Age, Origin and Functional Study of the Prevalent LDLR Mutation Causing Familial Hypercholesterolaemia in Gran Canaria

The p.(Tyr400_Phe402del) mutation in the LDL receptor (LDLR) gene is the most frequent cause of familial hypercholesterolaemia (FH) in Gran Canaria. The aim of this study was to determine the age and origin of this prevalent founder mutation and to explore its functional consequences. For this purpose, we obtained the haplotypic information of 14 microsatellite loci surrounding the mutation in one homozygous individual and 11 unrelated heterozygous family trios. Eight different mutation carrier haplotypes were identified, which were estimated to originate from a common ancestral haplotype 387 (110–1572) years ago. This estimation suggests that this mutation happened after the Spanish colonisation of the Canary Islands, which took place during the fifteenth century. Comprehensive functional studies of this mutation showed that the expressed LDL receptor was retained in the endoplasmic reticulum, preventing its migration to the cell surface, thus allowing us to classify this LDLR mutation as a class 2a, defective, pathogenic variant.


Introduction
Familial hypercholesterolaemia (FH, OMIM 144400) is an autosomal codominant disorder that affects 34 million people worldwide [1]. FH is characterised by increased low-density lipoprotein cholesterol (LDL-C) concentrations, which lead to premature atherosclerotic cardiovascular disease (ASCVD) and cholesterol deposits in the cornea and tendons [2].
FH is caused by an array of pathogenic variants affecting genes that regulate cholesterol metabolism [3]. Most of these pathogenic variants are located in the LDL receptor (LDLR) gene, resulting in 80% of the cases of FH, with more than 4000 variants described so far in the Human Gene Mutation Database. The heterozygous form of FH (HeFH) is the most common, with a prevalence of 1:200-250 people [4], whereas the more severe homozygous form (HoFH) occurs with a frequency of 1:250,000 to 360,000 [5].
The genetic isolation of certain populations has led to an increase in the frequencies of some variants via founder effects. This phenomenon has been reported in Afrikaners [6],  [13.5-85] Values correspond to average ± standard deviation unless otherwise specified. BMI, body mass index; ASCVD, atherosclerotic cardiovascular disease; HDL-c, high-density lipoprotein cholesterol; LDL-c, low-density lipoprotein cholesterol; IQR, interquartile range.
Fourteen autosomal microsatellite loci flanking the p.(Tyr400_Phe402del) mutation were analysed for all 82 individuals (34 mutation carriers and relatives, and 48 controls). All but one locus (trinucleotide) contained dinucleotide repeats and presented from 6 to 16 different alleles (Table 2). Significant deviations from HWE due to heterozygote deficiency were detected in two loci (L5 and R1). As expected, considering their genomic proximity, several locus combinations showed significant LD (Supplementary Table S1).

Age of the p.(Tyr400_Phe402del) Mutation
To estimate the age of the p.(Tyr400_Phe402del) mutation, haplotypic information for the 14 microsatellites analysed were deduced in carriers using their first-degree relatives. A total of eight different haplotypes were identified (Table 3).
Assuming a 'correlated' genealogy, which considers the possibility of the mutation age being more recent than the most recent common ancestor for the analysed population, the mutation arose 15.5 generations ago, with a confidence interval of 4.4-62.9. Cells in orange indicate the location of the mutation (see Figure 1, Panel B). Cells in blue indicate the different alleles that define a specific haplotype. Complete genotyping information is described in Supplementary Table S2.
.  Cells in orange indicate the location of the mutation (see Figure 1, Panel B). Cells in blue indicate the different alleles that define a specific haplotype. Complete genotyping information is described in Supplementary Table S2.

Expression of the p.(Tyr400_Phe402del) LDLR Variant in CHO-ldlA7 Cells
Expression of the p.(Tyr400_Phe402del) LDLR variant was analysed by Western blot and flow cytometry in CHO-ldlA7-transfected cells, as described in Section 4 (Materials and Methods). For surface expression analysis by flow cytometry, two variants were used as internal method controls, p.(Trp87)* (a null allele mutant) and the Ex3_4del LDLR variant that is expressed to a similar extent as wt LDLR but is a class 3 variant with 100% impaired binding activity [23]. As shown in Figure 2A To confirm whether p.(Tyr400_Phe402del) is not expressed in its mature form, LDLR expression was assessed 48 h post-transfection. As shown in Figure 2B, only the expression of immature p.(Tyr400_Phe402del) was detected by Western blot, confirming the flow cytometry results.

p.(Tyr400_Phe402del) LDLR Variant Classification by Confocal Microscopy
To further analyse the type of defect produced by the in-frame deletion of Tyr400_Phe402 residues, we studied whether the immature expressed form of the p.(Tyr400_Phe402del) LDLR variant colocalised with calregulin, an endoplasmic reticulum (ER) marker, using a confocal microscope. Confocal images show that the variant is expressed in transfected cells, but remains clearly retained in the ER, as indicated by the high colocalisation with calregulin (Figure 4), which corroborates the experimental data obtained by flow cytometry and Western blot. Accordingly, the p.(Tyr400_Phe402del) LDLR variant should be classified as a class 2a, defective, pathogenic variant.
transfection with the plasmids carrying the different LDLR variants. The values represent the mean of triplicates (n = 3) ± SD. * p < 0.001 compared to wt using Student's t-test.

p.(Tyr400_Phe402del) LDLR Variant Classification by Confocal Microscopy
To further analyse the type of defect produced by the in-frame deletion of Tyr400_Phe402 residues, we studied whether the immature expressed form of the p.(Tyr400_Phe402del) LDLR variant colocalised with calregulin, an endoplasmic reticulum (ER) marker, using a confocal microscope. Confocal images show that the variant is expressed in transfected cells, but remains clearly retained in the ER, as indicated by the high colocalisation with calregulin (Figure 4), which corroborates the experimental data obtained by flow cytometry and Western blot. Accordingly, the p.(Tyr400_Phe402del) LDLR variant should be classified as a class 2a, defective, pathogenic variant.

Discussion
In this study we aimed to reveal the age and origin of the p.(Tyr400_Phe402del) LDLR mutation and the functional consequences on the expressed LDLR variant. To this end, we selected an optimally distanced [24] set of 14 microsatellites spanning 8.86 cM around the variant (p. [Tyr400_Phe402del]), and applied a method based on ancestral segment lengths [22] using fine mapping with LD. In the analysed cohort, we identified eight different haplotypes. Considering a recurrent 9 bp deletion is extremely rare, we assumed the different variant carrier haplotypes detected in the Gran Canaria population derive from a common ancestral haplotype (i.e., correlated genealogy). Therefore, this scenario fits with a genetic signature of a founder effect, in which all the mutation carriers have inherited the variant from a common ancestor arising in the population about 387 years ago. Although the confidence interval obtained was rather wide (110 to 1572 years), this estimation postdates the one-century-long Spanish colonisation of the Canary Islands, which ended in 1496 with the surrender of Tenerife [15]. After this dramatic episode, the European colonisation of the Canary Islands involved a mix of Spanish, Portuguese, Italian and Flemish colonisers, who, in addition to the sub-Saharan Africans and Moorish slaves' contribution [25], have provided the genetic background of the contemporary Canarian population.

Discussion
In this study we aimed to reveal the age and origin of the p.(Tyr400_Phe402del) LDLR mutation and the functional consequences on the expressed LDLR variant. To this end, we selected an optimally distanced [24] set of 14 microsatellites spanning 8.86 cM around the variant (p. [Tyr400_Phe402del]), and applied a method based on ancestral segment lengths [22] using fine mapping with LD. In the analysed cohort, we identified eight different haplotypes. Considering a recurrent 9 bp deletion is extremely rare, we assumed the different variant carrier haplotypes detected in the Gran Canaria population derive from a common ancestral haplotype (i.e., correlated genealogy). Therefore, this scenario fits with a genetic signature of a founder effect, in which all the mutation carriers have inherited the variant from a common ancestor arising in the population about 387 years ago. Although the confidence interval obtained was rather wide (110 to 1572 years), this estimation postdates the one-century-long Spanish colonisation of the Canary Islands, which ended in 1496 with the surrender of Tenerife [15]. After this dramatic episode, the European colonisation of the Canary Islands involved a mix of Spanish, Portuguese, Italian and Flemish colonisers, who, in addition to the sub-Saharan Africans and Moorish slaves' contribution [25], have provided the genetic background of the contemporary Canarian population.
The most plausible scenarios supporting the high frequency of this variant in the Gran Canarian population is either the result of gene flow from any of the postcolonisation sources, or an isolated mutational event in the settled Gran Canarian population. Although gene flow has been proposed as the evolutionary process of introducing a different LDLR mutation (G197del) in Israel and Lithuania [26], in the case of Gran Canaria, several facts point to a mutational event in the population inhabiting the island after the Spanish colonisation: (i) the first reference to this mutation in the literature refers to participants from a local hospital (Hospital Universitario Dr. Negrín de Gran Canaria) [14]; (ii) the mutation has not been found in mainland Spain nor elsewhere; and (iii) the genetic characteristics and the geographic isolation of the population have been previously shown to facilitate the expansion of genetic variants, causing both recessive [18,19] and dominant disorders [27].
An additional evolutionary process that can facilitate the predominance of specific variants in a population is positive selection. Although a heterozygote advantage has been proposed in other disease-associated variants [28] and HeFH is the most common form of FH in Gran Canaria, this mechanism does not seem to have influenced the current incidence of the variant (p. [Tyr400_Phe402del]). Indeed, carriers of this variant present a higher than expected prevalence of type 2 diabetes [11], as opposed to the view of FH being protective against this disease [29,30]. In addition, as we demonstrate in this study, the p.(Tyr400Phe402del) LDLR variant leads to a defective protein. Specifically, the inframe deletion occurring in the p.(Tyr400Phe402del) LDLR variant causes the removal of a tyrosine residue from a highly conserved motif in the first YWTD domain of the LDLR polypeptide. This constitutes one of the six four-stranded beta-sheets ("blades") that maintain the domain structure, which is determinant for the correct folding of the β-propeller domain [31]. As a result, this in-frame deletion of three residues may trigger the "quality control" machinery of the ER that blocks the trafficking of misfolded proteins [32], thus preventing the migration of the expressed protein to the cell surface and leading to a very severe FH phenotype. Consequently, we can classify the p.(Tyr400Phe402del) LDLR mutation as a class 2a, defective, pathogenic LDLR variant.
Considering the high prevalence of this class 2a LDLR variant in the population of Gran Canaria, the establishment of a rapid diagnostic test to screen the population for the presence of this particular variant is paramount. This will clearly assist clinicians in the diagnosis of this important disease and will allow for the initiation of timely therapeutic interventions. Indeed, this population-based diagnostic strategy is the current routine, not only at our centre, which provides assistance to the Southern and Eastern regions of the island, but also in the other main hospital of Gran Canaria, thus providing full coverage for the island population.
We acknowledge that our study has some limitations. First, the geographic region of the cohort is restricted. However, the sample size surpasses that of other studies on dominant diseases. In addition, unrelated variant-carriers were selected, in order to maximise the representation of the population affected with HF in Gran Canaria. Second, we opted for a genotype-based method, which cannot assure the sequence of the analysed region is identical among subjects sharing the haplotypes identified in this study. However, this method has been widely applied in other studies dating mutations. Furthermore, the microsatellite markers were carefully selected to be optimally distanced and informative, as demonstrated by the identification of recombination points at both sides of the mutation. Third, the methodology applied may have underestimated the age of the variant under investigation, an artefact that is more evident in growing populations [33]. In this regard, we are currently conducting whole genome sequencing in a selected group of variant-carriers, which will not only help us corroborate or refine our estimation but also will provide an opportunity to identify potential modifier genes that may explain the phenotypic diversity observed in individuals affected with HF in Gran Canaria.

Subjects
The study population included families attending the Lipids Unit of the Complejo Hospitalario Universitario Insular Materno-Infantil de Gran Canaria. This cohort received a genetic diagnosis of FH, carried the p.(Tyr400_Phe402del) variant in LDLR and had both parents born on the island. We selected 11 unrelated family trios of p.(Tyr400_Phe402del)mutation carriers and a homozygous individual. The trios were either mother-fatherproband, or parent-proband-sibling.
In addition, 48 unrelated Canary Islanders not bearing the p.(Tyr400_Phe402del) mutation, who self-declared as having two generations of ancestors born in the Canary Islands, were included as controls.

Microsatellite Genotyping
Genomic DNA was extracted from whole blood samples preserved in EDTA using a salt precipitation protocol [34]. Fourteen microsatellite markers covering 5.4 Mbp (8.86 cM) flanking the p.(Tyr400_Phe402del) mutation (Table 2 and Figure 4) were genotyped in the cases and controls.
Amplifications were carried out in 10 µL volume PCRs containing 1× colourless GoTaq ® Flexi Buffer (Promega, Madison, WI, USA), 1.5 mm of MgCl 2 , 0.2 mm of each dNTP, 0.12 mm of each primer (see Table 2), and 0.1 U of Taq polymerase (Promega). The PCR programme consisted of 95 • C for 3 min, followed by 28 cycles (95 • C for 30 s, 58 • C for 15 s and 72 • C for 1 min) with a final extension at 72 • C for 10 min. Fluorescently labelled fragments were run on an ABI PRISM 3100 DNA sequencer (Applied Biosystems, Foster City, CA, USA) with the GeneScan-500 (LIZ) size standard. Alleles were scored using Peak Scanner™ Software v1.0 (Applied Biosystems).

Genetic Characterisation
Measures of genetic diversity, such as the total number of alleles per locus, mean observed (H O ) and mean expected (H E ) heterozygosities, were calculated using AR-LEQUIN version 3.5.2.2 [35]. The same resource was used to test for departures from the Hardy-Weinberg equilibrium (HWE) and deviations from the linkage equilibrium (LD) for all pairwise locus combinations. A sequential Bonferroni correction [36] was applied to the HWE and LD results.

Estimation of the Age of the Variant
To estimate the age of the p.(Tyr400_Phe402del) mutation we used the Gamma linkage disequilibrium method (with correlated genealogy) implemented in the R Shiny app Genetic Mutation Age Estimator (https://shiny.wehi.edu.au/rafehi.h/mutation-dating/ (accessed on 8 May 2023)), which is fully described by Gandolfo et al. in 2014 [22]. This method estimates the age of a genetic mutation based on the genetic length of ancestral haplotypes common to individuals who share the mutation. Furthermore, this method has the advantage of using the information of the genomic distances and recombination rates of the microsatellite markers used for genotyping the study cohort. In this study, haplotypes were reconstructed based on genotypic information from relatives of mutation carriers.

Analysis of LDLR Expression by Fluorescent Activated Cell Sorter (FACS)
LDLr expression at the cell membrane was assessed in a CytoFLEX Flow Cytometer (Beckman Coulter, Brea, CA, USA) using a mouse monoclonal antihuman-LDLR (C7) (1:100; 2.5 mg/L; Origene, Rockville, MD, USA) and an Alexa Fluor 488-conjugated goat antimouse IgG (1:200; Molecular Probes, Eugene, OR, USA) as primary and secondary antibodies, respectively, as previously described [37]. Each sample was performed in triplicate, and 10,000 events were acquired for data analysis.

Analysis of LDL Uptake by FACS
Forty-eight hours post-transfection, cells were incubated with FITC-LDL (20 µg/mL) for 4 h at 37 • C to determine LDL uptake, as previously described [37]. For determining LDLR expression, cells were washed out with PBS-1% BSA, fixed in 4% paraformaldehyde for 10 min at room temperature and washed again to remove residual fixative. To determine the amount of internalized LDL, Trypan blue solution (Sigma-Aldrich, Steinheim, Germany) was added directly to the samples to a final concentration of 0.2%. Each sample was performed in triplicate, and 10,000 events were acquired for data analysis.

Confocal Laser Scanning Microscopy
Confocal laser scanning microscopy was used to analyse LDLR expression and colocalization with the endoplasmic reticulum (ER)-specific marker calregulin. Cells transfected with the LDLR-containing plasmids were cultured for 48 h at 37 • C in 5% CO 2 . Then, the cells were washed twice with PBS-1% BSA, fixed with 4% paraformaldehyde for 10 min, washed and permeabilised with 1% TritonX-100 for 30 min at room temperature. Samples were blocked in PBS-10% FBS for 1h and incubated with the appropriate primary antibodies for 16 h at 4 • C, followed by incubation with the appropriate fluorescent secondary antibodies. Coverslips were mounted on a glass slide, and samples were visualised using a confocal microscope (Olympus IX 81, Tokyo, Japan) with sequential excitation and capture image acquisition with a digital camera (Axiocam NRc5; Zeiss, Jena, Germany). Images were processed using Fluoview v50 software (Olympus, Miami, FL, USA).

Statistical Analysis
All measurements were performed at least 3 times unless otherwise specified, and results represent the mean ± standard deviation (SD). The differences between LDLR variants and wild-type (wt) LDLR were tested by a two-tailed Student's t-test with a significance level of 0.05.

Conclusions
The evidence presented in this study suggests that the most prevalent mutation causing HF in the population of Gran Canaria, p.(Tyr400_Phe402del) in LDLR, was introduced or arose in the population after the Spanish colonisation of the Canarian Archipelago, which took place during the 15th century. This relatively recent mutation expresses a misfolded protein that is retained in the ER, preventing its expression at the cellular surface. Therefore, this in-frame deletion can be classified as a class 2a, defective, pathogenic LDLR variant.