Comparison of the Genetic Defect with LDL-Receptor Activity in Cultured Cells from Patients With a Clinical Diagnosis of Heterozygous Familial Hypercholesterolemia
Abstract In this study we have analyzed the genetic defect in 42 patients with a diagnosis of heterozygous familial hypercholesterolemia (FH) by Southern blotting, SSCP, and sequencing of PCR-amplified fragments of genomic DNA or sequencing of RT-PCR products from mRNA in cultured cells. The apoB Arg3500Gln mutation was identified in five patients. A molecular defect in the LDL-receptor gene was confirmed in 23 patients; 16 of these mutations have not been described before. No defect in the coding region, intron:exon junctions or proximal promoter of the LDL-receptor gene or in the region of the apoB gene coding for the LDL-receptor binding domain was found in the remaining 14 patients. LDL-receptor activity and protein content of cultured lymphoblasts from the patients was significantly lower in cells from patients with severe rather than mild LDL-receptor mutations. Cells from four patients with no detectable defect showed reduced LDL receptor activity compared with eight normal cell lines, whereas six others had reduced LDL-receptor activity but LDL-receptor protein content within the normal range. Cells from four patients appeared to have normal LDL-receptor function. Cells from two patients with a defined defect also had LDL-receptor activity within the normal range. The findings demonstrate the problems involved in the genetic diagnosis of FH in patients. .
- Received October 26, 1996.
- Accepted March 17, 1997.
Familial hypercholesterolaemia (FH) is caused by defects in the gene for the LDL receptor that result in reduced LDL-receptor activity and defective clearance of lipoproteins from the circulation.1 This causes, on average, a twofold increase in plasma LDL-cholesterol concentration in heterozygous FH individuals that markedly hastens the onset of atherosclerosis and increases their risk of premature coronary heart disease compared with the normolipidemic population. Although considered to be a fully penetrant, autosomal co-dominant genetic disorder, there is nonetheless a wide variation in the severity of both the hypercholesterolemia and its accompanying clinical phenotype in heterozygous FH patients. It has also been reported that FH patients vary widely in their response to lipid-lowering drug therapy (reviewed in reference 2).
The LDL receptor is a multifunctional cell-surface glycoprotein, comprising several structural domains, that mediates the specific binding and uptake by receptor-mediated endocytosis of lipoproteins containing either apoB or apoE.3 Many different mutations in the LDL-receptor gene have been identified, and their effect on the structure and function of the protein in cultured cells has shown that mutations in different regions of the protein can have very different effects on receptor function.3 This has led to a classification of mutations into five functional classes, including null, binding-defective, transport defective, internalization-defective, and recycling-defective,4 and supports the earlier conclusion from studies on receptor activity in cells that homozygous FH patients could be classified as “receptor-defective” or “receptor-deficient.”5 More recently it has been suggested that the heterogeneity of not only homozygous but also heterozygous FH derives largely from differences in the nature of the mutation in the LDL-receptor gene, and several studies have been published in support of this.6-8 However, because of the wide diversity of LDL-receptor gene mutations in most populations, most of the results published so far have come from comparisons between groups of patients who have inherited the same mutant allele from a recent common ancestor, as in the French Canadian study.9 Such groups will have many other genes in common and will probably also share key environmental factors so that the effect of the mutant LDL-receptor gene on clinical phenotype is difficult to disentangle from these other influences. In this study we have attempted to define the genetic defect in a heterogeneous group of 42 patients with a clinical diagnosis of heterozygous FH who had recently taken part in an angiographic trial of the effect of lipid-lowering therapy on coronary atherosclerosis,10 and we have compared the residual LDL-receptor activity in cultured cells with the nature of the LDL-receptor gene defect. In a subsequent paper, we will examine the relationship between LDL-receptor genotype and clinical features, including the response to lipid-lowering therapy in these patients.
Materials and Methods
Description of the Patients
This study comprises a group of 12 female and 30 male apparently unrelated patients, aged 35 to 60 years at the start of the study (mean 49.6±7.3 years), with a clinical diagnosis of definite or possible FH, as defined by the Simon Broome Study Group.11 The criteria are a plasma cholesterol concentration greater than 7.5 mmol/L in the proband, together with tendon xanthoma either in the patient or in a first degree relative (definite FH; 26/42 patients) or the presence of premature coronary heart disease or hypercholesterolemia in a first degree relative (possible FH; 16/42 patients). All patients had angiographic evidence of coronary artery disease and had participated in the Familial Hypercholesterolaemia Regression Study.10
Identification of Genetic Defects in the apoB Gene
A fragment of the apoB gene was amplified with allele-specific primers and digested with MspI for detection of mutations in codon 3500.12 The presence of the Arg3500Gln mutation was confirmed by hybridization with allele-specific primers or by nucleotide sequencing. A fragment of the gene was also amplified for nucleotide sequencing of the region encompassing codons 3458 to 3546 and for digestion with Nsi I to detect the Arg3531Cys mutation.13 Genotyping of the 3′-VNTR of the apoB gene was carried out as described by Boerwinkle et al.14
Identification of Genetic Defects in the LDL-Receptor Gene
The methods for Southern blotting, PCR-amplification and single-stranded conformational polymorphism (SSCP) analysis, manual sequencing, or restriction enzyme digestion of amplified fragments of the LDL-receptor gene have all been described elsewhere.15-17 Automated sequencing of fragments of LDL-receptor cDNA was carried out as described before.17
Amplification of larger fragments of the LDL-receptor gene encompassing each intron was carried out with primers in the exons of the LDL-receptor gene using the Expand Long Template PCR System (Boehringer Mannheim) according to the supplier’s guidelines. PCR products were cloned into the plasmid pGEM-T (Promega) and subjected to automated nucleotide cycle sequencing (Prism Ready Reaction DyeDeoxy Terminator Cycle Sequencing Kit; Applied Biosystems) with primers located in the vector or with the PCR primers. Details of the PCR reactions and nucleotide sequence obtained can be provided on request (E-mail firstname.lastname@example.org).
Genotyping of the D19S394 marker flanking the 5′-end of the LDL receptor gene was carried out as described by Day and co-workers.18
Measurement of LDL-Receptor Activity and Protein Content in Cultured Lymphoblasts
Lymphoblasts were maintained in culture as described previously.19 For experiments, cells (1×107 cells/10 mL in a 175 cm2 flask) were preincubated in medium containing compactin (1.08 μg/mL) and 10% lipoprotein-deficient serum for 40 hours to induce expression of the LDL-receptor gene, and then dispensed into triplicate dishes (1×106 cells/1 mL in a dish) for measurement of saturable degradation of 125I-labeled LDL.19 For immunoblotting of LDL-receptor protein, extracts of cells that had been preincubated as described above were subjected to SDS-PAGE on nonreducing gels, transferred to nitrocellulose membranes, and incubated with monoclonal antibodies to the LDL-receptor as described previously.16 Bound antibody was detected by chemiluminescence.19 For quantitation, different amounts of a standard extract of control cells (10 to 80 μg of cell protein/lane) were loaded onto each gel with the patient or control extracts (40 μg of cell protein/lane). The relative amount of LDL-receptor protein in extracts of cells from patients and normolipidemic controls was determined from densitometric scans of the blot by comparison with the standards. The values for the samples were always within the linear range of the standard curve.
Analysis of the AvaII Polymorphism in mRNA
A fragment of the LDL-receptor mRNA was amplified by RT-PCR for 25 cycles with one of the primers end-labeled with 32P, as described previously.19 The PCR products were digested with AvaII and the fragments separated by electrophoresis on 8% PAGE and detected by autoradiography. The relative proportion of each band was determined by analysis of the autoradiograph with a BioRad GS 690 Imaging Densitometer and Molecular Analyst software.
The significance of differences between mean values was determined by Student’s t test, and the significance of correlations by Fisher’s R to Z test, using the StatView statistical package (Abacus Concepts Inc, Berkeley, Calif).
Identification of the Genetic Defect
Genomic DNA was prepared from whole blood from all patients. Four of the patients (listed in Table 1⇓) were found to be heterozygous and one homozygous20 for the Arg3500Gln mutation in the apoB gene by amplification of genomic DNA with allele specific oligonucleotides. No carriers of the Arg3531Cys13 or Arg3500Trp21 mutations in the apoB gene were found.
To detect mutations in the LDL-receptor gene, genomic DNA was first analyzed by Southern blotting (data not shown), which revealed that two of the patients (Nos. 1 and 9 in Table 1⇑) were heterozygous for a deletion in the LDL-receptor gene. The extent of the deletion in each case was confirmed by PCR and nucleotide sequencing across the deletion joint (Fig 1⇓). The deletion of exon 5 apparently involved recombination between alu sequences in introns 4 and 5, as shown in the diagram, and is identical to that found in a patient of French origin,22 but different from that in a Danish patient.24 The precise position of the deletion joint in the allele with deletion of exons 13 and 14 could not be determined because the sequence was highly repetitive.
Each exon of the LDL-receptor gene from the patients was amplified from genomic DNA and analyzed by SSCP (data not shown); any PCR products that gave an abnormal pattern were subjected to nucleotide sequencing. This revealed the presence of a mutation in only nine additional patients (Nos. 6, 28, 15, 10, 20, 18, 39, 2, and 17), and one of these mutations was present in two patients (Nos. 28 and 15; Table 1⇑). Most of the mutations (7/9) detected by SSCP were small deletions, only one of which, a 2bp deletion in exon 4 that is relatively common in the UK,26 had been reported before our study commenced.
For the patients in whom a mutation had not yet been detected, Epstein-Barr virus-transformed lymphoblastoid cell lines were established, and the nucleotide sequence was determined of LDL-receptor cDNA produced by RT-PCR of total RNA isolated from the cells. Overlapping fragments of the LDL-receptor mRNA and a 600-bp region of genomic DNA encompassing the known 5′ regulatory elements were amplified. One of each PCR primer pair was biotinylated so that single-stranded DNA fragments could be immobilized for automated nucleotide sequencing, as indicated in the diagram in Fig 2⇓, employing Sequenase and dye terminators. This method is very reliable for the detection of heterozygous bases, and single base substitutions were identified in a further 11 patients (Nos. 11, 38, 27, 41, 4, 3, 23, 30, 32, 33, and 21), only two of which had been reported previously (Table 1⇑). Several of these were not detectable by SSCP, even with hindsight. In all these cases, the presence of the base substitution in genomic DNA was confirmed by nucleotide sequencing and, where informative, by restriction enzyme digestion of amplified fragments. Care was needed to avoid artifacts apparently arising from the poor fidelity of reverse transcription in some regions of the 3′ end of the mRNA (fragment 165/164 in Fig 2⇓), and the method was clearly not useful for detection of mutations that result in very low steady state levels of the mutant mRNA.
Analysis of mRNA in Cultured Lymphoblasts
The methods described above failed to detect any defect in the LDL receptor gene in 14 of the 42 patients. For these patients, we considered the possibility that mutations that disrupt transcription or mRNA stability could occur in an intron distant from the intron:exon junction or in the untranslated region. For example, we have recently reported the presence of a minor rearrangement 25bp upstream from the junction of intron 9 with exon 10 that disrupts mRNA splicing because it destroys the branch point consensus sequence.27 The majority of such mutations would not be detectable with the currently available sequence information. Therefore we determined sufficient nucleotide sequence of the introns of the LDL-receptor gene to allow primers to be designed for amplification of regions of genomic DNA encompassing each exon with approximately 100 bp of intron at the 5′ end, to include the possible branch point28 (Table 2⇓). Amplified genomic DNA from the 14 patients in whom no mutation had yet been identified was subjected to automated nucleotide cycle sequencing, using an improved enzyme system (New Amplitaq, Perkin Elmer Ltd) that produces peaks of sufficiently even height to detect heterozygous substitutions with confidence. A single base substitution was identified in the coding region of the gene in one patient (no. 8; Table 1⇑) that was not detected by sequencing of amplified mRNA from the patient’s cells because the mutant mRNA was present at too low a level. No potential mutations were found in the other patients.
Eight of the patients were heterozygous for the readily detectable AvaII RFLP in exon 13, and it was thus possible to assess whether their lymphoblasts contained mRNA from both alleles of the LDL-receptor gene. The appropriate fragment of the mRNA was amplified by RT-PCR with one of the primers end-labeled with 32P. Quantitation of the bands obtained after digestion of the labeled PCR product showed that the proportion of mRNA derived from the allele containing the AvaII site was within the range of that obtained with mRNA from five cell lines from normolipaemic individuals (55.05±6.84). For the remaining six patients, in whom no mutation had been found, this analysis was uninformative because the patient was homozygous at the AvaII site and at all other potentially informative polymorphic sites in the coding region of the gene.
Finally, nucleotide sequencing of a region of the apoB gene encompassing codons 3491 to 3552 amplified from genomic DNA did not reveal any potential mutations in the patients in whom no mutation had been detected (results not shown). Thus mutations in the apoB gene were found in five patients, of whom four had definite FH, and mutations in the LDL-receptor gene were found in 23 patients, 18 of whom had definite FH. In the remaining 14 patients, of whom four had definite FH, no genetic defect was detectable.
For one of the patients in whom no genetic defect had been detected, blood samples were available from sufficient members of her family to allow us to investigate whether hypercholesterolaemia co-segregated with an allele of the LDL receptor gene or the apoB gene. A recent report shows that the polymorphic microsatellite marker D19S394, which is located approximately 250Kb from the 5′ end of the LDL-receptor gene on chromosome 19, has a heterozygosity index of at least 90% and is in strong linkage disequilibrium with the LDL receptor gene. A mutant allele of the LDL receptor gene that is common in the south of England co-segregates exclusively with one comparatively rare allele at this locus.18 From the family tree shown in Fig 4⇓, it is possible to deduce that both parents of the index patient must have been heterozygous for different alleles of D19S394 and, by implication, for the LDL-receptor gene. The only allele of D19S394 (238) that is common to the index patient (IIii), her hypercholesterolemic daughter (IIIi), and her hypercholesterolemic sibling (IIv) is also carried by two normolipemic members of the family (IIiv and IVii). Similar results were obtained when the family were genotyped for alleles at the 3′-VNTR of the apoB gene. The index patient (IIii) was homozygous for an allele with 37 repeats, one of which was inherited by her hypercholesterolemic daughter (IIIi) and her normocholesterolemic granddaughter (IV ii); furthermore, her hypercholesterolemic sister (IIv) and normocholesterolemic brother (IIi) both inherited the same alleles from their parents. Thus we conclude that hypercholesterolemia in this family does not co-segregate with an allele of the LDL receptor gene or the apoB gene. We were unable to obtain blood samples from sufficient informative relatives to investigate co-segregation in any of the other families.
Measurement of LDL-Receptor Activity and Protein in Cultured Lymphoblasts
To determine whether LDL-receptor function was defective in cells from the patients in whom a mutation in the LDL-receptor gene could not be found, LDL-receptor activity and LDL-receptor protein content of cultured Epstein-Barr virus-transformed lymphoblasts was determined and compared with that of cells from patients with known mutations and from eight normolipidemic individuals. LDL-receptor activity was measured as the maximum rate of saturable degradation of 125I-labeled LDL by cells that had been preincubated with lipoprotein deficient serum and compactin to induce fully the expression of the LDL-receptor gene. First, the rate of saturable degradation with increasing concentrations of labeled LDL was determined for each cell line (for examples, see Fig 5⇓). The apparent affinity of the LDL receptor for LDL, as indicated by the concentration of LDL that gave the half-maximum rate of degradation, ranged from 1.5 to 3.0 μg of LDL protein/mL (average, 2.04±0.72), but there were no consistent differences between cell lines heterozygous for different mutations in the LDL-receptor gene nor between cell lines from the patients and those from eight normolipidemic controls. In subsequent experiments, the maximum rate of degradation was determined at a concentration of 25 μg of labeled LDL protein per mL (Table 1⇑).
Mean degradation (± SD) of 125I-labeled LDL by cells from patients with known mutations in the LDL-receptor gene was 63.0±14.4% (range 41.9 to 98.2) of the mean of the eight normal cell lines (100±10.9%, range 85 to 122) (significantly different in unpaired t test, P≤.0001). Degradation by cells from six of the patients with no detectable mutation was only 55.07±4.82% of the normal range (significantly different from normal in unpaired t test, P≤.0001). In contrast, cells from the remaining eight patients, degraded an amount of 125I-LDL that was greater than 75% (range 75.3% to 107.5%) of the normal mean (not significantly different from normal in unpaired t test). There was no correlation between the age of the patients and the degradation of LDL by their cells (r2=.076, NS in Fisher’s R to Z test), and there were insufficient female patients to determine whether gender had any significant influence on degradation of LDL by cells from patients with or without known mutations in the gene.
The cell content of LDL-receptor protein with electrophoretic mobility of the normal mature receptor was determined by semi-quantitative immunoblotting of solubilized cells with two different monoclonal antibodies to the bovine LDL receptor (Table 1⇑). Fig 6a⇓ shows that there was a high correlation between the amount of protein detected with the two antibodies for all the heterozygous cell lines, and, thus, none of these cells appeared to produce a mutant protein that was detected by one antibody but not the other. The amount of mature LDL-receptor protein detected per mg of cell protein varied from 38.4 to 107% of the mean of eight normal cell lines. Fig 6b⇓ shows the relationship between the amount of immunodetectable mature LDL-receptor protein in the cells and the ability of the cells to degrade labeled LDL. It is clear that some of the points fall outside the 95% confidence limits, suggesting that these cell lines contain LDL-receptor protein that does not make a functional contribution to receptor-mediated endocytosis and subsequent degradation of the ligand. These include cells from the patients heterozygous for the Trp66Gly and Asp200Asn mutations (Nos. 27 and 3), and from seven patients in whom a mutation has not been identified (No. 14, 22, 24, 34, 35, 36, and 37).
Several of the cell lines contained detectable amounts of an LDL-receptor protein with electrophoretic mobility similar to that of the normal precursor (listed in Table 1⇑), suggesting that the predicted amino acid substitutions in these individuals resulted in failure of the precursor form of the LDL-receptor protein to be processed to the mature form at the normal rate. A point mutation or a minor deletion in the LDL-receptor gene was detected in all these cell lines (Glu387Lys; Trp599Arg; Cys88Tyr; Cys660Ser; Pro587Arg; 24bp deletion in exon 7).
Effect of the Mutation in the LDL-Receptor Gene on LDL-Receptor Activity in Cultured Lymphoblasts
We next determined whether the type of mutation in the LDL-receptor gene predicted LDL-receptor activity in the cells. The cultured cells were divided into three groups, depending on the predicted or known effect of their mutation on LDL-receptor phenotype. Mutations predicted to produce little or no functional LDL receptor protein were defined as “severe.” Less deleterious or “mild” mutations were predicted to result in a single amino acid substitution in regions of the protein whose integrity is known to be less important for normal LDL-receptor function. “Severe” mutations included those predicted to introduce a premature stop codon, deletions that have been shown to result in an unstable mRNA or truncated protein, single amino acid substitutions in the EGF-precursor domain that result in a precursor that fails to form a mature protein, and a mutation in exon 4 that is predicted to result in substitution of Asp200, a highly conserved negatively charged residue in repeat 5 of the binding domain that is critical for binding activity (see Table 1⇑). “Mild” mutations included single amino acid substitutions in repeats 2 or 3 of the binding domain or in the O-linked sugars domain and a point mutation in the promoter that reduces, but does not abolish, the rate of transcription. The third group of mutations were those predicted to result in a single amino acid substitution in the EGF precursor domain of the protein but which did not result in the accumulation of a precursor protein to detectable levels; it was not possible to assign these to either of the other two groups on the basis of the experimental evidence available. These groups, singly or combined, were compared with a group comprising cells with no detectable mutation.
Mean degradation of LDL was significantly lower (P=.004) in cells from patients heterozygous for a severe mutation (mean 53.2±7.2% of normal) than in cells from patients heterozygous for a mild mutation (mean 76.1±13.7% of normal), as shown in Fig 7⇓. Cells from the patients heterozygous for a mutation in the EGF-precursor domain whose effect on receptor function is difficult to predict degraded 63.5±13.1% of the mean of that degraded by cells from normolipidemic individuals; this value did not differ significantly from the mean for other groups of patients. The values for three of this group were between 60 and 80%, but cells from the patient heterozygous for the Trp599Arg mutation degraded only 46.7% of normal, suggesting that the precursor protein produced from this mutant allele may be very unstable or that the mature mutant protein fails to recycle to the cell surface after internalization.
Mean degradation of LDL by cells from patients in whom we have failed to detect a mutation in the LDL receptor gene (73.9±20.0%) was significantly below the normal mean, even though degradation by several of the cell lines was apparently within the normal range, as described above.
In this study we have defined the nature of the underlying genetic defect in 28 of a group of 42 patients who fulfil the clinical criteria for heterozygous FH and who had been selected previously for a clinical trial because they had coronary heart disease documented by angiography. Despite rigorous analysis of the LDL-receptor gene, a defect in this gene has been identified in only 23 of the 42 patients; four additional patients were found to be heterozygous, and one homozygous, for the Arg3500Gln mutation in the apoB gene, but none for two other less common mutations in the apoB gene that are known to cause hypercholesterolemia.
There is little doubt that the mutations we have found in these patients are the underlying cause of their familial hypercholesterolemia. Four of the mutations have been characterized previously by us and three by others (ie, G+1A in intron 3,19 T-45C,25 deletion of exon 5,22 2bp deletion in exon 4,26 Glu207Stop,16 Glu387Lys,17 Trp66Gly31). The mutations that have not been described previously include a major deletion (exon 13 and 14) or minor deletions and insertions that are predicted to produce an abnormal or truncated protein (ie, in exons 7, 10, 13, and 14). The remainder are single base substitutions that occur in codons in which other mutations have been found that affect LDL receptor function (ie, Asp200Asn, Cys660Ser, Cys88Tyr32) or are predicted to substitute residues that are strongly conserved in the LDL-receptor protein in different species or between members of the human LDL-receptor gene family.33 The only exception is the mutation in exon 15 that is predicted to result in substitution of Arg723Gln, which lies in the O-linked sugars domain of the protein. Mutations in this domain, which has been considered to be relatively unimportant for LDL-receptor function,34 are an uncommon cause of FH. However, no other base substitutions were detected in mRNA from this patient and the mutation is not a common polymorphism because it was not detected during screening of 300 heterozygous FH patients attending clinics in the London area.17
We have shown in this study that the nature of the mutation in the LDL receptor gene influences the residual LDL-receptor activity in cultured lymphoblastoid cells from the patient, measured as the maximum rate of saturable degradation of 125I-labeled LDL by cells preincubated in compactin plus lipoprotein-deficient serum. Cells from patients heterozygous for a mutation that is predicted to essentially abolish LDL receptor protein production clearly degraded less LDL than cells from patients predicted to have a single amino acid substitution in a region of the protein that is known to be less important for its physiological function. Although this is not a surprising finding, reproducible measurement of LDL-receptor activity in cultured lymphoblastoid cells is not straightforward. Considerable variation can occur between cells from normolipidemic individuals, and differences between individuals heterozygous for a mutation in the LDL-receptor gene can be masked by the activity from the normal allele. Some of the variation between cell lines may reflect differences in the expression of the normal LDL-receptor resulting from different rates of growth of the cells although care was taken in this study to ensure that any such differences were kept to a minimum. It was surprising that a few of the cell lines with clearly defined mutations in the LDL-receptor gene apparently degraded LDL at a rate essentially within the normal range. Most notable in this respect were the cells from the patient with a point mutation in the promoter that reduces but does not abolish transcription of the gene, which degraded a mean of 98% of normal. Under the cholesterol-depleted cell culture conditions, the rate of transcription was presumably sufficient to produce the normal level of LDL-receptor activity from this allele. Other cell lines, mostly with no detectable genetic defect, degraded less LDL than expected from the amount of immunodetectable LDL-receptor protein in their cells, and it is possible that some of these cells were from individuals who have a defect other than in the LDL-receptor gene that reduces the ability of the cells to carry out LDL receptor-mediated endocytosis.
Several explanations are possible for our failure to find the underlying molecular defect in 14 patients. The first is that we simply failed to detect point mutations in the coding region, promoter, or 5′-flanking sequence of introns of the LDL-receptor gene with the methods employed here, but we consider this to be very unlikely. The second is that the patients do not have a mutation in the LDL-receptor gene, and we consider this to be the most probable explanation for patients 40, 16, 25, and possibly 7, whose cells degraded LDL at a rate within the normal range and contained approximately equal amounts of two RNA species for the LDL receptor. Furthermore, these four patients only fulfilled the clinical criteria for possible FH, because tendon xanthoma were not detected in the patient or a first degree relative (Table 1⇑). Patients 36 and 26 should probably also be included in this group although analysis of mRNA was not informative because they were homozygous at all known polymorphic sites in the transcribed part of the gene, and patient 26 did have tendon xanthoma. Furthermore, we were able to show that hypercholesterolemia did not co-segregate with an allele of either the LDL receptor or apoB genes in three generations of the family of patient 36. Nonetheless, LDL receptor activity in cultured cells apparently within the normal range does not necessarily exclude the presence of a defect in the LDL receptor gene and a diagnosis of FH, as shown in our study by cells from two patients heterozygous for a defective LDL receptor allele.
At the other extreme are four patients (nos. 14, 13, 24, and 34) with no apparent molecular defect whose cells showed defective LDL-receptor function. Although only one of these had tendon xanthoma, we believe it is possible that these patients are carriers of a mutation in a gene that affects LDL-receptor expression. In two further patients, no. 35 and 37, LDL receptor-mediated degradation of LDL was clearly reduced, but no informative polymorphisms were present to determine whether gene expression was normal from both alleles of the LDL receptor gene. The situation is less clear in the remaining three patients (Nos. 5, 22 and 40), whose cells degrade about 75% of normal and in whom analysis of mRNA was also uninformative. In some of these patients, the amount of mature LDL-receptor protein in their cells, determined by immunoblotting, was within the normal range, suggesting that LDL receptor-mediated uptake could be impeded by some unrelated gene defect.
Clearly some of the uncertainty about the likely presence or absence of a mutation in the LDL-receptor gene as the underling cause of the hypercholesterolaemia could be resolved if it was possible to establish or exclude linkage between an allele of the LDL-receptor gene with hypercholesterolemia in the family of the patient. Unfortunately, because of the ages of the subjects, insufficient family members of most of these patients were available to carry out this analysis, which normally requires a minimum of both parents and at least one sibling of the patient to be available for study. However, in one family in which the index patient had a diagnosis of possible FH, we were able to show that it was unlikely that hypercholesterolemia was attributable to defective allele of either the LDL receptor gene or the apoB gene.
Overall, our study highlights the difficulties involved in defining the nature of the genetic defect in patients with a clinical diagnosis of familial hypercholesterolemia attending a lipid clinic that specializes in this disorder, even after detailed genetic and biochemical analysis. Indeed, several conclusions that are relevant for the diagnosis of FH and for the identification of affected relatives of patients attending lipid clinics can be drawn from this study. First, measurement of LDL-receptor activity in isolated cells cannot always provide a sound basis for establishing the presence or absence of defective LDL-receptor function caused by a mutation in the LDL-receptor gene. Secondly, a surprisingly high proportion of typical patients with a clinical diagnosis of FH (10% to 15% in this study) may have defects in the LDL-receptor gene that are not amenable to detection by nucleotide sequence analysis of the coding region, the intron:exon junctions, and the proximal promoter region of the gene. Although the diagnostic criteria could be interpreted differently in the two lipid clinics involved in this study, there was no significant difference in the proportion of patients from each clinic, in whom we failed to detect a genetic mutation (χ2=1.20). Finally, a significant minority of patients (approximately 5% in this study) who appear to fulfil the clinical criteria for FH and who have angiographically documented premature coronary disease do not have defective LDL-receptor function or a detectable mutation in the LDL-receptor gene or a mutation in the region of the apoB gene that codes for the known LDL-receptor binding site. Clearly, any intention to identify and provide preventive treatment for all individuals with FH in the community will need to take these findings into account.
We are grateful to Dr B. Harding (MRC Molecular Medicine, Clinical Sciences Center) for isolating lymphoblastoid cell lines. This research was supported by the British Heart Foundation (project grant PG 93005).
Goldstein JL, Hobbs H, Brown, MS. Familial hypercholesterolaemia. In. Scriver CR, Beaudet AL, Sly WS, Valle D, eds. The Metabolic Basis of Inherited Disease. New York: McGraw-Hill; 1995:1215-1250.
Brown MS, Goldstein JL. A receptor-mediated pathway for cholesterol homeostasis. Science. 1986;232:34-47.
Koivisto PVI, Koivisto U-M, Kovanen PT, Gylling H, Miettinen TA, Kontula K. Deletion of exon 15 of the LDL receptor gene is associated with a mild form of familial hypercholesterolemia-FH-Espoo. Arterioscler Thromb. 1993;13:1680-1688.
Kotze MJ, De Villiers WJS, Steyn K, Kriek JA, Marais AD, Langenhoven E, Herbert JS, Graadt van Roggen JF, Van der Westhuyzen DR, Coetzee GA. Phenotypic variation among familial hypercholesterolemics heterozygous for either one of two Afrikaner founder LDL receptor mutations. Arterioscler Thromb. 1993;13:1460-1468.
Gudnason V, Day INM, Humphries SE. Effect on plasma lipid levels of different classes of mutations in the low-density- lipoprotein receptor gene in patients with familial hypercholesterolemia. Arterioscler Thromb.. 1994;14:1717-1722.
Moorjani S, Roy M, Torres A, Betard C, Gagne C, Lambert M, Brun D, Davignon J, Lupien P. Mutations of low-density-lipoprotein-receptor gene, variation in plasma cholesterol, and expression of coronary heart disease in homozygous familial hypercholesterolaemia. Lancet. 1993;341:1303-1306.
Scientific Steering Committee on behalf of the Simon Broome Register Group. Risk of fatal coronary heart disease in familial hypercholesterolaemia. BMJ. 1991;303:893-896.
Hansen PS, Rudiger N, Tybjaerg HA, Faergeman O, Gregersen N. Detection of the apoB-3500 mutation (glutamine for arginine) by gene amplification and cleavage with MspI. J Lipid Res. 1991;32:1229-1233.
Pullinger CR, Hennessy LK, Chatterton JE, Liu W, Love JA, Mendel CM, Frost PH, Malloy MJ, Schumaker VN, Kane JP. Familial ligand-defective apolipoprotein B. Identification of a new mutation that decreases LDL receptor binding affinity. J Clin Invest. 1995;95:1225-1234.
Boerwinkle E, Xiong WJ, Fourest E, Chan L. Inheritance of the 3′-VNTR TI: Rapid typing of tandemly repeated hypervariable loci by the polymerase chain reaction: application to the apolipoprotein B 3′ hypervariable region. Proc Natl Acad Sci U S A. 1989;86:212-216.
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.
Sun X-M, Patel DD, Webb JC, Knight BL, Fan L-M, Cai H-J, Soutar AK. Familial hypercholesterolemia in China: identification of mutations in the LDL receptor gene that result in a receptor negative phenotype. Arterioscler Thromb. 1994;14:85-94.
Webb JC, Sun X-M, McCarthy SN, Neuwirth C, Thompson GR, Knight BL, Soutar AK. Characterisation of mutations in the low density lipoprotein (LDL)-receptor gene in patients with familial hypercholesterolaemia (FH) and frequency of these mutations in FH patients in the UK. J Lipid Res. 1996;37:368-381.
Day INM, Haddad L, O’Dell SD, Day LB, Whittall RA, Humphries SE. Identification of a common low density lipoprotein receptor mutation (R329X) in the south of England: complete linkage disequilibrium with an allele of microsatellite D19S394. J Med Genet. 1997;34:111-116.
Sun X-M, Patel DD, Bhatnagar D, Knight BL, Soutar AK. Characterization of a splice-site mutation in the gene for the LDL receptor associated with an unpredictably severe clinical phenotype in English patients with heterozygous FH. Arterioscler Thromb Vasc Biol. 1995;15:219-227.
Gaffney D, Reid JM, Cameron IM, Vass K, Caslake MJ, Shepherd J, Packard CH. Independent mutations at codon 3500 of the apolipoprotein B gene are associated with hyperlipidemia. Arterioscler Thromb Vasc Biol. 1995;15:1025-1029.
Hobbs HH, Brown MS, Goldstein JL, Russell DW. Deletion of exon encoding cysteine-rich repeat of low density lipoprotein receptor alters its binding specificity in a subject with familial hypercholesterolemia. J Biol Chem. 1986;261:13114-13120.
Sudhof TC, Goldstein JL, Brown MS, Russell DW. The LDL receptor gene: a mosaic of exons shared with different proteins. Science. 1985;228:815-822.
Sun X-M, Neuwirth C, Wade D, Knight BL, Soutar AK. A mutation (T-45C) in the promoter region of the low-density-lipoprotein (LDL) receptor gene is associated with a mild clinical phenotype in a patient with heterozygous familial hypercholesterolaemia. Hum Mol Genet. 1995;4:2125-2129.
Gudnason V, King UL, Seed M, Sun X-M, Soutar AK, Humphries SE. Identification of recurrent and novel mutations in exon 4 of the LDL receptor gene in patients with familial hypercholesterolemia in the United Kingdom. Arterioscler Thromb. 1993;13:56-63.
Webb JC, Patel DD, Shoulders CC, Knight BL, Soutar AK. Genetic variation at a splicing branch point in intron 9 of the low density lipoprotein (LDL)-receptor gene: a rare mutation that disrupts splicing and a common polymorphism. Hum Mol Genet.. 1996;5:1325–1331.
Webb JC, Sun X-M, Patel DD, McCarthy SN, Knight BL, Soutar AK. Characterisation of two new point mutations in the low density lipoprotein receptor genes of an English patient with homozygous familial hypercholesterolemia. J Lipid Res. 1992;33:689-698.
Soutar AK, Patel DD, Knight BL. Identification of a point mutation in growth factor repeat C of the low density lipoprotein receptor-gene in a patient with homozygous familial hypercholesterolaemia that affects ligand binding and intracellular movement of receptors. Proc Natl Acad Sci U. S. A.. 1989;86:4166-4170.
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.
Webb JC, Patel DD, Jones MD, Knight BL, Soutar AK. Characterization and tissue-specific expression of the human ‘very low density lipoprotein (VLDL) receptor’ mRNA. Hum Mol Genet. 1994;3:531-537.
Davis CG, Elhammer A, Russell DW, Schneider WJ, Kornfeld S, Brown MS, Goldstein JL. Deletion of clustered O-linked carbohydrates does not impair function of low density lipoprotein receptor in transfected fibroblasts. J Biol Chem. 1986;261:2828-2838.