|Home | About | Journals | Submit | Contact Us | Français|
Interleukin-6 (IL-6) is a key inflammatory cytokine, signalling to most tissues by binding to a soluble IL-6 receptor (sIL-6r), making a complex with gp130. We used 1273 subjects (mean age 68 years) from the InCHIANTI Italian cohort to study common variation in the IL-6r locus and associations with interleukin 6 receptor (IL-6r), IL-6, gp130 and a battery of inflammatory markers. The rs4537545 single nucleotide polymorphism (SNP) tags the functional non-synonymous Asp358Ala variant (rs8192284) in IL-6r (r2 = 0.89, n = 343). Individuals homozygous for the rs4537545 SNP minor allele (frequency 40%) had a doubling of IL-6r levels (132.48 pg/ml, 95% CI 125.13–140.27) compared to the common allele homozygous group (68.31 pg/ml, 95% CI 65.35–71.41): in per allele regression models, the rs4537545 SNP accounted for 20% of the variance in sIL-6r, with P = 5.1 × 10−62. The minor allele of rs4537545 was also associated with higher circulating IL-6 levels (P = 1.9 × 10−4). There was no association of this variant with serum levels of gp130 or with any of the studied pro- and anti-inflammatory markers. A common variant of the IL-6r gene results in major changes in IL-6r and IL-6 serum levels, but with no apparent effect on gp130 levels or on inflammatory status in the general population.
Interleukin-6 (IL-6) is a major cytokine1 with complex clinical effects, including governing the resolution of acute innate immunity and steering transition to an acquired immune response and chronic inflammation.2,3 Membrane-bound IL-6 receptors (IL-6r) (mIL-6r, also called gp80) are expressed mainly on hepatocytes, neutrophils, monocytes/macrophages and some lymphocytes, mediating direct signalling. However, IL-6 is also active in many other tissues through an unusual ‘trans-signalling’ mechanism, which depends on soluble IL-6 receptor (sIL-6r). The IL-6 and sIL-6r complex combines with the ubiquitous membrane protein gp130, initiating intracellular signalling in cells lacking membrane-bound IL-6r.
Although some sIL-6r is clearly needed for trans-signalling, the significance of common clinically observed variation in serum levels of sIL-6r is less clear. The sIL-6r is present in the serum or plasma of healthy individuals and levels are raised in rheumatoid arthritis, multiple myeloma and T-cell abnormalities such as AIDS and adult T-cell leukemia.4 Also, increased serum levels of IL-6r have been reported in common acute conditions including acute myocardial infarction5 and asthma.6 An anti-human-IL-6 receptor monoclonal antibody has been developed (tocilizumab), which blocks the actions of IL-6 by binding to sIL-6r.7 Early clinical trials suggest efficacy in Castleman's disease,8 adult rheumatoid arthritis8 and systemic onset juvenile arthritis.9
The IL-6 receptor gene maps to human chromosome 1q21. Galacia et al.10 reported that the minor allele of a non-synonymous single nucleotide polymorphism (SNP), Asp358Ala, was associated with increased IL-6r serum concentrations in 70 Japanese subjects (P = 0.0001). Some studies have found that this Asp358Ala variant is associated with obesity,11,12 the metabolic syndrome13 and diabetes.14
As IL-6r is clearly of major clinical significance, identifying genetic factors influencing IL-6r levels is needed, for clinical interpretation of levels and perhaps for pharmacogenetic use. We aimed to comprehensively study common variation in the IL-6r locus and associations with IL-6, IL-6r and gp130 serum levels, other inflammatory markers, metabolic syndrome traits and common circulatory and other diseases. For this, we used six SNPs that capture a large proportion of common variation in the IL-6r gene and related these to phenotype data in the InCHIANTI study.15
The InCHIANTI study uses a representative population-based sample of Caucasian subjects from two small towns near Florence, Italy, to study risk factors for the onset of disability in older persons. The study includes 298 individuals of <65 age and 1155 individuals of age ≥65 years. The study design and protocol have been described in detail previously.15 The data collection started in September 1998 and was completed in March 2000. The INRCA Ethical Committee approved the study.
The details of the measurement of inflammatory markers are set out by Cesari et al.16 Venipuncture was performed in the morning after a 12-h fast. All cytokine assays were carried out at the INRCA central laboratory. IL-6, serum IL-6 receptor and tumor necrosis factor (TNF)-α were quantified with immunoassay kits (BioSource Cytoscreen human IL-6, human sIL-6r and human TNF-α Ultra-Sensitive kits; BioSource International Inc., Camarillo, CA, USA). The minimum detectable concentrations were 0.10 pg/ml for IL-6, 8 pg/ml for sIL-6r and 0.09 pg/ml for TNF-α. The interassay coefficient of variation was 7% for all three kits. Serum IL-1RA level was detected by enzyme-linked immunosorbent assay (ELISA) method using commercial kits (EASIATM ELISA Human IL-1RA, BioSource International Inc.); the minimum detectable concentration was 4.00 pg/ml and the mean interassay coefficient of variation was 8.2%. C-reactive protein (CRP) concentrations were measured with an immunoturbidimetric assay (Roche Diagnostics, GmbH, Mannheim, Germany). The lower detection limit was 0.3 mg/dl.
All assays were carried out in duplicate for all cytokine measures and the average of the two measures was used in the analyses.
Fasting blood glucose was determined by an enzymatic colorimetric assay using a modified glucose oxidase-peroxidase method (Roche Diagnostics) and a Roche-Hitachi 917 analyzer. Fasting insulin was determined using a commercial double-antibody, solid-phase radioimmunoassay (Sorin Biomedica, Milan, Italy) with an intra-assay coefficient of variation of 3.1%.17
Diseases were ascertained by an experienced clinician using pre-established criteria combining self-reported physician diagnoses, current pharmacologic treatment, medical records, clinical examinations and blood tests.18 The ‘diabetes’ group in this analysis had fasting blood glucose levels of ≥7.0 mmol/l or a physician diagnosis of diabetes with diabetes diet or drug treatment in the previous 2 weeks.
We used the modified Taqman assays at Kbiosciences to generate genotypes from SNPs for all samples including duplicates. Genotyping for rs4845371 and rs8192284 was performed in-house using conventional Taqman probes (Applied Biosystems, Foster City, CA, USA). Quality control included assessment of 6% blind duplicates.
Variables with modest non-normal distribution, including serum IL-6r and IL-6 concentrations, were log-transformed and all analyses were corrected for age and sex. Severely non-normal measures, including those with large numbers of subjects with undetectable low levels, were categorized into quantiles. Linear regression was used to test for association of the three genotypes at each SNP with serum IL-6r concentrations and other continuous traits. For single SNP analysis we used STATA version 9.1 and for the multiple marker haplotype analysis we used WHAP software (http://pngu.mgh.harvard.edu/purcell//whap/).
We successfully genotyped six tag SNPs in up to 1273 study participants, which, based on hapmap II data in European samples, captured 31 out of 38 SNPs (82%) with a minor allele frequency >10% at r2>0.8 and 37 out of 38 (97%) SNPs at an r2>0.5, across the IL-6r gene and including 10 kb of flanking sequence on either side. Call rates for these SNPs ranged from 88 to 95% and duplication errors ranged from 0 to 3.2%. All SNPs were in Hardy–Weinberg equilibrium (P>0.01). The positions of all the typed tag SNPs across the IL-6r gene along with an overview of linkage disequilibrium (LD, a measure of correlation of genotypes) in this region (including the previously studied Asp358Ala variant) are shown in Figure 1.
The mean age of the study respondents was 68 years, 44.6% were men (Table 1) and the mean sIL-6r concentration was 90.65 ng/ml. In age- and sex-adjusted regression analyses, we found that four SNPs were associated with serum sIL-6r concentrations at P-values <0.001 (Table 2). The strongest SNP IL-6r level association observed was for rs4537545; this SNP alone accounted for 19.9% of the variance in sIL-6r serum levels in the per allele regression model (P = 5.1 × 10−62). Individuals homozygous for the minor T allele of the rs4537545 marker SNP had a doubling in serum sIL-6r levels (132.48 pg/ml, 95% CI 125.13–140.27) compared to common C allele homozygous subjects (68.31 pg/ml, 95% CI 65.35–71.41).
To test the independence of association of rs4537545 with IL-6r, levels we performed a regression analysis of IL-6r levels with rs4537545 genotypes while adjusting for the three polymorphisms (rs4845371, rs952146 and rs2229238), age and sex. Here we found that rs4537545 was associated with IL-6r levels after adjusting for all these SNPs in two SNP regression analyses.
To explore if the effect was independent of diseases in which raised IL6r levels have been reported, we also tested associations of rs4537545 polymorphism with IL-6r levels in individuals who had no diagnosis of cardiovascular disease, hip arthritis, knee arthritis, diabetes or asthma (n = 708). Within this group we again found a highly significant association of rs4537545 with IL-6r levels (P = 3.8 × 10−33) such that individuals homozygous for the minor ‘T’ allele had IL-6r levels 63.56 ng/ml (95% CI 52.11–75.0 ng/ml) higher than individuals homozygous for the ‘C’ allele and rs4537545 explained 19% of variation in IL-6r levels. A similarly strong association was observed in 534 individuals diagnosed with any of the above disease states (P = 1.9 × 10−29) and informative for rs4537545 genotypes. Adjusting for CRP levels and erythrocyte sedimentation rate did not decrease the significance of the association (P = 1.1 × 10−61).
In addition, rs4537545 was associated with an increase in IL-6 concentrations (P = 1.9 × 10−4) although this effect was more modest, as illustrated in box plot format in Figure 2, comparing IL-6r and IL-6 levels by genotype.
The rs4537545 SNP minor allele (associated with higher IL-6r levels) was not associated with gp130 levels (Table 3). We next tested whether this SNP was associated with any other pro- and anti-inflammatory markers, to establish whether increased levels of IL-6 agonist and receptor resulted in a broader change in inflammatory profile. In age- and sex-adjusted regression analyses, there were strong associations between serum IL-6r levels and general markers of inflammation, including erythrocyte sedimentation rate and fibrinogen concentrations. However, neither white cell counts nor any of the studied pro- or anti-inflammatory markers were significantly associated with the marker SNP, indicating the absence of a general inflammatory effect.
We examined whether the rs4537545 SNP was associated with changes in metabolic syndrome-related measures, following previous reports (Table 4). We examined associations with the rs4537545 SNP and found nominally significant associations with fasting blood glucose levels, waist/hip ratio and adiponectin levels. However, effect sizes, especially for glucose and waist/hip ratios, were small. The rs4537545 SNP alone only accounted for 0.46 and 0.35% of the variation in these outcomes, respectively.
We then examined associations between the rs4537545 SNP and the available major common chronic conditions previously linked to changes in serum IL-6r levels (Table 5). Unfortunately, the InCHIANTI study has limited numbers of subjects with these conditions and therefore our analyses should be regarded as hypothesis generating. There were 138 subjects with diabetes, and rs4537545 rare allele carriers were significantly less likely to have diabetes (per minor allele OR 0.73; 95% CI: 0.55–0.95, P = 0.019), replicating a previous report.14 This association was stronger (P = 0.009) for a combined group of people with diabetes or impaired fasting glucose on ADA criteria.19 Similarly, for cardiovascular disease (including pervious myocardial infarction and angina, P = 0.059) and asthma (P = 0.070), we found possible reductions in prevalence in the minor allele groups. No associations were found with hip or knee arthritis.
As noted above, the minor allele of the non-synonymous Asp358Ala (rs8192284) polymorphism has previously been reported as altering IL6r levels.10 In Hapmap II data, the rs4537545 SNP is in strong LD (r2 = 0.96) with Asp358Ala. To check this in the InCHIANTI samples, we genotyped the rs8192284 polymorphism in 343 InCHIANTI participants and found these polymorphisms again to be in strong LD (r2 = 0.89). A highly significant association of the rs8192284 SNP with IL-6r levels was observed in the small number of typed individuals (P = 1.750 × 10−11).
In this study, we aimed to analyze common variation in the IL-6r locus in a large population sample. We related genetic variation directly to IL-6, IL-6r and gp130 serum levels, other inflammatory markers, metabolic syndrome traits and relevant common conditions. The minor allele of the rs4537545 SNP emerged as associated with a doubling of serum IL-6r levels (P = 5.1 × 10−62), and also raising IL-6 levels (P = 0.00019), although more modestly. The effect size seen is large for a single polymorphism, accounting for 20% of the variance in IL-6r on its own.
Our findings of higher serum IL-6r levels associated with the minor variant (essentially marking the Ala variant of Asp358Ala) are consistent with the findings of the study of 70 Japanese subjects by Galicia et al.10 and a very recent study of 1184 African Americans and 1789 European Americans by Reich et al.20 However, we have further shown that although the studied variant is associated with a dramatically lower serum concentration of IL-6r, there is little change in levels of the gp130 element of the IL-6r complex or in inflammatory status.
As soluble IL-6r acts with IL-6 in many tissues lacking membrane-bound IL-6 receptors, the raised levels of both IL-6r and IL-6 could be expected to result in broader inflammatory changes. However, we found no evidence for other inflammatory effects in our extensive battery of inflammatory markers. This is despite the large amount of evidence for the importance of IL-6 signalling in common major conditions of middle-aged and older people1 and even in the development of age-related frailty in the InCHIANTI study.21 The absence of broader inflammatory changes may suggest redundancy of higher levels of circulating sIL-6r, but the precise effects of the Asp358Ala polymorphism in the complex IL-6 signalling pathway clearly needs detailed investigation. Further work is also needed to establish the effects of this polymorphism in the less common and severe inflammatory diseases in which IL-6 and IL-6r levels are greatly increased: in these disorders, gross imbalances of the IL-6 system can occur, and the studied polymorphism may have different effects.
In terms of metabolic syndrome features, Wolford et al., studying 700 Pima Indians, reported that those carrying the minor Ala allele of Asp358Ala had higher mean body mass index (BMI),11 but Esteve et al.13 described greater metabolic syndrome prevalence and BMI with the more common allele (aspartic acid) in 390 Spanish Caucasian subjects, and Hamid et al.14 did not find an association of the same IL-6r polymorphism with BMI. In our older Caucasian sample, we found that minor allele carriers of the rs4537545 marker had very modestly larger waist/hip ratios, although any effect on adipose tissue was too small to alter BMI or lipids. The association with adiponectin found clearly needs to be replicated.
In our hypothesis generating screen of common diseases linked to changes in serum IL-6r levels, we found a decreased risk of type II diabetes (OR = 0.73 95% CI: 0.55–0.95, P = 0.019) although this is based on a relatively small number of cases. Our finding replicates the association of the Asp358Ala SNP with diabetes reported by Hamid et al.14 who found that the minor Ala allele was less frequent in diabetic subjects than controls (38.3 vs 41.2%, P = 0.007) in 1349 type II diabetic patients and 4596 glucose-tolerant control subjects in Denmark. However, we note that the IL-6r locus did not emerge among the most strongly associated polymorphisms in recent genome-wide association studies of diabetes.22–24
Our near-significant reductions in the prevalence of two other major common conditions linked to IL-6r, cardio-vascular disease and asthma, is intriguing but should be followed up in larger studies, with due care to examine associations in the somewhat older population studied here. If the association of this polymorphism with lower risks of the above conditions is confirmed in larger studies, it might be hypothesized that higher levels of sIL-6r might be the result of lower amounts of membrane-bound IL-6r (because of enhanced shedding or proteolytic cleavage) associated regulation of direct IL-6 signalling to leukocytes and hepatocytes. Artificially induced alterations at amino-acid positions 357 and 358 were shown by Mullberg et al.25 to change IL-6r membrane binding through alteration at a cleavage site, although the alanine for aspartic acid substitution of interest was not studied. The Asp358Ala SNP has also been shown to alter mRNA expression in transformed lymphocyte cells.26
The results presented have major implications for research on serum IL-6r as a clinical marker and a therapeutic target. The doubling of serum levels of sIL-6r and the presence of some increase in IL-6 levels in association with the minor allele of the Asp358Ala genetic variant will need to be accounted for in future studies. For example, comparisons of IL-6r levels in cases and controls for any condition will need to ensure appropriate genotypic matching, or preferably comparison within each genotype group. In addition, this variant will need to be studied in other ethnic groups, as existing evidence is from only Japanese10 and our Caucasian subjects. The presence of this polymorphism will also have implications for clinical trial design in therapeutic attempts to block IL-6r activity: investigators, clinicians and regulators may want to see evidence for dosing, efficacy and safety within each genetic variant group.
A common genetic variant of the IL-6 receptor locus (marking the Asp385Ala polymorphism rs8192284) is associated with a substantial change in serum levels of IL-6r and a more modest change in IL-6 levels. However, we have shown no association with gp130 levels or any of a wide range of pro- or anti-inflammatory markers. This absence of effect on inflammatory status may suggest redundancy of increased serum receptor presence in IL-6 signalling.
This work was supported in part by NIH/NIA Grant R01 AG24233-01 and by the Intramural Research Program, National Institute on Aging, NIH. DM is supported by an NHS Executive National Public Health Career Scientist Award, Ref: PHCSA/00/002. The InCHIANTI study was supported as a ‘targeted project’ (ICS 110.1\RS97.71) by the Italian Ministry of Health, by the US National Institute on Aging (Contracts N01-AG-916413 and N01-AG-821336, and 263 MD 9164 13 and 263 MD 821336). The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.
Authors' contributions: SR, TF, KS and DM contributed to the design of this study, and led the statistical analysis, writing and editing of the manuscript. AM participated in the coordination and the design of the study, undertook the genotyping and contributed to the editing of the manuscript. AJH and WH assisted with the statistical analysis and editing of the manuscript. MNW undertook the analysis of the genotyping for quality control. SB, LF and JG led the InCHIANTI study including design and data collection, and contributed to the editing of the manuscript. AMC prepared the InCHIANTI DNA and contributed to the editing of the manuscript. All authors read and approved the final manuscript.
Conflicts of interest: The authors have declared that no competing interests exist.