Common Mutations in the Low-Density-Lipoprotein–Receptor Gene Causing Familial Hypercholesterolemia in the Japanese Population
Abstract Familial hypercholesterolemia (FH) is a common genetic disorder caused by mutations of the LDL-receptor gene. In the present study, we investigated four Japanese FH homozygotes and identified five point mutations: a splice site mutation in intron 12 (the 1845+2 T→C mutation), a missense mutation in exon 7 (the C317S mutation), a nonsense mutation in exon 17 (the K790X mutation), a missense mutation in exon 14 (the P664L mutation), and a missense mutation in exon 4 (the E119K mutation). We developed simple methods for detecting these mutations. When we examined the presence of these mutations in 24 unrelated FH homozygotes, the 1845+2 T→C mutation was found in 7 of them, and the other four mutations were unique for each proband. We also screened 120 unrelated FH heterozygotes for these mutations and found that the frequencies of the 1845+2 T→C, C317S, K790X, P664L, and E119K mutations were 13.3% (16/120), 6.7% (8/120), 6.7% (8/120), 3.3% (4/120), and 1.7% (2/120), respectively. These mutations were found in more than 30% of unrelated Japanese FH patients. By using the detection methods developed in this study, the diagnosis of more than 30% of the genetic bases of Japanese FH heterozygotes is expected.
- Received April 24, 1995.
- Accepted July 19, 1995.
Familial hypercholesterolemia (FH) is a genetic disorder characterized by an elevated level of LDL cholesterol, xanthomas, and premature atherosclerosis.1 FH is an inherited disease in an autosomal dominant trait, and the genetic basis of FH is a lack of functional receptors for LDL on the cell surface.1 Various mutations of the LDL-receptor gene have been reported,2 3 including nonsense mutations, missense mutations, deletions, and insertions.
The frequency of heterozygous FH is estimated to be 0.2% in most populations, such as the European and North American populations.3 However, in selected populations, such as the Afrikaners of South Africa4 and the French Canadians,5 the frequencies of FH are fivefold and twofold higher, respectively, than average. Moreover, in these populations, limited mutations of the LDL-receptor gene predominate. For example, among Afrikaner FH patients, two mutations of the LDL-receptor gene comprise more than 95% of the mutant alleles.4 Among French Canadians, the “FH French Canadian-1” mutation comprises about 60% of the mutant alleles, and five mutations (FH French Canadian-1 to -5) comprise approximately 80% of the mutations of the LDL-receptor gene in this population.3 This phenomenon is known as the “founder’s effect,” which is observed in populations that have been isolated from other populations for a long period for geographical, religious, racial, or political reasons. Overall, the frequency of Japanese FH patients is also estimated to be 0.2%, which is not as high as in the Afrikaners or French Canadians. However, a characteristic distribution of the mutant alleles of the LDL-receptor gene might exist, since the Japanese people are uniracial, and Japan has been isolated geographically and politically for more than a thousand years. If there are limited kinds of mutations in the Japanese FH patients, it will be easier to diagnose heterozygous FH from their genetic backgrounds.
Here we describe five point mutations of the LDL-receptor gene in Japanese FH patients, four of which are common to FH patients in the Japanese population; these four mutations were found in more than 30% of the Japanese FH patients.
The genetic bases of four patients who were clinically diagnosed as FH homozygotes (Table 1⇓) were investigated. The clinical features and the LDL-receptor activities of patients K.I.K. and K.N. were reported previously.6
One hundred twenty unrelated Japanese patients clinically diagnosed as FH heterozygotes were also investigated in this study. If any of the following diagnostic criteria were satisfied, the proband was diagnosed with heterozygous FH7 ; (1) type II hyperlipidemia >6.7 mmol/L cholesterol with tendon xanthomas, (2) type II hyperlipidemia >6.7 mmol/L cholesterol and the presence of the subjects with type II hyperlipidemia >6.7 mmol/L cholesterol with tendon xanthomas in the proband’s first- or second-degree relatives, or (3) type II hyperlipidemia >6.7 mmol/L cholesterol and the LDL-receptor activity of the proband’s fibroblasts lower than that of a normal control. Although these criteria are less stringent than those in the Western countries,8 they are thought to be adequate for diagnosing heterozygous FH in the Japanese population because of the lower plasma cholesterol levels due to the traditional Japanese low-fat diet.9 In the total of 120 heterozygous FH patients apparently unrelated to each other, 48 were men and 72 were women. The mean age of the patients was 45.3 years (male, 43.2; female, 46.9). The mean value of plasma cholesterol of the patients was 9.5 mmol/L. The frequency of the involvement of coronary artery disease in the patients was 41.5%.
[α-35S]dCTP (1000 Ci/mmol) was purchased from New England Nuclear. [γ-32P]ATP (4500 Ci/mmol) was from ICN Biochemicals. Nucleic acids and hexadeoxynucleotides were obtained from Pharmacia. NA45 membrane was from Schleicher & Schuell. The restriction enzymes were from Bethesda Research Laboratories, New England Biolabs (NEB), Takara Shuzo, and Toyobo.
Amplification and Sequencing of Genomic DNA
The LDL-receptor genes of patients K.I.K. and K.N. (Table 1⇑) were amplified from 1 μg of genomic DNA extracted from leukocytes by the polymerase chain reaction (PCR) with 2.5 U Pfu DNA polymerase (Stratagene) using an automated thermal cycler from Perkin-Elmer Cetus. To amplify exons 1 to 17 and the coding region of exon 18 of the LDL-receptor gene, reported pairs of primers10 were used, except that the two nucleotides at the 3′ end of each oligonucleotide were omitted. Before amplification, the oligonucleotide primers were treated with T4 polynucleotide kinase (Toyobo). The amplification reaction comprised 25 cycles of incubation of the reaction mixture for (1) denaturation at 94°C for 1.5 minutes, (2) annealing at 45°C to 60°C for 2 minutes, and (3) extension at 72°C for 3 minutes. The amplified product was size fractionated by 2% agarose gel electrophoresis, and the band corresponding to the amplified sequence was adsorbed onto an NA45 membrane and then eluted as described previously.11 The purified DNA was blunt ended with T4 DNA polymerase (Toyobo), subcloned into the Sma I site of pUC118 with T4 DNA ligase (BRL), amplified, and then sequenced by the dideoxy method12 using T7 DNA polymerase, Sequenase (United States Biochemical).
Detection of the Mutations by the Single-Strand Conformation Polymorphism Method
In patients H.O. and E.O., all of exons 1 to 17 and the coding region of exon 18 were amplified as described above, except that polynucleotide kinase treatment of the primers was omitted. The PCR products were mixed with a dye solution (95% formamide, 20 mmol/L EDTA, 0.1% bromphenol blue, 0.1% xylene cyanol), heated at 80°C for 5 minutes, cooled on ice for 1 minute, and then electrophoresed in 5% to 20% gradient polyacrylamide gel (Pagel, Atto), followed by silver staining with a Silver Stain Plus kit (Bio-Rad Laboratories). The exons that showed different patterns from those in normal control subjects were subjected to gene amplification and sequencing analysis as described in the previous section.
Detection of the 1845+2 T→C Mutation
The reaction mixture for the PCR comprised a pair of oligonucleotide primers, 12SP76, which coded the 5′ flanking sequence of exon 12 (5′-TCTCCTTATCCACTTGTGTGTCT-3′),10 and 12SP77MM (5′-CTTGCATCTCGTACGTAAGCGAC-3′), which coded the antisense of the intron 12 sequence flanking the 3′ side of exon 12 with a mismatch base substitution of C to G; 1 μg of genomic DNA of each patient; 2.5 U Taq DNA polymerase (BRL); and 0.2 mmol/L of each dNTP in the buffer (20 mmol/L Tris-HCl, pH 8.4; 50 mmol/L KCl; 1.5 mmol/L MgCl2). The amplification reaction comprised 25 cycles of incubation of the reaction mixture for (1) denaturation at 94°C for 1.5 minutes, (2) annealing at 60°C for 2 minutes, and (3) extension at 72°C for 3 minutes. The amplified product was treated with 3 U restriction enzyme BsaHI (NEB) at 50°C for 1 hour, electrophoresed on 20% polyacrylamide gel, and then stained with ethidium bromide.
Detection of the C317S Mutation
A pair of oligonucleotide primers, 7SP66, which coded the intron sequence flanking the 5′ side of exon 7 (5′-AGTCTGCATCCCTGGCCCTGCGC-3′), and 7SP67 (5′-AGGGCTCAGTCCACCGGGGAATC-3′),10 which coded the antisense sequence of the intron 7 flanking the 3′ side of exon 7, were used for the PCR. The amplification reaction comprised 25 cycles of incubation of the reaction mixture for (1) denaturation at 94°C for 1.5 minutes, (2) annealing at 65°C for 2 minutes, and (3) extension at 72°C for 3 minutes. The PCR product was denatured by the alkaline method10 and then dotted onto a nylon membrane (GeneScreen, NEN). The DNA was fixed to the filter by UV-light irradiation. The allele-specific oligonucleotide homologous to either the normal sequence (7N-Cys; 5′-GGGGCACAGGCACTCGTAGCCGA-3′) or the mutant (7KN-Ser; 5′-GGGGCACAGGCTCTCGTAGCCGA-3′) was end labeled with [γ-32P]ATP.13 The filters were hybridized with end-labeled oligonucleotides as described.10 After being washed at 52°C, the filters were exposed to XAR-5 film (Eastman Kodak Co) at −70°C for 30 minutes with an intensifying screen.
Detection of the E119K, P664L, and K790X Mutations
Three pairs of oligonucleotide primers were used10 : 4SP61, which coded the intron 3 sequence flanking the 5′ side of exon 4 (5′-TGGTCTCGGCCATCCATCCCTGC-3′), and 3LR4N, which coded the antisense sequence of the midportion of exon 4 (5′-CATCTTCGCAGTCGGGGTCG-3′) for exon 4 amplification; 14SP80, which coded the intron 13 sequence flanking the 5′ side of exon 14 (5′-CCTGACTCCGCTTCTTCTGCCCC-3′), and 14SP81, which coded the antisense sequence of intron 14 flanking the 3′ side of exon 14 (5′-ACGCAGAAACAAGGCGTGTGCCA-3′) for exon 14 amplification; and 17SP86, which coded the intron 16 sequence flanking the 5′ side of exon 17 (5′-TGACAGAGCGTGCCTCTCCCTAC-3′), and 17SP87, which coded the antisense sequence of intron 17 flanking the 3′ side of exon 17 (5′-TGGCTTTCTAGAGAGGGTCACAC-3′) for exon 17 amplification. The genomic DNAs (1 μg) of the patients were amplified with each pair of exon-specific primers. The amplification reaction comprised 25 cycles of incubation of the reaction mixture for (1) denaturation at 94°C for 1.5 minutes, (2) annealing at 60°C for exons 4 and 17 and at 65°C for exon 14 for 2 minutes, and (3) extension at 72°C for 3 minutes. The amplified product was treated with restriction enzyme Mnl I (NEB) for exon 4, Pst I (Toyobo) for exon 14, and Mbo II (Toyobo) for exon 17 at 37°C for 1 hour, electrophoresed on 20% polyacrylamide gel, and then stained with ethidium bromide.
We analyzed four FH homozygotes belonging to independent families and identified four already known and one novel mutation. Table 1⇑ shows the clinical features of these FH homozygotes. Patient K.I.K. was found to be a true homozygote carrying a T-to-C base substitution in the 5′ splice donor site of intron 12. This mutation has been reported by Hobbs et al3 and named FH-Niigata. Recently, this mutation was found in another Japanese family by Funahashi et al.14 The mutation was designated as the 1845+2 T→C mutation according to the nomenclature suggested by Beaudet and Tsui.15 The second patient, K.N., was found to be a true homozygote for the missense mutation in codon 317 in exon 7, which was named FH-Wakayama.14 A base substitution (TGC to AGC) caused an amino acid change from Cys to Ser, and the mutation was designated as the C317S mutation. The third FH patient, H.O., was a compound heterozygote. One mutation was a missense mutation in exon 4, changing Glu119 (GAG) to Lys (AAG). This missense mutation has been found in a Philippine-Canadian3 and in a South African of Indian origin,16 was named FH-Philippines,3 and was designated as the E119K mutation. Sequencing of the PCR-amplified DNA fragments of exon 17 that were subcloned into the pUC plasmid showed that patient H.O. also had a nonsense mutation in exon 17. The mutation was in the codon of Lys790 (AAG to TAG) and was designated as the K790X mutation (Fig 1⇓). This mutation has not been reported previously. The mutation was identified in half of the subclones that were sequenced in patient H.O. The cholesterol level in the mother of the proband who did not have the E119K mutation but who had this mutation was 7.1 mmol/L (Fig 2A⇓). These results strongly suggest that the K790X mutation was on a different allele from the one that carried the E119K mutation in patient H.O. and was the cause of the hypercholesterolemia. The last patient, E.O., was a true homozygote having a mutation in exon 14. The mutation changed Pro664 (CCG) to Leu (CTG) and was designated as the P664L mutation. Interestingly, in patient E.O. there is no parental consanguinity. This mutation has been found among Asian-Indians,17 individuals of English and Norwegian descent,18 and the Dutch.19
To detect the 1845+2 T→C mutation in FH patients, we developed a method involving a mismatch primer for the PCR. We synthesized a pair of primers, 12SP76,10 which coded the 5′ flanking sequence of exon 12, and 12SP77MM, which coded the sequence in intron 12 (nucleotide numbers 25 to 3 from the beginning of intron 12), having a G-to-C substitution at nucleotide number 5 (Fig 3A⇓). With this pair of primers, the PCR product of the gene having the 1845+2 T→C mutation was expected to contain a new BsaHI site and to yield a restriction fragment of 168 bp by BsaHI digestion. As shown in Fig 3C⇓, the BsaHI digests of PCR products of proband K.I.K. and her brother, who were identified as homozygotes for this mutation, contained a 168-bp fragment, while those of normal subjects contained a 190-bp fragment. The sequence of the fragment was confirmed by the dideoxy method (data not shown). Furthermore, the BsaHI digests of the heterozygotes for this mutation contained both 190- and 168-bp fragments (Fig 3C⇓, III-2, III-3, IV-1, and IV-2).
To detect the second mutation, C317S, we developed an allele-specific oligonucleotide hybridization method.10 As shown in Fig 4⇓, the PCR-amplified fragment from the mutant allele could be clearly distinguished from the normal one.
The E119K mutation caused the loss of one of the Mnl I sites in exon 4 and the production of a 144-bp fragment on Mln I digestion of the PCR-amplified product (Fig 2B⇑). Similarly, the K790X mutation caused the loss of one of the Mbo II sites and the production of a 90-bp fragment (Fig 2C⇑). For the precise detection of these mutations, complete digestion of the amplified DNA with restriction enzymes is a prerequisite. As shown in Fig 2⇑, the mutation sites in the family members of proband H.O. (II-1), who was a compound heterozygote for the E119K and K790X mutations, were precisely determined by digestion with restriction enzymes (see Fig 2B⇑ and 2C⇑).
The P664L mutation was detected on the basis of the created Pst I site in exon 14.17 Pst I digested the PCR-amplified mutant fragment into two pieces, 95 and 109 bp. Fig 5⇓ shows Pst I digestion of the PCR-amplified exon 14 of patient E.O. and her parents, who were heterozygous for this mutation.
We have collected 24 FH homozygotes belonging to families unrelated to each other and found the five different point mutations in the four independent patients in the present study. Four of the mutations, E119K, C317S, P664L, and K790X, were each unique for one of the families of the probands, whereas the 1845+2 T→C mutation was found in 7 of the 24 families. The frequency of the 1845+2 T→C mutation was 29.2% in the FH homozygotes. This high frequency of the 1845+2 T→C mutation encouraged us to screen FH heterozygotes for the mutations. We have collected genomic DNA from 120 unrelated FH heterozygotes who were diagnosed as having FH from their plasma cholesterol levels and family studies7 and examined the presence of the five mutations by the procedures described above. We identified the 1845+2 T→C, C317S, K790X, P664L, and E119K mutations in 16, 8, 8, 4, and 2 patients, respectively (Table 2⇓). The frequency of the 1845+2 T→C mutation was 13.3% (16/120) for the FH heterozygotes. This mutation is a common one in the Japanese and, interestingly, has so far been found only among the Japanese population.3 14 As for the other four mutations, C317S, K790X, P664L, and E119K, each was unique for one proband among the 24 FH homozygotes. However, relatively high numbers of heterozygous patients carrying these mutations were identified: C317S (6.7%), K790X (6.7%), and P664L (3.3%) (Table 2⇓). Two heterozygotes for the E119K mutation were found. Accordingly, the genetic bases of 31.7%, ie, 38 of the 120, of Japanese FH heterozygotes were identified. Thus, surprisingly, only five point mutations comprised more than 30% of the mutant alleles in the Japanese FH patients. This could be a sort of “founder’s effect,” as the Japanese people are almost uniracial, and Japan is geographically isolated. The most important point is that 30% of the Japanese FH patients can be easily diagnosed by using the simple methods developed in the present study. At present, Japan has a population of about 120 million, and 0.2% (240 000) of them are expected to have mutations in the LDL-receptor gene. Therefore, about 72 000 (30%) who are expected to have one of the five mutations reported here can be easily diagnosed. As shown in Fig 6⇓, although the C317S mutation seemed to be localized to the Kansai district, the birthplaces of the patients carrying the other four mutations were widely distributed all over Japan. This again supports the idea that the frequencies of these mutations are meaningful values and the diagnoses can be applied to the entire Japanese population.
The plasma lipid levels in FH patients, as well as in the general population, are under genetic and environmental influences.1 The diagnostic criteria used in this study are less stringent than those suggested by Goldstein and Brown.1 This lower stringency is based on the fact that the Japanese have lower levels of plasma cholesterol than the Western population,8 mainly because of the traditional Japanese low-fat and high-carbohydrate diet.9 Although the diagnostic criteria have less stringency, the average clinical features of the patients selected as heterozygous FH in this study were similar to those in the United States and in the European countries1 (see “Methods”).
This work was supported in part by grants-in-aid for scientific research from the Ministry of Education, Science, and Culture of Japan and from the Ministry of Health and Welfare of Japan. We wish to thank Dr Yuko Yamamoto, Department of Pediatrics, Nishinomiya Municipal Central Hospital, and Drs Hiroko Tsukamoto and Tokuzo Harada, Department of Pediatrics, Osaka University Medical School, for providing the blood samples and skin fibroblast samples from FH patients and Dr Masaharu Kubo of Sumitomo Hospital for providing clinical data on the patients. We also thank Ikuko Okuno in our laboratory for her technical assistance.
Goldstein JL, Brown MS. Familial hypercholesterolemia. In: Scriver CR, Beaudet AL, Sly WS, Valle I, eds. The Metabolic Basis of Inherited Disease. 6th ed. New York, NY: McGraw-Hill International Book Co; 1989:1215-1250.
Leitersdorf E, van der Westhuyzen DR, Coetzee GA, Hobbs HH. Two common low density lipoprotein receptor gene mutations cause familial hypercholesterolemia in Afrikaners. J Clin Invest. 1989;84:954-961.
Tarui S. In: The annual report of the Research Group for Primary Hyperlipidemia [in Japanese]. Ministry of Health and Welfare of Japan; 1986:11-20.
Sun X-M, Webb JC, Gudnason V, Humphries S, Seed M, Thompson GR, Knight BL, Soutar AK. Characterization of deletions in the LDL receptor gene in patients with familial hypercholesterolemia in the United Kingdom. Arterioscler Thromb. 1992;12:762-770.
Mabuchi H, Haba T, Ueda K, Ueda R, Tatami R, Ito S, Kametani T, Koizumi J, Miyamoto S, Ohta M, Takeda R, Takegoshi T, Takeshita H. Serum lipids and coronary heart disease in heterozygous familial hypercholesterolemia in the Hokuriku district of Japan. Atherosclerosis. 1977;28:417-423.
Leitersdorf E, Tobin EJ, Davignon J, Hobbs HH. Common low-density lipoprotein receptor mutations in the French Canadian population. J Clin Invest. 1990;85:1014-1023.
Tajima S, Yamamura T, Yamamoto A. Analysis of apolipoprotein E5 gene from a patient with hyperlipoproteinemia. J Biochem (Tokyo). 1988;104:48-52.
Sanger F, Nicklen S, Coulson AR. DNA sequencing with chain-terminating inhibitors. Proc Natl Acad Sci U S A. 1977;74:5463-5467.
Sambrook J, Fritsch EF, Maniatis T. Molecular Cloning: A Laboratory Manual. 2nd ed. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory; 1989.
Funahashi T, Yamashita S, Maruyama T, Ueyama Y, Menju M, Nagai Y, Takemura K, Miyake Y, Tajima S, Matsuzawa Y. A compound heterozygote for familial hypercholesterolemia with a homozygous mother. J Intern Med. In press.
Soutar AK, Knight BL, Patel DD. Identification of a point mutation in growth factor repeat C of the low density lipoprotein-receptor gene in a patient with homozygous familial hypercholesterolemia that affects ligand binding and intracellular movement of receptors. Proc Natl Acad Sci U S A. 1989;86:4166-4170.
Defesche JC, van de Ree MA, Kastelein JJ, van Diermen DE, Janssens NW, van Doormaal JJ, Hayden MR. Detection of the Pro664-Leu mutation in the low-density lipoprotein receptor and its relation to lipoprotein(a) levels in patients with familial hypercholesterolemia of Dutch ancestry from The Netherlands and Canada. Clin Genet. 1992;42:273-280.