Genetic Variation Within the Interleukin-1 Gene Cluster and Ischemic Stroke
Background and Purpose—Evidence is emerging that inflammation plays a key role in the pathophysiology of ischemic stroke (IS). The aim of this study was to investigate whether genetic variation in the interleukin-1α, interleukin-1β, and interleukin-1 receptor antagonist genes (IL1A, IL1B, and IL1RN) is associated with IS and/or any etiologic subtype of IS.
Methods—Twelve tagSNPs were analyzed in the Sahlgrenska Academy Study on Ischemic Stroke (SAHLSIS), which comprises 844 patients with IS and 668 control subjects. IS subtypes were defined according to the Trial of Org 10172 in Acute Stroke Treatment criteria in SAHLSIS. The Lund Stroke Register and the Malmö Diet and Cancer study were used as a replication sample for overall IS (in total 3145 patients and 1793 control subjects).
Results—The single nucleotide polymorphism rs380092 in IL1RN showed an association with overall IS in SAHLSIS (OR, 1.21; 95% CI, 1.02–1.43; P=0.03), which was replicated in the Lund Stroke Register and the Malmö Diet and Cancer study sample. An association was also detected in all samples combined (OR, 1.12; 95% CI, 1.04–1.21; P=0.03). Three single nucleotide polymorphisms in IL1RN (including rs380092) were nominally associated with the subtype of cryptogenic stroke in SAHLSIS, but the statistical significance did not remain after correction for multiple testing. Furthermore, increased plasma levels of interleukin-1 receptor antagonist were observed in the subtype of cryptogenic stroke compared with controls.
Conclusion—This comprehensive study, based on a tagSNP approach and replication, presents support for the role of IL1RN in overall IS.
Ischemic stroke (IS) is a complex disease and involves a large array of biological processes, which together determine the susceptibility to develop and sustain ischemic events. Accumulating evidence supports a role for inflammation.1 One of the most potent a proinflammatory cytokine is interleukin-1 (IL-1), the activity of which can be inhibited by the endogenous receptor antagonist IL-1 receptor antagonist (IL-1RA). Interestingly, there are experimental data suggesting a role for IL-1/IL-1RA in IS. For instance, IL-1RA has been shown to reduce, whereas IL-1β exacerbates brain damage in a middle cerebral artery occlusion model in mice.2 IL-1 and IL-1RA may also contribute to the pathophysiology of IS by promoting vessel wall inflammation and atherosclerosis. In humans, mRNA levels of both IL-1β and IL-1RA have been found to be higher in atherosclerotic arteries than in normal arteries.3 Furthermore, it has been suggested that IL-1 has prothrombotic effects, because IL-1 can induce tissue factor and plasminogen activator inhibitor type 1 gene expression.4,5
Given the putative role of IL-1/IL-1RA both in the pathogenesis of IS and in ischemic injury, this pathway has been investigated in clinical studies on IS. These studies show that patients with IS have higher plasma levels of both IL-1β and IL-1RA compared with control subjects.6,7 A smaller study also revealed increased levels of IL-1β in the cerebrospinal fluid after IS, which may indicate an importance of IL-1 locally within the brain.8 Analysis of genetic variation within genes coding for inflammatory mediators can offer some advantage compared with analyses of the plasma protein levels, because genetic variation may affect local levels of the protein and reflect lifelong (rather than transient) inflammation status. Genetic variation in the IL-1 gene cluster has been investigated in relation to IS, the results of which are summarized in online-only Data Supplement Table I. No consistent picture emerges, which may be due to the fact that most previous studies are small and include a restricted number of single nucleotide polymorphisms (SNPs). However, even with regard to somewhat larger samples (>200 cases), both the presence and the absence of an association with IS have been reported.9–14 Because IS is a heterogenous disease with different etiologic subtypes, it is possible that these inconsistent results may reflect subtype specificity. Furthermore, because the influence of genetic factors in IS is more pronounced in younger subjects,15 the age of the participants may play a role.
Therefore, the aim of the present study was to investigate whether there is an association between genetic variation in the IL-1 gene cluster and overall IS and/or any etiologic subtypes of IS using the tagSNP approach in a sample of relatively young (≤70 years) stroke cases and control subjects. Two other Swedish samples were used for replication.
Subjects and Methods
The study population comprised participants in the Sahlgrenska Academy Study on Ischemic Stroke (SAHLSIS), the design of which has been reported.16 Briefly, white patients who presented with first-ever or recurrent acute IS before reaching the age of 70 years (n=844) were consecutively recruited between 1998 and 2008 at 4 stroke units in western Sweden. Healthy white community control subjects (n=668) from the same geographic area as the cases were randomly selected to match cases regarding to age and sex.16 The patients were classified into IS etiologic subtypes according to the Trial of Org 10172 in Acute Stroke Treatment criteria with minor modifications as described in the online-only Data Supplement.17
The Lund Stroke Register and the Malmö Diet and Cancer study were used as a replication sample. Sample characteristics, data collection, and clinical definitions have been described.18,19 Briefly, the Lund Stroke Register is a prospective, epidemiological register, which consecutively includes all patients with first-ever stroke from the area of Lund University Hospital. Control subjects were selected from the same region. The Malmö Diet and Cancer study is a prospective, population-based cohort study, which included 28 449 randomly selected persons at baseline examinations between 1991 and 1996 from which all incident cases of IS up to 2006 were included and matched for age and sex with stroke-free control subjects. At present, subtype data according to the same Trial of Org 10172 in Acute Stroke Treatment classification system as used in SAHLSIS are not available for all patients in the Lund Stroke Register and Malmö Diet and Cancer study samples.
Informed consent was obtained, and the studies were approved by the local Ethics Committees.
Functional outcome at 3 months and at 2 years after index IS was assessed according to the modified Rankin Scale for the first 600 patients in SAHLSIS.
To select a set of SNPs capturing common variation in the IL-1α, IL-lβ, and IL-1RA genes (IL1A, IL1B, and IL1RN, respectively), we used data from the HapMap project on the CEU population (Release 23). The “Tagger” program in HaploView was used to select a minimal set of tagSNP such that all alleles (minor allele frequency >0.1) to be captured were correlated at an r2 >0.8 threshold. This resulted in 12 tagSNP (including rs16944 in IL1B, also known as −511). In addition, rs4251961 was selected because it is located upstream of the IL1RN transcription start site and has been reported to associate with plasma levels of IL-1RA.20
Blood Sampling and IL-1RA Measurements
Standardized blood sampling and isolation of plasma were performed for the first 600 patients and 600 controls in SAHLSIS. In the present study, we selected patients with cryptogenic stroke, together with age- and sex-matched control subjects, to measure the plasma levels of IL1RA. This was performed as part of the analysis of a larger panel of cytokines using a Human Antibody Bead 25-plex kit for the Luminex system (Life Technologies, Carlsbad, CA).
Hardy-Weinberg equilibrium was assessed both in control subjects and cases. Associations between single SNPs and case–control status or functional outcome were investigated using an additive model in binary logistic regression, primarily adjusted for age and sex. In a second model, the vascular risk factors, hypertension, diabetes, and smoking were also included as covariates. Assuming a multiplicative genetic model, the ORs that can be detected with 80% power at the 5% level are in the range of 1.23 to 1.34, depending on the frequency (minor allele frequency, 0.4–0.12) of the high risk allele for overall IS in SAHLSIS. Correction for multiple testing in the genetic analyses without subsequent replication was conducted using Bonferroni correction.
Differences in IL-1RA plasma values between cases and control subjects were examined with Mann-Whitney U test. Time point differences in IL-1RA levels were compared using Wilcoxon signed ranks. The association between plasma IL-1RA values and case–control status was investigated using binary logistic regression, primarily adjusting for age and sex. In a second model, hypertension, diabetes, and smoking were also included as covariates. The reported ORs for IL-1RA were scaled to estimate the ORs associated with an increase of 1 SD in the log IL-1RA plasma level. Associations between genotypes and plasma levels of IL-1RA, high-sensitivity C-reactive protein, and fibrinogen were analyzed using logarithmically transformed plasma values in a linear regression analysis adjusting for age and sex.
For further details on study populations, genotyping, protein measurements, and statistical analyses, see the online-only Data Supplement.
Baseline characteristics of the samples are presented in Table 1. The genotype distribution for all SNPs conformed to Hardy-Weinberg equilibrium (P>0.05) in control subjects and cases, except for rs1143634 in IL1B, which was excluded from further analysis.
Association Between Overall IS and Genetic Variation in IL1RN, IL1A, and IL1B
The observed genotype frequencies in SAHLSIS are presented in online-only Data Supplement Table II. The minor allele (A) of rs380092 in IL1RN was significantly associated with an increased risk for overall IS when adjustments were made for age and sex (Table 2). This association remained after adjustment for hypertension, diabetes, and smoking (Table 2). No SNP in IL1A or IL1B showed an association with overall IS. In previous studies, genetic variation in IL1B has shown an association with obesity.21 However, including body mass index as a covariate in these analyses did not alter the results. Haplotype analyses did not add any further information (online-only Data Supplement Table III).
To investigate whether we could replicate the findings of an association between overall IS and genetic variation in IL1RN, we genotyped rs380092 in the Lund Stroke Register and Malmö Diet and Cancer study (online-only Data Supplement Table IV). A significant association was observed when adjusting for age and sex (Table 2). This association remained after adjusting also for hypertension, diabetes, and smoking (Table 2). Including study site as a covariate in the analysis did not alter the result. A significant association was also observed in a joint analysis of the discovery and the replication samples (Table 2).
Association Between IS Subtypes and Genetic Variation in IL1RN, IL1A, and IL1B
For genotype frequencies in IS subtypes in SAHLSIS, see online-only Data Supplement Table V. The minor alleles of rs380092, rs452204, and rs454078 in IL1RN were all nominally associated with cryptogenic stroke (Table 3). No association for IL1RN was detected for any of the other main IS subtypes. The SNP rs16944 in IL1B showed a nominal association with cardioembolic stroke (OR, 0.75; 0.57–0.99; P=0.04) and rs1143643 in IL1B showed a nominal association with large-vessel disease (OR, 1.37; 1.02–1.86; P=0.04). None of the subtype-specific results remained significant after adjustment for multiple testing.
Plasma IL-1RA Levels in the IS Subtype of Cryptogenic Stroke
We further elucidated the role of IL-1RA in cryptogenic stroke by analyzing plasma protein levels in this subsample of SAHLSIS. Plasma IL-1RA levels, both acutely and in the convalescent phase (3 months after index stroke), were elevated compared with control subjects (median and interquartile range 266 [183–448], 239 [180–347], and 188 [110–320], respectively). There was no significant difference in the plasma level between the 2 time points. Because approximately 13% of the samples had IL-1RA levels that were below the detection limit of the assay, we also tested whether the proportion of cases and control subjects with detectable IL-1RA plasma levels differed. IL-1RA was detected in a larger proportion of samples from patients (97% in the acute phase and 91% at convalescent phase) than from control subjects (72%). Binary logistic regression revealed a significant association between cryptogenic stroke and IL-1RA plasma levels both in the acute phase and in the convalescent phase (Table 4).
No association between genetic variation in IL1RN and plasma levels of IL-1RA was detected in cryptogenic stroke. In control subjects, an association between rs928940 in IL1RN and IL-1RA plasma levels was observed (P=0.02), but this association was not retained after including vascular risk factors in the model. Because high-sensitivity C-reactive protein and fibrinogen levels have been observed to associate with genetic variation in IL1RN,22 associations between IL-1 gene variants and plasma levels of high-sensitivity C-reactive protein and fibrinogen were analyzed in the whole sample in SAHLSIS. No association between SNPs in the IL-1 gene cluster and plasma levels of these variables was observed. Levels of high-sensitivity C-reactive protein and fibrinogen in relation to case–control status in SAHLSIS have been presented.23,24
Functional Outcome 3 Months and 2 Years After IS
No association between functional outcome 3 months or 2 years after index stroke and genetic variation in the analyzed genes was detected.
We report the largest case–control study investigating genetic variation in the IL-1 gene cluster in overall IS and IS subtypes, and our study suggests an association between genetic variation in IL1RN and overall IS.
In contrast to most previous genetic association studies on IS and IL-1, we used a tagSNP approach to capture the genetic variation within the IL-1 gene cluster. The main finding was that the SNP rs380092 in IL1RN was associated with overall IS in the primary sample and the replication sample. We did not genotype the commonly analyzed 86-bp variable number tandem repeat in IL1RN. However, SNP rs454078, which is in strong linkage with the variable number tandem repeat (r2=0.99),22 did not show an association with overall IS in the present sample. This is in line with previous studies on IS, in which a lack of association for the variable number tandem repeat (or for rs419598, a SNP in strong linkage with rs454078) has been observed.10–12 In contrast, in smaller, predominantly Asian studies, an association between the variable number tandem repeat and IS has been reported (see online-only Data Supplement Table I).14 This discrepancy could be due to the relatively small samples sizes or differences between populations.
This is the first study on genetic variation in IL1RN and IS that investigates the different subtypes separately in a larger sample of patients with stroke. The minor alleles of rs380092, rs454078, and rs452204 in IL1RN were all nominally associated with cryptogenic stroke in SAHLSIS. Interestingly, it has been previously suggested that inflammation plays a pathophysiological role in cryptogenic stroke. We therefore continued by analyzing plasma levels of IL-1RA in the subtype of cryptogenic stroke even though the associations for IL1RN did not withstand correction for multiple testing, and elevated IL-1RA levels compared with controls were observed. This is in concordance with studies on overall IS, in which an elevated plasma level of IL-1RA has been detected in patients compared with control subjects.7,25 Although not statistically significant, the median plasma levels in the present study are lower in the convalescent than in the acute phase. Furthermore, a larger proportion of samples was below the detection limit in the convalescent phase compared with the acute phase. Thus, the present and previous data indicate that IL-1RA levels decline over time after stroke25,26 and that increased IL-1RA levels in the acute phase reflect an inflammatory response. The elevated convalescent phase IL-1RA levels in our study may, however, indicate that levels were increased also before stroke onset. Results from prospective studies support the hypothesis that inflammation is involved in the pathogenesis of stroke as elevated C-reactive protein plasma levels are found to be a risk for stroke.27 To further elucidate the potential role of systemic IL-1RA levels in IS, subtype-specific prospective studies on IL-1RA are clearly warranted.
The potential functional role of the SNP that showed an association with overall IS in the present study, rs380092 in IL1RN, is unknown. This SNP is located in an intron and was not associated with plasma levels of IL-1RA in our study. For the exonic synonymous SNP rs315952, which is strongly linked to rs380092 (r2=0.8), there are conflicting data concerning its association with plasma IL-1RA levels.20,28 Furthermore, it cannot be ruled out that there are unknown SNPs that are linked to rs380092 that have an effect on the IL-1RA gene expression and/or protein function. It could also be speculated that the effect of the genetic variation in IL1RN on IS is mediated by increased plasma levels of C-reactive protein and fibrinogen, because associations between genetic variation in IL1RN and plasma levels of these proteins have been reported.22 However, results from our study do not support this hypothesis.
One of the strengths of the present study is that it includes well-characterized patients together with population-based control subjects and that both the initial sample and the replication sample are relatively homogenous because all participants are white and from the southwest of Sweden. There are also some limitations. First, this study is based on hospitalized cases, but the stroke admission rate in Sweden is high with >87% of cases aged <75 years being admitted to the hospital.19 Second, the discovery sample comprising 844 patients with IS has limited power to detect associations with low ORs. Additionally, we had limited power in the analyses of IS subtypes and therefore we cannot exclude a role for IL-1R in the subtypes small-vessel disease, large-vessel disease, and cardioembolic stroke. Third, with regard to IL-1RA plasma levels, a detailed history of recent infections was not available, and because the measurement was conducted as part of analyzing a larger panel of cytokines, the detection limit was not optimal and some samples were below this limit. Finally, the study lacks data on drinking habits.
To our knowledge this is the largest study analyzing genetic variation in the IL-1 gene cluster in IS to date, and it indicates an association between genetic variation in IL1RN and overall IS. In addition, this study adds novel information about the IS subtype of cryptogenic stroke, because genetic variation in IL1RN as well as plasma levels of IL-1RA were associated with this subtype. However, because the genetic association for cryptogenic stroke did not withstand correction for multiple testing, this finding is merely hypothesis-generating and requires replication.
Sources of Funding
This study was supported by the Swedish Research Council (K2011-65X-14605-09-6, K2010-61X-20378-04-3), the Swedish state (ALFGBG-148861), the Swedish Heart and Lung Foundation (20100256 and 20100228), Lund University, Region Skåne, the Freemasons Lodge of Instruction EOS in Lund, King Gustav V and Queen Victoria's Foundation, Promobilia, the Swedish Stroke Association, and the Rune and Ulla Amlöv, John and Brit Wennerström, Lars Hierta, Edit Jacobson, and Tore Nilsson Foundations. Genotyping was performed at the SNP&SEQ Technology Platform at Uppsala University and at the Genomics Core Facility at the Sahlgrenska Academy with support from the Knut and Alice Wallenberg foundation as well as from Uppsala University and University of Gothenburg. For the Lund Stroke Register, DNA isolation and biobank services were performed at Region Skåne Competence Center, Skåne University Hospital.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.111.647446/-/DC1.
- Received December 8, 2011.
- Revision received May 31, 2012.
- Accepted June 1, 2012.
- © 2012 American Heart Association, Inc.
- Touzani O,
- Boutin H,
- LeFeuvre R,
- Parker L,
- Miller A,
- Luheshi G,
- et al
- Olofsson PS,
- Sheikine Y,
- Jatta K,
- Ghaderi M,
- Samnegård A,
- Eriksson P,
- et al
- Tarkowski E,
- Rosengren L,
- Blomstrand C,
- Wikkelsö C,
- Jensen C,
- Ekholm S,
- et al
- Worrall BB,
- Brott TG,
- Brown RD Jr.,
- Brown WM,
- Rich SS,
- Arepalli S,
- et al
- Flossmann E,
- Schulz UG,
- Rothwell PM
- Jood K,
- Ladenvall C,
- Rosengren A,
- Blomstrand C,
- Jern C
- Adams HP Jr..,
- Bendixen BH,
- Kappelle LJ,
- Biller J,
- Love BB,
- Gordon DL,
- et al
- Li C,
- Engström G,
- Hedblad B,
- Berglund G,
- Janzon L
- Reiner AP,
- Wurfel MM,
- Lange LA,
- Carlson CS,
- Nord AS,
- Carty CL,
- et al
- Ladenvall C,
- Jood K,
- Blomstrand C,
- Nilsson S,
- Jern C,
- Ladenvall P
- Rost NS,
- Wolf PA,
- Kase CS,
- Kelly-Hayes M,
- Silbershatz H,
- Massaro JM,
- et al