Abstract

Recent genome-wide association (GWA) studies have identified new genetic determinants of complex quantitative traits, including plasma triglyceride (TG). We hypothesized that common variants associated with mild TG variation identified in GWA studies would also be associated with severe hypertriglyceridemia (HTG). We studied 132 patients of European ancestry with severe HTG (fasting plasma TG > 10 mmol/l), who had no mutations found by resequencing of candidate genes, and 351 matched normolipidemic controls. We determined genotypes for: GALNT2 rs4846914, TBL2/MLXIPL rs17145738, TRIB1 rs17321515, ANGPTL3 rs12130333, GCKR rs780094, APOA5 rs3135506 (S19W), APOA5 rs662799 (−1131T > C), APOE (isoforms) and LPL rs328 (S447X). We found that: (i) genotypes, including those of APOA5 S19W, APOA5 −1131T > C, APOE , GCKR , TRIB1 and TBL2/MLXIPL , were significantly associated with severe HTG; (ii) odds ratios for these genetic variables were significant in both univariate and multivariate regression analyses, irrespective of the presence or absence of diabetes or obesity; (iii) a significant fraction—about one-quarter—of the explained variation in disease status was associated with these genotypes. Therefore, common SNPs (single nucleotide polymorphisms) that are associated with mild TG variation in GWA studies of normolipidemic subjects are also associated with severe HTG. Our findings are consistent with the emerging model of a complex genetic trait. At the extremes of a quantitative trait, such as severe HTG, are found the cumulative contributions of both multiple rare alleles with large genetic effects and common alleles with small effects.

INTRODUCTION

Plasma lipoproteins are archetypal complex traits whose inter-individual variation is determined by both common and rare genetic variants ( 1 ). Recent genome-wide association (GWA) studies have identified new genetic determinants of several complex quantitative traits, including plasma lipoproteins ( 2–7 ). These studies evaluated large samples of normolipidemic (i.e. non-dyslipidemic) individuals and showed that multiple genetic determinants—single nucleotide polymorphisms (SNPs)—had replicable modest associations with plasma concentrations of total, low-density lipoprotein (LDL) and high-density lipoprotein (HDL) cholesterol and triglycerides (TG) ( 2 , 3 , 6 , 7 ). For instance, numerous common genomic variants, including several within well-established candidate genes contributed cumulatively to ∼12% of plasma LDL cholesterol variation in essentially normolipidemic samples ( 2 ), which was consistent with earlier findings in genetic isolates ( 8 ). In addition, associations were observed with many genes that had no previously known biochemical connection with lipoproteins ( 2 , 3 , 6 , 7 ). The potential diagnostic or prognostic utility of lipoprotein-associated markers identified in GWA studies as predictive of individual risk of cardiovascular disease or dyslipoproteinemia is unsettled, in part because the individual markers hold only a modest influence on lipoprotein traits ( 1 , 4 , 5 ). Furthermore, the association of these newly discovered common markers with severe dyslipidemia is unknown.

The genetic determinants of severe hypertriglyceridemia (HTG; MIM 144650), also called Fredrickson or World Health Organization hyperlipoproteinemia (HLP) type 5, are incompletely defined. Plasma TG > 10 mmol/l is found in 1:600 North Americans ( 9 ). Candidate gene resequencing showed that ∼10% of patients with plasma TG > 10 mmol/l together with fasting chylomicronemia had heterozygous loss-of-function missense mutations, primarily in the LPL gene encoding lipoprotein lipase, compared with only 0.2% of controls [carrier odds ratio (OR) 52, 95% confidence interval (CI) 8.6–319] ( 9 ). Furthermore, the common APOA5 S19W missense variant was associated with severe HTG (carrier OR 5.5 95% CI 3.3–9.1) ( 9 ). We hypothesized that common variants recently associated with relatively normal plasma TG identified in GWA studies ( 3 , 7 ) would also be associated with severe HTG. We evaluated polygenic determinants of severe HTG using multivariate linear and logistic regression analysis. We found significant contributions to severe HTG of common variants in several genes that included APOA5 , APOE , TRIB1 , TBL2/MLXIPL , GCKR and GALNT2 , underscoring this trait’s complex polygenic nature and indicating that genetic determinants of modest TG variation also underlie a related, but rarer and more extreme disease phenotype.

RESULTS

Clinical and biochemical features of study subjects

Baseline attributes of the study sample are shown in Table  1 . After excluding 16 patients with heterozygous loss-of-function mutation in LPL , APOC2 or APOA5 ( 9 ), 132 patients or cases with severe HTG remained for analysis. These were each matched with up to four normolipidemic controls based on age within 5 years and sex. By definition, severe HTG patients had markedly higher plasma TG and total cholesterol and significantly lower HDL cholesterol (Table  1 ). Plasma TG concentration in severe HTG patients ranged from 10.1 to 180 mmol/l. In addition, 37/132 severe HTG patients (28.0%) had been hospitalized on ≥1 occasion with pancreatitis and 90/132 (68.2%) had at least one first degree relative treated for dyslipidemia.

Table 1.

Clinical, biochemical and genetic attributes of study subjects

 Severe HTG cases Controls P -value  
Number 132 351  
Female 31.8% 40.7% NS (0.091) 
Diabetes 36.4% 1.1% <0.0001 
Age (years) 50.8 ± 13.1 47.3 ± 14.9 NS (0.14) 
Body mass index (kg/m 2 )  30.7 ± 4.8 27.1 ± 4.2 <0.0001 
Plasma cholesterol (mmol/l)    
 Total 11.9 ± 6.2 5.1 ± 0.8 <0.0001 
 High-density lipoprotein 0.8 ± 0.34 1.3 ± 0.34 <0.0001 
Plasma triglyceride (mmol/l) 31.2 ± 26.5 1.2 ± 0.41 <0.0001 
 Severe HTG cases Controls P -value  
Number 132 351  
Female 31.8% 40.7% NS (0.091) 
Diabetes 36.4% 1.1% <0.0001 
Age (years) 50.8 ± 13.1 47.3 ± 14.9 NS (0.14) 
Body mass index (kg/m 2 )  30.7 ± 4.8 27.1 ± 4.2 <0.0001 
Plasma cholesterol (mmol/l)    
 Total 11.9 ± 6.2 5.1 ± 0.8 <0.0001 
 High-density lipoprotein 0.8 ± 0.34 1.3 ± 0.34 <0.0001 
Plasma triglyceride (mmol/l) 31.2 ± 26.5 1.2 ± 0.41 <0.0001 

HTG, hypertriglyceridemia; NS, not significant.

Differences in distribution of DNA variants between severe HTG cases and controls

Genotype counts and frequencies in severe HTG patients and controls are shown in Table  2 . Minor allele frequencies (MAFs) for each genotype in severe HTG cases and controls are shown in Table  3 . Frequencies of each genotype did not deviate from Hardy–Weinberg equilibrium. The significance of the differences in genotype frequencies between severe HTG cases and controls are shown in Table  2 : in univariate χ 2 analysis, genotype frequencies of each evaluated marker had a significantly different distribution in HTG cases compared with controls (range of P -values 0.024 to 1.5 × 10 −12 ). The significance of the differences in MAFs between severe HTG cases and controls are shown in Table  3 : the MAF of almost every marker studied was significantly different between groups.

Table 2.

Genotype counts and frequencies of candidate genes evaluated

 Severe HTG cases Controls P -value  
APOA5 S19W  
S/S 88 (66.7%) 324 (92.8%)  1.5 × 10 −12 
S/W 37 (28.0%) 21 (6.0%) 
W/W 7 (5.3%) 4 (1.2%) 
APOA5 - 1131T > C  
TT 88 (66.7%) 316 (90.0%)  1.0 × 10 −10 
TC 36 (27.3%) 34 (9.7%) 
CC 8 (6.1%) 1 (0.3%) 
GCKR rs780094  
CC 30 (29.6%) 123 (35.0%)  8.24 × 10 −6 
CT 63 (47.7%) 174 (49.6%) 
TT 39 (22.7%) 54 (15.4%) 
TRIB1 rs17321515  
AA 56 (42.4%) 99 (28.2%)  4.5 × 10 −6 
AG 62 (47.0%) 167 (47.6%) 
GG 14 (10.6%) 85 (24.2%)  
GALNT2 rs4846914  
AA 38 (28.8%) 114 (32.5%)  5.3 × 10 −5 
AG 60 (45.5%) 191 (54.4%) 
GG 34 (25.8%) 46 (13.1%)  
TBL2/MLXIPL rs17145738  
CC 115 (87.1%) 265 (75.5%)  7.2 × 10 −4 
CT 16 (12.1%) 80 (22.8%) 
TT 1 (0.8%) 6 (1.7%)  
ANGPTL3 rs12130333  
CC 93 (70.5%) 215 (61.3%)  7.1 × 10 −4 
CT 38 (28.8%) 119 (33.8%) 
TT 1 (0.8%) 17 (4.8%)  
APOE isotype  
2/2 1 (0.8%) 0 (0.0%) 0.0002 
3/2 24 (18.2%) 33 (9.4%) 
3/3 70 (53.0%) 247 (70.4%) 
4/2 7 (5.3%) 3 (0.9%) 
4/3 25 (18.9%) 62 (17.7%)  
4/4 5 (3.8%) 6 (1.7%)  
LPL S447X  
SS 123 (93.2%) 302 (87.3%) 0.024 
SX 9 (6.8%) 44 (12.7%) 
 Severe HTG cases Controls P -value  
APOA5 S19W  
S/S 88 (66.7%) 324 (92.8%)  1.5 × 10 −12 
S/W 37 (28.0%) 21 (6.0%) 
W/W 7 (5.3%) 4 (1.2%) 
APOA5 - 1131T > C  
TT 88 (66.7%) 316 (90.0%)  1.0 × 10 −10 
TC 36 (27.3%) 34 (9.7%) 
CC 8 (6.1%) 1 (0.3%) 
GCKR rs780094  
CC 30 (29.6%) 123 (35.0%)  8.24 × 10 −6 
CT 63 (47.7%) 174 (49.6%) 
TT 39 (22.7%) 54 (15.4%) 
TRIB1 rs17321515  
AA 56 (42.4%) 99 (28.2%)  4.5 × 10 −6 
AG 62 (47.0%) 167 (47.6%) 
GG 14 (10.6%) 85 (24.2%)  
GALNT2 rs4846914  
AA 38 (28.8%) 114 (32.5%)  5.3 × 10 −5 
AG 60 (45.5%) 191 (54.4%) 
GG 34 (25.8%) 46 (13.1%)  
TBL2/MLXIPL rs17145738  
CC 115 (87.1%) 265 (75.5%)  7.2 × 10 −4 
CT 16 (12.1%) 80 (22.8%) 
TT 1 (0.8%) 6 (1.7%)  
ANGPTL3 rs12130333  
CC 93 (70.5%) 215 (61.3%)  7.1 × 10 −4 
CT 38 (28.8%) 119 (33.8%) 
TT 1 (0.8%) 17 (4.8%)  
APOE isotype  
2/2 1 (0.8%) 0 (0.0%) 0.0002 
3/2 24 (18.2%) 33 (9.4%) 
3/3 70 (53.0%) 247 (70.4%) 
4/2 7 (5.3%) 3 (0.9%) 
4/3 25 (18.9%) 62 (17.7%)  
4/4 5 (3.8%) 6 (1.7%)  
LPL S447X  
SS 123 (93.2%) 302 (87.3%) 0.024 
SX 9 (6.8%) 44 (12.7%) 

APOA5 , gene encoding apolipoprotein A–V; GCKR , gene encoding glucokinase receptor; TRIB1 , gene encoding homologue of Drosophila Tribbles 1 ; GALNT2 , gene encoding UDP- N -acetyl-alpha- d -galactosamine:polypeptide N -acetylgalactosaminyltransferase; TBL2/MLXIPL , locus containing genes encoding transducin-beta-like-2 and MLX interacting protein-like, also called carbohydrate response element binding protein (ChREBP); ANGPTL3 , gene encoding angiopoietin-like 3; APOE , gene encoding apolipoprotein E; LPL , gene encoding lipoprotein lipase.

Table 3.

Candidate gene MAFs

  Severe HTG cases ( N = 132)   Controls ( N = 351)  P -value  
APOA5 W19  0.193 0.042 <0.0001 
APOA5- 1131C  0.197 0.051 <0.0001 
GCKR rs780094 T  0.534 0.402 <0.0001 
TRIB1 rs17321515 G  0.341 0.481 <0.0001 
GALNT2 rs4846914 G  0.485 0.403 0.022 
TBL2/MLXIPL rs17145738 T  0.068 0.131 0.006 
ANGPTL3 rs12130333 T  0.152 0.218 0.024 
APOE non-E3  0.284 0.161 <0.0001 
LPL X447  0.034 0.064 NS (0.083) 
  Severe HTG cases ( N = 132)   Controls ( N = 351)  P -value  
APOA5 W19  0.193 0.042 <0.0001 
APOA5- 1131C  0.197 0.051 <0.0001 
GCKR rs780094 T  0.534 0.402 <0.0001 
TRIB1 rs17321515 G  0.341 0.481 <0.0001 
GALNT2 rs4846914 G  0.485 0.403 0.022 
TBL2/MLXIPL rs17145738 T  0.068 0.131 0.006 
ANGPTL3 rs12130333 T  0.152 0.218 0.024 
APOE non-E3  0.284 0.161 <0.0001 
LPL X447  0.034 0.064 NS (0.083) 

HTG, hypertriglyceridemia; NS, not significant and as of Table  2 .

Genetic risk of severe HTG: univariate ORs

Univariate ORs for severe HTG were determined for two clinical variables – namely diabetes and marked obesity [defined as body mass index (BMI) >33 kg/m 2 ] and for the HTG-risk genotype for each genomic variant. Both dominant and recessive models for each genotype were evaluated and the model that provided the strongest and most significant OR was chosen to serve as the nominal genotype variable for subsequent multivariate analyses. There was no significant linkage disequilibrium between the two APOA5 variants ( P = 0.23), so these were treated as independent variables for the purpose of subsequent analyses. For APOE , presence or absence of the common E3/3 genotype was evaluated.

Univariate ORs and 95% CIs are shown in Table  4 for the most significantly associated genetic model of each genotype: only the ANGPTL3 and LPL genotypes were not significant for either dominant or recessive model. However, both APOA5 variants, APOE non-E3/3 genotype, GCKR TT recessive genotype, TRIB1 AA recessive genotype and TBL2/MLXIPL CC recessive genotype each had significant ORs for severe HTG.

Table 4.

Univariate ORs for severe HTG

 OR (95% CI) 
Diabetes 49.6 (17.4–141) 
Obesity (BMI > 33 kg/m 2 )  6.04 (3.51–10.4) 
APOA5 W19 dominant  6.52 (3.78–11.2) 
APOA5 -1131C dominant  4.51 (2.73–7.46) 
GALNT2 G recessive  2.30 (1.40–3.79) 
GCKR T recessive  2.31 (1.44–3.70) 
TBL2/MLXIPL C recessive  2.20 (1.25–3.86) 
APOE non-E3 allele  2.04 (1.35–3.08) 
TRIB1 A recessive  1.88 (1.24–2.84) 
LPL S447 recessive  1.99 (0.94–4.20) 
ANGPTL3 C recessive  1.51 (0.98–2.32) 
Male sex 1.47 (0.96–2.25) 
APOA5 W19 dominant or -1311C dominant  6.93 (4.44–10.8) 
APOA5 plus 1 additional  6.93 (4.44–10.8) 
APOA5 plus 2 additional  6.79 (4.34–10.6) 
APOA5 plus 3 additional  7.58 (4.79–12.0) 
APOA5 plus 4 additional  8.92 (5.37–14.8) 
APOA5 plus 5 additional  25.0 (9.53–65.5) 
 OR (95% CI) 
Diabetes 49.6 (17.4–141) 
Obesity (BMI > 33 kg/m 2 )  6.04 (3.51–10.4) 
APOA5 W19 dominant  6.52 (3.78–11.2) 
APOA5 -1131C dominant  4.51 (2.73–7.46) 
GALNT2 G recessive  2.30 (1.40–3.79) 
GCKR T recessive  2.31 (1.44–3.70) 
TBL2/MLXIPL C recessive  2.20 (1.25–3.86) 
APOE non-E3 allele  2.04 (1.35–3.08) 
TRIB1 A recessive  1.88 (1.24–2.84) 
LPL S447 recessive  1.99 (0.94–4.20) 
ANGPTL3 C recessive  1.51 (0.98–2.32) 
Male sex 1.47 (0.96–2.25) 
APOA5 W19 dominant or -1311C dominant  6.93 (4.44–10.8) 
APOA5 plus 1 additional  6.93 (4.44–10.8) 
APOA5 plus 2 additional  6.79 (4.34–10.6) 
APOA5 plus 3 additional  7.58 (4.79–12.0) 
APOA5 plus 4 additional  8.92 (5.37–14.8) 
APOA5 plus 5 additional  25.0 (9.53–65.5) 

Variables entered into model are defined as follows: APOA5 W19 dominant had the test genotypes SW and WW and the reference genotype SS; APOA5 -1131C dominant had the test genotypes CC and TC and the reference genotype TT; GALNT2 G recessive had the test genotype GG and the reference genotypes AA and AG; GCKR T recessive had the test genotype AA and the reference genotypes GA and GG; TBL2/MLXIPL C recessive had the test genotype CC and the reference genotypes CT and TT; APOE non-E3 allele had the test genotypes 2/2, 3/2, 4/2, 4/3, 4/4 and the reference genotype 3/3; TRIB1 A recessive had the test genotype AA and the reference genotypes AG and GG; LPL S447 recessive had the test genotype SS and the reference genotype SX (there were no XX individuals); ANGPTL3 C recessive had the test genotype CC and the reference genotypes CT and TT. The last six rows show odds ratios (ORs) for individuals with combinations of at-risk genotypes, staring with either APOA5 genotype and then adding non- APOA5 at-risk genotypes. Abbreviation: BMI, body mass index.

Univariate ORs for severe HTG were also determined for combinations of genetic variables. Since both APOA5 variants were very strongly associated with HTG, the presence of either served as the primary genetic predictor: the OR was 6.93 (95% CI 4.44–10.8). Adding any one or two of the other genetic variables did not substantially change this OR. However, adding 3, 4 or 5 additional genetic markers to the presence of either APOA5 marker sequentially increased the OR from 7.58 to 8.92 to 25.0, so that when an individual had six genetic risk markers (i.e. either APOA5 risk marker plus any other five), the resulting OR for severe HTG was very high indeed.

Polygenic determinants of severe HTG: multivariate regression analysis

The multivariate ORs for severe HTG were calculated using the Wald statistic in multivariate logistic regression analysis with stepwise addition of variables and P < 0.05 for each step (Table  5 ). The first model, which included two clinical variables in addition to nine genetic variables, found that diabetes, obesity, two APOA5 markers, APOE non-E3 genotype and GCKR , TRIB1 and TBL2/MLXIPL genotypes were significantly associated with severe HTG. The C -statistic, which corresponds to the area under the receiver–operator curve for a diagnostic test, was 0.869 for this particular combination of clinical and genetic markers (Table  5 ). Hosmer and Lomeshow goodness of fit test showed that the models explained the observed data (χ 82 = 10.2; P = 0.25 and χ 82 = 7.47; P = 0.38). The second model assessed only genetic variables: the same genotypes from the first model remained significantly associated in the second model with one additional significantly associated genotype – namely GALNT2 , assuming a recessive effect for the G allele. The C -statistic was 0.800 for this combination of genetic markers (Table  5 ).

Table 5.

Multivariate ORs for severe HTG

 Model 1: all variables Model 2: genetic variables only 
Type 2 diabetes 35.9 (11.7, 110)  
Obesity (BMI > 33 kg/m 2 )  2.63 (1.30, 5.55)  
APOA5 W19 dominant  7.79 (3.98, 15.2) 7.36 (3.98, 13.6) 
APOA5 -1131C dominant  5.56 (2.93, 10.6) 5.57 (3.13, 9.90) 
APOE non-E3 allele  2.01 (1.16, 3.52) 2.14 (1.31, 3.49) 
GCKR T recessive  2.03 (1.08, 3.80) 2.11 (1.21, 3.67) 
TRIB1 A recessive  1.86 (1.07, 3.26) 2.02 (1.24, 3.30) 
TBL2/MLXIPL C recessive  2.67 (1.27, 5.62) 2.81 (1.46, 5.24) 
GALNT2 G recessive  NS 2.10 (1.15, 3.81) 
C -statistic  0.869 0.800 
 Model 1: all variables Model 2: genetic variables only 
Type 2 diabetes 35.9 (11.7, 110)  
Obesity (BMI > 33 kg/m 2 )  2.63 (1.30, 5.55)  
APOA5 W19 dominant  7.79 (3.98, 15.2) 7.36 (3.98, 13.6) 
APOA5 -1131C dominant  5.56 (2.93, 10.6) 5.57 (3.13, 9.90) 
APOE non-E3 allele  2.01 (1.16, 3.52) 2.14 (1.31, 3.49) 
GCKR T recessive  2.03 (1.08, 3.80) 2.11 (1.21, 3.67) 
TRIB1 A recessive  1.86 (1.07, 3.26) 2.02 (1.24, 3.30) 
TBL2/MLXIPL C recessive  2.67 (1.27, 5.62) 2.81 (1.46, 5.24) 
GALNT2 G recessive  NS 2.10 (1.15, 3.81) 
C -statistic  0.869 0.800 

Variable names defined as the legend to Table  4 ; NS (not significant). The model used backward elimination and had a nominal P -value of 0.05 for each variable.

The proportion of contribution of specific variables to severe HTG was calculated using partial r2 -values in multivariate linear regression analysis with stepwise addition of variables and P < 0.05 for each step (Table  5 ). The first model, which included two clinical variables in addition to nine genetic variables, found that diabetes, APOA5 markers, obesity, TBL2/MLXIPL genotype, APOE genotype, TRIB1 genotype and GCKR genotype were significantly associated with severe HTG. The model explained ∼43% of total variation in case versus control status, and of the explained variation (Table  6 ), the total contribution of the genetic variables was ∼40% (range ∼1–25%). The second model assessed only genetic variables: the same genotypes from the first model remained significantly associated in the second model with one additional significantly associated genotype – namely GALNT2 . The model accounted for ∼25% of total variation in case versus control status. Of explained variation, genetic markers accounted for ∼1–11% each.

Table 6.

Proportion of explained variation (PEV) of severe HTG

 Marginal (%) Partial (%) 
Model 1: all variables 
 Diabetes 25.65 14.06 
 Obesity (BMI > 33 kg/m 2 )  10.27 0.76 
APOA5 W19 dominant  11.14 6.54 
APOA5 -1131C dominant  7.92 4.34 
APOE non-E3 allele  2.43 1.59 
GCKR T recessive  2.56 0.81 
TRIB1 A recessive  1.84 1.21 
TBL2/MLXIPL C recessive  1.60 1.11 
 Model 43.87  
Model 2: genetic variables only 
APOA5 W19 dominant  11.14 8.17 
APOA5 -1131C dominant  7.92 6.17 
APOE non-E3 allele  2.43 2.07 
GCKR T recessive  2.56 0.83 
TRIB1 A recessive  1.84 1.72 
TBL2/MLXIPL C recessive  1.60 1.78 
GALNT2 G recessive  2.30 1.60 
 Model 25.93  
 Marginal (%) Partial (%) 
Model 1: all variables 
 Diabetes 25.65 14.06 
 Obesity (BMI > 33 kg/m 2 )  10.27 0.76 
APOA5 W19 dominant  11.14 6.54 
APOA5 -1131C dominant  7.92 4.34 
APOE non-E3 allele  2.43 1.59 
GCKR T recessive  2.56 0.81 
TRIB1 A recessive  1.84 1.21 
TBL2/MLXIPL C recessive  1.60 1.11 
 Model 43.87  
Model 2: genetic variables only 
APOA5 W19 dominant  11.14 8.17 
APOA5 -1131C dominant  7.92 6.17 
APOE non-E3 allele  2.43 2.07 
GCKR T recessive  2.56 0.83 
TRIB1 A recessive  1.84 1.72 
TBL2/MLXIPL C recessive  1.60 1.78 
GALNT2 G recessive  2.30 1.60 
 Model 25.93  

Variable names defined as the legend to Table  4 . The marginal percentages correspond to percent contribution without adjustment for other variables, while partial percentages reflect percent contribution when holding influences on other variables constant.

DISCUSSION

The principal novel findings in this study of newly identified genetic markers in patients with severe HTG were: (i) genotypes, including those of APOA5 S19W, APOA5 -1131T > C, APOE , GCKR rs780094, TRIB1 rs17321515, GALNT2 rs4846914 and TBL2/MLXIPL rs17145738, were significantly associated with severe HTG; (ii) ORs for these genetic variables were significant in both univariate and multivariate regression analyses, irrespective of the presence or absence of diabetes or obesity; (iii) a significant fraction—about one-quarter—of the attributable variation in disease status was associated with these genotypes. The findings further indicate that several genotypes that were found by GWA studies to be associated with moderate variation in plasma TG in samples without severe dyslipidemia are also associated with severe HTG. This confirms the complex, polygenic nature of severe HTG and also replicates the importance of loci identified in GWA as being more generally important in TG metabolism, especially in the pathogenesis of severe HTG with chylomicronemia and increased pancreatitis risk.

The current findings extend our previous results, which showed that a relatively small proportion (∼10%) of severe HTG subjects were carriers of rare, heterozygous loss-of-function mutations in candidate lipoprotein metabolism genes ( 9 ). In the current study, in which subjects with rare loss-of-function mutations were excluded, we found that common SNP alleles, including those in both known genes, such as APOA5 , LPL and APOE , and in genes recently implicated as positional candidates, such as TRIB1 , GCKR , TBL2/MLXIPL , GALNT2 and ANGPTL3 , are found in a substantial proportion—almost two-thirds—of individuals with severe HTG. Thus, the genetic component of this complex metabolic trait is composed of both rare and common variants.

We selected TBL2/MLXIPL rs17145738, TRIB1 rs17321515, GALNT2 rs4846914, ANGPTL3 rs12130333 and GCKR rs780094 because they were found by GWA studies to be associated with modest variation in TG in large normolipidemic population samples. In each case, the allele that was associated with severe HTG in our study was also associated in the GWA studies with higher plasma TG concentration: GALNT2 rs4846914 G, TBL2/MLXIPL rs17145738 C, TRIB1 rs17321515 A, ANGPTL3 rs12130333 C, and GCKR rs780094 T. Our study indicates that these common—and so far mechanistically undefined—markers and loci are strongly and cumulatively associated with severely disturbed TG metabolism. This further suggests that rare loss-of-function variants in these genes, or in proximal genes for which the SNPs are markers, might be determinants of severe HTG. Resequencing of genes marked by these SNPs appears thus to be indicated. But while the findings clearly link these genotypes with severe HTG, other factors must be important both in severe HTG patients with and without the genotypes evaluated here, since ∼30% of severe HTG patients had neither a rare dysfunctional variant nor an at-risk SNP genotype.

Thus, our findings are consistent with the emerging model that among individuals at the extremes of a complex genetic trait, such as severe HTG (HLP type 5), are found the cumulative contributions of both multiple rare alleles with large genetic effects and multiple common alleles with small effects. We do not suggest that the variants studied here are directly causative because serve HTG (HLP type 5) is a complex trait with no single simple genetic cause and additional factors, both genetic and non-genetic, are likely to be important determinants. However, the present study substantially increases the proportion of patients with severe HTG—now about two-thirds of patients—who have a significantly associated underlying genetic predisposition. The findings further confirm that the genetic contribution to severe HTG is complex and suggest that other genes or non-genetic factors may still have an important role to play. Also, the results show that significant associations can be identified by studying a relatively small number of subjects with extreme values of a quantitative lipoprotein trait.

MATERIALS AND METHODS

Study subjects

We studied 132 patients of European geographic ancestry with severe HTG, defined as having fasting (>12 h) plasma TG > 10 mmol/l documented on two occasions, from a single tertiary referral lipid clinic ( 9 ). Patients underwent complete medical history and examination; basic clinical, biochemical, and demographic variables were collected. Normolipidemic adult controls were taken from the European subgroup of the Study of Health Assessment and Risk in Ethnic groups (SHARE), a survey of cardiovascular risk factors in Canadian subpopulations ( 10 ) together with healthy population-based controls from the same region of Canada. No control had ischemic heart disease and there was no use of medications among these healthy control subjects. All patients provided informed consent for DNA (deoxyribonucleic acid) analysis.

DNA analysis

DNA was extracted as described ( 9 ). For SNP genotyping, we selected markers that were replicably associated with plasma TG in at least two studies ( 3 , 7 , 11 ) and that showed relatively strong association in each study ( 3 , 7 , 11 ). The selected genes and dbSNP identification numbers were: GALNT2 rs4846914, TBL2/MLXIPL rs17145738, TRIB1 rs17321515, ANGPTL3 rs12130333, GCKR rs780094, APOA5 rs3135506 (S19W) and LPL rs328 (S447X), were genotyped using validated genotyping assays (TaqMan ® SNP Genotyping Assays, Applied Biosystems, Foster City, CA, USA). APOA5 -1131T > C (dbSNP rs662799) was genotyped using a custom designed genotyping assay (TaqMan ® SNP Custom Genotyping Assays, Applied Biosystems). The custom probe uses primers as follows: 5′-CCC TGC GAG TGG AGT TCA-3′ and 5′-CTC TGA GCC CCA GGA ACT G. SNP genotyping was performed using an allelic discrimination assay using the 7900HT Fast Real-Time PCR System (Applied Biosystems) and genotypes were read using automated software (SDS 2.3, Applied Biosystems). Reactions were run in 5 µl volumes using an amplification protocol of 95°C for 10 min, followed by 42 cycles of 95°C for 15 s, then 60°C for 1.5 min. An established method was used to genotype APOE isoforms ( 12 ). For SNP analysis, we excluded patients with a known sequence-proven loss-of-function mutation in LPL , APOC2 or APOA5 , encoding lipoprotein lipase, apo C-II and apo A–V, respectively ( 9 ). Blinded between-day replicated genotypes of a random 3% of samples showed >99.9% concordance across all markers.

Statistical analysis

The two-sample t -test was used to compare the difference between case and control groups for quantitative traits, while Pearson’s χ 2 test was used to compare discrete traits with exact P -values obtained whenever cells contained <5 measurements. Deviations of genotype frequency from the Hardy–Weinberg assumption were assessed using a χ 2 test. Maximal likelihood linkage disequilibrium was estimated using PHASE v2.0 ( 13 ). To assess the relationship of SNPs with severe HTG, dominant and recessive models of minor allele genotypes were tested for each gene. A simple logistic regression model was used to assess univariate association between each SNP and severe HTG. A multiple logistic regression model with backward elimination procedure was adopted to assess the joint effects of genes and clinical variables such as presence of diabetes and marked obesity, i.e. BMI > 33 kg/m 2 . For a genotype with frequency 0.20, the study sample afforded statistical power (alpha error level = 0.05) to detect 1.4-, 1.6-, 1.8- and 2.0-fold increases in frequency of 59.1, 85.7, 96.9 and 99.9%, respectively. The adequacy of the final models was assessed using the Hosmer–Lemeshow goodness-of-fit test. Relative importance of genetic and clinical variables was quantified using the R2 computed with logistic regression raw residuals ( 14 ). Statistical significance was taken at nominal P -value < 0.05 for all comparisons. All analyses were performed using SAS version 9.1 (SAS Institute, Cary, NC, USA), with the exception of the exact tests which were performed using StatXact8 (Cytel Inc, Cambridge, MA, USA).

FUNDING

This study was supported by Jacob J. Wolfe Distinguished Medical Research Chair; the Edith Schulich Vinet Canada Research Chair (Tier I) in Human Genetics; the Jean Davignon Award for Cardiovascular Research (Pfizer, Canada); Career Investigator award from the Heart and Stroke Foundation of Ontario; Canadian Institutes for Health Research (MOP-13430, MOP-39533, MOP-39833); Heart and Stroke Foundation of Ontario (PRG-5967, NA-6059); Ontario Research Fund; Genome Canada through the Ontario Genomics Institute.

ACKNOWLEDGEMENT

Rebecca Provost and John Robinson provided outstanding technical assistance.

Conflict of Interest statement . All of the authors declare no conflict of interests.

REFERENCES

1
Pollex
R.L.
Hegele
R.A.
Genetic determinants of plasma lipoproteins
Nat. Clin. Pract. Cardiovasc. Med.
 , 
2007
, vol. 
4
 (pg. 
600
-
609
)
2
Kathiresan
S.
Melander
O.
Anevski
D.
Guiducci
C.
Burtt
N.P.
Roos
C.
Hirschhorn
J.N.
Berglund
G.
Hedblad
B.
Groop
L.
, et al.  . 
Polymorphisms associated with cholesterol and risk of cardiovascular events
N. Engl. J. Med.
 , 
2008
, vol. 
358
 (pg. 
1240
-
1249
)
3
Kathiresan
S.
Melander
O.
Guiducci
C.
Surti
A.
Burtt
N.P.
Rieder
M.J.
Cooper
G.M.
Roos
C.
Voight
B.F.
Havulinna
A.S.
, et al.  . 
Six new loci associated with blood low-density lipoprotein cholesterol, high-density lipoprotein cholesterol or triglycerides in humans
Nat. Genet.
 , 
2008
, vol. 
40
 (pg. 
189
-
197
)
4
Kathiresan
S.
Musunuru
K.
Orho-Melander
M.
Defining the spectrum of alleles that contribute to blood lipid concentrations in humans
Curr. Opin. Lipidol.
 , 
2008
, vol. 
19
 (pg. 
122
-
127
)
5
Krauss
R.M.
What can the genome tell us about LDL cholesterol?
Lancet
 , 
2008
, vol. 
371
 (pg. 
450
-
452
)
6
Sandhu
M.S.
Waterworth
D.M.
Debenham
S.L.
Wheeler
E.
Papadakis
K.
Zhao
J.H.
Song
K.
Yuan
X.
Johnson
T.
Ashford
S.
, et al.  . 
LDL-cholesterol concentrations: a genome-wide association study
Lancet
 , 
2008
, vol. 
371
 (pg. 
483
-
491
)
7
Willer
C.J.
Sanna
S.
Jackson
A.U.
Scuteri
A.
Bonnycastle
L.L.
Clarke
R.
Heath
S.C.
Timpson
N.J.
Najjar
S.S.
Stringham
H.M.
, et al.  . 
Newly identified loci that influence lipid concentrations and risk of coronary artery disease
Nat. Genet.
 , 
2008
, vol. 
40
 (pg. 
161
-
169
)
8
Hegele
R.A.
Brunt
J.H.
Connelly
P.W.
Multiple genetic determinants of variation of plasma lipoproteins in Alberta Hutterites
Arterioscler. Thromb. Vasc. Biol.
 , 
1995
, vol. 
15
 (pg. 
861
-
871
)
9
Wang
J.
Cao
H.
Ban
M.R.
Kennedy
B.A.
Zhu
S.
Anand
S.
Yusuf
S.
Pollex
R.L.
Hegele
R.A.
Resequencing genomic DNA of patients with severe hypertriglyceridemia (MIM 144650)
Arterioscler. Thromb. Vasc. Biol.
 , 
2007
, vol. 
27
 (pg. 
2450
-
2455
)
10
Anand
S.S.
Yusuf
S.
Vuksan
V.
Devanesen
S.
Teo
K.K.
Montague
P.A.
Kelemen
L.
Yi
C.
Lonn
E.
Gerstein
H.
, et al.  . 
Differences in risk factors, atherosclerosis, and cardiovascular disease between ethnic groups in Canada: the Study of Health Assessment and Risk in Ethnic groups (SHARE)
Lancet
 , 
2000
, vol. 
356
 (pg. 
279
-
284
)
11
Saxena
R.
Voight
B.F.
Lyssenko
V.
Burtt
N.P.
de Bakker
P.I.
Chen
H.
Roix
J.J.
Kathiresan
S.
Hirschhorn
J.N.
Daly
M.J.
, et al.  . 
Genome-wide association analysis identifies loci for type 2 diabetes and triglyceride levels
Science
 , 
2007
, vol. 
316
 (pg. 
1331
-
1336
)
12
Hixson
J.E.
Vernier
D.T.
Restriction isotyping of human apolipoprotein E by gene amplification and cleavage with HhaI
J. Lipid Res.
 , 
1990
, vol. 
31
 (pg. 
545
-
548
)
13
Stephens
M.
Donnelly
P.
A comparison of Bayesian methods for haplotype reconstruction from population genotype data
Am. J. Hum. Genet.
 , 
2003
, vol. 
73
 (pg. 
1162
-
1169
)
14
Heinze
G.
Schemper
M.
Comparing the importance of prognostic factors in Cox and logistic regression using SAS
Comput. Methods Programs Biomed.
 , 
2003
, vol. 
71
 (pg. 
155
-
163
)