Editor—Familial hypercholesterolaemia (FH) is an autosomal dominant inherited lipoprotein disorder characterised by raised plasma low density lipoprotein (LDL) levels, xanthomas, premature coronary heart disease, and a family history of one or more of these. Homozygous FH occurs in one in a million people and they are severely affected, while heterozygotes are moderately affected and occur at a frequency of 1 in 500 in genetically heterogeneous populations. FH is caused by a mutation in the LDL receptor gene (LDLR) and over 700 have been reported1 (http://www.ucl.ac.uk/fh). Among these, a missense mutation, T705I in exon 15 (FH Paris-9), was originally reported in a compound heterozygote (“homozygous” FH subject) of French-American origin,2 but has now been observed in several heterozygotes who also carry another mutation in the coding region of the LDL receptor protein.3 4 The presence of the I705 variant has also been reported in two normocholesterolaemic subjects in the heterozygous and homozygous form, which led to the suggestion that the T705I change is a non-FH causing variation.5 The most recent report of this variation was in a Spanish family where the hypercholesterolaemia segregated with the I705 substitution and no other mutation was identified.6Possible explanations for these contradictory findings have been that the exon 15 variant is only pathogenic when another environmental or genetic factor is present, or that in some subjects the I705 variant is in linkage disequilibrium with a second, as yet unidentified causative mutation.

We have set up a clinical genetic diagnostic service for FH7 8 and the I705 variant was identified in a subject with a clinical diagnosis of possible FH who was referred for FH genetic testing (data not shown). To investigate the pathogenicity of the amino acid substitution at codon 705, we have determined the frequency of the I705 variant in 2287 healthy UK men and examined the effect of this variant on plasma lipid levels.

An assay was designed for the T705I substitution where anNsiI restriction site was introduced into the rare C allele (I) by a mismatch in the sense primer (underlined): sense primer: 5′-CAG TGG CCA CCC AGG AGA CAT GCA-3′ and antisense primer: 5′-ATC TCC ACC GTG GTG AGC CCA-3′. PCR conditions were as described previously, the NsiI enzyme was added directly to the PCR product, and the products (139 bp uncut = T705 and 115 bp + (non-detected) 14 bp cut) were separated on a 7.5% MADGE9and stained with ethidium bromide. The I705 carriers were then analysed for the 1061-8C variation in intron 7 by a natural EarI restriction digest, after amplification of exon 8 (197 bp) with primers FH119 (FH website) and FH27 (FH website) and reaction conditions as previously described.8 A restriction site was lost in the rare C allele but a control constant cut site exists for confirmation of digestion. Fragments were analysed on a 10% acrylamide gel as above giving fragments of 183 bp (1061-8C) and 153 bp + 30 bp (non-detected) (1061-8T).

The sample studied consisted of 2287 white men from the Northwick Park Heart Study (NPHS-II).10 Exclusion criteria included non-whites, a history of unstable angina or myocardial infarction, regular medication with aspirin or anticoagulants, cerebrovascular disease, malignancy (except skin cancer other than melanoma), diseases exposing staff to risk of infection, mental disorder or other conditions precluding informed consent, or regular attendance for examination.10 Plasma cholesterol and triglycerides were measured as previously reported10 and genomic DNA was isolated by standard methods.11

As shown in table 1, 30 carriers were found, and thus the carrier frequency of the I705 variant was 1.3%. In a study from The Netherlands, 100 normolipidaemic controls were screened by DGGE and sequencing and the I705 variant was found in two subjects, one heterozygous carrier and one homozygote.3 Therefore, the carrier frequency of the I705 allele was 2% which is very similar to the estimate in the larger group of UK men. All but two I705 carriers from the NPHS-II group carried the 1061-8C variation in intron 7, which has been associated with this exon 15 variation.4 The 1061-8C variant was screened in 200 men from the NPSH-II group and the variant was only detected in the I705 carriers, so the carrier frequency was estimated to be ∼1.3% in the general population. No statistically significant differences were observed between the mean total cholesterol and triglyceride levels of the I705 carriers and the non-carriers.

These data strongly suggest that the I705 variant is not having a major effect on LDL receptor function. It is the second “non-functional” variant described to date, with only the A370T being previously known. T370 occurs at a frequency of 6% in the UK12 and is associated with, at most, only a modest effect on plasma lipid levels.13 Cell studies have not detected a significant impairment of LDL receptor function of the T370 substitution.13

In the original report of this mutation, the T705I substitution in the coding region of LDLR was thought to be one of the defective alleles in a compound heterozygote, while the second mutation remained undetected.2 It is reasonable to assume from the Dutch and our data that neither defect in this homozygote had been identified. The second case occurred in a 40 year old man who had a very high total cholesterol of 17.78 mmol/l.3 5 A splice site mutation (313+1G>A) was inherited from his hypercholesterolaemic mother, while the I705 variant came from his normolipidaemic father. A proposed explanation for these observations is that the I705 variant is only expressed when another LDLR defect is present, but two of the proband's younger sibs also had slightly raised cholesterol yet did not share an LDLR haplotype with their normocholesterolaemic sib, that is, they did not carry either of theLDLR variants. Thus, another variation inLDLR or in another gene may be responsible for the hypercholesterolaemia in these two sibs and would also explain the high cholesterol level in the proband. In the most recent report,6 the I705 substitution tracked with the hypercholesterolaemia phenotype, but again this could be explained by the presence of an unidentified mutation which may or may not be in linkage disequilibrium with the I705 variation. The intron 7 T1061-8C sequence change would be a candidate for such a mutation, since the rare alleles have been shown previously to occur together4and strong allelic association has been confirmed in this sample of men. However, since carriers of the 1061-8C allele do not have raised plasma cholesterol levels it appears unlikely that this sequence change is of functional importance.

Exon 15 consists of 171 nucleotides which encode 57 amino acids, of which 18 are threonine or serine residues,14 15 and most of the O linked sugars of the LDL receptor are attached to these threonine and serine residues.15 A similar region is also conserved in the LDL receptors of other mammals.15 The functional role of this domain was investigated by Davis et al 15 using site directed mutagenesis, where a portion of exon 15 was deleted and then expressed by transfection into fibroblast cell lines. The mutated cDNA coded for a receptor protein which was functionally indistinguishable from the normal receptor.15 Subjects carrying a similar natural deletion of exon 15 (FH Espoo) have LDL levels which are relatively low and a mild form of FH.16 Thus, a major rearrangement is actually a mild mutation, suggesting that mutations in this region may only have a mild effect on receptor function and therefore on lipid levels. Only six point mutations or single base deletions have been described in exon 15 (http://www.ucl.ac.uk/fh), two point mutations resulting in a stop codon, a minor deletion predicted to result in frameshift, and a splice donor site mutation, and all of these are highly likely to be pathogenic. In addition to the T705I substitution, two missense mutations have been reported, T721I17 and R723Q.18 19 LDL binding studies showed that the Q723 mutation had 70% of normal activity and is therefore a mildLDLR mutation. No details are known about the I721 mutation. It may be functional in that an O linked sugar may attach at this site but it may be non-pathogenic as with T705, which involves the same amino acid substitution.

Thus, from the available data on the reported exon 15 mutations, missense mutations, and FH Espoo, they appear to have a mild effect on the LDL receptor protein. Ideally, cellular studies should be carried out on all novel mutations but this is often not feasible. However, at the very least 100 normal subjects from the particular ethnic group should be screened for any novel mutation to determine frequency, and if any carriers are detected association studies should be performed. In particular, care should be taken in reporting missense mutations identified in exon 15 of LDLR as FH causing, as they appear to have a modest effect on LDL receptor function.