Introduction

Spinal muscular atrophy is one of the most common autosomal recessive disorders, with a carrier frequency of 1/40–1/60.1 SMA is caused by the degeneration of motor neurons in the anterior horn of the spinal cord and has been classified into three types on the basis of clinical severity and age of onset: type I (Werdnig–Hoffmann disease, OMIM 253300), the most severe form; type II, an intermediate form (OMIM 253550); and type III (Kugelberg–Welander disease, OMIM 253400), the milder form.2, 3

All three SMA types are associated with mutations in the survival motor neuron 1 gene (SMN1; OMIM 600354), located in the 5q13 region, containing a 500-kb duplication and inversion.4, 5, 6, 7, 8 SMN1 has a centromeric homolog, SMN2, which modulates the phenotype severity.9, 10 SMN1 and 2 genes differ by only five nucleotides, only one of which is set in the coding sequence.11 Nucleotide variations in exons 7 and 8 are commonly used to discriminate between SMN1 and 2 genes, using a polymerase chain reaction/restriction fragment length polymorphism assay.12 The majority of SMA patients (about 94%) carry a homozygous deletion of at least exon 7 of the SMN1 gene,13 whereas the absence of the SMN2 gene has no clinical consequence in individuals carrying at least one SMN1 copy. SMN dosage gene analyses, which can determine SMN1 and SMN2 copy numbers, have been developed over the past 10 years,10, 14, 15, 16 and show that (i) SMA patients who lack only one SMN1 gene carry allelic intragenic mutations (compound heterozygotes);15 and (ii) a CIS duplication of the SMN1 gene is found in 3–8% of putative carriers.16, 17, 18

A couple was referred to our clinic for preimplantation genetic diagnosis (PGD) because their first child was affected with SMA. DNA studies confirmed the diagnosis in the proband by demonstrating a homozygous deletion of the SMN1 exon 7 gene. SMN gene dosage analysis using a fluorescent multiplex PCR method16 indicated that the mother carried only one SMN1 copy, whereas the father was shown to carry two copies of SMN. Samples from the father's parents were not available, and there was no family history of SMA. Three situations might account for the father genotype (Figure 1): (i) a CIS duplication of the SMN1 gene on one chromosome with an absence of the SMN1 copy on the counterpart (‘2+0’ genotype); (ii) the occurence of a de novo deletion of a paternal chromosome carrying one SMN1 copy (one SMN1 copy on both chromosomes, called the genotype ‘1+1’); and (iii) a false paternity. The last hypothesis was ruled out by linkage analysis using microsatellites linked to the SMN1 gene (ie, D5S1988,19 D5S629,20 C212,8 C272,8 and D5S63720), which showed a paternal contribution to the child genotype for all tested markers (data not shown, available on request).

Figure 1
figure 1

Pedigree of the family showing linkage analysis and the SMN1 copy number. Two hypotheses can account for the results of SMN copy quantification and haplotyping from somatic cells of the nuclear family: (a) a CIS duplication of the SMN1 gene on one paternal chromosome with zero SMN1 copy on the counterpart; (b) occurrence of a de novo germ cell deletion on the paternal chromosome 5. Black bars represented the affected maternal and ‘at-risk’ paternal alleles. Haplotypes at loci D5S629 (AFM265wf5) and D5S1988 (AFMa245we9) are shown from centromere to telomere, along with the results of the SMN1 gene dosage analysis.

Full size image

De novo deletions may occur in 2% or more of families with SMA,21, 22 affecting preferentially the paternal chromosome.21 It has also been shown that a small proportion of parents carry the ‘2+0’ genotype, which is present in ≈3.2% of carriers.17 Using these data for Bayesian analysis as described in Ogino et al,1 we calculated a probability of 69% that the father was a ‘2+0’ carrier, and a probability of 31% that he was a ‘1+1’ noncarrier (Figure 1).

Distinguishing between these two hypotheses was important for reproductive decision making. The SMA recurrence risk in the offspring indeed decreased from 25% (genotype ‘2+0’) to a considerably lower figure in the second case (genotype ‘1+1’), reflected by the risk of germline mosaicism for the rearrangement in the father. Any application of the couple for PGD would have been prohibited in the last case. The French Law indeed authorizes PGD only when the genetic trait has previously been identified in the parents, preventing the use of PGD procedures for a risk of germline mosaicism.

In an attempt at clarifying this issue, we developed a strategy to determine the presence or absence of the SMN1 gene and haplotypes at the SMA locus from a single sperm cells.

Materials and methods

Single-sperm cell isolation

The technique was developed using single sperm cells obtained from freshly ejaculated sperm of the father. Spermatozoa were prepared using discontinuous gradient centrifugation. Separated sperms were rinsed and diluted in drops of PBS (Sigma Aldrich, Saint-Quentin Fallavier, France) supplemented with 0.1% polyvinyl alcohol (Sigma Aldrich) until single cells could be aspirated with a pulled glass micropipette in a clean laboratory. Single spermatozoa were transferred under visual control through an inverted microscope to PCR tubes containing lysis buffer (3 μl of 200 mM KOH, and 50 mM DTT).23 Lysis was performed by heating at 65°C for 10 min. A small volume of rinsing medium was transferred similarly and used as a negative template.

Single-sperm PCR amplification

The father was heterozygous for two linked [CA]n microsatellite markers (ie, D5S62920 and D5S198819), which are located on each side of the SMN1 gene (Figure 1). As these two markers showed a good amplification rate on single lymphocytes (data not shown), and because they flank the SMN1 gene, we selected them for sperm haplotype determination.

Polymerase chain reaction amplification conditions were similar to those used for blastomere analysis in SMA PGD,24 but amplification of D5S629 and D5S1988 was added to the detection of homozygous deletion of exon 7. Such amplification reactions enabled the mutant and wild-type alleles to be linked to individual haplotypes.

The multiplex ‘outer’ PCR reaction contained 3 μl of lysis buffer (lysed spermatozoa or negative control), 2.5 μl of primer mix 10 × (containing each primer at 2 μ M, Proligo, Evry, France), master mix 2 × (12.5 μl, QIAGEN Multiplex PCR kit, QIAGEN S.A, Courtaboeuf, France), and double-distilled water up to a final volume of 25 μl.

For the ‘inner’ reaction, aliquots (3 μl) from the multiplex ‘outer’ reaction were used as templates for individual PCR. Inner PCR primers for microsatellite amplification were 5′ fluorescently labeled with 6-FAM. ‘Inner’ amplification mixes contained primers 10 × (2.5 μl, containing each primer at 2 μ M, Proligo), master mix 2 × (12.5 μl, QIAGEN Multiplex PCR kit), and double-distilled water up to 25 μl.

Polymerase chain reaction programs were similar for the outer and inner amplifications: 15 min denaturation at 95°C, followed by 22 cycles (outer PCR) or 30 cycles (inner PCR) of 30 s at 94°C, 90 s at 60°C, and 60 s at 72°C, and terminated with a final extension of 30 min at 60°C.

Restriction enzyme digestion

An aliquot of 10 μl of the second round PCR product was incubated with 20 units of DraI (New England Biolabs, Ipswich, UK) at 37°C for 1 h. PCR primers were designed to introduce a mismatch in order to create a second restriction site for the DraI enzyme in the SMN2 gene, allowing its differentiation with SMN1. Digested SMN1 amplicons yielded two fragments of 203 and 33 bp, and SMN2 yielded three fragments of 178, 33, and 25 bp, respectively. The products of enzymatic restriction digestion were resolved by electrophoresis on a 4% agarose gel (Nalgene, Rosckland, ME, USA) stained with ethidium bromide for 60 min at 90 V, and visualized under UV light.

Analysis of microsatellite markers

Amplified ‘inner’ products (1 μl) were added to a mix containing formamide (15 μl, genetic analysis grade, Applied Biosystems, Courtaboeuf, France) and ROX 400HD (0.3 μl, Applied Biosystems). After denaturation for 2 min at 95°C, and fast cooling on ice, the products were electrophoresed in an automated genetic analyzer ABI 3130 (Applied Biosystems). Results were analyzed with Genescan and Genotyper softwares (Applied Biosystems).

Results and discussion

The proband and his parents were first haplotyped with D5S629 and D5S1988 markers, enabling the identification of the ‘at-risk’ paternal allele (Figure 1).

SMN1 exon 7 was successfully amplified in 7 out of 12 spermatozoa, whereas no amplification of SMN2 exon 7 was obtained in these cells (Figure 2 and Table 1). Conversely, a successful amplification of SMN2 exon 7 was obtained in 5 out of 12 spermatozoa, with no detection of SMN1 exon 7 (Table 1). No sperm carried both SMN1 and SMN2 genes, suggesting that SMN1 and SMN2 genes were carried on a different chromosome (Figure 2). Along the same line, no sperm carried two microsatellite alleles, supporting the fact that each tube contained only a single sperm cell.

Figure 2
figure 2

Analysis of SMN genes and detection of D5S1988 and D5S629 alleles on single spermatozoa. (a) Discrimination between SMN1 and SMN2 exon 7, using a PCR-Dra1 restriction approach (236, 203, and 178 bp depict undigested PCR products, SMN1 exon 7 and SMN2 exon 7, respectively). Lane 1: undigested PCR product. Lanes 2 and 3: proband's father and mother DNA are shown to carry both SMN1 and SMN2 exons 7. Lane 4: the proband carries SMN2 only. Lane 5: sperm 1 carrying only SMN1 copy. Lane 6: sperm 3 carrying only SMN2 copy. (b) Haplotype determination from a single sperm. Capillary electrophoregrams of amplified fluorescently labeled markers. Lanes 2–6 are as in (a). Sperm 1 harbors the ‘at-risk’ allele (hatched arrow), whereas sperm 3 harbors the wild-type allele (full arrow).

Full size image
Table 1 Results of 12 single-sperm analysis for SMN1 mutation and linked markers
Full size table

Taking together (i) the results of the SMN1 and SMN2 copy number in the father somatic cells and (ii) the presence or absence of SMN1 and SMN2 in germ cells, we hypothesized that the father was a compound heterozygous at the somatic level, with one allele carrying a CIS duplication of the SMN1 gene and an SMN2 deletion, whereas the other chromosome carried zero SMN1 copy and one SMN2 gene (Figure 1a). A comparison of the proband's and spermatozoa's haplotypes achieved by polymorphic marker analysis confirmed these results by showing that ‘SMN2 spermatozoa’ carried the ‘mutant haplotype’, whereas ‘SMN1 spermatozoa’ had the wild-type one (Figure 2 and Table 1).

These data ruled out the hypothesis of a de novo mutation and therefore gave a 25% recurrence risk for each pregnancy. PGD for this family thus became lawful. Amplification of SMN1 and SMN2 in sperm cells has been described for the purpose of diagnosing SMA carriers.25 In contrast to this approach, our method uses linked microsatellite markers in addition to the amplification of SMN1 and SMN2 to establish the sperm haplotype, and thereby decreases the number of single sperm cells that have to be tested. Amplification of SMN1 in both populations of spermatozoa (defined by different haplotypes) is indeed sufficient to define a ‘non carrier’ and to diagnose a ‘de novo’ deletion. In this case, our method additionally allows the detection of a germline mosaicism, provided that nonamplification of SMN1 in sperms carrying the at-risk allele is significantly more frequent than the 0.1 expected allele dropout rate.24

Single-sperm studies are useful for solving complex genetic situations. They have been used for the haplotype construction of de novo paternal mutations in dominant disorders, such as the neurofibromatosis type 126 and Currarino syndrome,27 in order to reduce the misdiagnosis risk of PGD procedures. We provide here a further illustration of the crucial impact of single germ cell studies on genetics counseling in SMA, a recessive disorder.