental), or the analysis and interpretation of data. Specifically, LVG and MS primarily performed sample and clinical data collection and TLR4 laboratory analyses. MS, AP, IB, DR, MG and MB participated in study design and supervised patient enrollment and clinical data monitoring. TB contributed to the study design and conception, performed the bioinformatics and performed and approved statistical analyses. AM and JH designed the study, supervised sample and data collection, performed analyses and drafted the manuscript. All authors were involved either in manuscript drafting or its revision. All authors have approved the final version of the manuscript.