Autosomal Dominant Hypercholesterolemia: Needs for Early Diagnosis and Cascade Screening in the Tunisian Population

Autosomal dominant hypercholesterolemia (ADH) is characterized by an isolated elevation of plasmatic low-density lipoprotein (LDL), which predisposes to premature coronary artery disease (CAD) and early death. ADH is largely due to mutations in the low-density lipoprotein receptor gene (LDLR), the apolipoprotein B-100 gene (APOB), or the proprotein convertase subtilisin/kexin type 9 (PCSK9). Early diagnosis and initiation of treatment can modify the disease progression and its outcomes. Therefore, cascade screening protocol with a combination of plasmatic lipid measurements and DNA testing is used to identify relatives of index cases with a clinical diagnosis of ADH. In Tunisia, an attenuated phenotypic expression of ADH was previously reported, indicating that the establishment of a special screening protocol is necessary for this population.


INTRODUCTION
Autosomal dominant hypercholesterolemia (ADH) was firstly reported by Carl Müller in 1938 with the study of families presenting tendon xanthomas and heart disease due to a hypercholesterolaemia dominantly inherited [1]. In the 1970's, several studies in patients and cultured cells revealed a defect in the receptor for the low-density lipoprotein and culminated in the Nobel Prize for Brown and Goldstein for their work on the regulation of the cholesterol metabolism [2,3]. ADH (OMIM # 143890) is one of the most frequent inherited disorders in humans, with a frequency of 1 in 500 for heterozygous in western populations [4]. In some populations, the frequency of heterozygous ADH is considerably higher because of founder effect.
A founder effect occurs when a subpopulation is formed through the immigration of a small number of "founder" subjects, followed by population expansion. If some of the founders had ADH, then genetic drift could lead to a high proportion of affected subjects who share specific mutations introduced by founders. Such founder effect was noted in French Canadians [5], South African Afrikaners [6], Jews [7], Indians [8], Christian Lebanese [9], Finns [10] and Tunisians [11].
In Tunisia, the frequency was estimated at 1/165 for heterozygous, and beside the founder effect, the high birth rate and consanguinity marriage influenced this frequency [11]. gous patients commonly have tendon xanthomas before the age of 10 years, and, if untreated, they develop severe atherosclerosis and CHD within their third decade [19]. In heterozygous patients tendon xanthomas are common after 25-30 years old, and the onset of CHD is mostly before 55 years old [20].
Diagnosis of ADH is mainly based on lipid levels, clinical signs, family history of dyslipidemia and/or premature CHD, and will be confirmed by genetic analysis. Three different diagnosis criteria were developed for ADH by the USMedPed (US Make early diagnosis Prevent early death) Program [18] (Table 1), the Simon Broome Register Group in the United Kingdom [21,22] (Table 2), and the Dutch lipid Clinic Network [23] (Table 3).
Actually, in Tunisia, the Simon Broome Register criteria for ADH are mostly used to determine potential patient of ADH. Particularly the cutoff LDL-cholesterol of 4.9 mmol/L (190 mg/dL) is commonly used to determine heterozygote ADH patients.

CLINICAL AND BIOLOGICAL ASPECTS OF ADH IN TUNISIA AND CASCADE SCREENING
Studies on ADH in Tunisia started in 1993 with the work of NM Slimane and coworkers. They estimated a high frequency of this disease for heterozygous (about 1/165). Beside they noted an attenuated phenotypic expression of ADH [11,24].
Indeed, the analysis of 91 ADH patients showed that the prevalence of CHD in Tunisian ADH heterozygous after 30 years old was 23.5% for men and 29.4% for women. All of them went through life without developing any tendon xanthomas (except one female aged 62). The mean total cholesterol level for heterozygous was 7.04 ± 1.40 mmol/L and was higher than the one reported in China (6.1 ± 1.2 mmol/L) [25], but lower than in Japan (8.8±2.0 mmol/L) [26], in UK (9.8± 1.7 mmol/L) [27], in Afrikaners (10.8±1.8 mmol/L) [28], or in Italy (8.49±1.66 mmol/L) [29]. The same observation was made concerning LDL-cholesterol levels.
Concerning homozygous patients, xanthomas were present for all of them, CHD was present for 10% of them before 9 years old, for 71% between 10 and 19 years old. and for 100% above 20 years old. Therefore, CHD in Tunisian ADH homozygous appears to have a later onset than in other homozygous populations. Indeed, CHD occurs for 50% of the Afrikaners ADH homozygous patients before 9 years old. [6] and for 25% in Japan before 10 years old. [26]. Their mean life expectancy was 13 years old. compared with 17 years old in Japan (26) and 21 years old in Lebanon [29]. The mean total cholesterol level for homozygotes reported was 17.52±3.12 mmol/L [24], similar to those reported in other populations.
A recent study in Tunisia showed that 24% (9 out 38) of the ADH patients carrying an heterozygous mutation in the LDLR gene have a LDL-cholesterol level under the 60 th percentile of an age-and gender-matched reference population [30]. This discrepancy between the clinico/biological and molecular phenotype observed reveals the existence of factors that decrease the severity of the disease. In a previous study, we identified one of these factors as the traditional Tunisian diet which is enriched in polyunsaturated fats [11]. This type of diet has been shown to have long-term beneficial therapeutic effects by reducing the incidence of recurrent cardiovascular events. To conclude, 15 years after the first study in 1993 [11] similar characteristics of a mild phenotype of ADH in Tunisia was reported, particularly for heterozygous patients [24]. Thus, it appears clearly that despite the change in the diet habit to a more western diet, the Tunisian population still has the same mild clinical expression of ADH. According to these characteristics of the Tunisian population, the establishment of specific cutoff point seems to be necessary.

MOLECULAR DEFECTS
The known genetic bases of the ADH phenotype are mutations in the LDLR, APOB, or PCSK9 genes.

In the LDL Receptor Gene (LDLR)
The discovery of the LDL receptor and its defective function led to a great advance in the understanding of the pathophysiology of familial hypercholesterolemia (FH).
The LDL receptor is produced in the endoplasmic reticulum (ER) where the 21 amino acid signal peptide is cleaved and the protein glycosylated to give rise to a mature receptor [31]. The 160kDa transmembrane receptor (glycoprotein of 839 amino acids) is present at the surface of most cell types and mediates endocytosis thus playing a pivotal role in cholesterol homeostasis [31].
More than 1741 allelic variants have been identified in the LDLR gene and are distributed as presented in Fig. (1). All gene variants for LDLR are compiled online at two web sites: http://www.ucl.ac.uk/fh/ and http:// http://www.umd.be /LDLR/.
Functional LDLR mutations have been classified into five classes based on biosynthetic and functional studies of fibroblast cell [19,32]. Class 1 mutations are due to disruption of the promoter sequence, nonsense, frameshift or splicing mutations, all resulting in an absence of protein synthesis (null alleles). Class 2 mutations, that primarily occur in the ligand binding and epidermal growth factor precursor domains, disrupt transport of the LDL receptor from the endoplasmic reticulum to the Golgi apparatus. Class 3 mutations interfere with cell surface binding of the receptor to LDL, and these mutations are also primarily found in the ligand-binding and epidermal growth factor precursor domains. Class 4 mutations appear in the cytoplasmic and membrane-spanning domains. They inhibit the clustering of the LDL receptor at the cell surface and the LDL internalization. Class 5 mutations disrupt the recycling of the LDL receptor to the cell surface [19,32].
The first few defects in LDLR gene to be characterized were large deletions identified by southern blotting [33,34]. Once amplification by PCR and direct automated sequencing of PCR products became possible the number of point mutation and minor deletions/insertions has greatly increased. The expanding use of multiplex ligation dependent probe amplification (MLPA), contributed to the evaluation of the exact contribution of major rearrangements. Recent sequencing into further intronic sequences has allowed identification of a large population of splice site mutations [35].      Fig. (1). Distribution of molecular defects reported in the LDLR gene.

In the Apolipoprotein B 100 Gene (APOB)
The interaction between LDL and its receptor is fundamental for the regulation of plasma cholesterol in humans [31]. The only protein component of LDL is ApoB-100, which is the major ligand for the LDL receptor [36].
ApoB-100, a large protein of 550 kDa, is encoded on chromosome 2 and has 26 exons. The binding region is rich in positively charged amino acids and interacts with the binding domains of the LDL receptor [20]. The domain of apoB-100 that interacts with the LDL receptor has been defined using several approaches. The proposed model of this binding region, comprising two clusters [A (3147-3157) and B (3359-3367)] of basic amino acids that are linked through a disulfide bond between residues 3167 and 3297 [37,38] has been further expanded though the discovery of the ADH causative mutation at residue 3500 [39,40]. This has led to the general view that residues 3130-3630 are important for the binding of apoB-100 to the LDL receptor [41].
With the development of immune-electron microscopy studies, it was demonstrated that normal receptor binding involves an interaction between Arginine 3500 and Tryptophan 4369 in the carboxy-tail of apoB100 [42].
In contract to the numerous ADH causative mutations in the LDLR gene, only a very few mutations have been reported in the APOB gene. To date, 10 mutations in APOB gene were described. The most frequent one is p.Arg 3500Gln. This form of ADH, due to APOB gene mutations, was previously called FDB for Familial ligand-Defective apolipoprotein B (OMIM #144010).
Compared with individual mutation in the LDLR gene, each of which is rare, the p.Arg3500Gln APOB mutation is common in Europe, where 2-5% of hyperholesterolemic are heterozygous carriers. The penetrance of the mutant APOB allele, however, is not 100%, so patients with familial ligand-defective apoB have a less-severe phenotypes than FH (Familial Hypercholesterolemia) patients with a LDLR mutation [43,44].

In the Proprotein Convertase Subtilisin Kexine Type 9 Gene
The third locus causing ADH was identified to be a gene located at chromosome 1p32.3, and named proprotein convertase subtilisin kexine type 9 gene (PCSK9) [45].
PCSK9 encodes the ninth member of the subtilisine-like protein convertase family (PCs). PCs are implicated in limited proteolysis of protein precursors going through the secretory pathway such as prohormones or precursors of neuropeptides [46]. This gene comprises 12 exons transcribed into a complementary DNA that spans 3617 bp. The preproPCSK9 is synthesized as a 694 amino acid long, that undergoes autocatalytic cleavage between the prodomain and catalytic do-main [47,48]. The prodomain (~14 kDa) remains bound to the mature protein throughout the secretory pathway [49]. The mature PCSK9 and the associated prodomain both undergo tyrosine sulfation in the late Golgi complex before secretion [50]. The role of this post-translational modification in PCSK9 has not been defined [51].
The observation that PCSK9 mutations cause dominant hypercholesterolemia suggests that mutations confer a gainof-function [45]. This hypothesis was confirmed by studies in which wild-type and mutant PCSK9 (S127R, F216L) were expressed at high levels in the mice liver; hepatic LDL receptor fell dramatically in the mice receiving either the wildtype or mutant PCSK9 [50,53]. No associated reduction in LDL receptor mRNA levels were observed. Thus overexpression of PCSK9, whether mutant or wild type, reduces the number of receptors through a post-transcriptional mechanism. However, the existence of a direct effect of PCSK9 on LDL receptor degradation has never been reported [54)].
To confirm the hypothesis that loss-of-function mutant of PCSK9 would causes hypocholesterolemia, Cohen et al. [55] sequenced the coding region of PCSK9 in individuals with low levels of plasma LDL-cholesterol (<5 th percentile). Surprisingly, one out of 50 African-Americans in the population had a nonsense mutation in PCSK9 that lowered LDLcholesterol levels by ~ 40% [55]. Subsequently, additional PCSK9 mutations associated with a reduction in plasma levels of LDL-cholesterol have been found, including in-frame, and missense mutations [56,57].
Until now, a total number of 101 unique variants were reported, covering the entire gene of PCSK9. [45, 50, 52 56, 57; http://www.ucl.ac.uk/ldlr/Current/index.php?select_db= PCSK9] PCSK9, which interacts with the LDL receptor, is a promising therapeutic target for hypercholesterolemia and coronary artery disease. A clear link between PCSK9 and LDL-cholesterol is also observed in animal studies. Indeed PCSK9 knockout mice have decreased plasma LDLcholesterol [58]. In non-human primates, PCSK9 knockdown by siRNA or inhibition by a monoclonal antibody also leads to decreased plasma LDL [59,60].
A recent study showed that antibody 1B20, which binds to PCSK9 with high affinity, disrupts the PCSK9-LDL receptor interaction, and inhibits the effect of PCSK9 on cellular LDL uptake [61]. Moreover, treatment with the 1B20 antiPCSK9 monoclonal antibody in mice and rhesus monkeys led to robust LDL-cholesterol lowering in plasma decreased liver PCSK9 and LDL mRNAs, and transient increases in total plasma levels of PCSK9 [61].
To conclude, understanding the physiology of PCSK9 is important, and this protein has become a major new target for lipid lowering therapy.

Other Possible Genes for ADH
ADH has proven to be genetically heterogeneous and associated with defects in other still unknown genes. Indeed, in a Mexican population, no PSCK9 mutations were found in one large ADH family that showed positive linkage to the 1.p34-32 locus. This indicates that genes other than PCSK9 in the locus may be involved [13]. Marques-Pinheiro et al. [14] showed in a large family with ADH phenotype, but with no mutations in the three known genes, the implication of a fourth loci that was named HCHOLA4. This locus is located at 16q22.1 in a 7.89 Mb interval containing 154 genes.
In the Chinese ADH population, after performing a genome-wide linkage analysis of a family pedigree without mutations in LDLR, APOB and PCSK9 genes, a two suggestive linkage loci were identified on chromosome 3q25.1-26.1 and 21q22.3 [15].
In ADH families from Spain, with no mutation in the known ADH-causing genes, Cenarro et al. demonstrated the implication of a new locus located in 8p.24-22 through linkage analyses [16].
Finally, a Portuguese ADH study found only 48% of its total received cases with clinical diagnosis of ADH had genetic defects on LDLR, APOB or PCSK9, leaving the other 52% of ADH cases with possible undiscovered genes mutations [62].
These studies confirm complex etiologies and suggest new genetic causal factors for the ADH disorder.

MOLECULAR DEFAULTS THAT CAUSE ADH IN TUNISIA
In Tunisia, ADH is one of the most frequent genetic disorders with a frequency of 1/165 for heterozygous [11]. This population had a mild phenotype of ADH, in particularly for heterozygous carriers.
Primary genetic studies were focused on LDLR gene. Recently, we started research on PCSK9 gene variation. Concerning APOB gene, studies were realized to search for the p.Arg3500Gln mutation. Studies were carried on 102 patients from 19 unrelated ADH families.
Mutations identified in Tunisian ADH patients are presented in (Table 4).
In the LDLR gene, we identified 11 mutations in the different exons of the gene, from them 7 were novels. Mutations were nonsense, frame shift, missense and major rearrangement. The mutation p.Ser493ArgfsX44 in exon 10 appears to be the most frequent mutation. [11,24,30,63,64].
Concerning the PCSK9 gene, our team identified a novel missense mutation named c.520C>T (p.Pro174Ser) localized in exon 3. Study indicates that this new PCSK9 variant is able to reduce the severity of FH, very probably acting as a loss-of-function variant. This finding should be confirmed by in vitro experiments [30].
Moreover, four common polymorphisms of PCSK9 were identified in a sample of 13 unrelated FH patients: L10, L11 p.474Val and pGlu670. Their frequencies were similar to those reported in previous study for different population [30].
No APOB Arg3500Gln was identified in all patients genotyped until now. This observation was also noted for the Lebanese [65] and Moroccan [66] population.
To conclude, the clinical expression of ADH in heterozygous patients is influenced by environmental factors as well as genetic factors in particular genes affecting lipoprotein metabolism such as APOE, MTP, HL, and ABCA1 genes. In unrelated ADH patients, the plasma LDL-cholesterol level is influenced by APOE, MTP and APOB polymorphisms, and the plasma High Density Lipoprotein (HDL)-cholesterol level is influenced by HL, FABP-2 and LPL polymorphisms [67]. The sequence analysis of these genes in Tunisian ADH patients may reveal genetic factors that are responsible of the mild clinical and biological phenotype of heterozygous ADH actually observed.

CONCLUSION
Special efforts are required to identify individuals with ADH in Tunisia as they are at high risk of premature coronary heart disease. The condition is seriously under diagnosed and the diagnosis is often made too late, in particular for heterozygous subjects, restricting the benefits of the treatments [68].
These patients can be treated to lower their cholesterol levels, before the installation of CHD, and thus avoid the complications and early death.
The attenuated phenotype of ADH in Tunisia was demonstrated and the establishment of a special diagnosis protocol is essential for the Tunisian population.