Oligogenic familial hypercholesterolemia, LDL cholesterol, and coronary artery disease

Journal of Clinical Lipidology (2018) -, -–-

Original Article

Oligogenic familial hypercholesterolemia, LDL cholesterol, and coronary artery disease Hayato Tada, MD*,1, Masa-aki Kawashiri, MD1, Akihiro Nomura, MD1, Ryota Teramoto, MD1, Kazuyoshi Hosomichi, PhD, Atsushi Nohara, MD, Akihiro Inazu, MD, Hiroshi Mabuchi, MD, Atsushi Tajima, PhD, Masakazu Yamagishi, MD Department of Cardiovascular and Internal Medicine, Kanazawa University Graduate School of Medicine, Kanazawa, Japan (Drs Tada, Kawashiri, Nomura, Teramoto, Nohara, Mabuchi, and Yamagishi); Innovative Clinical Research Center, Kanazawa University, Kanazawa, Japan (Dr Nomura); Department of Bioinformatics and Genomics, Graduate School of Advanced Preventive Medical Sciences, Kanazawa University, Kanazawa, Japan (Drs Hosomichi and Tajima); and Department of Laboratory Science, Molecular Biochemistry and Molecular Biology, Graduate School of Medical Science, Kanazawa University, Kanazawa, Japan (Dr Inazu) KEYWORDS: Familial hypercholesterolemia; PCSK9; Genetics; Low-density lipoprotein cholesterol

BACKGROUND: The genetic background of severe familial hypercholesterolemia (FH) has yet to be determined. OBJECTIVE: We tested if genetic variants associated with low-density lipoprotein (LDL)–altering autosomal recessive diseases influenced LDL cholesterol levels and the odds for coronary artery disease in patients with high LDL cholesterol. METHODS: We recruited 500 individuals with elevated LDL cholesterol levels ($180 mg/dL or $140 mg/dL for subjects ,15 years). We sequenced the exons of 3 FH genes (LDLR, apolipoprotein B, and proprotein convertase subtilisin/kexin type 9) and 4 LDL-altering accessory genes (ABCG5, ABCG8, APOE, and LDL receptor adaptor protein 1). In addition, 4 single nucleotide polymorphisms associated with polygenic FH in East Asian subjects were genotyped. Oligogenic FH patients were defined as those who harbored damaging variants of both conventional FH genes and LDL-altering accessory genes. RESULTS: We identified damaging variants of conventional FH genes in 248 participants (50%). We also detected damaging variants in accessory genes in 57 patients (11%) and identified oligogenic FH in 27 of these patients (5%). Polygenic score in the subjects without any FH mutations was significantly higher than those in any other groups. Compared with monogenic FH, oligogenic FH exhibited significantly higher LDL cholesterol (265 mg/dL, 95% confidence interval [CI] 216–312, and 210 mg/dL, 95% CI 189–243; P 5.04). Oligogenic FH exhibited higher odds for coronary artery disease when compared with monogenic FH, although it did not reach statistical significance (odds ratio 1.41, 95% CI 0.68–2.21, P 5 .24). CONCLUSIONS: Among patients with elevated LDL cholesterol, those with oligogenic FH had higher LDL cholesterol than monogenic FH. Ó 2018 National Lipid Association. All rights reserved.

1

These authors contributed equally. * Corresponding author. Department of Cardiovascular and Internal Medicine, Kanazawa University Graduate School of Medicine, 13-1 Takara-machi, Kanazawa, 920-8641, Japan.

1933-2874/Ó 2018 National Lipid Association. All rights reserved. https://doi.org/10.1016/j.jacl.2018.08.006

E-mail address: [email protected] Submitted June 26, 2018. Accepted for publication August 15, 2018.

2

Introduction Familial hypercholesterolemia (FH) is characterized by the clinical triad of primary hyper low-density lipoprotein (LDL) cholesterolemia, tendon xanthomas, and premature coronary artery disease (CAD).1 FH is caused by damaging mutations in genes associated with LDL metabolism such as LDL receptor (LDLR), apolipoprotein B (APOB), and proprotein convertase subtilisin/kexin type 9 (PCSK9).2 In addition to the classical monogenic FH, the idea of polygenic FH caused by the accumulation of common LDL raising genetic alleles have been introduced.3 On the other hand, the concept of ‘‘oligogenic’’ disease was established based on the recent advancement of comprehensive genetic analyses, through which multiple lowfrequent damaging variants could be associated with Mendelian inherited diseases.4,5 Moreover, recent studies revealed that even a heterozygous damaging variant of genes that cause lipid-related autosomal recessive diseases, such as autosomal recessive hypercholesterolemia caused by mutations in the LDL receptor adaptor protein 1 (LDLRAP1) gene, or sitosterolemia caused by mutations in the adenosine triphosphate-binding cassette subfamily G member 5/8 (ABCG5/8) gene, might additively contribute to the oligogenic inheritance.6-8 The concept of severe FH was recently introduced with the aim of identifying individuals at an extremely high-risk for CAD among patients with FH. We previously demonstrated that the clinical signs of FH as well as the genetic status of FH additively increased the odds of CAD beyond known traditional risk factors.9 Under these conditions, we tested the hypothesis that oligogenic FH who harbored damaging variants of both conventional FH genes and LDL-altering accessory genes is associated with a higher serum LDL cholesterol level and increased odds for CAD in patients with extremely high LDL cholesterol.

Materials and methods Study population We retrospectively analyzed data in a cross-sectional manner pertaining to 500 Japanese unrelated subjects (mean age: 45 years, 203 males [41%], CAD patients 5 123 [25%]) with serum LDL cholesterol level $180 mg/dL or $140 mg/dL for subjects ,15 years of age. The subjects had no known secondary cause for the increase in LDL level, had complete data, and had visited the Kanazawa University Hospital between April 2012 and March 2016. We used data assessed before the introduction of lipid-lowering therapies for all of the subjects.

Clinical signs of FH In addition to the significant elevation of LDL cholesterol, we assessed the subjects for any of the clinical signs of FH: a) tendon xanthoma (tendon xanthoma on the backs

Journal of Clinical Lipidology, Vol -, No -, - 2018 of the hands, elbows, knees, etc., or Achilles tendon hypertrophy; X-ray for the assessment of Achilles tendon thickness $9 mm) or xanthoma tuberosum; and b) family history of FH or premature CAD among the patient’s second-degree relatives, according to the criteria recommended by the Japan Atherosclerosis Society.10

Sequencing of target genes We isolated the genomic DNA for each participant from peripheral white blood cells using a standard DNA extraction protocol. DNA was pooled, selected for size, ligated to sequencing adapters, and amplified to enrich for targets that were sequenced using the KAPA DNA Library Preparation. A custom NimbleGen in-solution DNA capture library (Roche NimbleGen Inc, Madison, WI) was designed to capture all coding exons in 21 dyslipidemia-related Mendelian genes, including 3 FH genes (LDLR, PCSK9, and APOB) and 4 LDL-altering accessory genes (ABCG5, ABCG8, APOE, and LDLRAP1). Details are described in Supplementary Table 1. Target-enriched products were sequenced using the Illumina MiSeq. The target coverage for each subject was $20-fold in $98% of all targeted exons. For each participant, paired-end reads were aligned using the Burrows–Wheeler Aligner Maximum Exact Match algorithm on the human reference genome build GRCh37 using quality score calibration, soft clipping, and adapter trimming. After excluding PCR, duplicate reads using Picard software, single-nucleotide variants and insertions/deletions (indels) were called using the Genome Analysis Toolkit (GATK).11 Variants were filtered based on the Phred-scaled genotype quality score. Indel realignment was performed and the calling algorithm merged the output of the GATK HaplotypeCaller. The GATK Variant Quality Score Recalibration tool was used to update the quality score for each variant. All variants were annotated using the Variant Effect Predictor (version 82)12 and the dbNSFP (ver. 2.9.1) software13 was used to predict their functional consequences. Allele frequencies for the East Asian population were referenced using the Exome Aggregation Consortium database.14 We also assessed copy number variations (CNV) at the LDLR gene using the eXome Hidden Markov Model software.15 The detailed CNV analysis pipeline and commands have been described elsewhere.16 We validated putative CNVs with eXome Hidden Markov Model using Multiplex Ligation-dependent Probe Amplification (MLPA) (FALCO Biosystems Ltd, Kyoto, Japan).

Determination of causative variants for FH in LDLR, PCSK9, and APOB genes We defined a causative variant for FH if it fulfilled any of the following criteria: a) rare (minor allele frequency ,1% among the East Asian population) protein truncating variants (premature stop, insertions or deletions that shift frames, or canonical splice-sites) at the LDLR gene; b) rare damaging missense variants at the LDLR gene, defined as those

Tada et al

Oligogenic FH, LDL-C, and CAD

predicted as damaging by all 5 in silico software (SIFT, Polyphen2-HDIV, Polyphen2-HVAR, MutationTaster-2, and LRT) as previously described17; c) ClinVar-registered pathogenic or likely pathogenic variants that cause FH in LDLR, PCSK9, or APOB18; or d) PCSK9 p.Val4Ile and p.Glu32Lys variants previously reported to cause FH in Japanese.19,20 In addition, we evaluated if those variants are classified as pathogenic, at least with a supporting evidence based on the standard American College of Medical Genetics (ACMG) criteria.21

Involvement of LDL-altering accessory genetic variants We also defined a causative ‘‘accessory’’ variant for the FH phenotype in ABCG5, ABCG8, APOE, and LDLRAP1 genes if it fulfilled any of the following criteria: a) rare protein truncating variants; b) rare damaging missense variants; c) ClinVar-registered pathogenic or likely pathogenic variants that cause hypercholesterolemia. In addition, we evaluated if those variants are classified as pathogenic, at least with a supporting evidence based on the standard ACMG criteria.21

Assessment of polygenic hypercholesterolemia Four single nucleotide polymorphisms (SNPs) validated in assessing polygenic cause of FH in East Asian patients were sequenced (Supplemental Table 2,22). Weighted mean SNP scores were calculated based on LDL cholesterol–raising alleles and their effect sizes shown in the literature22,23

Classification of subjects with significantly elevated LDL cholesterol levels based on comprehensive genotyping We divided the subjects into 5 groups based on the presence/number of FH mutation(s) and LDL-altering

3 accessory genetic variants (groups 1–5, Fig. 1). We defined the patients with a mutation in conventional FH genes and variant(s) in LDL-altering accessory genes as having oligogenic FH (group 4).

Biochemical analysis Blood samples were drawn for assays after overnight fasting. Serum levels of total cholesterol, triglycerides, and HDL cholesterol were determined enzymatically (Qualigent, Sekisui Medical, Tokyo, Japan) using automated instrumentation. LDL cholesterol levels were calculated using the Friedewald formula if triglyceride levels were ,400 mg/dL; otherwise, they were determined enzymatically.

Clinical evaluation Hypertension was defined as systolic blood pressure $40 mm Hg, diastolic blood pressure $90 mm Hg, or prior use of antihypertensive medication. The presence of diabetes was defined as previously described by the Japan Diabetes Society24 or was based on the use of diabetes medication. CAD was defined as the presence of angina pectoris, myocardial infarction, or severe stenotic region(s) in the coronary artery ($75% stenosis), identified either on angiography, computed tomography, or by electrocardiogram in infantile cases.

Ethical considerations The study was approved by the Ethics Committee at the Kanazawa University. All procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1975, as revised in 2008. Informed consent for genetic analyses was obtained from all of the subjects for inclusion in the study.

Figure 1 Genotyping and phenotyping in subjects with severe hypercholesterolemia. We divided these subjects into 5 groups based on the presence/numbers of FH mutation(s), other LDL-altering accessory genetic variants (group 1–5). White indicates the subjects without mutations. Light blue indicates the subjects with an LDL-altering accessory variant. Pink indicates the subjects with a single mutation in conventional FH genes. Purple indicates the subjects with a single mutation in conventional FH genes, and an LDL-altering accessory variant. Gray indicates the subjects with Homo FH, who has double mutations in conventional FH genes. FH, familial hypercholesterolemia; LDL, low-density lipoprotein.

Journal of Clinical Lipidology, Vol -, No -, - 2018

4

Statistical analysis

Quality control of targeted sequencing

Categorical variables are expressed as percentages. Fisher’s exact test or the chi-squared test was used to assess between-group differences, as appropriate. Continuous variables with a normal distribution are presented as mean 6 standard deviation. For nonnormally distributed variables, the median and the interquartile range are reported. Mean values of continuous variables were compared with the Student’s t-test for independent data, and median values were compared with the nonparametric Wilcoxon rank sum test. LDL cholesterol levels were compared among the groups, separated according to the presence of FH mutation(s) and/or other LDL-altering accessory genetic variants, using Tukey’s honest significance test. Odds ratios (ORs) for CAD were calculated using stepwise logistic regression model (backwards) including age, sex, hypertension, diabetes, smoking, LDL cholesterol, and variants in LDL-altering accessory genes. Analyses by ANOVA followed by Tukey post hoc test were conducted to test the differences among the groups. All statistical analyses were conducted using R statistical software. All 2-tailed P-values ,.05 were considered indicative of a statistically significant between-group difference.

We performed targeted sequencing in 3 conventional FH genes (LDLR, APOB, and PCSK9), and 4 LDL-altering accessory genes (ABCG5, ABCG8, APOE, and LDLRAP1). All the study participants had more than 20 coverages in 98% of targeted exons or more. The median read depth in study participants was 216 [interquartile range, 153–292] (Supplementary Fig. 1).

We identified ‘‘classical FH’’ with causative mutations in LDLR or PCSK9 genes in 248 patients (50%). The presence of CNV was validated using MLPA (Supplemental Fig. 2). As we previously described,25 a nonsense mutation in the LDLR gene (c.2431A.T, p.Lys811X) was frequently observed in our cohort (n 5 81, 31%, Supplemental Table 3 and Fig. 3). In addition, none of the subjects in our cohort harbored a causative mutation in the APOB gene, whereas a gain-of-function mutation in the PCSK9 gene (c.94G.A, p.Glu32Lys) was observed relatively frequently (n 5 36, 14%, Supplemental Table 3 and Fig. 3). Supplemental Table 4 illustrates the combination of double mutations found in group 5 (among Supplemental Table 3).

Identified LDL-altering accessory genetic variants

Results Characteristics of study subjects The clinical characteristics of the study subjects are summarized in Table 1. A total of 324 (65%) subjects exhibited either of the 2 clinical signs of FH (xanthoma or family history). As expected, there were significant differences between patients with and without CAD in most, but not all, parameters.

Table 1

Identified classical FH mutations

Deleterious variant(s) in other LDL-altering accessory genes, including ABCG5/8, APOE, and LDLRAP1 were identified in 57 patients (11%), among whom 27 oligogenic FH patients (5%) were found. Some of the above variants are known to cause sitosterolemia or autosomal recessive hypercholesterolemia, whereas others are related to APOE7.26 Identified variants are summarized in Supplemental Table 5 and in Supplemental Figure 4.

Baseline characteristics of the subjects All

CAD

Variable

(n 5 500)

YES (n 5 123)

NO (n 5 377)

P-value

Age (y) Male Hypertension Diabetes Smoking Total cholesterol (mg/dL) Triglyceride (mg/dL) HDL cholesterol (mg/dL) LDL cholesterol (mg/dL) Clinical signs of FH (family history and/or tendon xanthoma) FH mutation

45 6 19 203 (41%) 105 (21%) 33 (7%) 119 (24%) 303 [276–346] 117 [78–168] 54 [45–64] 223 [197–263] 324 (65%)

58 6 13 76 (62%) 64 (52%) 24 (20%) 76 (62%) 298 [268–352] 132 [90–210] 45 [38–55] 221 [190–280] 104 (85%)

40 6 19 125 (33%) 41 (11%) 9 (2%) 43 (11%) 304 [279–345] 111 [71–156] 56 [48–65] 223 [198–261] 220 (58%)

,2 ! 10216 3.4 ! 1028 1.5 ! 10214 1.2 ! 10210 ,2 ! 10216 .2934 6.5 ! 1025 1.9 ! 10212 .351 2.2 ! 1027

248 (50%)

64 (52%)

184 (48%)

CAD, Coronary artery disease; FH, familial hypercholesterolemia; HDL, high-density lipoprotein; LDL, low-density lipoprotein.

.3812

Tada et al

Oligogenic FH, LDL-C, and CAD

5

Clinical significance of LDL-altering accessory genetic variants The characteristics of the patients in each group are summarized in Table 2. There was no significant difference in the proportions of mutation types among those groups (Table 2). Interestingly, the weighted polygenic score in the group 1 was significantly higher than those in any other groups, whereas those in the group 3 and group 4 were not different (Table 2). Next, we compared the LDL cholesterol levels among the groups and separated them based on the presence of FH mutation and LDL-altering accessory genetic variants. Compared with participants without damaging variants, those with oligogenic FH exhibited significantly higher LDL cholesterol (265 mg/dL, 95% confidence interval [CI] 216–312, and 210 mg/dL, 95% CI 189–243; P 5 3.7 ! 10212, Fig. 2). In addition, we found that the LDL cholesterol levels for patients with oligogenic FH

Table 2

were significantly higher than those for patients with ‘‘classical FH,’’ defined as a single FH mutation (group 4 vs group 3, 265 mg/dL, 95% CI 216–312 vs 233 mg/dL, 95% CI 207– 275, respectively; P 5 .04, Fig. 2). In addition, we could see a general trend of increased odds for CAD with the accumulation of FH mutations as well as LDL-altering accessory genetic variants (Fig. 3). Compared with participants without damaging variants, those with oligogenic FH exhibited significantly increased odds for CAD (OR 5.0; 95% CI 2.2–7.4; P 5 1.1 ! 1024). In addition, oligogenic FH exhibited higher odds for CAD when compared with monogenic FH, although it did not reach statistical significance (OR 1.41, 95% CI 0.68–2.21, P 5 .24). The association between the presence of such accessory variants and CAD did not reach statistical significance (OR 1.29; 95% CI 0.16– 2.28; P 5 .16, Table 3), although most of the established risk factors, including age, sex, diabetes, and smoking as well as clinical signs of FH and FH mutation were significantly associated with CAD (Table 3).

Baseline characteristics of the subjects according to the FH classification Classification Group 2 (n 5 30) LDL-altering accessory variants

Group 3 (n 5 210) monogenic FH

Group 4 (n 5 27) oligogenic FH

Group 5 (n 5 11) ‘‘Homo’’ FH

Variable

All (n 5 500)

Group 1 (n 5 222) no mutation

Age (y) Male Hypertension Diabetes Smoking Total cholesterol (mg/dL) Triglyceride (mg/dL) HDL cholesterol (mg/dL) LDL cholesterol (mg/dL) CAD Protein truncating mutation (LDLR) Missense mutation (LDLR) Missense mutation (PCSK9) Variant (LDL-altering accessory genes) Weighted polygenic score

45 6 19 203 (41%) 105 (21%) 33 (7%) 119 (24%) 303 [276–346]

50 6 17 93 (42%) 57 (26%) 20 (9%) 64 (29%) 295 [271–325]

46 6 21 10 (33%) 7 (23%) 2 (7%) 6 (20%) 300 [277–344]

41 6 19 85 (40%) 35 (17%) 10 (5%) 44 (21%) 307 [280–352]

37 6 15 11 (41%) 6 (22%) 1 (4%) 5 (19%) 351 [296–393]

25 6 19 4 (36%) 0 (0%) 0 (0%) 0 (0%) 423 [362–446]

117 [78–168] 54 [45–64] 223 [197–263] 123 (25%) 126

130 [91–192] 53 [45–64] 210 [189–243] 51 (23%) 0 (0%)

107 [74–147] 60 [47–66] 215 [190–236] 4 (13%) 0 (0%)

106 [70–151] 53 [45–63] 233 [207–275] 58 (28%) 106 (50%)

86 [66–118] 54 [46–62] 265 [216–312] 9 (33%) 14 (52%)

85 [58–119] 51 [41–59] 341 [225–378] 1 (9%) 6 (55%)

90

0 (0%)

0 (0%)

70 (33%)

10 (37%)

10 (91%)

43

0 (0%)

0 (0%)

34 (16%)

3 (11%)

6 (55%)

67

0 (0%)

35* (100%)

0 (0%)

32† (100%)

0 (0%)

0.46 6 0.10x

0.47 6 0.10ǁ

0.48 6 0.12

0.52 6 0.11‡

0.46 6 0.17{

0.48 6 0.20

CAD, coronary artery disease; FH, familial hypercholesterolemia; HDL, high-density lipoprotein; LDL, low-density lipoprotein; LDLR, low-density lipoprotein receptor; PCSK9, proprotein convertase subtilisin/kexin type 9. Square brackets indicate interquartile range. *5 individuals exhibited double mutations. †5 individuals exhibited double mutations. ‡P-values for group 1 versus group 2 5 .026, group 1 versus group 3 5 .034, group 1 versus group 4 5 .04, group 1 versus group 5 5 .09. xP-values for group 2 versus group 3 5 .339, group 2 versus group 4 5 .68, group 2 versus group 5 5 .246. ǁ P-values for group 3 versus group 4 5 .45, group 3 versus group 5 5 .196. {P-value for group 4 versus group 5 5 .188.

6

Discussion In the present study, we found that w5% of subjects harbored deleterious mutations in FH genes as well as in LDL-altering accessory genes, and that such ‘‘accessory’’ variants significantly affected the LDL cholesterol level among individuals referred to our hospital with a significant elevation in LDL cholesterol. Our results demonstrated the idea of ‘‘oligogenic FH’’ for the first time, and that comprehensive genotyping in patients with severe hyper LDL cholesterolemia might help to identify extremely high-risk individuals among those with elevated LDL cholesterol level. A definition of severe heterozygous FH was recently proposed by the International Atherosclerosis Society, mainly focusing on the classical coronary risk factors.27 On the other hand, genome-wide association studies have repeatedly shown that common genetic variations in the ABCG5/8, APOE, and LDLRAP1 genes are associated with plasma LDL cholesterol levels. It leads to the notion of polygenic FH caused by the accumulation of such LDL-raising alleles.3 Moreover, the concept of oligogenic disease has been established as a condition in which the accumulation of multiple low-frequent damaging variants mimics the Mendelian disease.5 In this study, we performed comprehensive genetic analyses among subjects with significantly elevated LDL cholesterol level and found that deleterious ‘‘accessory’’ variants affected their LDL cholesterol levels, rather than the accumulation of common SNPs. It is true that polygenic, monogenic, and oligogenic diseases are all categorized as Mendelian disease; however, the severity of disease, at least that of FH, seems to be different. Our data, as well as others, suggest that monogenic FH exhibit more severe phenotype than polygenic FH (in our data, group 3 vs group 1). In addition to those observations, the current data suggest that oligogenic FH

Journal of Clinical Lipidology, Vol -, No -, - 2018 (group 4) exhibits more severe phenotype than monogenic (group 3) as well as polygenic FH (group 1). Several conclusions may be drawn from these observations. First, comprehensive genetic analysis was quite useful for determining the genetic status of such high-risk individuals. We could successfully determine the genotypes of 96 individuals simultaneously within 5 days after the extraction of DNA using this scheme. Second, checking for genetic backgrounds, including LDL-altering accessory genetic variants, is quite important in patients with significant elevation of LDL cholesterol. Recent studies have suggested that identifying individuals with a genetic highrisk could lead to greater benefits with the implementation of statin therapy28 and a healthy life style.29 Genetic information could be determined long before the development of classical coronary risk factors such as hypertension, diabetes, and smoking. Therefore, genetic testing could be more useful for younger individuals, who have yet to develop clinical signs of FH.30 Third, the polygenic FH also seem to exist among Japanese severe hypercholesterolemic patients.

Study limitations The retrospective cross-sectional observational study design is a key limitation of our study. However, our findings are based on one of the largest samples of Japanese patients with significantly elevated LDL cholesterol and can contribute to our understanding of this subject across ethnicities. Moreover, this is the first study investigating the prevalence as well as the impact of oligogenic FH on LDL cholesterol and CAD. Second, we only included individuals with significantly elevated LDL cholesterol levels, ignoring findings from the general population with normal LDL cholesterol levels. In addition, we recruited participants at the outpatient clinic of our University Hospital, leading to a selection bias. However, we believe that

Figure 2 Impact of comprehensive genetic status of FH on LDL cholesterol level. LDL cholesterol levels are plotted according to the mutation status. *P , .05, **P , .001. FH, familial hypercholesterolemia; LDL, low-density lipoprotein.

Tada et al

Oligogenic FH, LDL-C, and CAD

7

OR (95% CI) Group 1 (n = 222) No Mutation

P value

Reference

Reference

Group 2 (n = 30) LDL-altering accessory variant(s)

1.0 (0.2-2.0)

0.46

Group 3 (n = 210) Monogenic FH

3.1 (1.7-6.0)

0.00038

Group 4 (n = 27) Oligogenic FH

5.0 (2.2-7.4)

0.00011

Group 5 (n = 11) “Homo” FH

2.5 (0.1-24)

0.47

0.1

1

10

100

OR for coronary artery disease

Figure 3 Impact of genetic status of FH on CAD. Odds ratio for CAD were calculated using multivariate logistic regression after adjustment for age, sex, hypertension, diabetes, smoking, and LDL cholesterol levels. CAD, coronary artery disease; FH, familial hypercholesterolemia; LDL, low-density lipoprotein.

our intensive assessments of genotype and phenotype in the subjects with severely elevated LDL cholesterol could aid in identifying the individuals at the highest risk. Third, the pathogenicities of the variants were not fully assessed in this study, potentially leading to underestimation or overestimation of the effects of the variants. However, most of the genetic variants identified in this study have already been shown to be associated with the elevation of LDL cholesterol and/or with the FH phenotype itself. In addition, we used stringent criteria to determine the causative variants, and found that all the variants in this study had any supporting evidence of pathogenicities of the variants according to ACMG criteria. Accordingly, we believe that misclassification of the pathogenicity of the variants is likely to be minimal. Fourth, we did not perform MLPA analyses for all the subjects in this study, although we validated CNV Table 3

Factors associated with CAD

Variable

OR (95% CI)

P value

Age (y) Male Hypertension Diabetes Smoking LDL cholesterol (per 10 mg/dL) Clinical signs of FH A mutation in conventional FH-genes Variant(s) in LDL-altering accessory genes

1.11 2.46 1.70 3.49 6.54 1.02

4.3 ! 10213 .019 .15 .025 4.8 ! 1027 .03

(1.08–1.14) (1.16–5.24) (0.82–3.55) (1.20–10.88) (3.96–15.44) (1.01–1.04)

8.34 (3.96–18.32) 2.62 (1.36–5.16)

1.29 (0.16–2.28)

1.1 ! 1028 .0045

.16

FH, familial hypercholesterolemia; OR, odds ratio. Multivariate logistic regression analysis was performed including the variables listed in the table.

identified by our platform using MLPA (Supplemental Fig. 2). Thus, we may miss the subjects, especially, those with single exon deletion. Fifth, the variants identified in our scheme were not confirmed by Sanger sequencing. However, the median read depth in this study was as much as 216. That fact could make us quite confident of our results. Sixth, mutation detection rate as well as the proportion of oligogenic FH harbored damaging variants of both conventional FH genes and LDL-altering accessory genes are quite high in our study. This is probably due to a selection bias referred to our hospital. However, the importance lies not in the mutation rate or in the proportion of oligogenic FH, but in the disease severity of this situation. Accordingly, we believe that the essential messages would not be affected. Finally, the association between the presence of such accessory variants and CAD did not reach statistical significance, probably due to the lack of power. In this regard, the increase of LDL cholesterol by 18 mg/dL caused by LDL-altering accessory variants could be estimated to be associated with an increase in the odds for CAD by 30%.31 And we found that 29% increase by the presence of such accessory variants. The power analysis showed that w2000 individuals were needed to establish an independent association between the presence of such accessory variants and CAD. Nevertheless, our results suggest for the first time that these oligogenic FH with deleterious mutations in FH genes and in ‘‘accessory’’ genes could partly explain the development of severe FH, at least in the presence of high LDL cholesterol levels. To support this notion, recent studies suggest that deleterious variants in ABCG5/8 genes may contribute to hypercholesterolemia much more than previously considered.32,33

Conclusion Oligogenic FH, which partially explains the development of severe FH, was associated with increased LDL

8 cholesterol level as well as increased odds for CAD among patients with significantly elevated LDL cholesterol. Comprehensive genotyping could aid in the identification of extremely high-risk individuals among those with significantly elevated LDL cholesterol levels.

Acknowledgments The authors express their special thanks to Ms. Kazuko Honda, Mr. Sachio Yamamoto, and Ms. Yoko Iwauchi (staff of the Kanazawa University) for their outstanding technical assistance. They thank Advanced Preventive Medical Sciences Research Center, Kanazawa University for the use of facilities. This work has been supported by scientific research grants from the Japan Agency for Medical Research and Development (AMED), Sakakibara Memorial Research Grant from the Japan Research Promotion Society for Cardiovascular Diseases, and Astellas Foundation for Research on Metabolic Disorders. Authors’ contributions: All authors contributed to article preparations and revisions. H.T., M-a. K., A.N., R.T., A.N., A.I., H.M., and M.Y. were involved in the recruitment of patients. H.T., M-a.K., and M.Y. were involved in the conception and design of the study/analyses. H.T. obtained the data and performed statistical analyses. H.T., M-a.K., A.N., A.I., H.M., and M.Y. were involved in data interpretation and article drafting and reviewing. H.T., A.N., R.T., K.H., and A.T. were involved in genetic analyses.

Supplementary data Supplementary data related to this article can be found online at https://doi.org/10.1016/j.jacl.2018.08.006.

References 1. Goldstein JL, Hobbs HH, Brown MS. Familial hypercholesterolemia. In: Scriver CR, Beaudet AL, Sly WS, Valle D, editors. The Metabolic and Molecular Bases of Inherited Disease. New York: McGraw-Hill, 2001. p. 2863.e913. 2. Soutar AK, Naoumova RP. Mechanisms of disease: genetic causes of familial hypercholesterolemia. Nat Clin Pract Cardiovasc Med. 2007; 4:214–225. 3. Talmud PJ, Shah S, Whittall R, et al. Use of low-density lipoprotein cholesterol gene score to distinguish patients with polygenic and monogenic familial hypercholesterolaemia: a case-control study. Lancet. 2013;381:1293–1301. 4. Posey JE, Harel T, Liu P, et al. Resolution of disease phenotypes resulting from multilocus genomic variation. N Engl J Med. 2017;376: 21–31. 5. Blue GM, Kirk EP, Giannoulatou E, et al. Advances in the genetics of congenital heart disease: a clinician’s guide. J Am Coll Cardiol. 2017; 69:859–870. 6. Tada H, Kawashiri MA, Ikewaki K, et al. Altered metabolism of lowdensity lipoprotein and very-low-density lipoprotein remnant in autosomal recessive hypercholesterolemia: results from stable isotope kinetic study in vivo. Circ Cardiovasc Genet. 2012;5:35–41.

Journal of Clinical Lipidology, Vol -, No -, - 2018 7. Perez-Campo FM, De Castro-Oros I, Noriega A, et al. Functional analysis of new 3’ untranslated regions genetic variants in genes associated with genetic hypercholesterolemias. J Clin Lipidol. 2017;11:532–542. 8. Willer CJ, Schmidt EM, Sengupta S, et al. Discovery and refinement of loci associated with lipid levels. Nat Genet. 2013;45:1274–1283. 9. Tada H, Kawashiri MA, Nohara A, Inazu A, Mabuchi H, Yamagishi M. Impact of clinical signs and genetic diagnosis of familial hypercholesterolaemia on the prevalence of coronary artery disease in patients with severe hypercholesterolaemia. Eur Heart J. 2017;38: 1573–1579. 10. Teramoto T, Sasaki J, Ishibashi S, et al. Familial hypercholesterolemia. J Atheroscler Thromb. 2014;21:6–10. 11. DePristo MA, Banks E, Poplin R, et al. A framework for variation discovery and genotyping using next-generation DNA sequencing data. Nat Genet. 2011;43:491–498. 12. McLaren W, Pritchard B, Rios D, Chen Y, Flicek P, Cunningham F. Deriving the consequences of genomic variants with the ensembl API and SNP effect predictor. Bioinformatics. 2010;26:2069–2070. 13. Liu X, Jian X, Boerwinkle E. dbNSFP v2.0: a database of human nonsynonymous SNVs and their functional predictions and annotations. Hum Mutat. 2013;34:E2393–E2402. 14. Lek M, Karczewski KJ, Minikel EV, et al. Analysis of protein-coding genetic variation in 60,706 humans. Nature. 2016;536:285–291. 15. Fromer M, Moran JL, Chambert K, et al. Discovery and statistical genotyping of copy-number variation from whole-exome sequencing depth. Am J Hum Genet. 2012;91:597–607. 16. Fromer M, Purcell SM. Using XHMM software to detect copy number variation in whole-exome sequencing data. Curr Protoc Hum Genet. 2014;24:7–23. 17. Khera AV, Won HH, Peloso GM, et al. Diagnostic yield and clinical utility of sequencing familial hypercholesterolemia genes in patients with severe hypercholesterolemia. J Am Coll Cardiol. 2016;67:2578–2589. 18. Landrum MJ, Lee JM, Riley GR, et al. ClinVar: public archive of relationships among sequence variation and human phenotype. Nucleic Acids Res. 2014;42:D980–D985. 19. Mabuchi H, Nohara A, Noguchi T, et al. Genotypic and phenotypic features in homozygous familial hypercholesterolemia caused by proprotein convertase subtilisin/kexin type 9 (PCSK9) gain-of-function mutation. Atherosclerosis. 2014;236:54–61. 20. Ohta N, Hori M, Takahashi A, et al. Proprotein convertase subtilisin/kexin 9 V4I variant with LDLR mutations modifies the phenotype of familial hypercholesterolemia. J Clin Lipidol. 2016;10:547–555. 21. Richards S, Aziz N, Bale S, et al. Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and genomics and the association for molecular pathology. Genet Med. 2015;17:405–424. 22. Kwon M, Han SM, Kim DI, et al. Evaluation of polygenic cause in Korean patients with familial hypercholesterolemia - A study supported by Korean Society of lipidology and atherosclerosis. Atherosclerosis. 2015;242:8–12. 23. Kim YJ, Go MJ, Hu C, et al. Large-scale genome-wide association studies in East Asians identify new genetic loci influencing metabolic traits. Nat Genet. 2011;43:990–995. 24. Committee of the Japan Diabetes Society on the Diagnostic Criteria of Diabetes Mellitus, Seino Y, Nanjo K, Tajima N, et al. Report of the committee on the classification and diagnostic criteria of diabetes mellitus. J Diabetes Investig. 2010;1:212–228. 25. Yu W, Nohara A, Higashikata T, Lu H, Inazu A, Mabuchi H. Molecular genetic analysis of familial hypercholesterolemia: spectrum and regional difference of LDL receptor gene mutations in Japanese population. Atherosclerosis. 2002;165:335–342. 26. Yamamura T, Dong LM, Yamamoto A. Characterization of apolipoprotein E7 (Glu(244)–.Lys, Glu(245)—.Lys), a mutant apolipoprotein E associated with hyperlipidemia and atherosclerosis. J Lipid Res. 1999;40:253–259. 27. Santos RD, Gidding SS, Hegele RA, et al. Defining severe familial hypercholesterolaemia and the implications for clinical management: a consensus statement from the international atherosclerosis society

Tada et al

Oligogenic FH, LDL-C, and CAD

severe familial hypercholesterolemia panel. Lancet Diabetes Endocrinol. 2016;4:850–861. 28. Natarajan P, Young R, Stitziel NO, et al. Polygenic risk score identifies subgroup with higher burden of atherosclerosis and greater relative benefit from statin therapy in the primary prevention setting. Circulation. 2017;135:2091–2101. 29. Khera AV, Emdin CA, Drake I, et al. Genetic risk, adherence to a healthy lifestyle, and coronary disease. N Engl J Med. 2016;375:2349–2358. 30. Nordestgaard BG, Benn M. Genetic testing for familial hypercholesterolaemia is essential in individuals with high LDL cholesterol: who does it in the world? Eur Heart J. 2017;38:1580–1583.

9 31. Tada H, Kawashiri MA, Yamagishi M. Comprehensive genotyping in dyslipidemia: mendelian dyslipidemias caused by rare variants and mendelian randomization studies using common variants. J Hum Genet. 2017;62:453–458. 32. Lamiquiz-Moneo I, Baila-Rueda L, Bea AM, et al. ABCG5/G8 gene is associated with hypercholesterolemias without mutation in candidate genes and noncholesterol sterols. J Clin Lipidol. 2017;11: 1432–1440. 33. Brinton EA, Hopkins PN, Hegele RA, et al. The association between hypercholesterolemia and sitosterolemia, and report of a sitosterolemia kindred. J Clin Lipidol. 2018;12:152–161.