The NFKB1  Promoter Polymorphism (−94ins/delATTG) Alters Nuclear Translocation of NF-κB1 in Monocytes after Lipopolysaccharide Stimulation and Is Associated with Increased Mortality in Sepsis

• The nuclear factor-κB and activator protein-1–coupled signaling pathway of the innate immune system is known to amplify and perpetuate inflammatory and coagulatory mechanisms in sepsis

• Genetic variations that alter nuclear factor-κB gene expression could affect pathophysiologic mechanisms and thereby influence mortality in severe sepsis

• A functional insertion–deletion polymorphism in the promoter of nuclear factor-κB and activator protein-1 affected expression and regulation of mediators of inflammation and coagulation in vitro 

• Expression of the same polymorphism was associated with increased 30-day mortality in patients with severe sepsis

THERE is wide variability regarding the outcome in severe sepsis, which in part may be caused by genetic variations.1–3A potential candidate for such variations is the gene encoding nuclear factor-κB and activator protein-1 (NF-κB1), because the ubiquitous transcription factor NF-κB1 binds to recognition elements in the promoter regions of several genes encoding the innate immune and the coagulation system. The nuclear factor-κB family comprises five proteins, NF-κB1, NF-κB2, RelA, RelB, and c-Rel. Nuclear factor-κB family members share structural homology with the retroviral oncoprotein v-Rel, resulting in their classification as NF-κB/Rel proteins. The NFKB1  gene encodes the NF-κB-p105 subunit. NF-κB1 and NF-κB2 proteins are translated as large precursors, p105 and p100, which undergo processing, regulated by the ubiquitin–proteasome pathway, to build the mature nuclear factor-κB subunits, p50 and p52.

In most cells, NF-κB1 is found in the cytoplasm in its inactive form, bound to inhibitory proteins. Many extracellular stimuli, including bacterial lipopolysaccharide, viruses, oxidants, inflammatory cytokines, and immune stimuli, can activate NF-κB1.4,5Once activated, it binds to regulatory DNA elements in the promoter regions of inflammatory and immune response genes, such as those encoding proinflammatory cytokines, chemokines, enzymes relevant for inflammation, and adhesion molecules.6,7Furthermore, this pathway induces tissue factor in the blood stream and, thus, activates coagulation.7–10Consequently, disseminated intravascular coagulation is a common finding during severe sepsis, which in turn can evoke ischemia and multiple organ dysfunction.8,9In conclusion, the NF-κB1–coupled pathway is known to amplify and perpetuate inflammatory and coagulatory mechanisms prevailing in sepsis.10Accordingly, genetic variations that alter NFKB  gene expression could impact key mechanisms in sepsis and, therefore, influence mortality in severe sepsis.

Recently, a functional insertion–deletion polymorphism in the promoter of NFKB1  (−94ins/delATTG in relation to the transcription initiation site or −24.219ins/delATTG in relation to the A as +1 of the initiation codon ATG), which encodes the major isoform of NF-κB1, was reported to be associated with altered NFKB1  gene expression and an increased risk for ulcerative colitis.11Furthermore, we found that the deletion allele influences severity but not mortality of patients suffering from acute respiratory distress syndrome.12In conclusion, there are many reasons to suspect that the NFKB1  insertion–deletion (−94ins/delATTG) polymorphism could impact key mechanisms in sepsis and, therefore, influence mortality in severe sepsis.

Accordingly, we prospectively tested the hypotheses that the NFKB1  insertion–deletion (−94ins/delATTG) polymorphism (1) alters nuclear translocation of NF-κB1 in monocytes after lipopolysaccharide stimulation; (2) affects lipopolysaccharide-induced NFκB1 messenger RNA (mRNA) expression, tumor necrosis factor-α (TNFα) concentration, and tissue factor activity; and (3) may be associated with increased 30-day mortality in patients with severe sepsis.13 

After ethics committee approval (ethics committee of the University Hospital Essen, Essen, Germany) and written informed consent, venous blood (24 ml) was withdrawn from 10 healthy individuals with the homozygous insertion (II) or deletion (DD) genotype, respectively, and centrifuged at 1800g  for 20 min using Ficoll density gradient centrifugation tubes (Vacutainer CPT tubes; Becton Dickinson, Franklin Lakes, NJ). Monocytes were resuspended in RPMI 1640 medium (Gibco Products Invitrogen Corporation, Grand Island, NY) containing 5% fetal calf serum (Biochrom AG, Berlin, Germany) and antibiotics (100 U/ml penicillin and 100 μg/ml streptomycin; Invitrogen Corporation, Carlsbad, CA). The cell suspension was added to cell culture tubes, and monocytes were allowed to adhere to the surface of the tubes for 2 h. Finally, the supernatant was discarded, fresh RPMI 1640 medium was added, and the cells were allowed to rest for 48 h (37°C; 5% CO²in air) before the experiments. Meanwhile, 12 mm glass plates were coated with 400 μl fibronectin (1 mg/ml; Sigma-Aldrich, St. Louis, MO) inside 24-well plates for 24 h. Afterward, the supernatant was discarded and 500 μl cell suspension (1 × 10⁶monocytes ml^-1) was added. Cells were incubated for 60 min either with lipopolysaccharide (10 μg/ml, serotype 0111:B4, Sigma-Aldrich) or without lipopolysaccharide. After incubation, cells were fixed with ice-cold methanol–acetone (1:1) for 10 min at −20°C. Immunofluorescence staining was performed using a primary NF B rabbit antihuman polyclonal p65 antibody (1:200 dilution; Santa Cruz Biotechnologies, Santa Cruz, CA) followed by a Immunoglobulin G - Alexa-flour 568 coupled goat anti-rabbit antibody (1:400 dilution; Molecular Probes, Eugene, OR), as described previously.14 

An independent investigator, blind for the specific NFKB1  promoter polymorphism and stimulus applied, processed all immunofluorescence slides. Fluorescence microscopy was performed on a Nikon Eclipse E1000 microscope (Nikon GmbH, Düsseldorf, Germany) using the NIS-Elements F.30.0 imaging software (Laboratory Imaging, Prague, Czech Republic). All slides were analyzed in a standardized order, that is, a representative image of each quadrant was captured at 20-fold magnification, and nuclear NFκB-positive cells were counted using an Image J software (ImageJ; National Institute of Health, Bethesda, MD).

After approval by the ethics committee of the University Hospital Essen and written informed consent, venous blood was drawn from the antecubital vein of 105 healthy volunteers. After discarding the first 2 ml, blood was collected in one-tenth volume of citrate (Sarstedt Monovette, Nümbrecht, Germany) and samples were immediately used for the experiments. Subsequently, whole blood samples were incubated with lipopolysaccharide at final concentrations of 0, 25, 75, and 225 µg/ml for 4 h at 37°C. Thereafter, samples were subjected to thromboelastometry, and mRNA and DNA were isolated as described in DNA Genotyping. In addition, TNFα concentration was determined in plasma obtained from whole blood samples after incubation with lipopolysaccharide (225 µg/ml; n = 60).

To further elucidate the mechanism of the lipopolysaccharide-induced activation of coagulation, lipopolysaccharide- and vehicle (NaCl 0.9%)-treated whole blood samples were incubated with the protein synthesis inhibitor cycloheximide (35 µM) and with an inhibitor of tissue factor effects, active site–inhibited factor seven (50 µg/ml), for 30 min, subsequently challenged with lipopolysaccharide (225 µg/ml) or vehicle (NaCl 0.9%) for 4 h, and then subjected to thromboelastometry as recently described.15,16To assess the tissue factor concentrations induced by lipopolysaccharide, a standard curve was generated by addition of tissue factor standard to whole blood samples and the resulting clotting time was determined.

Citrate blood samples were subjected to rotational thromboelastometry (Roteg 5^TM, Pentapharm, Munich, Germany), a modification of the original thromboelastography described in 1948 by Hartert,17and coagulation was initiated by recalcification with calcium chloride.18,19Clotting times were determined from the thromboelastogram as described in References 18 and 19. Kinetics of clot formations are measured by the clot formation time and the angle α. The maximum amplitude reflects the strength of the clot and is dependent on both the number and function of platelets and their interaction with fibrin.18,19 

TNFα plasma levels were determined by using a microsphere-based multiplexing system kit (Invitrogen, Paisley, United Kingdom). As previously shown, the assays were performed according to the manufacturer’s protocol and concentrations measured with a Luminex 100 system (Luminex, Austin, TX).20 

RNA from whole blood (n = 105) was extracted using the RNeasy kit (Qiagen, Hilden, Germany). First-strand complementary DNA was synthesized from 0.6 µg of total RNA. Quantification of NF-κB1 mRNA was performed by using the intron spanning primer (forward primers) 5′-GTGAAGGCCCATCCCATGGT-3′ and (reverse primer) 5′-TGTGACCAACTGAACAATAACC-3′ resulting in a 122-bp fragment. PCR and primers for the housekeeping gene hβ-actin were used as described.12,21The real-time PCR reaction was performed as described using the Quantitect SYBR Green Kit (Qiagen).21,22A complementary DNA dilution series confirmed a PCR efficiency greater than 95%, which was comparable to the efficiency of hβ-actin. Relative NF-κB1  mRNA expression was measured by two-step real-time PCR with hβ-actin as internal control and calculated as 2^{−[Ct(NFKB1 )-Ct(β-Actin)]}.

Genomic DNA of patients was extracted from whole blood using standard methods (QIAamp, Qiagen). Genotypes of the insertion–deletion polymorphism were determined by pyrosequencing. A 200-bp PCR fragment was amplified using primer NFKB1  _del/ins_f(5′-ATGGACCGCATGACTCTATCAG-3′) and biotinylated primer NFKB1  _del/ins_BIO_r(5′-GGGGCGCGCGTTAGGCGG-3′). PCR was performed at an annealing temperature of 60°C in a 50-μl reaction mixture applying a commercially available PCR master mix (Eppendorf, Hamburg, Germany). Pyrosequencing was performed on a PSQ96 MA (Pyrosequencing, Uppsala, Sweden) using sequencing primer NFKB1  _del/ins_seq (5′-CGTTCCCCGACCAT-3′). Randomly chosen samples were reanalyzed with a different nucleotide injection order for genotype confirmation.

Lipopolysaccharide (Escherichia coli ; serotype 0.111:B4) was obtained from Sigma-Aldrich. Tissue factor standard was obtained from Thrombinoscope BV, Maastricht, Netherlands. Active site–inhibited factor VIIa was a generous gift from Novo Nordisk, Zurich, Switzerland. All other reagents were of analytical grade.

This study was reviewed and approved by the ethics committee of the University Hospital Essen. Over a period of 2 yr, 143 patients (93 men and 50 women, mean age: 57 ± 16 yr) admitted to an intensive care unit of the University Hospital of Essen were considered eligible for the study if they fulfilled the criteria for severe sepsis as defined by Bone et al .23Informed consent was obtained for all patients from the guardian of the patient. Exclusion criteria were age more than 18 yr and no affiliation to the Caucasian ethnicity or no informed consent was obtained. All patients were white Germans of Caucasian ethnicity. Clinical and demographic data on study entry, including Simplified Acute Physiology Score II24and the Sequential Organ Failure Assessment score,25were calculated over the first 24 h after the patient met severe sepsis criteria (table 1). All patients were followed up for 30-day mortality. There were no dropouts from the procedures or data lost in any fashion. Patients were treated with a multimodal concept, which included analgosedation, fluid administration, and protective mechanical ventilation, hemodynamic, antibiotic, and diagnostic management. Continuous hemofiltration/dialysis was initiated and technically performed by the Department of Nephrology according to standardized protocols.

DNA was isolated and genotyped for the NFKB1  promoter polymorphism (−94ins/delATTG) as described in DNA Genotyping.

After incubation, the percentage of nuclear NFκB-positive cells was normalized to the percentage of nuclear NFκB-positive cells in unstimulated controls. The Kolmogorov–Smirnov test and Shapiro–Wilk normality test revealed a normal distribution for nuclear NFκB-positive cells, and the Student t  test for unpaired samples was used to analyze differences between groups. Semiquantitative analysis was performed using the Graphpad Prism 5 statistics program (Graph Pad Software, San Diego, CA).

To test the effect of the NFKB1  promoter polymorphism (−94ins/delATTG) on NF-κB1 mRNA expression, TNFα concentration, and tissue factor activity after lipopolysaccharide stimulation, heterozygous deletion (ID) and DD genotypes were combined and then tested against the II genotype, because there were only five volunteers with the DD genotype. Potential deviation from the Hardy–Weinberg equilibrium was tested with the statistic program Excel of Microsoft Office 2010 (Microsoft Deutschland, Unterschleißheim, Germany).

Statistical analyses were performed using two-way ANOVA followed by post hoc  testing with Bonferroni Holm, respectively, using SPSS 13.0 (SPSS, Chicago, IL) or Graphpad Prism 5.0. Data are presented as mean and standard error or SD of the mean, as indicated. For the determination of tissue factor concentrations from clotting time, a four-parameter Hill function was used for the curve fit (SigmaPlot Software, San Jose, CA).

The NFKB1  promoter polymorphism (−94ins/delATTG) genotype distributions were tested for deviations from Hardy–Weinberg equilibrium (exact two-sided P  value 1.00) in the sepsis cohort. Explorative comparisons by NFKB1  promoter polymorphism (−94ins/delATTG) genotypes (ID/DD vs.  II) were performed for several clinical characteristics of the sepsis patients (table 1). ID and DD were combined because of the low frequency of the DD genotype.

The clinical endpoint was survival over the first 30 days dependent on NFKB1  promoter polymorphism (−94ins/delATTG) genotypes. Survival probabilities were graphically assessed by the Kaplan–Meier method, and the log-rank test was used to evaluate the univariate relationship between NFKB1  genotypes and clinical outcome. Thereafter, we performed multivariate Cox regression analyses to assess the joint impact of NFKB1  promoter polymorphism (−94ins/delATTG) genotypes, sex, age, Simplified Acute Physiology Score II, Sequential Organ Failure Assessment score, requirement of continuous hemofiltration/dialysis, C-reactive protein, and interleukin 6 concentration as predictors for the clinical outcome (30-day survival). Model diagnostic of the proportional-hazards assumption for the NFKB1  promoter polymorphism (−94ins/delATTG) genotypes comprised both graphical and formal investigations—none of which indicated strong evidence for a deviation from the proportional-hazards assumption. The multivariate analyses included two steps with a focus on NFKB1  promoter polymorphism (−94ins/delATTG) genotypes and 30-day survival. In the initial model, all main effects were simultaneously investigated (table 2). To avoid overfitting, a restricted model with only four variables was assessed afterward using only those predictors with a P  value ⩽0.05 in either the univariate or the initial multivariate comparison (table 2).

CIs were calculated with coverage of 95% (abbreviated 95% CI). All reported P  values are nominal, two-sided, and we applied a significance level α of 5%.

After lipopolysaccharide stimulation, the DD genotype was associated with a 5.5-fold (± 2.9 SD) increase in nuclear NFκB-positive cells normalized to controls (incubation with vehicle), compared with a 2.8-fold (±0.9 SD) increase in II genotype (P =  0.001, fig. 1).

NF-κB1/actin mRNA ratios were determined in the absence and presence of lipopolysaccharide (225 µg/ml). The mRNA for NF-κB1 increased nearly threefold from 0.013 ± 0.025 to 0.034 ± 0.025 (P = 0.002) on lipopolysaccharide stimulation (fig. 2). Moreover, the NF-κB1/actin ratio was markedly associated with genotypes. As shown in figure 2, lipopolysaccharide-induced expression of NF-κB1 was three times as high in the presence of the deletion allele compared with the homozygote II genotype (0.018 ± 0.015 vs.  0.051 ± 0.001, P =  0.0001). In the absence of lipopolysaccharide, NF-κB1 expression was not associated with genotypes.

Clotting time was determined in 105 volunteers both in the absence of lipopolysaccharide and after 4 h of incubation with lipopolysaccharide. Lipopolysaccharide shortened the clotting time in a concentration-dependent manner from 612 ± 10 to 325 ± 10 s (25 µg/ml lipopolysaccharide), 282 ± 10 s (75 µg/ml lipopolysaccharide), and 220 s ± 8 s (225 µg/ml lipopolysaccharide), respectively.

Determination of NFKB1  genotypes revealed 42 subjects homozygous for the insertion allele, 45 were heterozygous, and 5 were homozygous for the deletion allele. As shown in figure 3, the clotting times of II and combined deletion allele carriers were not different in the absence of lipopolysaccharide (618 ± 17 vs.  607 ± 15 s; P = 0.58). After stimulation with lipopolysaccharide, however, the deletion allele overall lipopolysaccharide stimulation concentrations were associated with a 20% lower clotting time compared with the II genotype (25 µg/ml, P =  0.047; 75 µg/ml, P =  0.003; and 225 µg/ml, P =  0.021). Neither maximum clot formation nor clot formation time was associated with genotypes (data not shown).

To unravel the mechanism underlying the lipopolysaccharide-induced shortening of clotting time, the effects of the protein synthesis inhibitor cycloheximide and of the tissue factor inhibitor, active site–inhibited factor seven, were determined. Both drugs abolished the effect of lipopolysaccharide on clotting time, demonstrating that lipopolysaccharide exerts its effect via de novo  protein synthesis of tissue factor (fig. 4).

To determine the biologically active tissue factor concentration responsible for the observed lipopolysaccharide-induced shortening of clotting time, a concentration response curve was constructed using tissue factor standards. As shown in figure 5A, exogenous tissue factor shortened the clotting time over a broad concentration range from 20 fM to 20 pM. Correlation by nonlinear regression analysis in six experiments using a four-parameter Hill function was 0.98. Figure 5Bdepicts the effects of the genotypes of the NFKB1 insertion–deletion polymorphism on tissue factor concentration in the presence of lipopolysaccharide as calculated from the standard curve shown in figure 5A. In comparison to the homozygous II genotype, the lipopolysaccharide-induced tissue factor concentration was two- to three-fold increased in the presence of the deletion allele (25 µg/ml, P =  0.039; 75 µg/ml, P =  0.003; and 225 µg/ml, P =  0.021; fig. 5B)

Because TNFα release is affected by the activity of NF-κB1, we determined the effect of lipopolysaccharide on the concentration of this cytokine (fig. 6). Lipopolysaccharide (225 µg/ml) increased the TNFα concentration more than 1,000-fold (P <  0.001). The effect of lipopolysaccharide was not modulated by the genotypes of the NFKB1  promoter polymorphism, because testing with a two-way ANOVA after post hoc  testing with Bonferroni Holm revealed a P  value of 0.08. In addition, in the absence of lipopolysaccharide, TNFα concentrations were not different.

Thirty-day mortality was significantly associated with NFKB1 insertion–deletion (−94ins/delATTG) polymorphism genotypes (P = 0.034). Mortality was 25% for II genotypes but 41% for combined ID/DD genotypes (fig. 7). ID/DD genotypes displayed a significantly higher risk for death than patients with homozygous II genotypes (hazard ratio, 1.4; 95% CI, 1.03–1.9; P = 0.034).

Multivariate proportional hazard analysis, including sex, age, Sequential Organ Failure Assessment score, Simplified Acute Physiology Score II, procalcitonin, C-reactive protein, or interleucin-6 concentrations, and requirement for continuous hemofiltration/dialysis as covariates revealed the NFKB1  insertion–deletion (−94ins/delATTG) polymorphism as an important and independent prognostic factor for 30-day mortality. ID/DD genotypes had an almost twofold greater risk for death (hazard ratio, 2.3; 95% CI, 1.13–4.8; P =  0.022) compared with II genotypes (table 2).

The requirement of continuous hemofiltration/dialysis was also an independent and strong prognostic factor for 30-day survival (hazard ratio, 3.4; 95% CI, 1.38–8.22; P =  0.008). Genotype frequencies and the clinical characteristics of the patients grouped by NFKB1  insertion–deletion (−94ins/delATTG) polymorphism genotypes are displayed in table 1. We found no significant associations of NFKB1  insertion–deletion (−94ins/delATTG) polymorphism genotypes with sex, age, body mass index, procalcitonin, C-reactive protein, or interleucin-6 concentrations, necessity for continuous hemofiltration/dialysis, type of infection, or primary diagnosis at hospital admission. In contrast, the NFKB1  insertion–deletion (−94ins/delATTG) polymorphism was associated with a genotype-dependent increase of the Sequential Organ Failure Assessment score (P =  0.015) (table 1). The deletion allele carriers had higher Sequential Organ Failure Assessment score compared with the II genotype carriers.

Our study shows for the first time that the ID/DD genotypes of the NFKB1  insertion–deletion (−94ins/delATTG) polymorphism are significantly associated with markedly increased 30-day mortality in patients with severe sepsis. In addition, our study unravels the molecular mechanism by which the D allele may amplify and perpetuate inflammatory mechanism prevailing in sepsis.

Genotypes of the NFKB1  insertion–deletion (−94ins/delATTG) polymorphism turned out to represent both an important and independent prognostic factor for 30-day mortality. The hazard ratio of 2.3 not only suggests that the D allele of the NFKB1  insertion–deletion (−94ins/delATTG) polymorphism has important effects on NF-κB1 expression but also underscores the potential relevance of NF-κB1 expression in severe sepsis. Because wide variability exists regarding the outcome in sepsis,1,26part of this variability may be explained by the genetic variation in the promoter of NFKB1. 

In addition, our results show that the D allele is also associated with initial disease severity, because it was associated with an increased Sequential Organ Failure Assessment score. These findings are in accordance with our previous investigation, showing that the deletion allele of this polymorphism was also associated with increased severity of adult respiratory distress syndrome.12 

Of note, this study unravels molecular mechanism by which the D allele of the NFKB1  insertion–deletion (−94ins/delATTG) polymorphism seems to amplify and perpetuate inflammatory mechanism prevailing in sepsis, which may consecutively influence severity and mortality of severe sepsis. The DD genotype of the NFKB1  insertion–deletion (−94ins/delATTG) polymorphism is associated with increased nuclear translocation of NF-κB1.

This also suggests that increased nuclear translocation of NF-κB alters the expression of various genes involved in the inflammatory response such as cytokines or the hypoxia inducible factor 1 α. Hypoxia inducible factor 1 α, a well-characterized NF-κB target, is the key regulator linking inflammatory and hypoxic signaling pathways.27Therefore, polymorphisms of the NF-κB1 gene play a pivotal role in mediating both an excessive inflammatory response in humans and an induction of the hypoxic signaling. Thus, a variety of pathways are altered, resulting in disturbed vascular tone or cellular adhesion and changes in cellular receptor expression that, in concert, are likely to be associated with an unfavorable outcome.28 

Because NF-κB1 binds to recognition elements in the promoter regions of inflammatory genes, it can be speculated that the D allele potentially alters the consequences of sepsis via  the coordinated activation of several inflammatory genes of the innate and adaptive immunity. However, NF-κB1 also influences the coagulation system via  tissue factor. During sepsis, tissue factor is highly expressed on monocytes, resulting in disseminated intravascular coagulation, ischemia, and organ dysfunction.29,30In the current study, we also show for the first time that lipopolysaccharide-induced tissue factor expression was affected by the D allele of the NFKB1  promoter polymorphism. As the development of disseminated intravascular coagulation, which is mediated by tissue factor expression, is an independent risk factor in sepsis, it can be speculated that the deletion allele may also affect the course of sepsis via  increased tissue factor expression.

In the current study, whole blood from volunteers was incubated with lipopolysaccharide in vitro  to investigate the effects of the NFKB1  promoter polymorphism on toll-like receptor 4 agonist-evoked responses under well-controlled conditions. Specificity of our clotting assay for tissue factor was demonstrated by the use of active site–inhibited factor seven, because the lipopolysaccharide-induced shortening of clotting time was abolished in presence of this inhibitor. We decided to determine clotting time as a measure of tissue factor using thromboelastometry instead of protein, because this functional assay has several important advantages compared with other tissue factor assays.31,32The sensitivity of our clotting assay is high, because lipopolysaccharide-induced tissue factor in whole blood samples exceeds the detection limit of the assay more than 100-fold. Furthermore, the functional assay measures biologically relevant tissue factor activity on the surface of intact cells. In contrast, measurements in homogenates, which are commonly used for enzyme-linked immunosorbent assay, also determine the functionally irrelevant intracellular tissue factor.31,33,34Moreover, enzyme-linked immunosorbent assay and chromogenic assays, but not a clotting assay, may also detect soluble tissue factor, which is believed to be of minor physiologic relevance.31,33,34Tissue factor concentration was calculated from clotting time using a calibration curve obtained by the addition of tissue factor standard to whole blood samples. In theory, this method might be influenced by a direct effect of the NFKB1  genotypes on platelets or coagulation factors. However, no differences in platelet count or coagulation factor activity was detected, as judged by clot firmness using thromboelastometry.

Because of the study design, the question arises, whether the lipopolysaccharide concentration used in our in vitro  experiments is of clinical relevance. To study biological effects, lipopolysaccharide has been used in a wide concentration range covering the ng/ml up to the mg/ml range. In most whole blood studies, lipopolysaccharide ranged from µg/ml to mg/ml.35–38In contrast, most cell culture studies were performed with lipopolysaccharide concentrations in the ng/ml range,38–40that is, concentrations much lower in comparison to whole blood. The use of these different concentration ranges can be explained by the physicochemical properties of the lipopolysaccharide molecule. Lipopolysaccharide, as an amphiphilic molecule in biological fluids, is always bound to other molecules such as peptides and proteins.41It is well recognized that human plasma neutralizes lipopolysaccharide up to 6 µg/ml42and that lipopolysaccharide diluted in plasma before application is less effective than lipopolysaccharide diluted in crystalloid solution.43,44Accordingly, we used 10 µg /ml for the cell culture studies and 225 µg/ml lipopolysaccharide for the measurements of NFKB1 mRNA expression and TNFα protein concentration in whole blood, respectively. In our study, the EC⁵⁰for the lipopolysaccharide-induced activation of coagulation was 18 µg/ml. The EC⁵⁰of lipopolysaccharide in our ex vivo  assay can also be achieved by intravenous administration of a 2-mg/kg dose of lipopolysaccharide if an intravascular volume of 10% of the body weight is assumed. Notably, this lipopolysaccharide dosage was used in primates,45,46and injection of 1 mg lipopolysaccharide has been described in humans.47 

Limitations of this investigation should also be mentioned. Unrecognized selection bias, inherent to many genetic association studies, cannot ultimately be excluded. Moreover, although all sepsis patients were treated with a rather standardized multimodal regimen, because of the multifactorial nature of this disorder, we cannot exclude that unknown potentially confounding factors exist. Nevertheless, the study population was rather large, and multivariate proportional hazard analysis revealed the NFKB1  insertion–deletion (−94ins/delATTG) polymorphism to be an important and strong independent prognostic factor for survival. This underscores the potential relevance of NF-κB1 expression in severe sepsis, regardless even of mechanisms involved.

In conclusion, after lipopolysaccharide stimulation, the D allele of the NFKB1  insertion–deletion (−94ins/delATTG) polymorphism is associated with increased nuclear translocation of NF-κB1, boosted coagulation, and increased 30-day mortality in patients with severe sepsis. Accordingly, part of the variability in the outcome may be explained by the genetic variation in the promoter of NFKB1  and its molecular consequences mediated by altered nuclear translocation of NF-κB1.