The impact of interleukin-6 promoter −597/−572/−174genotype on interleukin-6 production after lipopolysaccharide stimulation

Authors


Michael Müller-Steinhardt MD, Institute of Transfusion Medicine and Immunology, Faculty of Medicine Mannheim, University of Heidelberg, Friedrich-Ebert-Straße 107, 68167 Mannheim, Germany.
E-mail: m.mueller-steinhardt@blutspende.de

Summary

Interleukin (IL)-6 is a pleiotropic cytokine, produced by different cells. There is accumulating evidence that IL-6 promoter polymorphisms impact substantially on various diseases and we identified kidney transplant recipients carrying the IL-6 GGG/GGG −597/−572/−174genotype to have superior graft survival. To prove a functional impact on gene expression, we analysed systematically IL-6 production in healthy individuals with respect to the IL-6 −597/−572/−174genotype. IL-6 was determined in 100 healthy blood donors at protein and mRNA levels upon specific stimulation in monocytes and T lymphocytes under whole blood conditions. GGG/GGG individuals showed a lower IL-6 secretion upon lipopolysaccharide (LPS)-stimulation versus all others (P = 0·039). This link was even stronger when −597 and −174GG genotypes were reanalysed separately (P = 0·008, P = 0·017). However, we found neither a difference at the mRNA level or percentage of CD14+ cells nor after T cell stimulation. We found evidence for the IL-6 −597/−572/−174genotype to affect IL-6 synthesis, i.e. lower levels of IL-6 protein upon LPS-stimulation in GGG/GGG individuals. Further studies are needed in kidney transplant recipients to investigate the potential link between the GGG/GGG genotype and graft survival. In line with this, determination of the genetic risk profiles might be promising to improve the transplant outcome in the individual patient.

Introduction

Interleukin (IL)-6 is a pleiotropic cytokine with a broad range of effects [1]. It regulates the release of other cytokines such as IL-1β and tumour necrosis factor (TNF)-α, the production of adhesion molecules and induces fever and hepatic acute phase proteins. IL-6 is produced by a variety of different cells such as monocytes and T lymphocytes, playing a crucial role at the interface of adoptive and innate immunity. The increasing knowledge about individual genetic susceptibility of the immune system triggered the identification of three single nucleotide polymorphisms (SNP) within the promoter region of the IL-6 gene at positions −597(G→A), −572(G→A) and −174(G→C)[2–3].

There is accumulating evidence that IL-6 promoter polymorphisms have substantial clinical impact on the prevalence, incidence and the prognosis of different major diseases, such as Alzheimer's disease [4,5], atherosclerosis [6,7], cardiovascular disease [8,9], cancer [10,11], non-insulin-dependent diabetes mellitus (NIDDM) [12,13], osteoporosis [14,15], sepsis [16] and systemic-onset juvenile chronic arthritis [2], and as a risk factor for preterm birth [17]. In clinical transplantation the IL-6 −174genotype of the kidney donor was found to be a major risk factor for the occurrence of acute rejection episodes [18], while our own group identified kidney allograft recipients of the GG −174genotype to have a better long-term graft outcome [19].

Recent disease association studies focused mainly on the −174(G→C) polymorphism. However, a recent investigation revealed evidence that the three SNPs are in linkage disequilibrium and described naturally occurring IL-6 −597/−572/−174haplotypes and their frequencies among Caucasians [3]. Subsequently, association studies considering a possible cooperative impact of all three SNPs have been published [20,21]. Reanalysis of our own data revealed an even stronger advantage for carriers of the GGG/GGG −597/−572/−174genotype with respect to long-term allograft survival than isolated analysis of the −174genotype [22].

Because the promoter region of a gene is crucial for transcriptional regulation, a functional impact of the IL-6 genotype on gene expression and thus a potential pathogenetic relevance was postulated [2,3]. Until now, mainly cell line experiments [2] or the determination of IL-6 plasma levels that focused on the −174genotype have been published. While two studies investigating patients with abdominal aneurysm (n = 466) [23] and carotid atherosclerosis (n = 1000) [24] suggested the CC−174 genotype to be associated with higher IL-6 levels, two other investigations on healthy subjects (n = 32) [25] (n = 25) [21] found the opposite, whereas another study on cardiac patients (n = 942) [26] failed to establish any association at all. Thus we hypothesized that a possible functional relevance of the IL-6 genotype may be revealed only after stimulation of a IL-6 response in specific cells.

It was the aim of the present study to analyse systematically IL-6 production in healthy individuals upon specific stimulation in both monocytes and T lymphocytes and clarify as to whether GGG/GGG individuals differ from other −597/−572/−174genotypes. For this purpose we quantitatively analysed the IL-6 response under whole blood conditions at the mRNA and protein level.

Materials and methods

Determination of IL-6 −597/−572/−174haplotypes

A polymerase chain reaction system with sequence-specific primers (PCR-SSP) was used for direct haplotyping of the three SNPs. Details of the method, sequences of eight different allele-specific primers (AS1-AS8) and two control primers (control 63 + control 64) have been described elsewhere [3]. Using a 12-reaction PCR-SSP system, mismatches at the 3′ end of forward and reverse primers of each PCR enabled us to establish unequivocally both haplotypes for the three bi-allelic sites in each individual. PCR amplifications were carried out in 96-well plates using a Perkin-Elmer 9600 Thermocycler System (Perkin-Elmer, Norwalk, USA). PCR conditions were: 95°C for 60 s; five cycles of 95°C for 35 s, 70°C for 45 s and 72°C for 35 s; 21 cycles of 95°C for 25 s, 65°C for 50 s and 72°C for 40 s; followed by four cycles of 95°C for 35 s, 55°C for 60 s and 72°C for 90 s. Primer pairs were combined as follows: AS1 + AS8 (−597G/−174C), AS1 + AS7 (−597G/−174G), AS2 + AS8 (−597A/−174C), AS2 + AS7 (−597A/−174G), AS3 + AS8 (−572G/−174C), AS3 + AS7 (−572G/−174G), AS4 + AS8 (−572C/−174C), AS4 + AS7 (−572C/−174G), AS1 + AS6 (−597G/−572C), AS1 + AS5 (−597G/−572G), AS2 + AS6 (−597A/−572C), AS2 + AS5 (−597A/−572G). In order to avoid false negative results, control primers were used in each reaction.

Whole blood assay

Whole blood samples were obtained from 100 healthy regular blood donors at the Institute of Immunology and Transfusion Medicine, University of Lübeck School of Medicine. Participants gave informed consent to participate in the study, which was approved by the local Ethics Committee (file no. 04-051). The donors (74 male and 26 female, mean age: 39·2 ± 9·5 years) showed no signs of infection or other malignancies. Only non-smokers and non-obese individuals (body mass index < 25) were included in the study. Additional exclusion criteria were regular or recent intake of any medications or clinically relevant abnormal findings in the medical history or laboratory values. Blood samples were collected in lithium–heparin tubes (Sarstedt, Nürnbrecht, Germany) and processed within 2 h. Human whole blood assays were performed in six-well-cluster tissue culture dishes (diameter 35 mm, Costar, Cambridge, MA, USA). For each sample, 1 ml aliquots in duplicate were diluted in 9 ml Iscove's modified Dulbecco's medium (IMDM) supplemented with penicillin (100 U/ml), streptomycin (100 µg/ml) and 10 mM l-glutamine. Consecutively, whole blood assays were stimulated with 1 µg/ml anti-CD3 monoclonal antibody (MoAb) (CLBT3/4E) and 1 µg/ml anti-CD28 MoAb (CLB-CD28/1; Hiss Diagnostics, Amsterdam, the Netherlands). In parallel assays, whole blood cultures were stimulated with lipopolysaccharide (LPS) (30 ng/ml; Sigma, Deisenhofen, Germany). Unstimulated whole blood assays served as controls. At the end of the incubation periods, whole blood cultures were either processed further for intracellular IL-6 staining (4 h LPS-stimulation, addition of monensin prior to incubation) or 2 ml of supernatants were withdrawn for enzyme immunoassay (EIA) analysis (4 h LPS-simulation, 24 h anti-CD3/anti-CD28-stimulation). Residual culture material was processed for RNA isolation and quantitative PCR (4 h LPS-stimulation, 24 h anti-CD3/anti-CD28-stimulation). In addition basal IL-6 mRNA expression was analysed from peripheral blood leucocytes.

RNA isolation

Total RNA from whole blood leucocytes was isolated using the Purescript RNA isolation kit (Gentra Systems, Minneapolis, USA) according to the manufacturer's protocol. The resulting RNA was resuspended in 300 µl diethylpyrocarbonate (DEPC) water and stored at −80°C until use.

Taqman reverse transcriptase–PCR

The RT–PCR protocol for the quantitative detection of cytokine mRNA has been described previously [27,28]. In brief, the PCR reaction mixture contained 25 µl 2× TaqMan Mastermix (Perkin Elmer Foster City, CA, USA), 100 nM of forward and reverse primer, 100 nM fluorogenic probe, 20 U RNAse-inhibitor (Gibco brl, Eggenstein, Germany), 25 U murine leukemia virus (MuLV)–RT (Perkin Elmer, Foster City, CA, USA), 1·25 U Ampli-Taq gold-DNA-polymerase (Perkin Elmer) and 20 µl of water control, diluted standards or unknown RNA-template in a total volume of 50 µl. Sequence-specific primer pairs and fluorogenic probes were obtained from TIB Molbiol (Berlin, Germany) (IL-6) or Perkin Elmer Cetus (Foster City, USA) (β-actin kit for cDNA samples). PCR conditions were 2 min at 50°C and 30 min at 48°C for RT, 10 min at 95°C for DNA–polymerase activation, followed by 40 cycles of 15 s at 95°C and 1 min at 30 s at 60°C with a final 25°C hold. Standardized IL-6 mRNA quantities (IL-6 copies/106β-actin copies) were determined by dividing the interpolation-derived values from the cytokine standard curve by the normalization factor (β-actin content).

EIA-analysis of IL-6 protein secretion

Supernatants were assayed after incubation for determination of IL-6 protein concentrations by a sandwich EIA technique using unlabelled and enzyme-coupled MoAbs against different IL-6 epitopes according to the manufacturer's instructions (Laboserv, Giessen, Germany).

Flow-cytometric analysis of intracellular IL-6 production in CD14+ cells

During the 4 h incubation period with LPS, 3 μM monensin (Sigma) was added to the whole blood cultures followed by a fixation with 4% paraformaldehyde (Riedel de Haen, Seelze, Germany), as described previously [23]. An unstimulated control was run in parallel. For intracellular staining, cells were subsequently washed in Hanks's balanced salt solution (HBSS) and resuspended in a buffer consisting of HBSS, 0·1% saponin (Riedel de Haen) and 0·01 M HEPES buffer (Seromed Biochrome, Berlin, Germany). Aliquots (200 µl) of cells were added to tubes containing 0·5 μg/10 μl of MaBs (BD Pharmingen, Heidelberg, Germany) against CD14 [M5E2, phycoerythrin (PE)-conjugated] and IL-6 [MQ2-13A5, fluorescein isothiocyanate (FITC)-conjugated]. Preincubation with a surplus of unconjugated anti-cytokine MoAbs (5 μg/10 μl; Pharmingen) served as a negative control for intracellular staining to each sample. Isotype-specific antibodies were used to detect irrelevant specificity for surface molecule staining. All flow cytometric analyses were performed on an EPICS XL MCL (Coulter-Immunotech, Krefeld, Germany) equipped with a single air-cooled argon laser with an excitation line at 488 nm. Green fluorescence (FITC) was detected through a 525-nm band pass filter, orange emission (PE) through a 575-nm band pass filter and deep red fluorescence from 7-AAD was detected through a 675-nm band pass filter. All measurements were performed for at least 10 000 cells at low sample pressure.

Statistical analysis

Statistical analyses were performed with commercially available software for personal computers (spss 10·0 for Windows; SPSS Inc., Chicago, IL, USA). IL-6 −597/−572/−174haplotype and −597/−572/−174genotype frequencies between healthy blood donors and kidney transplant recipients [22] were compared using Fisher's exact test. Due to the limited number of cases, data were considered not to be normally distributed. Therefore statistical differences between groups were tested by the Mann–Whitney U-test. P-values < 0·05 were considered significant. Linkage disequilibrium between the −597 and −174 polymorphism was calculated according to Morton's method [29].

Results

Distribution of −597/−572/−174haplotypes and −597/−572/−174genotypes

Among the investigated blood donors we found four different −597/−572/−174haplotypes (Table 1). The relative frequency of these haplotypes was comparable to a previously published cohort of kidney allograft recipients [22] (P-values for the correlation of −597/−572/−174haplotype frequencies were: AGC = 0·113, GGG = 0·194, GCG = 0·419, GGC = 0·332, Fisher's exact test). The −597/−572/−174genotype distribution is given in Table 2. Relative genotype frequencies were comparable to kidney transplant patients (P-values for the correlation of −597/−572/−174genotype frequencies were: AGC/AGC = 0·226, AGC/GGG = 0·298, GGG/GGG = 0·203, AGC/GCG = 0·428, GGG/GCG = 0·556, AGC/GGC = 0·624 and GGG/GGC = 0·386, Fisher's exact test). In contrast to the transplant patients, we identified one carrier of the GGG/GGC −597/−572/−174genotype in our blood donor cohort. Comparison with the −174-allele distribution revealed a linkagedisequilibrium between the −174G and −597G-allele and between the −174C and −597A-allele (r = 0·977, δ = 1·0) except for two individuals (−174CC/−597AG and −174GC/−597GG).

Table 1.  Interleukin (IL)-6 −597/−572/−174haplotype frequencies in kidney transplant recipients [22] and in healthy blood donors.
HaplotypeKidney transplant
recipients, n = 158 [22]
Healthy
Caucasian blood
donors, n = 100
No.(%)No.(%)
AGC13542·77437·0
GGG16552·211356·5
GCG154·7115·5
GGC10·321
Total316100200100
Table 2.  Interleukin (IL)-6 −597/−572/−174genotype distribution in kidney transplant recipients [22] and in healthy Caucasian blood donors.
GenotypeKidney transplant
recipients, n = 158 [22]
Healthy Caucasian blood donors, n = 100
No.(%)No.(%)
AGC/AGC3119·61515
AGC/GGG6641·83838
GGG/GGG4528·53434
AGC/GCG63·855
GGG/GCG95·766
AGC/GGC10·611
GGG/GGC0011
Total158100100100

IL-6 mRNA expression

With regard to assessment of basal IL-6 mRNA expression in peripheral blood leucocytes, we failed to prove any significant differences between the different −597/−572/−174haplo- or genotypes (GGG: 410 ± 423 IL-6 copies/106β-actin copies versus all other haplotypes: 388 ± 427, P = 0·696; GGG/GGG: 409 ± 387 versus all other genotypes: 396 ± 445, P = 0·596).

After 4 h LPS stimulation of whole blood cultures, we observed a marked IL-6 mRNA expression even though a difference between various haplo- or genotypes was not found (GGG: 1·9 ± 2·0 × 106 IL-6 copies/106β-actin copies versus all other haplotypes: 2·4 ± 3·1 × 106, P = 0·553; GGG/GGG: 2·1 ± 2·2 × 106versus all other genotypes: 2·1 ± 2·7 × 106, P = 0·930). Similarly, after 24 h of anti-CD3/anti-CD28 stimulation the IL-6 mRNA expression did not differ between the groups (GGG: 3·6 ± 5·2 × 104 IL-6 copies/106β-actin copies versus all other haplotypes: 4·8 ± 7·8 × 104, P = 0·781; GGG/GGG: 3·3 ± 4·6 × 104versus all other genotypes: 4·5 ± 7·3 × 104, P = 0·942). Notably, we observed a relatively high interindividual variability for all mRNA values.

IL-6 protein secretion

While individuals of the GGG −597/−572/−174haplotype showed no inferior IL-6 secretion upon LPS-stimulation (P = 0·058), homozygous carriers of the GGG/GGG −597/−572/−174genotype had significantly lower IL-6 concentrations versus all other genotypes (P = 0·039) (Fig. 1). Interestingly, individuals of the GGG/GCG −597/−572/−174genotype (homozygous −597GG and −174GG, but heterozygous −572GC) had comparably low IL-6-values (Fig. 1). This prompted us to investigate additionally the IL-6 secretion for the −597 and −174genotypes, respectively. An even stronger link to low IL-6 secretion was detected for both the −597 and the −174GG-genotype (−597GG: 835 ± 716 versus 1148 ± 683 pg/ml, P = 0·008; −174GG: 848 ± 720 versus 1134 ± 686 pg/ml, P = 0·017) (Fig. 2).

Figure 1.

Interleukin (IL)-6 secretion after 4 h of lipopolysaccharide (LPS) stimulation (mean ± s.d.) (n); left side: GGG/GGG −597/−572/−174genotype versus all others; right side: all other −597/−572/−174genotypes in detail.

Figure 2.

Interleukin (IL)-6 secretion after 4 h of lipopolysaccharide (LPS) stimulation (mean ± s.d.) (n); left side: GG versus GA/AA −597genotype; right side: GG versus GC/CC −174genotype.

IL-6 secretion after anti-CD3/anti-CD28 stimulation showed no significant differences between the various haplo- and genotypes (GGG: 582 ± 247 pg/ml versus all other haplotypes: 567 ± 274, P = 0·650; GGG/GGG: 585 ± 241 versus all other genotypes: 571 ± 270, P = 0·721).

Intracellular IL-6 production in CD14+ cells

The percentage of IL-6-producing CD14+ cells after LPS stimulation did not differ with respect to the IL-6 597/−572/−174haplotype or −597/−572/−174genotype (GGG: 55·7 ± 17·3% versus all other haplotypes: 56·5 ± 18·3, P= 0·712; GGG/GGG: 55·0 ± 16·7 versus all other genotypes: 56·6 ± 18·3, P = 0·164).

Discussion

In a previous study we found evidence for an exceptional clinical impact of the IL-6 −597/−572/−174genotype on the success of kidney transplantation [22]. In a multi-factorial Cox regression analysis, transplant recipients of the GGG/GGG −597/−572/−174genotype were identified as having a superior allograft survival versus all other genotypes who had an 8·0-fold increased risk of premature graft loss. As recently published cell line experiments indicated IL-6 promoter SNPs to affect IL-6 synthesis [2,3], we hypothesized the IL-6 promoter genotype to be of potential pathogenetic relevance and GGG/GGG carriers to differ from other individuals with respect to IL-6 production.

The current study indicates an association of the IL-6 −597/−572/−174genotype with IL-6 secretion upon LPS stimulation under whole blood conditions for the first time. As the transferability of cell line or cell culture experiments to in vivo conditions is questionable [30], we performed a whole blood assay [27] that obviates the need for cell purification and therefore closely approximates the state of responsiveness of circulating monocytes and T cells. We were able to show that healthy carriers of the GGG/GGG −597/−572/−174genotype show a reduced IL-6 secretion upon LPS stimulation. Interestingly, individuals of the GGG/GCG −597/−572/−174genotype being homozygous −597GG and −174GG, but heterozygous −572GC, also showed relatively low IL-6 secretion values. Thus, an additional, separate analysis of the −597 and −174genotypes, respectively, revealed an even stronger link for both the −597GG and the −174GG-genotypes to low IL-6 secretion. Consistent with recent investigations [21,22], we identified a close linkage disequilibrium between the −174G and −597G-allele and between the −174C and −597A-allele, respectively. Consequently, both SNPs have a similar strong association with IL-6 secretion, when investigated separately. While the −174GG and the −597GG genotypes are associated with lower IL-6 levels, the −174C- and −597A-alleles carried either hetero- or homozygously are associated with increased IL-6-values.

To explain the potential pathophysiological link between the two SNPs and IL-6 synthesis, we suggest further functional studies such as surface plasmon or electrophoretic mobility shift analyses. Different transcription factors, transactivators or even inhibitor elements may exist, each requiring two distinct docking sites at positions −597 and −174. While one factor may require the −597G- and −174G-allele for binding and results in a lower IL-6 transcription rate, the other may require the −597A- and −174C-allele resulting in an increased transcriptional activity [2]. The position of −597 upstream from the glucocorticoid responsive elements (GRE2 and GRE1) and of −174 upstream from the second messenger responsive enhancer region (MRE1 and MRE2) is intriguing. Both GRE and MRE are involved in binding of glucocorticoid receptors [2,31], suggesting that IL-6 promoter polymorphisms may interact with glucocorticosteroids. Interestingly, there is no linkage disequilibrium between the −572 and either the −597 and −174 polymorphism. We therefore suggest that transcription factor binding at position −572 must occur independently from binding at positions −597 and −174. However, at this stage a functional relevance of the IL-6 promoter polymorphisms cannot be proven. They might also be in tight linkage disequilibrium with other flanking, functionally relevant polymorphisms.

However, we failed to confirm this result at the mRNA level and found no evidence for lower IL-6 mRNA values in GGG/GGG individuals. We propose the following explanations: as we observed high interindividual variability and standard deviations for all mRNA values, our real-time PCR protocol may not be appropriate to detect subtle differences caused by genotype variations. Perhaps the investigation of larger cohorts will be necessary to prove an impact at the mRNA level. Moreover, it cannot be excluded that the determination of IL-6 mRNA levels at other time-points could possibly confirm our results at the protein level. However, this would require extensive kinetic studies. Alternatively, further functional studies such as reporter gene assays maybe appropriate to identify an impact of the −597/−572/−174genotype on the IL-6 mRNA expression, in order to verify a direct regulation of the transcription by IL-6 promoter polymorphisms. Interestingly, we also failed to show an impact of the IL-6 genotype on the percentage of IL-6 producing CD14+ cells. Thus the genotype may affect the amount of IL-6 per CD14+ cell rather than the number of IL-6-producing cells.

Furthermore, we found no impact of the −597/−572/−174genotype on IL-6 synthesis in T lymphocytes upon anti-CD3/anti-CD28-stimulation. Neither IL-6 mRNA expression nor protein secretion differed between the various genotypes. Even though allograft rejection is an adoptive immune response mediated primarily by T cells, our results suggest, rather, a possible indirect impact of the IL-6 genotype on T cell functions. An altered IL-6 secretion of monocytes may also affect their antigen-presenting cell (APC) function and co-stimulatory capacity on T cells, respectively, thereby affecting the allogenic T cell response.

The impact of the −597/−572/−174genotype on the synthesis of IL-6 in randomly selected healthy individuals under whole blood conditions has not been subjected previously to large trials. Previous investigations concentrated on the isolated impact of the −174genotype in cell line experiments [2,3] on selected individuals or the determination of IL-6 plasma concentrations, respectively [21,23–26]. As these studies revealed conflicting results, we hypothesized that a possible functional relevance of the IL-6 genotype may be discovered only after stimulation of a IL-6 response considering the −597/−572/−174genotype as a whole.

However, the following potential limits of our study should be addressed: first, the use of monoclonal antibodies and LPS for cytokine induction are supraphysiological stimuli. Thus transferability to the in vivo situation remains incomplete, even though we performed a whole blood assay. Moreover, we only considered smoking habits and obesity as parameters that potentially impact on IL-6 production. Whether additional parameters might be of functionalrelevance is currently not known. Nevertheless, our data may be used as reference of a healthy population, and prospective studies of different patient cohorts will be necessary to test the influence of the IL-6 −597/−572/−174genotype on IL-6 synthesis. This may provide not only new insights into the role of IL-6 in the prevalence, incidence and prognosis of different major diseases, it may also help to provide more individualized therapeutic strategies for patients with an increased genetic susceptibility.

In conclusion, our results indicate an association of the IL-6 −597/−572/−174genotype on IL-6 synthesis in healthy individuals. Carriers of the GGG/GGG genotype show a decreased IL-6 response upon LPS stimulation under whole blood conditions. Confirmatory studies and microarray analyses will be required to clarify the involvement of additional potentially relevant genes. Furthermore, the impact of IL-6 promoter polymorphisms on IL-6 synthesis should be investigated in kidney transplant patients before and after transplantation. In a recent investigation we found evidence for an individually different impact of immunosuppressive drugs on IL-6 synthesis [32]. Thus IL-6 promoter polymorphisms are highly interesting candidates to explain these differences.

The determination of risk profiles based on clinical, laboratory and genetic factors (e.g. IL-6 −597/−572/−174genotype) might be a promising strategy to prevent early and chronic kidney allograft loss and optimize immunosuppressive treatment in the individual patient.

Acknowledgements

The authors are grateful to Ms Daniela Griffiths for carefully editing the manuscript.

Ancillary