Editor—With a prevalence of 1/6000 live births, spinal muscular atrophy (SMA) represents the second most common fatal autosomal recessive disorder after cystic fibrosis.1 2 SMA is characterised by the degeneration of anterior horn cells of the spinal cord, resulting in progressive, symmetrical limb and trunk paralysis associated with muscular atrophy. This condition is clinically heterogeneous and has been subdivided into three types according to age of onset and clinical course3: type I (Werdnig-Hoffmann disease, MIM 253300), type II (intermediate form, MIM 253550), and type III SMA (Kugelberg-Welander disease, MIM 253400). TheSMA locus has been mapped to chromosome 5q11.2-q13.3 within a region characterised by the large inverted duplication of a 500 kb element.4-6 The survival motor neurone (SMN) gene, which lies within this element, is duplicated and both copies are expressed. The telomeric gene (SMN1) has been shown to be deleted or mutated in all three types of SMA.4 SMN1 encodes almost the full length transcript whereas the centromeric copy (SMN2) generates alternatively spliced variants lacking the C-terminal sequence.5 7 TheSMN region contains low copy repeats triggering homologous recombination events. Indeed, approximately 95% of SMA patients lack both SMN1 genes owing to either deletion or gene conversion.4 In SMA patients who lack only one SMN1 gene, allelic intragenic mutations have been identified, confirming the involvement of SMN1 in the pathogenesis of SMA.5 8-10

The heterozygote frequency has been estimated to be 1/40. However, the duplication of the SMA locus makes the detection of SMA carriers in the general population difficult, and this has hampered genetic counselling in affected families. Initial attempts to estimate theSMN copy number were based on the measurement of theSMN1/SMN2ratios,11-13 but the broad variability ofSMN2 copy number hinders reliable quantification. For this reason, subsequent studies have included two internal standards in the PCR reaction, corresponding to the modifiedSMN1 and CFTRsequences, respectively.10 14 15 In these methods, the quantification of SMN copies is based on the ratio between the PCR amplification of the specific genomic DNA and that of an internal standard for each subject tested. The results are normalised to the mean of control samples. Although these methods can efficiently detect heterozygous SMN1deletions,10 14 15 overlaps between carriers and non-carriers have been observed.10

In the present study, we describe a novel method which allows easy detection of heterozygous SMN1 deletions in SMA carriers and SMA patients without homozygousSMN1 deletions. We devised a multiplex PCR assay of fluorescent fragments based on the approach that we initially developed for the detection of mismatch repair gene rearrangements in hereditary non-polyposis colorectal cancer.16 We simultaneously amplified exon 7 of the SMN1and SMN2 genes using a mismatch primer X7-Dra, which introduced a DraI restriction site into amplified SMN1 exon 7,17 BRCA1 exon 11, andMLH1 exon 18, which contains a natural internal DraI restriction site (table 1). The PCR reaction was performed in a final volume of 50 μl, using 0.75 μmol/l SMN primers, 0.5 μmol/lBRCA1 primers, 0.35 μmol/lMLH1 primers, 0.2 mmol/l dNTP, 1.5 mmol/l MgCl₂, 1 unit of Taq polymerase (Eurobio, Les Ulis, France), and 100 ng of genomic DNA. The PCR consisted of 20 cycles of 94°C for 15 seconds, 55°C for 15 seconds, and 72°C for 15 seconds, preceded by an initial denaturation step of five minutes at 94°C and followed by a final extension of five minutes at 72°C. The entire PCR reaction was then digested using 4 units of DraI (New England Biolabs) in a total volume of 150 μl for at least four hours. After purification using the Qiagen Gel Extraction Kit, PCR products were resuspended in a mix containing 2.5 μl of deionised formamide, 0.5 μl of GeneScan-500 Rox (PE Applied Biosystems, Perkin Elmer), and 1 μl of loading buffer. After denaturation for two minutes at 90°C, 2 μl of each sample was loaded onto a 4.25% denaturing polyacrylamide gel (Sequagel). Electrophoresis was performed for three hours on an Applied Biosystems model 377 automated sequencer (PE Applied Biosystems, Perkin Elmer). Data were analysed using the Gene Scanner Model 672 Fluorescent Fragment Analyser (PE Applied Biosystems, Perkin Elmer) and electropherograms generated from different samples were superimposed.

Each multiplex PCR yielded a pattern composed of four fluorescent peaks corresponding to exonic fragments of BRCA1,MLH1, SMN1, andSMN2 respectively and the patterns generated from two control samples could be easily superimposed (fig 1A). For validation, we studied the SMN1 andSMN2 copy numbers (fig 1) in a SMA family in which linkage analysis, using the C212 and C272 microsatellite markers,18 and analysis of theSMN1 and SMN2genes by PCR digestion had previously shown a homozygousSMN1 gene deletion in the affected child and a homozygous SMN2 gene deletion in an unaffected sib, which was suggestive of a large deletion encompassing both SMN1 andSMN2 on the paternal allele (fig 2). The relatives of this family were therefore predicted to harbour a variable number of SMN1 andSMN2 copies. Fig 1 shows that the multiplex PCR, using as a control a subject predicted to carry two copies ofSMN1 and two of SMN2,easily detected no, one, or two copies ofSMN1 or SMN2within this family. This technique confirmed the large paternal deletion and showed a gene conversion event on the mutant maternal allele. We then tested 86 parents of SMA patients carrying a homozygousSMN1 deletion (50 parents of SMA type I, 28 parents of SMA type II, two parents of SMA type III, and six parents of SMA patients of undetermined type). An approximate 0.5 reduction of theSMN1 peak area, indicative of a heterozygous deletion, was clearly observed in 80 parents (93%). TwoSMN1 copies were detected in six putative carriers. In four out of these six families, linkage analysis with the C212 and C272 microsatellite markers and quantification ofSMN1 in relatives allowed us to show the existence of two de novo deletions and twoSMN1duplications.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Quantification of SMN1 and SMN2 copies using a non-competitive fluorescent multiplex PCR assay. Genomic fragments were PCR amplified using dye labelled primers (table 1). After DraI digestion, fragments were separated by electrophoresis on an ABI 377 DNA automated sequencer and electropherograms from two subjects were superimposed. In each electropherogram, the y axis displays fluorescence intensity in arbitrary units and the x axis indicates the size (in bp) of the fragments. The expected sizes are indicated. The superimposition of two distinct samples, a control sample with two copies of SMN1 and two of SMN2 (blue) and the analysed sample (red), indicates the number of SMN1 and SMN2 copies. The result of this quantification for one normal subject and five relatives of a SMA family (fig 2) are presented. (A) Normal subject with two copies of SMN1 and two copies of SMN2. (B) Index case's father (I.1) with one copy of SMN1 and one copy of SMN2. (C) Mother (I.2) with one copy of SMN1 and two copies of SMN2. (D) Unaffected sib (II.1) with one copy of SMN1 and no copy of SMN2. (E) Affected sib (II.2) with no copy of SMN1 and two copies of SMN2. (F), Unaffected sib (II.4) with two copies of SMN1 and one copy of SMN2.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Pedigree of the SMA family used to validate the multiplex PCR. Filled symbols, affected subjects; open symbols, asymptomatic subjects. For each subject, the haplotype analysis using the C212 and C272 microsatellite markers18 and the schematic representation of the SMN locus (1: SMN1, 2: SMN2) are indicated.

In contrast to the previously reported methods,10 14 15the estimation of SMN1 copy number in this assay is based on the comparison of the fluorescence levels between theSMN1 peak generated from different samples rather than between the different peaks generated from the same sample. In order to keep PCR amplification within an exponential range, we tested various numbers of cycles (18, 20, 22, and 24) and found that 20 cycles, with shorter times of annealing and extension than those previously described, were optimal.10 14 15 The simultaneous amplification of two other fragments (BRCA1 and MLH1) allowed an accurate comparison of electropherograms generated from different samples. The absence of the 244 bp MLH1PCR product and the appearance of a 209 bp peak (fig 1), expected from DraI digestion, indicated that the enzymatic digestion was complete, a feature which is essential to distinguish between SMN1 andSMN2 amplified fragments.

The simplicity of this assay should facilitate its development in molecular diagnostic laboratories and hopefully aid in genetic counselling in SMA families. However, one must keep in mind the existence of (1) small intragenic mutations within theSMN1 gene, (2)SMN1 duplications in cis (on one chromosome) masking a heterozygous deletion on the other chromosome, (3) de novo deletions, and (4) germline mosaicism. Small intragenicSMN1 mutations account for 1.3-3.4% of the mutant SMN1 alleles and have been identified in SMA patients carrying heterozygous SMN1deletions.5 8-10 On the other hand, de novoSMN1 deletions have been shown to be involved in approximately 2% of SMA cases.15 19 In order to estimate the error risk resulting from duplication or de novo deletion, we counted SMN1 copies in 86 parents of SMA children carrying a homozygousSMN1 deletion and found that six out of 86 putative carriers (7%) had more than oneSMN1 copy. These data are in complete agreement with the results of Chen et al,15 who detected 5/60 putative SMA carriers with two copies of SMN1 (8.3% including one carrier with a small intragenic SMN1mutation, two putative carriers with a de novo deletion, and two carriers with a SMN1 duplication). Finally, germline mosaicism has to be considered.20 Despite this error risk (less than 10%), the determination ofSMN1 copy number in relatives of SMA patients, harbouring homozygous SMN1deletions, will make genetic counselling easier and hopefully limit prenatal screening. For example, for a couple with an a priori risk of 1/320 of having an affected child (corresponding to the situation of the index case's uncle or aunt), detection of twoSMN1 copies in both the relative and his/her spouse will reduce the probability of having an affected child to 1/32 000 ([1/2 × 1/10] × [1/40 × 1/10] × 1/4), which is lower than the risk of the general population. Detection of one copy in the relative and two copies in his/her spouse will decrease the risk to 1/1600 (1 × [1/40 × 1/10] × 1/4). This assay will also facilitate the detection of heterozygousSMN1 deletion in SMA patients without a homozygous SMN1 deletion who must be screened for small SMN1 mutations on the other allele, as previously shown by Wirth et al.10 Finally, this assay will allow the study of the influence of SMN2 copy numbers on the SMA phenotype for research purposes, a feature previously suggested by both the observation of an increased number ofSMN2 copies in patients with a milder phenotype5 12 14 15 21 22 and by the effect of the expression of human SMN2 inSmn ^-/- mice.23 24