Introduction

Stroke is the third cause of death and the first cause of permanent disability in developed countries. Up to 15% of ischemic events are associated with atherosclerotic carotid stenosis.1 Toll-like receptor-4 (TLR4), the primary receptor of the innate immune system, is crucial for endotoxin signaling and is a limiting factor for the innate response to bacterial liposaccharide (LPS).2 Besides LPS various endogenous ligands for TLR4 have been identified and some have been proposed to function in the process of atherogenesis.3, 4, 5 TLR4 is expressed in circulating monocytes,6 and in a number of different cell types present in the plaques of atherosclerotic patients.7 There is evidence that overexpression of this gene is involved in initiation and progression of atherosclerosis.8 Activation of TLR4 initiates an intracellular signaling cascade, leading to the activation of nuclear factor (NF)-κB transcriptional factor, which promotes the induction of a proinflammatory response, including the release of antimicrobial peptides, inflammatory cytokines and chemokines.9 Genomic variations that alter TLR4 gene expression may represent genetic risk factors involved in the predisposition to atherosclerosis and ischemic events.10 The aim of this study is to investigate if polymorphisms in the TLR4 gene promoter region may influence gene expression in peripheral blood of atherosclerotic patients.

Results and Discussion

Consecutive patients undergoing carotid endarterectomy at Verona University Hospital were enrolled, with either a recently symptomatic or an asymptomatic internal carotid artery (ICA) stenosis, as described elswhere.11 Controls were individuals presenting no sign of atherosclerotic plaques at ecodoppler examination. DNA and total RNA were extracted from peripheral blood samples. ICA stenosis degree was stratified into 50–69%, 70–80% and >80%. Clinical severity of stroke was rated through the National Institute of Health Stroke Scale (NIHSS).12 Functional polymorphisms −1607T>C, −2026A>G and −2604G>A13 were genotyped by PCR-restriction fragment length polymorphism (Supplementary Table 2), and gene expression was performed by Real-time PCR as described11 (Supplementary Table 3). By in silico analysis, other single-nucleotide polymorphisms present in the region do not alter transcription-binding sites or are in strong linkage disequilibrium with the selected ones.13 Gene expression differences between genotypes of polymorphisms were assessed by analysis of variance test. Dominant and recessive models were tested by the Student’s t-test. Single analysis was adjusted for multiple tests (P<0.02). Fisher’s exact test showed no significant difference between symptomatic and asymptomatic patients for clinical variables (except for sex; Supplementary Table 1). Allele frequencies and genotype distribution are shown in Supplementary Table 4.

TLR4 expression was significantly higher in symptomatic patients compared with asymptomatic and controls. No difference between asymptomatic and controls was observed (Supplementary Figure 1).

TLR4 gene expression showed no correlation with age or NHISS (data not shown).

A trend of association of TLR4 gene expression with stenosis degree was observed in symptomatic patients (P=0.05, Figure 1). The most severe grade of stenosis showed the highest TLR4 gene expression, suggesting that TLR4 expression could represent a marker of disease severity in this group of patients.

Figure 1
figure 1

TLR4 gene expression in relation to stenosis degrees in symptomatic (a) and in asymptomatic patients (b). (1) ICA stenosis degree of 50–69%, (2) 70–80% and (3) >80%. Bars indicate mean values for each group. P-value is indicated in the graph.

Full size image

Gene expression showed no significant differences between the different genotypes of variants 1607T>C and −2026A>G (data not shown). Polymorphism −2604G>A showed a lower level of TLR4 expression of the genotype −2604AA (0.86±0.58 (AA) vs 1.58±1.9 (GA+GG), P<0.02; Figure 2). This difference was also observed among groups when patients were divided into symptomatic, asymptomatic and control group. Consistently, linear regression analysis demonstrated that this effect was independent from disease status, indicating that both disease and genotype may influence TLR4 expression.

Figure 2
figure 2

(a) TLR4 gene expression in relation to −2604G>A genotypes AA, AG and GG, and (b) in relation to individuals carrying AA versus the counterpart GA/GG. Bars indicate mean values+s.d. for each group. P-value is indicated in the graph.

Full size image

The lower gene expression observed in individuals carrying the −2604AA genotype is consistent with previous studies,14 in which vector constructs containing the −2604A variant, which creates a putative GATA-2 transcription-binding site, showed lower promoter activity when transfected into A498 renal carcinoma cells. To determine whether this polymorphism alters the transcription-binding site for GATA-2, electrophoretic mobility shift assay (EMSA) was performed as described.14 In brief, we used radiolabeled oligonucleotide probes containing either −2604G or −2604A probes (Figure 3). On incubation of the probe −2604A with cell nuclear extracts, two distinct prevalent complexes (indicated as C1 and C2, in Figure 3) were revealed. The specificity of the binding reaction was confirmed by full competition of the non-labeled GATA oligonucleotides compared to the almost absence competition of non-labeled NF-κB consensus binding site oligonucleotides. In the presence of excess of the non-labeled −2604A fragment, the complexes C1 and C2 were competed away too, demonstrating the specificity for the probe. Compared to A allele, the 2604G probe showed a different electrophoretic mobility of the C1 complex (Figure 3), indicating that the −2604G>A polymorphism can affect the binding affinity of the TLR4 promoter to transcription factors. The pattern of complexes resulted similar to that obtained by a consensus GATA sequence, suggesting that members of the GATA family participate in the complex. The supershift analyses in presence of the antibody directed to GATA-2 transcription factor (Santa Cruz Biotechnology, Santa Cruz, CA, USA) failed to detect a mobility shift of the complexes with all the used probes (Figure 3), indicating that the sequence might be recognized by GATA family transcription factors other than GATA-2.

Figure 3
figure 3

Electrophoretic mobility shift assay (EMSA). 32P-end-labeled oligonucelotides encompassing −2604G>A polymorphism or GATA consensus were incubated with nuclear extracts in the presence or absence of a 200-fold molar excess of unlabeled competitor DNA. Supershift was analyzed in the presence of antibody against GATA-2 (right panel). C1 and C2 and asterisks indicate protein–DNA complexes.

Full size image

Considering that −2604A allele shows a lower level of TLR4 gene expression, and that EMSA results showed a different mobility pattern of −2604G>A variants, we can infer that nuclear factor/s binding to the region covering −2604G>A polymorphism may function as modulator(s) for TLR4 transcription, probably due to the contribution of an undetermined member of the GATA zinc-finger transcription factors. Indeed, GATA transcription factors have been demonstrated to have a crucial role in the regulation of epithelial innate immune responses.15, 16, 17

In conclusion, TLR4 is upregulated in stroke patients, indicating that TLR4 may be involved in the inflammation process leading to the ischemic event.

The presence of allele −2604A is associated with lower levels of TLR4 gene expression. This difference could be due to different transcription factors binding capacity of variants of the −2604G>A polymorphism.

Further case/control studies in larger populations should verify if TLR4 variant −2604A>G represents a genetic marker for susceptibility to atherosclerotic plaque development and progression.