UvA-DARE (Digital Academic Repository) Familial hypercholesterolemia: the Dutch approach Huijgen, R. Link to publication

Transcription

1 UvA-DARE (Digital Academic Repository) Familial hypercholesterolemia: the Dutch approach Huijgen, R. Link to publication Citation for published version (APA): Huijgen, R. (2012). Familial hypercholesterolemia: the Dutch approach. General rights It is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), other than for strictly personal, individual use, unless the work is under an open content license (like Creative Commons). Disclaimer/Complaints regulations If you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, stating your reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Ask the Library: or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam, The Netherlands. You will be contacted as soon as possible. UvA-DARE is a service provided by the library of the University of Amsterdam ( Download date: 07 Apr 2019

2 Genetic variation in APOB, PCSK9, and ANGPTL3 in carriers of pathogenic autosomal dominant hypercholesterolemic mutations with unexpected low LDL-C levels Roeland Huijgen, Barbara Sjouke, Kelly Vis, Janine S.E. de Randamie, Joep C. Defesche, John J.P. Kastelein, G. Kees Hovingh, Sigrid W. Fouchier Human Mutation 2012;33(2):

3 Abstract Autosomal Dominant Hypercholesterolemia (ADH) is caused by LDLR and APOB mutations. However, genetically diagnosed ADH patients do not always exhibit the expected hypercholesterolemic phenotype. Of 4,669 genetically diagnosed ADH patients, identified through the national identification screening program for ADH, 75 patients (1.6%) had LDL-C levels below the 50 th percentile for age and gender prior to lipid-lowering therapy. The genes encoding APOB, PCSK9 and ANGPTL3 were sequenced in these subjects to address whether monogenic dominant loss-of-function mutations underlie this paradoxical phenotype. APOB mutations, resulting in truncated APOB, were found in 5 (6.7%) probands, reducing LDL-C by 56%. Rare variants in PCSK9, and ANGPTL3 completely correcting the hypercholesterolemic phenotype were not found. The common variants p.n902n, c t>a, p.d2312d, and p.e4181k in APOB, and c a>g in PCSK9 were significantly more prevalent in our cohort compared to the general European population. Interestingly, 40% of our probands carried at least one minor allele for all four common APOB variants compared to 1.5% in the general European population. While we found a low prevalence of rare variants in our cohort, our data suggest that regions in proximity of the analyzed loci, and linked to specific common haplotypes, might harbor additional variants that correct an ADH phenotype. 184

4 Loss of function APOB and PCSK9 mutations in FH patients without phenotype Introduction Cholesterol, a precursor molecule for many bioactive substances and a constituent of the plasma membrane, is essential for the functioning of all eukaryotic organisms. Approximately 70% of the total amount of cholesterol in plasma is transported in the low-density lipoprotein (LDL) fraction. In normal physiology, LDL particles are cleared by the liver via LDL receptor (LDLR; MIM ) mediated endocytosis (LDLR; MIM ). The interaction between the LDL particle and the LDLR is mediated by apolipoprotein B (APOB; MIM ), the apolipoprotein bound to LDL. 1 Mutations in LDLR and APOB have been found to cause autosomal dominant hypercholesterolemia (ADH; MIM #143890), clinically characterized by elevated levels of total (TC) and LDL-cholesterol (LDL-C) levels, resulting in a major risk for the development of premature cardiovascular disease (CVD). A large array of mutations in the LDLR gene, which result in dysfunctional LDLRs, has been found in subjects with familial hypercholesterolemia (FH, MIM #143890). 1 Familial defective apolipoprotein B (FDB, MIM #144010) is caused by specific mutations in APOB resulting in decreased affinity for the LDLR, and consequently increased levels of plasma LDL-C. 2 Proprotein convertase subtilisin kexin 9 (PCSK9, MIM ) plays a crucial role in the post-transcriptional regulation of the LDLR via the epidermal growth factor (EGF) repeat A region. 3, 4 Gain-of-function mutations in PCSK9 are also a rare cause of ADH. 5 The prevalence of heterozygous ADH is about one in 500 individuals in most western countries and the initial diagnosis is based on clinical features. Some of the physical stigmata, such as xanthomas, however, develop later in life, and therefore establishing the diagnosis in younger patients is often difficult. While focusing on these clinical hallmarks, identification of a causative mutation provides the only unequivocal diagnosis. Maximum health benefit can be obtained in ADH patients if identification and treatment are initiated as early as possible, and the World Health Organization supports this statement. 6 An efficient approach to identify subjects with ADH and a consequently high CVD risk is through genetic screening of family members of ADH index cases. Such an identification screening program was instituted in 1994 in the Netherlands and efforts to identify all ADH patients are ongoing ever since. 7 Identification of genetically defined ADH in previously undiagnosed relatives provides the opportunity to initiate cholesterol-lowering treatment as primary prevention to reduce cardiovascular morbidity and mortality. 8,

5 However, genetically diagnosed ADH patients do not always exhibit a hypercholesterolemic phenotype. In fact, 15% of the individuals who carry a mutation identified by the Dutch identification screening program show LDL-C levels below the 75 th percentile for age and gender without lipid lowering therapy. 8 Non-penetrance of an FH or FDB mutation may be the result of co-existing mutations responsible for a hypocholesterolemic phenotype. 10, 11 Loss-of-function mutations in APOB have been reported to result in an autosomal co-dominant form of hypocholesterolemia. 12, 13 Carriers of such APOB mutations are diagnosed with familial hypobetalipoproteinemia (FHBL; MIM #605019) and TC and LDL-C levels are typically below the 5 th percentile for age and gender. Loss-of-function mutations in APOB mainly underlie the formation of prematurely truncated forms of APOB resulting in impaired secretion of very low-density lipoprotein (VLDL), the precursor of LDL. 14 We have previously shown that in FDB patients, who also carry a truncating mutation in APOB, the hypercholesterolemia phenotype is completely abolished. 15 Loss-of-function mutations in PCSK9 have also been described to result in FHBL. 16 Mutations interfering with normal activity of PCSK9 in binding to the LDLR EGF repeat A domain prevent LDLR from degradation, which result in more LDLR available at the hepatocyte cell s surface. 17 Additionally, it has been shown that patients carrying nonsense mutations in ANGPTL3 (MIM *604774) also show a hypocholesterolemic phenotype, due to both decreased production rates of VLDL and increased fractional catabolic rates for LDL. 18 Loss-of-function mutations in PCSK9, ANGPTL3, as well as mutations in EGF repeat A region of the LDLR gene, could theoretically correct the hypercholesterolemic phenotype in FH and FDB patients. Here, we set out to explore the prevalence and effect of mutations in APOB, PCSK9, ANGPTL3, and the EGF repeat A region of LDLR in carriers of pathogenic ADH mutations with unexpected low LDL-C levels. MATERIALS AND METHODS Patients and Probands Selection of the patients of interest was based on data collected by the Dutch identification screening program for ADH between May 2003 and September Recruitment of our patients has been described previously. 7 In short, clinically diagnosed ADH index cases are referred for molecular analysis. Once a mutation is identified, additional family members are approached for molecular and biochemical ascertainment. Lipoprotein levels were measured with the LDX-analyzer. 19 LDL-C 186

6 Loss of function APOB and PCSK9 mutations in FH patients without phenotype was calculated by the Friedewald formula. 20 Age and sex specific percentiles of LDL-C were calculated using the reference values of the Caucasian population. 21 The dataset contained demographic and clinical information, including lipid profiles, medication use and specific FH or FDB mutation carrier status. Carriers of a pathogenic FH or FDB mutation who had LDL-C levels below the 50 th percentile for sex and age prior to lipid lowering therapy were selected as probands. To avoid selection of carriers of mutations that are associated with modest LDL-C elevations, we specifically selected carriers of mutations that exhibited LDL-C levels above the 90 th percentile for age and gender, as we have reported previously. 22 Carriers of the two most prevalent ADH mutations in the Netherlands, i.e. p.r3527q in APOB and p.[n564h;l799_f801del] in LDLR have mean LDL-C levels at the 85 th and 84 th percentile, respectively. 22 Given this mild phenotype, carriers of these mutations were only included in case untreated LDL-C levels were below the 40 th percentile for age and sex. In case several individuals from the same family fulfilled the inclusion criteria, the subject with the most pronounced hypocholesterolemic phenotype was selected for DNA analysis. Upon identification of a putative cholesterol-lowering mutation, relatives of the proband were screened for the specific familial mutation and co-segregation with low LDL-C levels was assessed. Data regarding minor allele frequencies (MAF) were preferentially extracted from the general European 1000G population ( or from data available at Minor allele frequencies of novel common variants were determined in a randomly selected cohort of 200 unrelated healthy individuals from Dutch descent. The p.e2566k in APOB was evaluated in 44 p.r3527q carriers with LDL-C levels above the 95 th percentile for age and gender. Written informed consent was obtained from all individuals participating in this study. 11 DNA analysis Genomic DNA was prepared from 5 ml whole blood on an AutopureLS apparatus according to the protocol provided by the manufacturer (Gentra Systems, Minneapolis, USA). Primer sets were designed to analyze all exons and flanking intronic regions of APOB, PCSK9, and ANGPTL3 and exon 7 of LDLR. Primer sequences and conditions for PCR are available upon request. Sequence analysis was performed by direct sequencing using the Big Dye Terminator ABI Prism Kit, version 1.1 (Applied Biosystems, Foster City, CA, USA). Products of sequence reactions were run on a 187

7 Genetic Analyzer 3730 (Applied Biosystems, Foster City, CA, USA) and sequence data were analyzed by the use of the Sequencer package (GeneCodes Co, Ann Arbor, MI, USA). Nomenclature Mutations were described according to the nomenclature as proposed by den Dunnen and Antonarakis. 23 Thus, for all genes the numbering was based on the cdna with nucleotide c.1 being A of the ATG initiation codon p.1. The following reference sequences were used: APOB:NM_ , LDLR:NM_ , PCSK9: NM_ , and ANGPTL3:NM_ Statistical Analysis All data were analyzed using SPSS software (version 10.1, SPSS, Chicago, MI, USA) by ANOVA and by univariate general linear model analyses with adjustment for age and sex. A p-value below 0.05 was considered statistically significant. Triglyceride levels were log transformed prior to analysis. Differences in allele frequencies were determined with the one-tailed Fisher s exact test. Linkage disequilibrium was calculated using the CubeX program. 24 RESULTS Lipid levels were evaluated in a cohort comprising 4,669 untreated FH and FDB subjects. Seventy-five unrelated probands (1.6%) met our inclusion criteria: 31 with p.r3527q in the APOB gene, 14 with the p.[n564h;l799_f801del] in the LDLR gene and 30 with other pathogenic LDLR mutations. The clinical characteristics of these 75 probands are presented in Table 1. Table 1. Clinical characteristics of the 75 selected genetically diagnosed ADH patients without an ADH phenotype. Age (y) [IQR]* 38 [23-52] Male gender (% ) 45 [58%] Lipid profile (mmol/l) Total cholesterol [SD] 4.36 [0.88] p30 LDL-cholesterol [SD] 2.53 [0.66] p22 HDL-cholesterol [SD] 1.14 [0.35] p34 Triglycerides [IQR] * 1.52 [ ] p62 Values are described as mean levels plus standard deviation [SD] and *in case of skewed distribution values are described as median plus interquartile range [IQR]. P represents percentile for age and sex followed by mean percentile. 188

8 Loss of function APOB and PCSK9 mutations in FH patients without phenotype APOB rare variants A total of five different APOB loss-of-function mutations were found in four FH and one FDB patient (Table 2). The p.q1336x in exon 25, resulting in a truncated APOB-28, has been identified previously in Dutch hypocholesterolemic patients. 12 Three novel truncated variants were all located in exon 26: p.q2157x (APOB-46), c.6541delt (APOB-47), and p.e3545x (APOB-76). One putative functional splice mutation, resulting in a G to C substitution in intron 1 (c.82+1g>c), was predicted by the Alamut software package (V2.0, Interactive Biosoftware, Rouen, France) to delete the donor splice site in intron 1. Screening all probands relatives (N=51) showed co-segregation of all five APOB mutations with the hypocholesterolemic phenotype (data not shown). Characterization based on molecular diagnosis clearly showed that the hypercholesterolemia phenotype, as being caused by pathogenic variants in either LDLR or APOB, was absent once an APOB loss-of-function mutation was present (Table 2). The nine subjects who carried a pathogenic FH or FDB mutation as well as an APOB loss-of-function mutation had LDL-C levels similar to that of family members without any mutation (2.49 ± 0.97 versus 2.72 ± 0.89 mmol/l; p=0.228), and LDL-C levels were 56% lower compared to subjects who carried a FH or FDB mutation only (2.49 ± 0.97 versus 5.67 ± 1.96 mmol/l, p<0.001). Family members only carrying an APOB loss-of-function mutation were characterized by 63% lower LDL-C levels (1.02 ± 0.32 versus 2.72 ± 0.89 mmol/l; p=0.001) compared to those without any mutation. Comparable results were seen for TC, while no significant differences in HDL-C or triglyceride levels were observed. Other rare variants (MAF<5%) identified in our cohort mainly involved variants as frequently present in the general 1000G population (Table 3), with the exception of p.i2716i in exon 26. This synonymous variant was significantly more prevalent in the general population, and thus unlikely to explain the normolipidemic phenotype in its carriers. Additionally, several novel missense and intronic variants were identified with unknown frequency in the general population. Since these missense variants were predicted to be non-functional by the in silico prediction algorithms SIFT 25 and Polyphen-2, 26 and the intronic variants were predicted not to affect mrna splicing by the splice site prediction algorithm Human Splicing Finder, 27 these variants also are unlikely to explain the normolipidemic phenotype in its carriers

9 Table 2. Molecular and clinical characteristics of putative hypocholesterolemic mutation carriers and their relatives. N age TC LDL HDL TG Family ADH mutation FHBL mutation total M F mean SD mean SD mean SD mean SD mean SD APOB1 WT WT (18.4) 4.49** (0.96) 2.80** (1.02) 1.31 (0.31) 0.83** (0.31) LDLR-c.313+1G>C/313+2T>C a c.82+1g>c (17.9) 4.57** (0.93) 2.97** (0.77) 1.31 (0.40) 0.63** (0.19) LDLR-c.313+1G>C/313+2T>C a WT (19.3) 7.83 (1.59) 6.19 (1.94) 1.19 (0.37) 1.80 (0.86) APOB2 WT p.q1336x (2.12) 2.93 (0.10) 1.26 (0.02) 1.44 (0.06) 0.53 (0.02) APOB-p.R3527Q a p.q1336x (2.12) 3.04 (0.37) 1.77 (0.52) 0.92 (0.05) 0.77 (0.21) APOB3 WT p.q2157x WT WT (26.3) 4.51 (1.04) 2.56 (0.91) 1.52 (0.25) 0.94* (0.25) LDLR-p.N564H;L799_F801del a p.q2157x LDLR-p.N564H;L799_F801del a WT (10.4) 5.63 (2.08) 3.80 (2.15) 1.46 (0.15) 0.81* (0.19) APOB4 WT WT (12.0) 4.74** (0.74) 2.87** (0.63) 1.23 (0.38) 1.41 (0.63) LDLR-p.V429M a c.6541delt (0.0) 5.78 (0.68) 3.53** (0.10) 1.21 (0.37) 2.29** (2.50) LDLR-p.V429M a WT (17.8) 7.73 (1.64) 6.17 (1.48) 1.28 (0.28) 0.62 (0.19) APOB5 WT p.e3545x (2.12) 3.25 (0.18) 1.05 (0.03) 1.29 (0.24) 1.99 (1.00) WT WT LDLR-p.W44X p.e3545x APOB all WT YES (11.4) 2.99** (0.29) 1.02** (0.32) 1.33 (0.16) 1.39 (0.94) WT WT (20.3) 4.55** (0.89) 2.72** (0.89) 1.33 (0.32) 1.08 (0.58) YES YES (18.5) 4.15** (1.31) 2.49** (0.97) 1.10 (0.35) 1.22 (1.14) YES WT (16.3) 7.33 (1.82) 5.67 (1.96) 1.28 (0.30) 1.17 (0.79) PCSK9 WT WT (26.2) 5.00** (1.08) 2.89** (0.96) 1.45 (0.34) 1.44 (0.49) LDLR-c.313+1G>C/313+2T>C a p.r46l ** 2.95** LDLR-c.313+1G>C/313+2T>C a WT (19.8) 7.69 (1.51) 6.11 (1.76) 1.02 (1.91) 1.23 (0.97) ANGPTL3 WT p.l127f (14.4) 4.67** (0.93) 2.67** (0.78) 1.19 (0.30) 1.75 (0.93) WT WT (19.2) 4.11** (0.86) 2.43** (0.84) 1.16 (0.34) 1.16 (0.59) YES p.l127f (18.4) 4.64** (0.79) 2.30** (0.81) 1.52** (0.39) 1.79 (0.61) YES WT (15.6) 5.55 (1.14) 3.69 (1.22) 1.06 (0.25) 1.73 (1.19) *p<0.05 versus carriers of FHBL mutation solely, **p<0.05 versus carriers of ADH mutation solely. TC: total cholesterol; LDL: LDL-cholesterol; HDL: HDL-cholesterol; TG: triglycerides; ADH: autosomal dominant hypercholesterolemia; M: male; F: female; SD: standard deviation; NA: not applicable. a All FH and FDB mutations have been published before [Fouchier et al., 2001; Fouchier et al., 2005a] 190

10 Loss of function APOB and PCSK9 mutations in FH patients without phenotype APOB common variants The p.e2566k variant, frequently seen in our cohort, was only found in individuals also carrying p.r3527q in APOB, implying that p.e2566k is tightly linked to p.r3527q (D =1, r 2 =0.9112), rather than causing a hypocholesterolemic phenotype. Indeed, the analysis of 44 p.r3527q carriers with LDL-C levels above the 95 th percentile for age and gender, revealed 43 carriers of p.e2566k. Four common variants, p.n902n, c t>a, p.d2312d and p.e4181k, were each significantly more present in our cohort compared to the general European population (Table 3). Interestingly, 30 (40%) of the 75 probands carried at least one minor allele for all four variants, compared to 4 (1.5%) of 283 individuals in the European cohort genotyped for the 1000G project. PCSK9 rare mutations We identified two missense mutations in PCSK9. One was the most common lossof-function mutation p.r46l in exon 1. Eight relatives of the p.r46l proband were available for analysis (Table 2). The carrier of both, the FH mutation p.ldlrc.313+1g>c/313+2t>c and the p.r46l variant in PCSK9, showed 52% lower mean LDL-C levels compared to relatives only carrying p.ldlr-c.313+1g>c/313+2t>c (2.95 vs ± 1.76 mmol/l, p=0.013). The plasma LDL-C concentration in this individual did not differ from family members not carrying both mutations (2.95 vs ± 0.96 mmol/l; p=0.393). The second missense mutation was a potentially novel loss-of-function mutation, p.g394s in exon 7, involving a G to A substitution changing the highly conserved amino acid Gly into Ser at position 394. Since this variant was introduced by a married-in, co-segregation analysis to evaluate functionality was not possible. The remaining novel rare variants, two synonymous and two intronic variants were predicted not to affect mrna splicing by the splice site prediction algorithm Human Splicing Finder PCSK9 common variants One common variant, c a>g, in intron 11 was significantly more present in our cohort compared to the general European 1000G population (MAF:0.461 vs , p<0.0001, Table 3). 191

11 Table 3. Genetic variants identified in APOB, PCSK9 and ANGPTL3. gene exon cnomen pnomen Splice effect SIFT PolyPhen-2 rs number MAF cohort MAF p-value 1000G a APOB 5 -UTR c.-644c>t p.= NO rs UTR c.-430c>a p.= NO rs UTR c.-393t>c p.= NO rs UTR c.-115c>g p.? YES rs exon 1 c.34_42del p.l12_l14del NO rs intron 1 c.82+1g>c p.? YES intron 2 c t>a p.? NO rs exon 4 c.293c>t p.t98i NO TOLERATED benign rs intron 8 c g>a p.? NO intron c t>a p.? NO rs intron 12 c t>c p.? NO intron 12 c c>t p.? NO exon 14 c.1853c>t p.a618v NO DELETERIOUS probably damaging rs exon 15 c.2188g>a p.v730i NO DELETERIOUS probably damaging rs intron 17 c g>c p.? NO rs exon 18 c.2706c>t p.n902n NO rs exon 19 c.2841c>t p.t947t NO rs exon 19 c.2981c>t p.p994l NO DELETERIOUS probably damaging rs intron 20 c g>a p.? NO rs intron 20 c t>c p.? NO rs intron 21 c a>g p.? NO intron 23 c c>t p.? NO rs intron 24 c t>a p.? NO rs intron 24 c g>a p.? NO rs exon 25 c.4006c>t p.q1336x NO exon 26 c.4838g>c p.s1613t NO TOLERATED benign rs b exon 26 c.5768a>g p.h1923r NO DELETERIOUS probably damaging rs exon 26 c.6469c>t p.q2157x NO exon 26 c.6543del p.f2181lfsx14 NO exon 26 c.6936c>t p.d2312d NO rs exon 26 c.7545c>t p.t2515t NO rs exon 26 c.7696g>a p.e2566k NO DELETERIOUS benign rs exon 26 c.8148c>t p.i2716i NO rs exon 26 c.8216c>t p.p2739l NO DELETERIOUS probably damaging rs exon 26 c.10131g>a p.l3377l NO rs exon 26 c.10633g>t p.e3545x NO exon 26 c.10913g>a p.r3638q NO TOLERATED benign rs exon 26 c.11503a>c p.i3835l NO TOLERATED benign intron 27 c g>c p.? NO rs exon 29 c.12541g>a p.e4181k NO TOLERATED benign rs exon 29 c.12809g>c p.r4270t NO TOLERATED benign rs exon 29 c.12940a>g p.i4314v NO TOLERATED benign rs exon 29 c.13013g>a p.s4338n NO TOLERATED benign rs exon 29 c.13441g>a p.a4481t NO TOLERATED benign rs exon 29 c.13444a>g p.i4482v NO TOLERATED benign

12 Loss of function APOB and PCSK9 mutations in FH patients without phenotype PCSK9 5 -UTR c.-64c>t p.? NO rs exon 1 c.61_63dup p.l23dup NO rs b exon 1 c.137g>t p.r46l NO TOLERATED probably damaging rs exon 1 c.141c>t p.s47s NO rs exon 1 c.158c>t p.a53v NO TOLERATED benign rs intron 1 c g>a p.? NO rs intron 2 c t>c p.? NO rs intron 3 c g>a p.? NO rs intron 4 c.657+9g>a p.? NO rs intron 4 c c>a p.? NO rs intron 4 c g>a p.? NO rs intron 4 c c>a p.? NO rs intron 4 c dup p.? NO rs intron 4 c g>a p.? NO rs intron 4 c.658-7c>t p.? NO rs exon 5 c.720c>t p.g240g NO rs exon 5 c.753c>t p.r251r NO rs intron 5 c.799+3a>g p.? NO rs intron 6 c a>g p.? NO rs exon 8 c.1180g>a p.g394s NO DELETERIOUS probably damaging intron 8 c t>c p.? NO rs exon 9 c.1380g>a p.v460v NO rs exon 9 c.1420a>g p.v474i NO TOLERATED benign rs exon 9 c.1491c>t p.g497g NO exon 9 c.1503g>a p.e501e NO intron 9 c g>a p.? NO intron 10 c c>t p.? NO rs intron 10 c g>a p.? NO rs intron 11 c c>t p.? NO intron 11 c a>g p.? NO rs exon 12 c.2009a>g p.g670e NO TOLERATED benign rs UTR c.*75c>t p.? NO rs ANGPTL3 exon 1 c.379c>t p.l127f NO DELETERIOUS benign rs intron 2 c _607-46del p.? YES rs b UTR c.*50_*59del p.? NO rs c MAF: minimal allele frequency, a Allele counts in individuals from European descent extracted from b Allele count extracted from c Analyzed in 200 unrelated healthy individuals

13 ANGPTL3 rare variants The analysis of ANGPTL3 revealed one missense variant, p.l127f, in exon 1 (Table 3). Albeit not statistically significant, this variant was more prevalent amongst our cases compared to the frequency as being described in the European 1000G population (MAF:0.023 vs , p=0.052, Table 2). Two genetically diagnosed FDB patients with p.r3527q and two genetically diagnosed FH patients with p.a705p, carrying the p.l127f in ANGPTL3, showed normal TC and LDL-C levels, which was significantly lower than FH and FDB relatives without the ANGPTL3 mutation (TC:4.64 versus 5.55 mmol/l, p=0.015 and LDL-C:2.30 versus 3.69 mmol/l, p=0.043). On the other hand, three p.l127f carriers not carrying the familial FH or FDB mutation and one molecularly diagnosed FDB patient did not have lower than expected TC or LDL-C levels (data not shown). ANGPTL3 common variants One common variant, c.*50_59del, in the 3 -UTR of ANGPTL3 was found with unknown frequency in the general population (Table 3). Genotyping of a cohort comprising 200 unrelated healthy individuals from Dutch descent revealed similar allele frequencies. This finding suggests that this variant is unlikely to be associated with the normolipidemic phenotype in our probands. LDLR Mutations in the EGF A repeat region of the LDLR, encoded by exon 7, hypothetically preventing the degradation of the LDLR by PCSK9 were not detected in our probands. DISCUSSION In the current study we describe the prevalence of functional monogenic co-dominant inheritance of hypocholesterolemic mutations in APOB, PCSK9, ANGPTL3 and exon 7 of LDLR, in genetically diagnosed ADH patients with paradoxically low plasma LDL-C levels. Only a small proportion, 75 (1.6%) patients of a total of 4,669 FH and FDB patients met our criteria, clearly showing the rarity of this paradoxically phenotype. In five (6.7%) of the 75 cases a loss-of-function mutation in APOB was found. The reported prevalence of loss-of-function mutations in APOB, as determined in normocholesterolemic and hypocholesterolemic individuals show a broad range between 1.9% and 52%. 12, 28 With regard to the selected FH and FDB subjects in our study, it has been demonstrated that 4 of 23 unrelated subjects (17.3%) with 194

14 Loss of function APOB and PCSK9 mutations in FH patients without phenotype a molecular diagnosis of FH or FDB but unusual low TC levels, had a mutation in APOB that lead to reduced LDL-C levels. 29 Thus, the prevalence of APOB loss-offunction mutations in our cohort does not significantly differ with the findings by others. The prevalence of loss-of-function mutations in PCSK9 ranges from 3.2% 16 to 6.5% 29 in normocholesterolemic and hypocholesterolemic individuals, respectively. PCSK9 loss-of-function mutations in FH and FDB patients correcting the ADH phenotype, however, have yet not been described. 29 We did identify two candidate variants, p.r46l and p.g394s, but it remains questionable which variants are fully responsible for the corrected ADH phenotype. The hypocholesterolemic effect of p.r46l in PCSK9 has been intensively studied. In the general population this variant has a frequency of 1.3 to 3.2% and yields significant 11 to 16% lower LDL-C levels, 16, 30 and as such might be considered a minor determinant of LDL-C levels. Despite these modest reductions in LDL-C levels, however, the minor L46 allele is associated with a pronounced reduction of CVD risk. 31 One study assessed the consequences of the p.r46l in FH patients. 32 The analysis of 1130 unrelated patients affected by functional LDLR mutations revealed 30 (2.7%) individuals carrying p.r46l, which is similar to the frequencies in the general population. Total cholesterol levels were 6% lower in the 30 p.r46l carriers compared to non-carriers, however, this difference did not reach statistical significance. The p.r46l carrier identified in our study exhibited a much more pronounced reduction in TC and LDL-C levels. This effect, however, was only seen in our proband and given the fact that no additional carriers of both mutations were found, no firm conclusions can be drawn. The large impact of the p.r46l mutation could be due to ascertainment bias. The finding of six p.r46l mutation carriers among 644 clinically diagnosed ADH patients, who were negative for LDLR and APOB ADH mutations, is in line with this (unpublished data). The hypocholesterolemic potential of the novel missense variant, p.g394s could not be assessed. However, this variant has also been identified in a hypercholesterolemic patient during routine diagnostics for ADH (unpublished data), which suggests that this variant is unlikely to explain the normolipidemic phenotype in the proband. The role of loss-of-function mutations in PCSK9 potentially correcting the ADH phenotype in genetically diagnosed ADH patients remains unclear. Since cohorts with these patients are small, we can only speculate on whether heterozygous lossof-function mutations in PCSK9 alone are not sufficient enough to induce a major reduction in LDL-C or that the prevalence of PCSK9 mutations is rare in general

15 Recently, it became evident that loss-of-function mutations in ANGPTL3 cause a hypocholesterolemic phenotype. 18 The analysis of ANGPTL3 in our cohort revealed only one missense variant, p.l127f, in exon 1. Although this variant showed a trend toward a significant higher prevalence in our cohort, several carriers of p.l127f without the familial ADH mutation had normal cholesterol levels, while one FH patient showed the expected hypercholesterolemic phenotype. This suggests that p.l127f is unlikely to cause a monogenic co-dominant hypocholesterolemic phenotype. In an attempt to explain the corrected ADH phenotype in our cohort we mainly focused on rare variants located in commonly accepted important regions of a gene; i.e. the promoter region, coding sequence and intronic boundaries. Interestingly, we found several common variants in APOB and PCSK9 that were significantly overrepresented in our cohort compared to the general European 1000G population. 33 Given the frequency and position of these synonymous and intronic variants it is unlikely that these variants cause the normolipidemic phenotype, but this finding does suggest that important variants located outside the analyzed regions are in linkage disequilibrium with these common variants. Several studies have shown that specific regions in intron two and three of APOB interact with the basal transcriptional machinery of this gene. 34, 35 Additionally, it has been suggested that a negative regulatory region is located downstream of the APOB gene. 36 Additional studies are warranted to address the role of these non-transcribed regions in abolishing the hypercholesterolemic phenotype in genetically diagnosed ADH patients. In 12 of the 75 probands none of the overrepresented common variants were present, and these subjects might carry mutations in genes that hitherto have not been described to have a role in LDL-C metabolism. A recent large-scale GWAS revealed that at least 95 loci across the human genome harbor common variants associated with serum lipid traits, of which 22 were associated with LDL-C. 37 Screening of ANGPTL3, one of the 22 LDL-C associated genes, resulted in the identification of rare variants responsible for familial combined hypolipidemia, inherited in a monogenic co-dominant manner. 18 It therefore seems rational to search for mutations in these candidate genes with a putative effect on LDL-C. In addition one could speculate that epi- and nutrigenomic aspects contribute to variable expression in ADH mutation 38, 39 carriers. Recently, several studies have addressed the role of a gene dose score (i.e. the number of common variants with additive effects) on variations in LDL-C levels. One might argue whether carriership of a number of common variants could explain the absence of the ADH phenotype in our cohort. The relatively small effect on LDL-C 196

16 Loss of function APOB and PCSK9 mutations in FH patients without phenotype variance (~12.2%), however, argues against this assumption. 18 Recently, fine mapping of 5 loci also identified in GWAS, harboring APOB, APOC1, APOC2, APOE, PCSK9, LDLR and SORT1, showed that the addition of less common variants resulted in a 2-fold increase of the proportion of heritability 40 then this would not explain the complete absence of the ADH phenotype in our strictly selected cohort. However, these individuals represent only 10% of all individuals with unexpectedly low LDL-C levels. It is tempting to speculate that the common and less common variants could explain a less severe unexpectedly low LDL-C (50 th -75 th percentile) levels in carriers of an ADH mutation. Carriers of hypocholesterolemic mutations completely abolishing the ADH phenotype might be regarded as a natural life-long exposure model for lipidlowering therapy. Genes in which such mutations have found are considered targets for therapeutic interventions to lower LDL-C and CAD risk. Indeed, inhibition of hepatic APOB synthesis is in an advanced stage of clinical development. A 6 week treatment with antisense inhibition of APOB resulted in a LDL-C reduction 36% in heterozygous FH patients. 41 Similar results might be expected from PCSK9 targeted therapy. Liver-specific sirna silencing of PCSK9 in mice and rats reduced PCSK9 mrna levels by 50-70% and cholesterol levels by 60%. 42 An important clinical question is whether genetically diagnosed FH or FDB patients without the ADH phenotype should be treated with cholesterol-lowering therapy. Recently, the risk for CVD in ADH patients was shown to be attributable to the LDL-C levels and not to the presence of a mutation per se, 43 emphasizing the importance to focus on the clinical phenotype of ADH. Mutations causing ADH and those compensating for the FH or FDB phenotype are not necessarily linked and therefore transmission to the offspring is independent. This implies that offspring of carriers of both rare entities should be analyzed for ADH causing mutations, since absence of a compensating mutation will invariably result in high LDL-C levels and high CVD risk

17 References 1. Goldstein JL, Hobbs HH, Brown MS. In: The metabolic and molecular bases of inherited disease. 8 th ed. New York: McGraw-Hill; 2001: Innerarity TL, Weisgraber KH, Arnold KS et al. Familial defective apolipoprotein B-100: low density lipoproteins with abnormal receptor binding. Proc Natl Acad Sci U S A 1987;84(19): Park SW, Moon YA, Horton JD. Post-transcriptional regulation of low density lipoprotein receptor protein by proprotein convertase subtilisin/kexin type 9a in mouse liver. J Biol Chem 2004;279(48): Zhang DW, Lagace TA, Garuti R et al. Binding of proprotein convertase subtilisin/kexin type 9 to epidermal growth factor-like repeat A of low density lipoprotein receptor decreases receptor recycling and increases degradation. J Biol Chem 2007;282(25): Abifadel M, Varret M, Rabes JP et al. Mutations in PCSK9 cause autosomal dominant hypercholesterolemia. Nat Genet 2003;34(2): World Health Organisation, human genetics programme: FH: Report of a WHO consultation, WHO/HGN/FH/CONS/98.7. Geneva; Umans-Eckenhausen MA, Defesche JC, Sijbrands EJ, Scheerder RL, Kastelein JJ. Review of first 5 years of screening for familial hypercholesterolaemia in the Netherlands. Lancet 2001;357(9251): Huijgen R, Kindt I, Verhoeven SB et al. Two years after molecular diagnosis of familial hypercholesterolemia: majority on cholesterol-lowering treatment but a minority reaches treatment goal. PLoS ONE 2010;5(2):e Versmissen J, Oosterveer DM, Yazdanpanah M et al. Efficacy of statins in familial hypercholesterolaemia: a long term cohort study. BMJ 2008;337:a Emi M, Hegele RM, Hopkins PN et al. Effects of three genetic loci in a pedigree with multiple lipoprotein phenotypes. Arterioscler Thromb 1991;11(5): Sass C, Giroux LM, Ma Y et al. Evidence for a cholesterol-lowering gene in a French-Canadian kindred with familial hypercholesterolemia. Hum Genet 1995;96(1): Fouchier SW, Sankatsing RR, Peter J et al. High frequency of APOB gene mutations causing familial hypobetalipoproteinaemia in patients of Dutch and Spanish descent. J Med Genet 2005;42(4):e Tarugi P, Averna M, Di LE et al. Molecular diagnosis of hypobetalipoproteinemia: an ENID review. Atherosclerosis 2007;195(2):e19-e Aguilar-Salinas CA, Barrett PH, Parhofer KG et al. Apoprotein B-100 production is decreased in subjects heterozygous for truncations of apoprotein B. Arterioscler Thromb Vasc Biol 1995;15(1): van der Graaf A, Fouchier SW, Vissers MN et al. Familial defective apolipoprotein B and familial hypobetalipoproteinemia in one family: two neutralizing mutations. Ann Intern Med 2008;148(9): Cohen JC, Boerwinkle E, Mosley TH, Jr., Hobbs HH. Sequence variations in PCSK9, low LDL, and protection against coronary heart disease. N Engl J Med 2006;354(12):

18 Loss of function APOB and PCSK9 mutations in FH patients without phenotype 17. Lagace TA, Curtis DE, Garuti R et al. Secreted PCSK9 decreases the number of LDL receptors in hepatocytes and in livers of parabiotic mice. J Clin Invest 2006;116(11): Musunuru K, Pirruccello JP, Do R et al. Exome sequencing, ANGPTL3 mutations, and familial combined hypolipidemia. N Engl J Med 2010;363(23): Cobbaert C, Boerma GJ, Lindemans J. Evaluation of the Cholestech L.D.X. desktop analyser for cholesterol, HDL-cholesterol, and triacylglycerols in heparinized venous blood. Eur J Clin Chem Clin Biochem 1994;32(5): Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem 1972;18(6): Gotto AM, Jr., Bierman EL, Connor WE et al. Recommendations for treatment of hyperlipidemia in adults. A joint statement of the Nutrition Committee and the Council on Arteriocslerosis. Circulation 1984;69(5):1065A-1090A. 22. Huijgen R, Kindt I, Fouchier SW et al. Functionality of sequence variants in the genes coding for the low-density lipoprotein receptor and apolipoprotein B in individuals with inherited hypercholesterolemia. Hum Mutat 2010;31(6): den Dunnen JT, Antonarakis SE. Mutation nomenclature extensions and suggestions to describe complex mutations: a discussion. Hum Mutat 2000;15(1): Gaunt TR, Rodriguez S, Day IN. Cubic exact solutions for the estimation of pairwise haplotype frequencies: implications for linkage disequilibrium analyses and a web tool CubeX. BMC Bioinformatics 2007;8: Kumar P, Henikoff S, Ng PC. Predicting the effects of coding non-synonymous variants on protein function using the SIFT algorithm. Nat Protoc 2009;4(7): Adzhubei IA, Schmidt S, Peshkin L et al. A method and server for predicting damaging missense mutations. Nat Methods 2010;7(4): Desmet FO, Hamroun D, Lalande M, Collod-Beroud G, Claustres M, Beroud C. Human Splicing Finder: an online bioinformatics tool to predict splicing signals. Nucleic Acids Res 2009;37(9):e Welty FK, Lahoz C, Tucker KL, Ordovas JM, Wilson PW, Schaefer EJ. Frequency of ApoB and ApoE gene mutations as causes of hypobetalipoproteinemia in the framingham offspring population. Arterioscler Thromb Vasc Biol 1998;18(11): Leren TP, Berge KE. Identification of mutations in the apolipoprotein B-100 gene and in the PCSK9 gene as the cause of hypocholesterolemia. Clin Chim Acta 2008;397(1-2): Benn M, Nordestgaard BG, Grande P, Schnohr P, Tybjaerg-Hansen A. PCSK9 R46L, low-density lipoprotein cholesterol levels, and risk of ischemic heart disease: 3 independent studies and metaanalyses. J Am Coll Cardiol 2010;55(25): Kathiresan S. A PCSK9 missense variant associated with a reduced risk of early-onset myocardial infarction. N Engl J Med 2008;358(21): Strom TB, Holla OL, Cameron J, Berge KE, Leren TP. Loss-of-function mutation R46L in the PCSK9 gene has little impact on the levels of total serum cholesterol in familial hypercholesterolemia heterozygotes. Clin Chim Acta 2010;411(3-4): Mills RE, Walter K, Stewart C et al. Mapping copy number variation by population-scale genome sequencing. Nature 2011;470(7332):

19 34. Brooks AR, Blackhart BD, Haubold K, Levy-Wilson B. Characterization of tissue-specific enhancer elements in the second intron of the human apolipoprotein B gene. J Biol Chem 1991;266(12): Levy-Wilson B, Paulweber B, Nagy BP, Ludwig EH, Brooks AR. Nuclease-hypersensitive sites define a region with enhancer activity in the third intron of the human apolipoprotein B gene. J Biol Chem 1992;267(26): Paulweber B, Brooks AR, Nagy BP, Levy-Wilson B. Identification of a negative regulatory region 5 of the human apolipoprotein B promoter. J Biol Chem 1991;266(32): Teslovich TM, Musunuru K, Smith AV et al. Biological, clinical and population relevance of 95 loci for blood lipids. Nature 2010;466(7307): Ordovas JM. Genetic interactions with diet influence the risk of cardiovascular disease. Am J Clin Nutr 2006;83(2):443S-446S. 39. Shirodkar AV, Marsden PA. Epigenetics in cardiovascular disease. Curr Opin Cardiol 2011;26(3): Sanna S, Li B, Mulas A et al. Fine mapping of five Loci associated with low-density lipoprotein cholesterol detects variants that double the explained heritability. PLoS Genet 2011;7(7):e Akdim F, Visser ME, Tribble DL et al. Effect of mipomersen, an apolipoprotein B synthesis inhibitor, on low-density lipoprotein cholesterol in patients with familial hypercholesterolemia. Am J Cardiol 2010;105(10): Frank-Kamenetsky M, Grefhorst A, Anderson NN et al. Therapeutic RNAi targeting PCSK9 acutely lowers plasma cholesterol in rodents and LDL cholesterol in nonhuman primates. Proc Natl Acad Sci U S A 2008;105(33): Huijgen R, Vissers MN, Kindt I et al. Assessment of carotid atherosclerosis in normocholesterolemic individuals with proven mutations in the low-density lipoprotein receptor or apolipoprotein B genes. Circ Cardiovasc Genet 2011;4(4): Fouchier SW, Defesche JC, Umans-Eckenhausen MW, Kastelein JP. The molecular basis of familial hypercholesterolemia in The Netherlands. Hum Genet 2001;109(6): Fouchier SW, Kastelein JJ, Defesche JC. Update of the molecular basis of familial hypercholesterolemia in The Netherlands. Hum Mutat 2005;26(6):