Background and objectives Vascular disease is common in systemic lupus erythematosus (SLE) and patients with antiphospholipid antibodies (aPL) are at high risk to develop arterial and venous thrombosis. Since HLA class II genotypes have been linked to the presence of pro-thrombotic aPL, we investigated the relationship between HLA-DRB1 alleles, aPL and vascular events in SLE patients.
Methods 665 SLE patients of Caucasian origin and 1403 controls were included. Previous manifestations of ischaemic heart disease, ischaemic cerebrovascular disease (ICVD) and venous thromboembolism (together referred to as any vascular events (AVE)) were tabulated. aPL were measured with ELISA. Two-digit HLA-DRB1 typing was performed by sequence-specific primer-PCR.
Results HLA-DRB1*04 was more frequent among SLE patients with ICVD compared to unaffected patients. This association remained after adjustment for known traditional cardiovascular risk factors. HLA-DRB1*13 was associated with AVE. All measured specificities of aPL—cardiolipin IgG and IgM, β2-glycoprotein-1 IgG, prothrombin (PT) IgG and a positive lupus anticoagulant test were associated with HLA-DRB1*04—while HLA-DRB1*13 was associated with IgG antibodies (β2-glycoprotein-1, cardiolipin and PT). In patients with the combined risk alleles, HLA-DRB1*04/*13, there was a significant additive interaction for the outcomes AVE and ICVD.
Conclusions The HLA-DRB1*04 and HLA-DRB1*13 alleles are associated with vascular events and an aPL positive immune-phenotype in SLE. Results demonstrate that a subset of SLE patients is genetically disposed to vascular vulnerability.
- Systemic Lupus Erythematosus
- Antiphospholipid Antibodies
- Cardiovascular Disease
Statistics from Altmetric.com
Systemic lupus erythematosus (SLE) is a chronic autoimmune disease characterised by the production of autoantibodies, a heterogeneous clinical presentation, a remarkable female predominance (90%) and a very high relative risk for both cardiovascular disease (CVD)1 and venous thromboembolism (VTE).2 With better treatment for lupus itself, cardiovascular co-morbidity has become the major cause of the shorter life expectancy seen in these patients.3 ,4 The mechanisms behind the enhanced CVD risk in SLE are still obscure. It is only to a minor part explained by abundance of traditional ‘CVD risk factors’,5 and it seems to be unaffected by modern treatment as it has not declined since the 1960s.3 Thus, the high incidence of CVD seems to be associated with SLE per se. Pro-thrombotic antiphospholipid antibodies (aPL) occur in 30–50% of SLE patients and prospective studies have demonstrated that their occurrence predict both VTE2 and CVD.6 ,7
In rare cases inherited defects of the early components of the complement cascade seem to cause SLE,8 but in most cases genetic predisposition is assumed to play a contributory role. Common loci, within the major histocompatibility complex (MHC) region on chromosome 6 are known to predispose to SLE, and to other autoimmune diseases.9 In Caucasians SLE has in particular been associated with HLA-DRB1*03 and HLA-DRB1*15.10 Recently, a growing number of other SLE susceptibility genes11–13 have been identified in large international collaborations, but the strongest evidence of association with SLE remains in the MHC region also in light of these genome-wide association studies.12 ,13
Previously genetic variations in mannose-binding lectin,14 Fcgamma receptor IIA genes15 and CRP16 have been associated with thrombotic vascular disease in SLE. We recently reported a surprisingly strong association between a variant of the autoimmune risk gene signal transducer and activator of transcription 4 (STAT4), and occurrence of stroke in SLE patients. We also found that the STAT4 genotype was associated with pro-thrombotic aPL.17
In SLE and in the primary antiphospholipid syndrome (PAPS), aPL occurrence has previously been associated with HLA-DRB1 genotypes, in particular with HLA-DRB1*04 and HLA-DRB1*13 and in a few studies with HLA-DRB1*07.18–21 Our objective was to investigate whether these or other genetic variants in the HLA-DRB1 region are associated with manifest vascular disease in SLE. Prothrombotic aPL predict both arterial and venous events in SLE.2 ,6 ,7 ,22 ,23 As HLA-DRB1 genotypes have been reported to contribute to autoantibody specificities including aPL,18–21 we investigated whether a possible association between HLA-DRB1 alleles and vascular events in SLE was linked to the occurrence of a set of aPL.
SLE patients of European Caucasian origin from three Swedish hospitals were included. All fulfilled at least four of the American College of Rheumatology (ACR) classification criteria for SLE.24 If related only, the first case in each family was included. The first group was from the Karolinska University Hospital (n=364), the second group was from Lund University Hospital (n=161) and the third group was from Uppsala University Hospital (n=140). All included patients were interviewed and examined by a rheumatologist and medical records were scrutinised. The number of ACR 1982 revised SLE criteria,24 treatment for hypertension, occurrence of diabetes and ever habitual smoking were tabulated. Blood samples were consecutively collected and stored at −70°C. A total of 1403 European Caucasian controls from the mid-south of Sweden were genotyped to give a comparison to the local genetic background. These were derived from a population-based study,25 from which controls were selected throughout the middle and southern parts of Sweden. All study participants gave informed consent to participate and the regional ethics boards approved the study.
Vascular events were objectively verified in each case, and defined as follows:
Ischaemic heart disease (IHD): myocardial infarction (MI), confirmed by electrocardiography and a rise in plasma creatine kinase, muscle and brain fraction (CK-MB) or troponine T and/or angina pectoris confirmed by exercise stress test.
Ischaemic cerebrovascular disease (ICVD): stroke including cerebral infarction, confirmed by CT or MRI and/or transitory ischaemic attacks, defined as transient focal symptoms from the brain or retina with a maximum duration of 24 h.
Ischaemic peripheral vascular disease (IPVD): intermittent claudication and/or peripheral arterial thrombosis or embolus confirmed by angiogram or Doppler flow studies.
VTE: deep vein thrombosis, confirmed by venography or ultrasonography and/or pulmonary embolism, confirmed by radionucleotide lung scanning or angiogram.
With any arterial event (AAE) we refer to the occurrence of one or more of 1–3, and with any vascular event (AVE) we refer to the occurrence of one or more of 1–4.
Antibodies and lupus anticoagulant
Antibodies against (a) cardiolipin (CL, IgG and IgM), β2-glycoprotein-1 (β2GP-1, IgG) were determined in 661 patients and anti-prothrombin (aPT, IgG) were analysed in 593 patients by ELISA (Orgentec, Mainz, Germany). All analyses were performed in one laboratory. The cut-off levels corresponded to the 99th percentile of healthy blood donors. Lupus anticoagulant (LAC) was determined with a modified Dilute Russel Viper Venom method (Biopool, Umeå, Sweden) using Bioclot LAC in 363 patients from Stockholm.
HLA-typing was performed by sequence-specific primer PCR assay (SSP-PCR) (DR low-resolution kit; Olerup SSP, Saltsjöbaden, Sweden) and the PCR products were loaded into 2% agarose gels for electrophoresis. An interpretation table was used to determine the specific genotype according to the manufacturer's instructions.26 The HLA-DRB1 allelic groups studied were DRB1*01, DRB1*03, DRB1*04, DRB1*07, DRB1*08, DRB1*09, DRB1*10, DRB1*11, DRB1*12, DRB1*13, DRB1*14, DRB1*15 and DRB1*16.
The STAT4 SNP rs10181656 had previously been genotyped using the GoldenGate assay (Illumina, San Diego, California, USA) or the SNPstream system (Beckman-Coulter, Fullerton, California, USA).17
Patient characteristics and allele frequencies between cases and controls were compared with χ2 tests. Continuous variables were analysed using analysis of variance. OR and 95% CI were calculated from 2×2 contingency tables. Meta-analyses were done with RevMan 5.1.4 (Review Manager (RevMan), V.5.1; The Nordic Cochrane Centre, The Cochrane Collaboration, Copenhagen, 2011). Interaction between HLA-DRB1*04 and HLA-DRB1*13 alleles was estimated through attributable proportion (AP) due to interaction, which is an estimate of the proportion of excess risk when two factors are present simultaneously.27 Patients without both alleles were used as the reference group. Univariable and multivariable-adjusted logistic regression analyses were used to estimate the impact of known CVD risk factors for having presented with previous ICVD. In the multivariable model HLA-DRB1*04, 1 or 2 versus 0 STAT4 risk alleles17 and covariates with p≤0.05 were entered. Due to co-linearity between aPL of different specificities we chose, based on lowest p value in univariable anaysis, aCL IgG as a representative of aPL. Data processing was performed using JMP’ software (SAS Institute, Carey, North Carolina, USA).
Patient characteristics are shown in table 1. A total of 211 patients had presented with at least one vascular event (venous and/or arterial, AVE), 139 patients had experienced AAE and 78 of these patients presented with ICVD; 67 with stroke. Sixty-nine patients had a history of IHD; 55 of them were diagnosed with MI, the remaining patients had isolated symptoms of angina pectoris (table 1). Associations between aPL and vascular events are reported in supplementary table S1.
Associations between HLA-DRB1 alleles and SLE
First we investigated the associations between HLA-DRB1 alleles and the risk of developing SLE. The HLA-DRB1*03 allele was enriched among SLE patients as compared to controls. Due to the strong influence of HLA-DRB1*03 we stratified for the HLA-DRB1*03 allele in all further calculations. This means that DRB1*03 positives were excluded from the testing of DRB1*01, DRB1*04, DRB1*07, DRB1*08, DRB1*13 and DRB1*15. In these calculations we could confirm the association between HLA-DRB1*15 and SLE (table 2, see supplementary table S2).
Associations of HLA-DRB1 alleles and vascular events among patients with SLE
We next explored the HLA-DRB1 genotypes in subgroups of SLE patients as defined by occurrence of previous vascular events. In meta-analysis for the whole patient material, HLA-DRB1*13 was associated with increased risk for AVE, while HLA-DRB1*04 conferred enhanced risk of ICVD (table 3). HLA-DRB1*13 was associated with stroke and VTE in the larger Stockholm cohort, though this association was not significant in meta-analysis (see supplementary table S3). Having the combined risk alleles HLA-DRB1*04/*13 was associated with a high risk for AVE (AP: 30.52, 95% CI 0.10 to 0.94), AAE (AP: 30.65, 95% CI 0.32 to 0.98) and ICVD (AP: 0.63, 95% CI 0.27 to 0.99) and detailed analysis showed that this high risk is due to a significant gene–gene interaction between HLA-DRB1*04 and HLA-DRB1*13 alleles (table 4).
Associations of HLA-DRB1 alleles and aPL antibodies among patients with SLE
A possible association between aPL and the HLA-DRB1 alleles, which were associated with vascular events, was then investigated. In meta-analyses, HLA-DRB1*04 was associated with aPL of all measured specificities and with a positive test for LAC in the patient group from Stockholm. HLA-DRB1*13 was also associated with IgG antibodies targeting β2GP-1, aCL and PT (table 5). Carriers of the combined alleles (*04/*13) were associated with higher ORs for LAC positivity, aCL IgG and aβ2GP-1 IgG and the simultaneous occurrence of ≥positive aPL tests (≥2aPL, table 5, see supplementary table S4). No significant interaction was detected.
Risk factors for ischaemic cerebrovascular events among SLE patients
We finally analysed occurrence of HLA-DRB1*04 alleles and the risk of ICVD in the context of other known risk factors for ICVD. Age at inclusion in this study, gender, diabetes, treatment for hypertension and aCL IgG antibodies, investigated in the majority of patients, were evaluated. We asked if the previously reported association between the SLE risk allele rs10181656(G) in STAT4 and ICVD/aPL,17 was independent of HLA-DRB1*04. In multivariable logistic regression we therefore entered presence of one or two STAT4 rs10181656 risk alleles (yes/no) together with the traditional risk factors, which were associated with ICVD in univariable analysis (table 6). Results demonstrate that age, treatment for hypertension, aCL IgG, STAT4 rs10181656(G) and HLA-DRB1*04 all remained independently associated with a history of ICVD (table 6).
The major finding in the present investigation is that the HLA-DRB1*04 and the HLA-DRB1*13 alleles confer increased risk for vascular events among SLE patients. The observed association between HLA-DRB1*04 and ICVD was independent of available traditional risk factors and of the previously reported association with a STAT417 risk allele. In patients with the combined, *04/*13, genotype the risk alleles were associated with a gene–gene interaction regarding risk of vascular events. Positivity in the functional LAC test was associated with both the HLA-DRB1*04 and the HLA-DRB1*13 alleles. These alleles and aPL of different specificities were consistently associated, extending previous reports.18 ,20 ,28
It has been suggested that HLA-DRB1*04 confers protection with respect to development of SLE29; on the other hand a positive signal of association to SLE from HLA-DRB1*0401 has previously been reported.9 The impact of HLA-DRB1*04 has hitherto not been in focus of genetic studies in SLE, but associations with other autoimmune conditions, in particular rheumatoid arthritis (RA) and type 1 diabetes, are well recognised.9 ,30 Several subtypes of HLA-DRB1*04 belong to the group of HLA-DRB1 alleles, collectively referred to as the shared epitope (SE), which confers the largest known genetic contribution to RA.9 ,30 One cannot exclude that genetic risk may correspond to only one or a limited number of alleles from the HLA-DRB1*04 group. However, it most likely associates with the allele most common in Caucasians, the *0401 allele. To get sufficient power to analyse different HLA-DRB1*04 alleles, one should increase the number of observations dramatically with phenotypes under investigation or study non-Caucasian populations with different patterns of HLA-DRB1*04 alleles.
Three studies have to date demonstrated that the HLA-DRB1*04 is in a dose dependent manner associated with premature mortality in RA, and in particular with mortality caused by CVDs.31–33 Despite the fact that HLA-DRB1*04 is not a risk gene for SLE per se, it was associated with vascular events and with aPL among SLE patients in this study. HLA-DRB1*04 thus seems to be associated with vascular vulnerability both in RA31–33 and in SLE. In the general population the HLA region has not been recognised as a risk region for ICVD, IHD or VTE,30 ,34 with the exception of a recently reported weak association between IHD and HLA-DRB1*01.35 Thus the HLA-DRB1 region seems to be of greater importance for vascular events among patients with autoimmune conditions than in the general population.
We have previously noted that the risk profiles for ICVD and IHD differ in SLE.17 We therefore split arterial events into subgroups. A history of ICVD and IHD was approximately equally common. In a longitudinal design we recently demonstrated that 43% of deaths among SLE patients in Stockholm were caused by IHD, but only 2% by ICVD.22 Consequently a cross-sectional design, which is limited to include only survivors of CVD, as used in this and in many other genetic studies, may systematically underestimate the occurrence of IHD. Longitudinal studies are needed to reliably evaluate the genetic impact on IHD in SLE.
The positive associations between HLA-DRB1*04/*13 and aPL suggest that aPL is one underlying mechanism, which contributes to vascular vulnerability among carriers of these genotypes. In RA the SE has been associated with the occurrence of both anti-citrullinated protein antibodies and rheumatoid factor,36 ,37 erosive disease and extra-articular manifestations,38 suggesting an association with a more pro-inflammatory phenotype. It remains to ascertain whether this is also the case in SLE.
Among SLE patients aPL are well-known risk factors for stroke.17 ,39 In the general population several studies,40 ,41 but not all,40 report a positive association between stroke and aPL, in particular among younger stroke patients.41 ,42 Early studies reported that the occurrence of aPL was linked to certain HLA genotypes. Several of these found that HLA-DRB1*0418 ,19 ,28 ,43 and/or HLA-DRB1*1318 were associated with aPL, both among SLE patients and among patients with the PAPS.28 ,44 We confirm the association between the HLA-DRB1*04 allele and aCL/aβ2GP1 antibodies reported in a large multicentre study by Galeazzi et al.20 We also corroborate the association between HLA-DRB1*04 and aPT reported by Bertolaccini et al and Sebastiani et al.19 ,21 The association with aCL IgM and with positivity in the LAC test is an extension of previous reports. The functional LAC test is generally considered to have higher clinical significance than the specific aPLs.41 ,45 Taken together we demonstrate consistent associations between HLA-DRB1*04 and all aPL specificities investigated and a slightly weaker association between aPL and HLA-DRB1*13, preferentially with antibodies of the IgG isotype targeting CL and protein co-factors β2GP1 and PT. The association between HLA-DRB1*07 and aPL20 was not confirmed in our study (data not shown).
In multivariable analyses, which adjusted for aCL IgG and available traditional risk factors, both HLA-DRB1*04 and STAT4 remained significantly associated with ICVD. Age and hypertension were also predictive as expected. Disease activity, hyperlipidaemia and body mass index have previously been associated with ICVD,42 ,46 but these data were not available in our patients.
The association between SLE and HLA-DRB1*03 and HLA-DRB1*15 is well established.10 ,29 ,47–49 In the investigated Swedish SLE patients, these associations were robust and of similar strength as previously reported. These SLE risk alleles did however not correlate with the occurrence of vascular events; rather there was a negative association between HLA-DRB1*15 and previous IPVD, but this figure should be interpreted with caution due to few cases.
Strength of this study is that all vascular events were confirmed clinically and all aPL were measured in one laboratory. The Stockholm and Uppsala cohorts are similar (table 1). The majority of included patients from Lund belong to a longitudinal cohort from a defined catchment area but consecutive patients from adjacent regions were also included. Disparities in collection procedures have previously been described.4 ,50 Despite minor cohort differences our results remained significant in meta-analyses and they were even stronger in the patient group with the combined risk alleles, the HLA-DRB1*04/*13 genotype. Limitations are that aPL measurements were only performed once, LAC was only determined in the Stockholm cohort, aβ2-GP1 IgM were not measured and we did not evaluate obstetric manifestations of antiphospholipid syndrome (APS). Consequently, we could not assess how many patients formally fulfilled APS criteria.51 As the initial hypothesis was to investigate the association between vascular events and a few candidate HLA-DRB1 alleles we did not correct for multiple comparisons. The investigated patients were all of European Caucasian origin. Thus our results cannot be generalised to patients of other ethnicities.
To conclude, we demonstrate that SLE patients carrying HLA-DRB1*04 and/or HLA-DRB1*13 alleles have an increased risk for vascular events. Presence of the combined risk alleles, HLA-DRB1*04/*13, was associated with interaction regarding risk of a previous vascular event. Both risk alleles were furthermore associated with an aPL positive immune phenotype. The HLA-DRB1*04 allele remained as an independent risk factor for prior ICVD, even after adjustment for aPL. Our results link genotype to both immune and clinical phenotypes. They illustrate that when investigating genetic susceptibility in complex diseases it is important not only to analyse genetic frequencies in the present diagnostic entities, but also to look in more detail at clinical symptoms and subgroups of patients.
We thank Anna-Britta Johansson for analysing aPL, and Eva Jemseby, Gull-Britt Almgren and Julia Boström for handling of blood samples. The authors are also grateful to Orgentec for providing the test reagents.
Review history and Supplementary material
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Contributors All authors of this article have contributed in the following three aspects: conception and design, or analysis and interpretation of data; drafting the article or revising it critically for important intellectual content; and final approval of the version to be published.
Funding ALF funding from Stockholm County Council, the Swedish Research Council, the Swedish Heart-Lung Foundation, the Swedish Rheumatism Foundation, The King Gustaf V 80th Birthday Fund, The Swedish Society of Medicine, The Åke Wiberg Foundation, The Foundation in memory of Clas Groschinsky, Karolinska Institutet Foundations, the Medical Faculty at Lund University, Greta and Johan Kock's Foundation, the Foundation of the National Board of Health and Welfare and Skåne University Hospital, the Uppsala University Hospital Research and Development Fund, the Knut and Alice Wallenberg Foundation, Ragnar Söderbergs Foundation and COMBINE.
Competing interests None.
Ethics approval Regional ethics boards in Stockholm, Lund and Uppsala.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.