Journal of the American Society of Nephrology
2007 JASN IMPACT FACTOR 7.111 HOME   AUTHOR INFO   EDITORIAL BOARD   SUBSCRIBE   FEEDBACK   ALERTS   HELP 
    advanced
CURRENT ISSUE ARCHIVES JASN Express ONLINE SUBMISSION


Published ahead of print on September 20, 2006
J Am Soc Nephrol 17: 3158-3166, 2006
© 2006 American Society of Nephrology
doi: 10.1681/ASN.2006030299
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
ASN.2006030299v1
17/11/3158    most recent
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Liu, Y.
Right arrow Articles by Coresh, J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Liu, Y.
Right arrow Articles by Coresh, J.
Related Collections
Right arrowRelated Article

Human Genetics

Functional Variants in the Lymphotoxin-{alpha} Gene Predict Cardiovascular Disease in Dialysis Patients

Yongmei Liu*, Yvette Berthier-Schaad{dagger},{ddagger}, Laura Plantinga{dagger}, Nancy E. Fink{dagger}, Russell P. Tracy§, Wen Hong Kao{dagger}, Michael J. Klag{dagger}, Michael W. Smith{ddagger},|| and Josef Coresh{dagger}

* Wake Forest University School of Medicine, Winston-Salem, North Carolina; {dagger} Johns Hopkins Medical Institutions, Baltimore, Maryland; {ddagger} Laboratory of Genomic Diversity; || Basic Research Program, SAIC-Frederick, National Cancer Institute, Frederick, Maryland; and § University of Vermont, Burlington, Vermont

Address correspondence to: Dr. Josef Coresh, Johns Hopkins University, 2024 East Monument Street, Suite 2-600, Baltimore, MD 21205. Phone: 410-955-0495; Fax: 410-955-0476; E-mail: coresh{at}jhu.edu

Received for publication April 5, 2006. Accepted for publication August 7, 2006.


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
TNF-beta that is encoded by lymphotoxin-{alpha} gene (LTA) regulates adhesion molecules and IL-6. Previously, a genome-wide case-control study showed that LTA gene variants predisposed to cardiovascular disease (CVD). In a prospective study of 775 dialysis patients, LTA and IL-6 gene variants were tested as independent predictors of CVD risk. Four polymorphisms in the LTA gene and one in the IL-6 gene were genotyped. CVD events were ascertained from medical records. During a mean follow-up of 2.6 yr, 294 first-incident CVD events occurred. The LTA 26Asn variant predicted higher adjusted CVD risk (hazard ratio HR 1.33 for each additional copy of Asn allele; 95% confidence interval 1.14 to 1.55; P = 0.0003). Two other nonsynonymous polymorphisms in the LTA, 13Agr and 51Pro, were associated with lower inflammatory activity and CVD risk. LTA haplotypes (based on all four single-nucleotide polymorphisms) were associated with inflammatory markers and predicted CVD risk (P = 0.005) after adjustment. These LTA genotype associations were independent of the IL-6 –174G/C genotype association that was reported recently. LTA and IL-6 gene variants independently predicted risk for CVD among dialysis patients, suggesting that susceptibility in multiple inflammatory pathways contribute to the development of CVD.


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
Serum measures of the inflammatory process have been associated with increased cardiovascular disease (CVD) risk in many studies, but it is unclear whether the systemic inflammatory response is causative or just a marker of the local vascular inflammatory response that is seen in all phases of atherosclerosis. Proinflammatory cytokines, such as TNF and IL-6, have been implicated as important in atherogenesis and thrombogenesis. We showed previously that a common IL-6 gene polymorphism at position –174 (G/C) in the promoter region (on chromosome 7) predicts incident CVD in incident dialysis patients (1). TNF-beta is a primary proinflammatory cytokine that stimulates the expression of the "messenger" cytokine, IL-6 (2). Lymphotoxin-{alpha} (LTA) gene, encoding TNF-beta, is located on chromosome 6p21. The joint effect of LTA and IL-6 gene variation on CVD risk, if any, is unclear.

A recent experimental study reported that LTA gene knockout mice showed a 62% reduction in fatty streak lesion size (3). A genome-wide screen of 92,788 markers in a Japanese case-control study recently identified the LTA gene as a candidate locus associated with susceptibility to myocardial infarction (MI) (4). In following up on significant findings, 10 common single-nucleotide polymorphisms (SNP) of the LTA gene were studied, and 252A/G in intron 1 was found to be almost completely linked to a missense mutant (Thr26Asn) in exon 3 and associated with MI; subsequent functional assays showed a higher transcriptional activity associated with the 252G allele and higher mRNA expression of vascular cell-adhesion molecule 1 and E-selectin associated with the 26Asn allele. The association of the 252G allele or 26Asn allele with coronary artery disease or MI was confirmed in another Japanese case-control study (5) and a trio family study of white individuals (6) but not in three other case-control studies in white individuals (7–9). Because most of those studies were retrospective, the selection of case patients and control subjects could bias the association, and prospective studies are warranted. Moreover, two nonsynonymous polymorphisms in the coding region, Cys13Arg and His51Pro, have not yet been examined in an association study. Examination of LTA in dialysis patients is helpful in assessing genetic effects in a proinflammatory setting.

In a prospective study, we investigated the association of LTA variants with levels of inflammatory markers at baseline and with incident CVD and cause-specific mortality during 2.6 yr of follow-up in a cohort of multiethnic dialysis patients. We further examined whether the LTA variants predicted risk for CVD independent of IL-6 –174G/C variant. Confirmation of the predisposition to CVD associated with the LTA gene in a prospective study will provide a better understanding of the role of TNF-beta as a key mediator in atherogenesis and/or thrombogenesis.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
Study Design and Population
The Choices for Healthy Outcomes in Caring for ESRD (CHOICE) study is a national, prospective cohort study of 1041 incident dialysis patients who were aged 19 to 95 yr and recruited from 81 dialysis clinics associated with Dialysis Clinic, Inc. (DCI, Nashville, TN), New Haven CAPD (New Haven, CT), or Saint Raphael’s hospital (New Haven, CT). The study design and enrollment criteria have been described elsewhere (10). Participants were enrolled from October 1995 to June 1998, a median of 45 d after initiation of dialysis (95% within 3.5 mo), and were followed up through November 2000. Only those in the DCI facilities (n = 923) were able to participate in the specimen bank (n = 898). Genotype information on IL-6 and LTA was available on 775 of 898 participants for whom blood was drawn before a dialysis session at a median of 5.0 mo from the initiation of dialysis (95% within 8.7 mo). Their baseline demographic and clinical data were obtained from questionnaires as well as hospital and clinic records. The level of comorbidity was assessed by a trained nurse on the basis of medical records and clinic staff reports using the Index of Co-Existent Disease, a standardized and validated four-level scale that has been tested in multiple studies (11). The study was approved by the Institutional Review Board at the Johns Hopkins Medical Institutions and the dialysis clinics, and participants provided written informed consent.

DNA Analysis
TaqMan assays were used for genotyping the four SNP—252A/G in intron 1 (rs909253), Cys13Arg in exon 2 (rs2229094), His51Pro in exon 3 (rs2229092), and Thr26Asn (rs1041981) in exon 3 (at amino acid position 60)—of the LTA gene. A length-modified single base extension protocol (12) was used for genotyping the –174G/C polymorphism in the IL-6 gene. The {kappa} statistics, based on 45 blindly split samples from the CHOICE cohort, were 100% for the LTA SNP and 96% for the IL-6 –174G/C polymorphism. The call rates for the five SNP were 99% or above.

Biochemical Measurements
IL-6 was measured in serum by an ultrasensitive ELISA method (R&D Systems, Minneapolis, MN) with a coefficient of variation (CV; interassay CV) of 7%. Serum high-sensitivity C-reactive protein (hs-CRP) was measured using a colorimetric competitive ELISA (interassay CV 8.9%). Plasma fibrinogen was measured using an automated clot-rate assay (interassay CV 2.9%). Serum P-selectin was measured using an ultrasensitive, solid-phase sandwich ELISA (interassay CV 8.9%).

Outcome Ascertainment
Death events were ascertained through the dialysis clinics and Center for Medicare & Medicaid Services (CMS) data. CVD events included any of the following events that occurred during the follow-up period: MI, cerebrovascular accident, coronary artery bypass graft, percutaneous coronary angioplasty, peripheral artery bypass, amputation, abdominal aortic aneurysm repair, carotid endarterectomy, and sudden coronary death. Medical records from hospitalizations were requested and, when available, reviewed and adjudicated by two members of the study’s outcomes committee using uniformly applied criteria that were modified from the Cardiovascular Health Study (13) and HEMO study (11). The {kappa} statistic for the event adjudication was 95%.

Of 303 deaths, 141 (69% of 204 in-hospital deaths) terminal hospitalization medical records were available. For patients whose terminal medical records were unavailable, the underlying cause of death was determined using the Center for Medicare & Medicaid Services death notification form (Form 2746), completed at the time of death by the dialysis unit staff. The underlying cause of death was considered to be CVD when the death was adjudicated as being due to coronary heart disease, cerebrovascular accident, or peripheral vascular disease.

Statistical Analyses
All regression analyses were performed using STATA 7.0 statistical software (StataCorp, College Station, TX). P ≤ 0.05 was marked as statistically significant and an analysis of the false-positive reporting probability (14) is discussed. The distributions of hs-CRP, IL-6, and P-selectin levels were highly skewed to the right, and logarithmic transformations were applied for normalization. Regression coefficients from a linear regression model of the log-transformed levels were converted to percentage change in the geometric mean by calculating 1 – exp(coefficient). P value for trend was estimated by assigning genotypes with the numeric values 1, 2, and 3.

Survival analysis was conducted from initiation of dialysis to the first CVD event or death. Individuals were censored as a result of renal transplant (n = 129), loss to follow-up (n = 3), or death attributed to causes other than CVD (n = 113). Kaplan-Meier analysis and the log-rank test were used to analyze the associations adjusted for race. Several groups of covariates were selected a priori for multivariate Cox proportional hazards regression models. The first group included the following variables: Age, race, gender, dialysis modality (hemodialysis versus peritoneal dialysis), cigarette smoking, dialysis clinics, diabetes, and comorbidity score. The second group added seven other risk factors: The –174G/C polymorphism of the IL-6 gene, prevalent CVD, congestive heart failure, body mass index (BMI), systolic BP (SBP), serum cholesterol, and albumin. The third group added the four inflammatory makers: hs-CRP, IL-6, fibrinogen, and P-selectin. For all survival analyses, the proportionality assumption of the Cox model was confirmed by inspection of log (–log[survival function]) curves and Schoenfeld residuals.

Haplotypes were constructed separately for black and white patients using the PHASE program, version 2.0.2 (15). Individuals were assigned the most likely pair of haplotypes (when the probability of assignment was >90%). Haplotype dose was modeled 0, 1, and 2 for LTA. In the haplotype analyses, to use the common haplotype as a reference and adjust for other haplotypes, all of the haplotypes except for the common haplotype were included in the regression models. To minimize the type 1 error rate as a result of multiple comparisons procedure, we tested overall significance at the gene using an omnibus likelihood ratio test.

To assess for potential population stratification, we genotyped a panel of 87 ancestry-informative SNP to measure admixture. The degree of individual genetic white-to-black admixture was estimated using Bayesian methods implemented in the STRUCTURE program (version 2) (16). We also tested the null hypothesis that the LTA gene effect is similar across race by stratifying the analyses according to race and evaluating the gene-by-race interaction terms. In addition, we explored the potential gene–environment and gene–gene interactions as part of a full presentation of the data.


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
Table 1 shows the characteristics of the 775 individuals by race. Black patients tended to be younger and were more likely to be female, current smokers, and on hemodialysis. Black patients also less frequently presented with history of CVD and congestive heart failure and had higher BMI and SBP and lower levels of IL-6, CRP, fibrinogen, and P-selectin.


View this table:
[in this window]
[in a new window]

  		Table 1. Patient characteristics by race (n = 775)a

 
The genotype frequencies of the –174G/C polymorphisms and the four LTA SNP were consistent with Hardy-Weinberg expectations. Allele frequencies in both subgroups were similar to those reported previously (17). The variants –174C in the IL-6 gene and 51Pro in the LTA gene were less frequent in black than in white patients, but the two variants 252G and 26Asn in LTA were more common in black patients (Table 1). Because the two polymorphisms 252A/G and Thr26Asn were in almost total linkage disequilibrium (LD; r2 = 1), data are not presented further for 252A/G. The pair-wise LD, r2, for the remaining three SNP ranged from 0.02 to 0.17 in black patients and from 0.04 to 0.20 in white patients. A four-SNP haplotype had the following common variants: Hap1 as carriers of all conserved alleles, Hap2 as carriers of 13Arg but not 51Pro, Hap3 as carriers of both 13Arg and 51Pro, and Hap4 as carriers of 26Asn. Hap1 and Hap3 were less common and Hap4 was more common in black than in white patients.

LTA Polymorphisms and Levels of Inflammatory Markers
The two novel nonsynonymous polymorphisms, 13Arg and 51Pro, were associated with lower IL-6, CRP, and fibrinogen levels (Figure 1). The 26Asn allele was marginally associated with higher P-selectin levels in an additive mode of inheritance, and this association was limited to white patients (P = 0.047 for trend).


Figure 1
View larger version (17K):
[in this window]
[in a new window]
[as a PowerPoint slide]
  		Figure 1. Adjusted levels of inflammatory markers in 775 dialysis patients according to lymphotoxin-{alpha} gene (LTA) genotypes (after adjustment for age, gender, race, dialysis modality, cigarette smoking, dialysis clinics, diabetes, and comorbidity score).

 
LTA Polymorphisms and the Risk for CVD and CVD Mortality
During a mean of 2.6 yr of follow-up in 775 individuals, 294 individuals developed a new event, which comprised 155 individuals with coronary heart disease events, 52 with cerebral vascular disease events, and 87 with peripheral vascular disease events. During this period of follow-up, 303 deaths occurred, 152 of which were attributed to CVD. The Cys13Arg genotype was not statistically associated with risk for CVD (Figure 2A), but those who carried 51Pro allele (Figure 2B) had lower risk for CVD. Compared with the genotype 26Thr/Thr, 26Thr/Asn and 26Asn/Asn were associated with higher risk for CVD (Figure 2C). After adjustment for age, race, gender, dialysis modality (hemodialysis versus peritoneal dialysis), cigarette smoking, dialysis clinics, diabetes, and comorbidity score, these associations remained significant (Figure 2D). Compared with 26Thr homozygotes, the hazard ratios (HR) for CVD were 1.42 (95% confidence interval [CI] 1.09 to 1.85) for 26Thr/Asn heterozygotes and 1.75 (95% CI 1.27 to 2.40) for 26Asn/Asn homozygotes. Under an additive mode of inheritance, the HR associated with the 26Asn allele was 1.33 (95% CI 1.14 to 1.55; P = 0.0003) for risk for incident CVD overall. The similar pattern was seen for CVD mortality (HR 1.31; P = 0.006). This association was not apparent for non-CVD mortality and was weaker for all-cause mortality (data not shown). Table 2 shows the results stratified by race; the 26Asn variant association with CVD/CVD mortality remained significant in white but not in black patients. This may be due to lower power in black patients; interaction with race was not statistically significant.


Figure 2
View larger version (31K):
[in this window]
[in a new window]
[as a PowerPoint slide]
  		Figure 2. Cumulative incidence of cardiovascular disease (CVD; adjusted to white patients) according to LTA genotypes Cys13Arg (A), His51Pro (B), Thr26Asn (C) and adjusted hazard ratio (HR; adjusted for age, gender, race, dialysis modality, cigarette smoking, dialysis clinics, diabetes, and comorbidity score) of CVD risk by LTA genotypes (D) in 775 dialysis patients.

 

View this table:
[in this window]
[in a new window]

  		Table 2. Adjusted HR of incident CVD and CVD mortality associated with the Thr26Asn in the LTA genea

 
LTA Thr26Asn, IL-6 –174G/C, and the Risk for CVD and CVD Mortality
The previously seen associations of the IL-6 –174C allele with higher CVD risk remained after adjustment for the Thr26Asn polymorphism of the LTA gene (HR 1.36; P = 0.01 for –174C carriers; HR 1.31; P = 0.0002 per one-copy increment of 26Asn). The combined effect of the two at-risk SNP, Thr26Asn and –174G/C, is displayed in Figure 3. The effects of LTA 26Asn allele on both incident CVD and CVD mortality are consistent across the individuals with the genotype –174G/G and the –174C carriers. The greatest joint effect, compared with those who were homozygotes for both LTA 26Thr/Thr and IL-6 G/G, was seen for LTA 26Asn/Asn and IL-6 –174C carriers on incident CVD (HR 2.17; 95% CI 1.41 to 3.89; P = 0.001) and CVD mortality (HR 2.75; 95% CI 1.38 to 6.06; P = 0.005). Carriers of either the LTA 26Asn or IL-6 –174C allele were at higher risk for CVD (HR 1.75 [95% CI 1.16 to 2.63; P = 0.008] overall; HR 2.15 [95% CI 1.33 to 3.47; P = 0.002] for white patients; HR 1.09 [95% CI 0.59 to 1.99; P = 0.79] for black patients) and higher risk for CVD mortality (HR 2.62 [95% CI 1.49 to 4.62; P = 0.001] overall; HR 2.30 [95% CI 1.22 to 4.31; P = 0.01] for white patients; HR 6.69 [95% CI 1.01 to 44.43; P = 0.049] for black patients), compared with noncarriers at both loci. There was no statistical interaction between the two gene variants.


Figure 3
View larger version (26K):
[in this window]
[in a new window]
[as a PowerPoint slide]
  		Figure 3. HR of CVD and CVD mortality in 757 dialysis patients according to the LTA (Thr25Asn) and IL-6 (–174G/C) genotypes (adjusted for age, gender, race, cigarette smoking, dialysis modality, dialysis clinics, history of diabetes, and comorbidity score). Genotypes combined multiplicatively with no interaction, P values are for the combined effect of both genotypes: *P < 0.05; **P < 0.01.

 
LTA Haplotypes, Levels of Inflammatory Markers, Incident CVD, and CVD Mortality
Haplotype analyses revealed patterns that were similar to the single-locus analyses (Table 3). Haplotypes other than the four previously mentioned haplotypes were omitted from the analyses because of their small size, and adjusting for them as one group did not change the results. Compared with Hap1 homozygotes, each additional copy of Hap2 was associated with 14% lower CRP levels (P = 0.037) and 19 mg/dl lower fibrinogen (P = 0.029) levels; Hap3 was associated with 24% lower IL-6 (P = 0.029), 24% lower CRP (P = 0.07), and 45 mg/dl lower fibrinogen (P = 0.005) levels; and Hap4 was not associated with inflammatory markers, except for a marginal association with higher serum P-selectin levels (P = 0.08). The four haplotypes explained 1 to 2% of the variation in IL-6, CRP, fibrinogen, and P-selectin levels overall. In white patients, all of the observed associations in overall analyses remained significant; in addition, Hap4 was significantly associated with higher P-selectin levels (P = 0.023). In black patients, only the relationship between Hap3 and lower IL-6 levels was significant (P = 0.008).


View this table:
[in this window]
[in a new window]

  		Table 3. Association of LTA Haplotypes with levels of inflammatory markers, incident CVD, and CVD mortality in 756 dialysis patientsa

 
Compared with the homozygous Hap1 haplotype, Hap3 was marginally associated with lower (P = 0.081), whereas Hap4 was associated with higher (P = 0.029), risk for CVD (Table 3). A similar pattern was seen for CVD mortality (Table 3). These estimates remained after additional adjustment for the –174G/C polymorphism of the IL-6 gene, prevalent CVD, congestive heart failure, BMI, SBP, serum cholesterol, and albumin. Risk estimates were slightly attenuated and were marginally significant after additionally controlling for any one or all four of the inflammatory markers (data not shown).

Secondary Analyses
We also checked for potential gene–environment interaction by running stratified analyses. Similar results were seen in the various subgroups stratified by age (<60 versus ≥60 yr), gender, smoking status, and presence of diabetes or CVD. In addition, all regression models were rerun with adjustment of genetic admixture. No significant changes were observed with this additional adjustment (data not shown).

On the basis of DNA variations in the LTA gene (23 polymorphisms) that were retrieved from the resequencing data of 24 African American and 23 European individuals, (17) a single-haplotype block can be defined using the CI method (18) implemented in Haploview 3.2. Five more polymorphisms were selected to capture common variation in the LTA gene, on the basis of two primary criteria: (1) The minor allele frequency of an SNP >10% and (2) the minimum level of association between assayed and unassayed SNP, measured by the LD statistic r2 of 0.65 (19). These polymorphisms included polymorphisms at sites 306 (rs2009658), 2490 (rs746868), 2756 (rs3053942), 3842 (rs3093544), and 4371 (rs1799964). Although these polymorphisms added information on haplotype heterozygosity, the additional haplotypes did not increase the predictive value of LTA variants for CVD risk.


   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
Ours is the first prospective study to replicate the effect of the haplotype 252G in intron 1 and 26Asn (Hap4) on increased CVD risk reported by the large-scale, genome-wide Japanese study (4). In our study, the predictive value of the variants for CVD risk was independent of the IL-6 –174G/C polymorphism and traditional cardiovascular risk factors. The risk was explained partially by systemic inflammatory markers, such as IL-6, CRP, fibrinogen, and P-selectin. Our results also are consistent with another Japanese case-control study (5) and a trio family study of 460 families recruited from the United Kingdom, Sweden, Germany, and Italy (6) but not with three case-control studies in white individual, in which no association was found (7–9). This difference may be related to the smaller sample size (148 to 199 cases) and exclusion of fatal cases in the case-control studies.

Several studies have suggested that the haplotype 252G in intron 1 and 26Asn (Hap4) influence gene function. The 252G allele is associated with higher TNF-beta levels in human, and greater production of TNF-{alpha} and TNF-beta in in vitro studies, compared with the 252A allele, presumably through higher transcriptional regulatory activity (4,20–24). Ozaki et al. (4) further showed that binding of an unknown nuclear factor to the 252G allele is tighter than to the 252A allele, pointing to a novel regulatory sequence element in intron 1. The 26Asn allele is almost in total LD with the 252G allele, which makes it almost impossible to study its individual independent effect in an epidemiologic study. In an in vitro setting, however, biologic activity of a threonine-to-asparagine substitution at codon 26 was suggested; the 26Asn allele was associated with a higher expression of mRNA coding for vascular cell-adhesion molecule 1 and E-selectin (4). These proteins have a major role in leukocyte rolling and attachment in atherogenesis (25). Expression of P-selectin, an adhesion receptor that is expressed on activated endothelial cells and a member of the adhesion protein family, also is regulated by TNF. The higher level of P-selectin associated with 26Asn allele that was observed in our study suggests that the high-risk haplotype may be atherogenic partially through the adhesion protein pathway.

We also report associations of the two nonsynonymous polymorphisms of LTA, Cys13Arg and His51Pro, with circulating inflammatory markers and CVD risk. 13Arg and 51Pro were associated with lower inflammatory states (as assessed by lower levels of IL-6, CRP, and fibrinogen) and lower CVD risk and CVD mortality, which was particularly prominent for carriers of both 13Arg and 51Pro (defining Hap3). Given that these variants are in coding regions, our findings suggest that they may alter TNF-beta’s structural stability and decrease inflammation. However, this first report of the possible risk implications of these polymorphisms needs to be replicated.

This study has several limitations. We measured products of activated TNF-beta, IL-6, CRP, fibrinogen, and P-selectin levels on only one occasion; the surrogate nature of the measurement and unmeasured variability within individuals may underestimate the relationship between the LTA variants and inflammatory activities. Given the long duration of the processes that lead up to ESRD, people with variant alleles may have died of CVD before developing ESRD. Such selective mortality can lead to an underestimation of the risk that is associated with variant alleles. Because terminal medical records for approximately half of the patients were unavailable for detailed review of their underlying cause of death, nondifferential misclassification bias might occur, which could have underestimated the risk for CVD death or non-CVD death associated with variant alleles. In addition, this genotyping study did not include all of the study participants as a result of missing genotype information. The baseline characteristics of patients who were not included in the genotyping study were similar to those who were included in terms of age, gender, race, smoking, BMI, BP, serum total cholesterol, history of diabetes, and CVD event rate (all comparisons, P > 0.38). Furthermore, we have limited power to examine racial difference, so the ability to generalize these findings to black individuals is uncertain. Sensitivity analysis that adjusted for percentage of admixture among black patients and was estimated using 87 markers did not change the main results. Finally, our study has a limited power to detect a small effect for rare polymorphisms. However, our overall sample would have at least 80% power to detect an HR of 1.29 in CVD risk for the variants with a minor allele frequency >25% at a two-tailed {alpha} of 0.05.

The observed association could be due to a highly linked disease variant in or near LTA. Some resequencing efforts have concluded that three polymorphisms, 10G/A, 252A/G, and Thr26Asn, occur on the same haplotype in general and that the 10G/A did not express a biologic activity (4,17,26). Furthermore, Ozaki et al. (4) also screened the flanking region of the LTA gene and identified two more variants, –63T/A in the promoter region of NFKBIL1 and –23G/C in the promoter region of BAT1, that were in tight association with the LTA 252A/G variant within one extended block of intense LD (also known as a haplotype block). The two variants in the NFKBIL1 (encoding nuclear factor of k light polypeptide gene enhancer in B cells, inhibitor-like 1) and BAT1 gene (encoding HLA-B associated transcript 1), however, did not show biologic activity that would explain their association with higher risk for MI. Furthermore, TNF-{alpha} and LTA loci are located in the MHC. A recent sequence analysis of 69 samples (27) found no association between TNF-{alpha} or LTA alleles and specific human leukocyte antigen (HLA) genes (HLA-B and HLA-DR) alleles. Although TNF-{alpha} alleles had some linkage with LTA alleles, it was not genotyped because no observed polymorphism in the former locus was significantly associated with MI in the Japanese genome-wide association study.

Unlike the measurements of serum levels of IL-6 and TNF that are strongly correlated, the variants in the LTA gene (252A/G and Thr26Asn) and the variant in the IL-6 gene (–174G/C) are uncorrelated because they reside on different chromosomes and thereby provide an option to examine their independent effects. The multiplicative effect of the two gene variants suggests that TNF-beta has additional proatherogenic effects (perhaps mediated through adhesion molecules) in addition to its effect on IL-6. Their independent predictive values for clinically evident CVD and CVD mortality in our study, together with the reported biologic functions of these variants in vitro and in vivo, support the importance of cytokines in determining CVD risk. Therefore, combined evaluation of both genes ought to be more useful in identifying CVD high-risk groups than screening for either alone.

Interpretation of genetic association studies in the face of multiple comparisons can be aided by an analysis of the false-positive report probability suggested by Wacholder et al. (14). If we assume a most likely relative hazard (substituted for odds ratio) worthy of clinical reporting to be 1.5 for CVD in dialysis patients, then we can estimate the false-positive report probability as a function of the previous probability that a given SNP is associated with CVD. For the LTA 26Asn variant (P = 0.0003), the association reported here is likely to have a false-positive report probabilities are 0.002 if a high (0.1) previous probability of association with CVD is assumed. Assigning a high previous probability is supported by previous human, animal, and expression studies of LTA 26Asn (3,4,20–24). The false-positive report probabilities are 0.001, 0.03, and 0.22 if previous probabilities of 0.25 (very high), 0.01 (medium), and 0.001 (low) are assumed. For the omnibus test of LTA haplotypes with CVD (P = 0.0053), the false-positive report probability is 0.02, 0.05, and 0.35 for previous probabilities of 0.25 (very high), 0.1 (high), and 0.01 (medium). The associations with intermediate phenotypes are less statistically significant and may reflect a higher false-positive report rate and therefore should be interpreted more cautiously. We explored these pathways as the most likely intermediates and therefore think that finding associations in the expected directions is supportive of inflammation’s playing a role in the risk that is associated with LTA, but confirmation of the specific associations of these and other intermediate markers is needed.

This study was limited to dialysis patients, who experience both a high level of inflammation and high risk for CVD and mortality and may have different pathophysiology of CVD from the general population. The findings suggest that genetic susceptibility plays an important role in the response of different patients to the proinflammatory and proatherogenic state that is induced by kidney failure and dialysis treatment. More prospective studies and clinical trials in both the high-inflammation populations and other populations are needed to understand fully the interplay between genetic susceptibility at LTA and IL-6 and CVD risk. Confirmation of the independent effect of this genetic susceptibility will strengthen the understanding of the multiple causal pathways that lead to CVD risk and could prompt strategies for targeting effectors of inflammation directly in prevention and therapy of CVD.


   Acknowledgments
 
CHOICE is supported by RO1-HL-62985 (National Heart, Lung, and Blood Institute), RO1-DK-59616 (National Institute of Diabetes and Digestive and Kidney Diseases), R01-HS-08365 (Agency for Healthcare Research and Quality), and a grant from Baxter Healthcare Corp. J.C. is supported in part as an American Heart Association established investigator (01-4019-7N). M.J.K. is supported by K24-DK-02856 (National Institute of Diabetes and Digestive and Kidney Diseases. R.P.T. is supported by HL 46696 and HL 58329. This research was supported by the Intramural Research Program of the National Institutes of Health, National Cancer Institute, Center for Cancer Research. This publication has been funded in part with federal funds from the National Cancer Institute, National Institutes of Health under contract no. NO1-CO-12400.

We thank the patients, staff, laboratory, and physicians who participated in the CHOICE Study at Dialysis Clinic, Inc., and Johns Hopkins University and the Cardiovascular Endpoint Committee: Current members: Bernard G. Jaar, MD, MPH; Yongmei Liu, MD; Joseph A. Eustace, MD, MHS; Richard M. Ugarte, MD; Melanie H. Katzman, MD, MHS; and J. Craig Longenecker, MD, PhD. Former members of the Committee include Michael Klag, MD, MPH; Neil R. Powe, MD, MPH, MBA; Michael J. Choi, MD; Renuka Sothinathan, MD, MHS; and Caroline Fox, MD, MPH. Cardiovascular events adjudicators are Nancy E. Fink, MPH, and Laura C. Plantinga, ScM.


   Footnotes
 
Published online ahead of print. Publication date available at www.jasn.org.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Liu Y, Berthier-Schaad Y, Fallin MD, Fink NE, Tracy RP, Klag MJ, Smith MW, Coresh J: IL-6 haplotypes, inflammation, and risk for cardiovascular disease in a multiethnic dialysis cohort. J Am Soc Nephrol 17 : 863 –870, 2006[Abstract/Free Full Text]
  2. Libby P, Ridker PM: Novel inflammatory markers of coronary risk: Theory versus practice. Circulation 100 : 1148 –1150, 1999[Free Full Text]
  3. Schreyer SA, Vick CM, LeBoeuf RC: Loss of lymphotoxin-alpha but not tumor necrosis factor-alpha reduces atherosclerosis in mice. J Biol Chem 277 : 12364 –12368, 2002[Abstract/Free Full Text]
  4. Ozaki K, Ohnishi Y, Iida A, Sekine A, Yamada R, Tsunoda T, Sato H, Sato H, Hori M, Nakamura Y, Tanaka T: Functional SNPs in the lymphotoxin-alpha gene that are associated with susceptibility to myocardial infarction. Nat Genet 32 : 650 –654, 2002[CrossRef][Medline]
  5. Iwanaga Y, Ono K, Takagi S, Terashima M, Tsutsumi Y, Mannami T, Yasui N, Goto Y, Nonogi H, Iwai N: Association analysis between polymorphisms of the lymphotoxin-alpha gene and myocardial infarction in a Japanese population. Atherosclerosis 172 : 197 –198, 2004[CrossRef][Medline]
  6. A trio family study showing association of the lymphotoxin-alpha N26 (804A) allele with coronary artery disease. Eur J Hum Genet 12 : 770 –774, 2004[CrossRef][Medline]
  7. Padovani JC, Pazin-Filho A, Simoes MV, Marin-Neto JA, Zago MA, Franco RF: Gene polymorphisms in the TNF locus and the risk of myocardial infarction. Thromb Res 100 : 263 –269, 2000[CrossRef][Medline]
  8. Braun J, Marz W, Winkelmann BR, Donner H, Henning UK, Badenhoop K: Tumour necrosis factor beta alleles and hyperinsulinaemia in coronary artery disease. Eur J Clin Invest 28 : 538 –542, 1998[CrossRef][Medline]
  9. Koch W, Kastrati A, Bottiger C, Mehilli J, von Beckerath N, Schomig A: Interleukin-10 and tumor necrosis factor gene polymorphisms and risk of coronary artery disease and myocardial infarction. Atherosclerosis 159 : 137 –144, 2001[CrossRef][Medline]
  10. Longenecker JC, Coresh J, Powe NR, Levey AS, Fink NE, Martin A, Klag MJ: Traditional cardiovascular disease risk factors in dialysis patients compared with the general population: The CHOICE Study. J Am Soc Nephrol 13 : 1918 –1927, 2002[Abstract/Free Full Text]
  11. Rocco MV, Yan G, Gassman J, Lewis JB, Ornt D, Weiss B, Levey AS: Comparison of causes of death using HEMO Study and HCFA end-stage renal disease death notification classification systems. The National Institutes of Health-funded Hemodialysis. Health Care Financing Administration. Am J Kidney Dis 39 : 146 –153, 2002[Medline]
  12. Lindblad-Toh K, Winchester E, Daly MJ, Wang DG, Hirschhorn JN, Laviolette JP, Ardlie K, Reich DE, Robinson E, Sklar P, Shah N, Thomas D, Fan JB, Gingeras T, Warrington J, Patil N, Hudson TJ, Lander ES: Large-scale discovery and genotyping of single-nucleotide polymorphisms in the mouse. Nat Genet 24 : 381 –386, 2000[CrossRef][Medline]
  13. Tracy RP, Lemaitre RN, Psaty BM, Ives DG, Evans RW, Cushman M, Meilahn EN, Kuller LH: Relationship of C-reactive protein to risk of cardiovascular disease in the elderly. Results from the Cardiovascular Health Study and the Rural Health Promotion Project. Arterioscler Thromb Vasc Biol 17 : 1121 –1127, 1997[Abstract/Free Full Text]
  14. Wacholder S, Chanock S, Garcia-Closas M, El GL, Rothman N: Assessing the probability that a positive report is false: An approach for molecular epidemiology studies. J Natl Cancer Inst 96 : 434 –442, 2004[Abstract/Free Full Text]
  15. Stephens M, Smith NJ, Donnelly P: A new statistical method for haplotype reconstruction from population data. Am J Hum Genet 68 : 978 –989, 2001[CrossRef][Medline]
  16. Pritchard JK, Stephens M, Donnelly P: Inference of population structure using multilocus genotype data. Genetics 155 : 945 –959, 2000[Abstract/Free Full Text]
  17. Carlson CS, Eberle MA, Rieder MJ, Smith JD, Kruglyak L, Nickerson DA: Additional SNPs and linkage-disequilibrium analyses are necessary for whole-genome association studies in humans. Nat Genet 33 : 518 –521, 2003[CrossRef][Medline]
  18. Gabriel SB, Schaffner SF, Nguyen H, Moore JM, Roy J, Blumenstiel B, Higgins J, DeFelice M, Lochner A, Faggart M, Liu-Cordero SN, Rotimi C, Adeyemo A, Cooper R, Ward R, Lander ES, Daly MJ, Altshuler D: The structure of haplotype blocks in the human genome. Science 296 : 2225 –2229, 2002[Abstract/Free Full Text]
  19. Carlson CS, Eberle MA, Rieder MJ, Yi Q, Kruglyak L, Nickerson DA: Selecting a maximally informative set of single-nucleotide polymorphisms for association analyses using linkage disequilibrium. Am J Hum Genet 74 : 106 –120, 2004[CrossRef][Medline]
  20. Knight JC, Keating BJ, Rockett KA, Kwiatkowski DP: In vivo characterization of regulatory polymorphisms by allele-specific quantification of RNA polymerase loading. Nat Genet 33 : 469 –475, 2003[CrossRef][Medline]
  21. Messer G, Spengler U, Jung MC, Honold G, Blomer K, Pape GR, Riethmuller G, Weiss EH: Polymorphic structure of the tumor necrosis factor (TNF) locus: An NcoI polymorphism in the first intron of the human TNF-beta gene correlates with a variant amino acid in position 26 and a reduced level of TNF-beta production. J Exp Med 173 : 209 –219, 1991[Abstract/Free Full Text]
  22. Lee SH, Park SH, Min JK, Kim SI, Yoo WH, Hong YS, Park JH, Cho CS, Kim TG, Han H, Kim HY: Decreased tumour necrosis factor-beta production in TNFB*2 homozygote: An important predisposing factor of lupus nephritis in Koreans. Lupus 6 : 603 –609, 1997[Abstract/Free Full Text]
  23. Warzocha K, Ribeiro P, Bienvenu J, Roy P, Charlot C, Rigal D, Coiffier B, Salles G: Genetic polymorphisms in the tumor necrosis factor locus influence non-Hodgkin’s lymphoma outcome. Blood 91 : 3574 –3581, 1998[Abstract/Free Full Text]
  24. Temple SE, Cheong KY, Almeida CM, Price P, Waterer GW: Polymorphisms in lymphotoxin alpha and CD14 genes influence TNFalpha production induced by Gram-positive and Gram-negative bacteria. Genes Immun 4 : 283 –288, 2003[CrossRef][Medline]
  25. Libby P: Inflammation in atherosclerosis. Nature 420 : 868 –874, 2002[CrossRef][Medline]
  26. Abraham LJ, Du DC, Zahedi K, Dawkins RL, Whitehead AS: Haplotypic polymorphisms of the TNFB gene. Immunogenetics 33 : 50 –53, 1991[Medline]
  27. Posch PE, Cruz I, Bradshaw D, Medhekar BA: Novel polymorphisms and the definition of promoter 'alleles’ of the tumor necrosis factor and lymphotoxin alpha loci: inclusion in HLA haplotypes. Genes Immun 4 : 547 –558, 2003[CrossRef][Medline]

Related Article

This Month’s Highlights
J. Am. Soc. Nephrol. 2006 17: 2947-2948. [Full Text] [PDF]




This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
ASN.2006030299v1
17/11/3158    most recent
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Liu, Y.
Right arrow Articles by Coresh, J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Liu, Y.
Right arrow Articles by Coresh, J.
Related Collections
Right arrowRelated Article

HOME CURRENT ISSUE ARCHIVES JASN Express ONLINE SUBMISSION AUTHOR INFO
EDITORIAL BOARD SUBSCRIBE FEEDBACK ALERTS HELP