Introduction

The Hutchinson-Gilford progeria syndrome (HGPS; MIM 176670) has existed as a named clinical entity for almost 100 years (Hutchinson 1886; Gilford 1904). Organ systems degenerate to such an extent that the affected subject resembles an old person. Clinical features include short stature, micrognathia, alopecia, prominent scalp veins, prominent joints, hyperlipidemia and atherosclerosis, often with premature death from coronary artery disease (de Paula Rodrigues et al. 2002). HGPS was proposed to have autosomal recessive inheritance (Gabr et al. 1960), but other mechanisms, such as chromosomal rearrangements or uniparental disomy, appear to be important in some instances (Brown et al. 1990; Luengo et al. 2002; Eriksson et al. 2002). No mutations have yet been reported in affected subjects, but there are clues to the molecular genetic basis of HGPS. For instance, HGPS patients can have an absence of subcutaneous fat reminiscent of lipodystrophy (Paterson 1922). In some HGPS patients (Parkash et al. 1990) the clinical presentation resembled mandibuloacral dysplasia with partial lipodystrophy (MIM 248370), which is caused by mutant LMNA, encoding nuclear lamin A/C on chromosome 1q21 (Novelli et al. 2002). Also, some HGPS patients had an inverted insertion of 1q (Brown et al. 1990) and another had an interstitial deletion involving 1q23 (Luegno et al. 2002). Yet another HGPS patient had uniparental disomy of 1q, involving the region flanked by markers D1S498 and D1S2836 (Eriksson et al. 2002), which harbours LMNA. Finally, there have been direct suggestions that LMNA mutations underlie HGPS (http://grants.nih.gov/grants/guide/pa-files/PA-03-069.html and Eriksson et al. 2003). Thus, LMNA is an excellent candidate gene for HGPS.

The Wiedemann-Rautenstrauch syndrome (WRS; MIM 264090) is another extremely rare phenotype that is characterized by progeroid features from birth (Rautenstrauch et al. 1977; Wiedemann 1979). This is associated with a spectrum of clinical features, including delayed psychomotor development and physical growth, alopecia, macrocephaly and lipoatrophy (Arboleda et al. 1997; Bitoun et al. 1995; Castineyra et al. 1992; Devos et al. 1981; Martin et al. 1984; Pivnick et al. 2000; Toriello 1990). No genomic DNA mutations have yet been reported in WRS, but because lipodystrophy is a central feature, genes that cause lipodystrophy can be considered as candidates. Based on these lines of evidence, we sequenced LMNA from genomic DNA of cell lines from HGPS and WRS patients.

Materials and methods

Study subjects

Genomic DNAs from HGPS probands NG07091, NG03506, NG03259, NG03344, NG10579, NG10587 and NG10801, and the fibroblast cell lines from WRS probands AG09233 and AG13207 were obtained from the Coriell Cell Repositories (Camden, N.J.). The clinical attributes of the HGPS and WRS probands were described in detail on the Coriell Cell Repository website (http://locus.umdnj.edu/ccr/) and the key attributes are summarized in Table 1. All HGPS and WRS probands studied were unrelated Caucasians. All samples had been deposited at the Coriell Cell Repositories between 1978 and 1990.

Table 1. Clinical and molecular features of patients with Hutchinson-Gilford progeria (HGPS) and Wiedemann-Rautenstrauch progeroid (WRS) syndromes
Full size table

Mutation-containing exons were subsequently screened in samples from 100 clinically normal Caucasian subjects in order to determine whether any observed genomic variant was also present in the normal population. The University of Western Ontario Ethics Review Panel had approved the study.

Screening LMNA for DNA variants

Established methods were used to extract genomic DNA, then to amplify genomic DNA and to directly sequence the LMNA coding sequence and promoter (Cao and Hegele 2000). Mutations were identified using Sequence Navigator software (PE-Applied Biosystems, Mississauga, Ont.).

Determining the phase of LMNA mutations in HGPS proband NG07091

Because two mutations were seen in the subject NG07091, subcloning was performed to separate the alleles in order to determine whether the mutations were in a cis or trans relationship. A 1,069-bp fragment containing LMNA exons 8–10 was amplified from the genomic DNA of HGPS proband NG7091 using primers F-5′-GCAAGATACACCCAAGAGCC-3′ and R-5′-ACACCTGGGTTCCCTGTTC-3′. The amplified product was subcloned into the pCR2.1 vector (TA cloning kit; Invitrogen, Carlsbad, Calif.) and purified after EcoRI digestion of plasmid DNA. Five gel-purified plasmid inserts were each split into two aliquots. Because the R471C (c.1623C>T) predicted loss of an HhaI site and R527C (c.1791C>T) predicted the loss of a BsiWI site, each aliquot was digested either with endonuclease HhaI or BsiWI. Fragments after digestion were resolved on 2% agarose gels.

Statistical analysis

Fisher's exact test was used to compare allele frequencies between affected and normal subjects (SAS version 6.12, SAS Institute, Cary, N.C.).

Results

Identification of LMNA genomic DNA variants

Results of the analysis of genomic DNA sequence of LMNA in HGPS and WRS probands are shown in Table 1. We found that: (1) all seven genomes from HGPS probands had novel LMNA coding sequence variants; (2) neither genome from the WRS probands had a LMNA sequence variation; (3) none of the 100 normal control genomes had any of these novel LMNA variants. The finding that each HGPS proband had at least one of these rare LMNA mutations indicates a strong, non-random association between the disease and the mutations (P<4×10−11, Fisher's exact test). One HGPS proband (NG10801) was a simple heterozygote for a novel missense mutation in exon 11, namely G608S (c.2034G>A). Five HGPS probands were simple heterozygotes for a novel synonymous mutation in exon 11, namely c.2036C>T.

In order to determine whether mutations R471C and R527C in HGPS proband NG07091 were in a cis or trans relationship, the genomic region spanning LMNA exons 8–10 was amplified and then subcloned. Aliquots of five plasmid inserts were digested with HhaI: four had the expected fragment sizes of 585, 245, 97, 91, 39 and 12 bp, while the fifth had fragment sizes of 585, 245, 188, 39 and 12 bp, consistent with a lost HhaI site resulting from R471C (c.1623C>T). Aliquots of these inserts were digested separately with BsiWI: the insert that showed the abnormal HhaI fragment pattern had the expected BsiWI fragment pattern of 721 and 348 bp, while the four fragments that showed the normal HhaI pattern had only a single 1,069-bp fragment, consistent with the loss of the BsiWI recognition site. This confirmed that HGPS proband NG07091 was a compound heterozygote for R471C and R527C missense mutations. Neither parental phenotype data nor DNA was available for further study.

Discussion

We report four novel and rare LMNA coding sequence variants, namely R471C, R527C, G608S and c.2036C>T, which were detected by sequencing genomic DNA from seven unrelated HGPS probands. The genomes of all seven HGPS probands had a variant LMNA sequence, while no genome of 100 normal control subjects had a rare LMNA variant, indicating a strong non-random statistical association between LMNA variation and HGPS (P<4×10−11). In contrast, neither genome from the two WRS probands had any LMNA sequence abnormality. The findings support the idea that HGPS is associated with genomic DNA variation of LMNA, while WRS is not.

Two HGPS probands had missense mutations in LMNA. One of these probands, NG10801, a simple heterozygote for G608S, was a six-year-old female Caucasian with mandibular dysplasia in addition to the classic features of HGPS. The other proband, NG07091, a compound heterozygote for R471C and R527C, was the oldest HGPS patient that we studied. While no DNA or clinical data from parents or family members could be analysed for this study, the findings from these two HGPS indicate that HGPS inheritance may be more complicated than classical Mendelian dominant or recessive patterns. This would be consistent with prior observations that some HGPS patients have specific cytogenetic abnormalities (Brown et al. 1990; Luengo et al. 2002) and uniparental disomy of 1q in some cases (Eriksson et al. 2002).

Since each parent of NG07091 had contributed one LMNA allele with a novel missense mutation, it would have been of interest to determine whether each also expressed any clinical features of a laminopathy. Unfortunately, this information was not readily available to us. In each of these two mutations, cysteine replaces wild-type arginine, which is a common type of substitution that is seen in other autosomal dominant laminopathies (Burke and Stewart 2002). At present, it cannot be determined whether milder laminopathies are seen in the heterozygous state for R471C or R527C, with compound heterozygosity leading to the more severe phenotype of HGPS. Alternatively, one or both of these mutations may have no associated clinical phenotype, with progeria representing a true autosomal recessive trait.

Five HGPS probands were simple heterozygotes for the rare synonymous LMNA c.2036C>T mutation in exon 11. The statistical significance of the absence of c.2036C>T from the genomes of 100 normal subjects (P<10−8, Fisher's exact test) strongly indicates mechanistic importance that is specific to HGPS. This importance is amplified when considering the complete absence of LMNA c.2036C>T from the genomes of hundreds of other subjects that were sequenced during mutational screening for disease. While silent at the amino acid level, it would be important to consider the possibility that LMNA c.2036C>T might have some other functional impact, such as an affect on RNA splicing, although functional studies to assess this possibility are beyond the scope of this preliminary report. An alternative explanation is that there might be linkage disequilibrium between c.2036C>T and another unmeasured LMNA mutation. However, this possibility is made somewhat less likely by our sequencing strategy, which captured ~700 bp of the LMNA promoter, ~500 bp of the LMNA 3′-untranslated region and ~100 bp of intronic sequence at each intron-exon boundary.

The distribution of the HGPS mutations in LMNA suggests that lamin A, specifically, is the affected isoform in this disease (Burke and Stewart 2002). There are no previously reported mutations at LMNA codons 471 and 608; however, R527P was found in a patient with autosomal dominant Emery-Dreifuss muscular dystrophy (Bonne et al. 1999). This indicates that different mutations at codon 527 are associated with different phenotypes. Furthermore, it seems an interesting coincidence that the nucleotide sequence of codon 608 is altered in both the G608S and c.2036C>T mutations. Codon 608 is specific to the lamin A isoform and the G608S and c.2036C>T mutations are among the most 3′ reported in LMNA. These two mutations have apparently been observed in other HGPS subjects (Eriksson et al. 2003).

The coexistence of lipodystrophy and progeroid features in WRS made LMNA a reasonable candidate gene to have considered in this syndrome. Assuming no genetic heterogeneity, our findings suggest that LMNA mutations are not associated with WRS, although the number of WRS probands that we screened was admittedly small. However, WRS is even rarer than HGPS, so while obtaining more substantial numbers of subjects for additional genomic screening of LMNA and other candidate genes would be desirable, it might be difficult.

In conclusion, our findings suggest that HGPS is a laminopathy, bringing to seven the number of inherited diseases that result from LMNA mutations (Burke and Stewart 2002). The mechanism(s) by which the observed LMNA mutations in HGPS might cause disease, like the LMNA mutations in all laminopathies, is unclear. The LMNA missense mutations might affect interactions with transcription factors that would determine the specific clinical outcomes (Burke and Stewart 2002). The exon 11 c.2036C>T mutation might affect RNA splicing, but an in vitro assay of this function would be required to prove this. In contrast, the related WRS phenotype does not appear to be a laminopathy. The growing list and clinical relevance of laminopathies will hopefully intensify the efforts to understand how mutations in LMNA lead to the specific clinical phenotypes.