Introduction

Spinal muscular atrophy (SMA) is an autosomal recessive neuromuscular disorder with an estimated incidence of 1:6,000 to 1:10,000 live births and a carrier frequency of 1/35–1/60 [1, 2]. Clinically, SMA is a continuous spectrum of phenotypes ranging from severely compromised neonates and infants to adults with minimal manifestations. Patients are classified into four main groups based on age of onset and motor milestones [3].

Survival motor neuron 1 (SMN1) has been identified as the SMA disease-determining gene [4]. SMN2 is a highly homologous copy of SMN1, which has been described as an SMA modifier [5, 6]. In 90% of SMA cases, the molecular pathology is absence of SMN1 by deletion or gene conversion and in 5% SMN2–SMN1 hybrid genes [7,8,9,10]. The remaining SMA cases are compound heterozygous [7, 8].

Current methods of SMN1 dosage do not discriminate between 1/1 non-carriers and 2/0 carriers (individuals with two SMN1 copies in cis [2]). Familial haplotype analysis with polymorphic markers of the SMA locus is helpful to detect these 2/0 carriers within blood relatives. However, diagnosis in partners of SMA carriers from the general population with two SMN1 copies is challenging [2].

Two SMN1 variants have been recently associated to chromosomes carrying two SMN1 copies in cis in the Ashkenazi Jewish population [11]. These variants, c.*3 + 80 T > G corresponding to g.27134 T > G in intron 7 and c.*211_*212del corresponding to g.27706_27707delAT in exon 8 of the SMN1 gene, had been previously described by Luo and co-workers [11] according to the first nucleotide of the gene, 5000 bases off from the NG_008691.1 reference sequence. Here, we tested these variants in a large set of Spanish individuals to confirm their utility to improve identification of 2/0 SMA carriers.

Materials and methods

A total of 270 Spanish individuals were analysed for the presence of the c.*3 + 80 T > G and c.*211_*212del variants in the SMN1 gene. Carriers were divided into classical 1/0 SMA carriers (n = 41) and 2/0 confirmed SMA carriers (n = 32). An exceptional 3/0 carrier was included in the study. Non-carriers had two (1/1; n = 99) or more than two copies of the gene (3 SMN1, n = 58; 4 SMN1, n = 3). Finally, we studied a subset of 16 SMA patients who lack SMN1 and another 20 with hybrid SMN2–SMN1 genes [9, 10]. All individuals signed their informed consent.

Determination of SMN1 copy number, sequencing methods and haplotype studies have been described elsewhere [2, 12, 13]. For detection of the variants, we used primers R111 in intron 6 and 541C1120 in exon 8 of the SMN genes [8, 10]. All variants were numbered according to the first translated base of the SMN gene (NM_000344.3 or NG_008691.1 for intronic changes) according to standard nomenclature guidelines [14]. Exons are numbered as in ref. [15]. All variants detected in this study were submitted to the LOVD database (http://databases.lovd.nl/shared/genes/SMN1) with patients’ IDs 00150115, 00150116, 00150120–00150129, and 00150239–00150250. Χ2 statistics were calculated with SPSS package, and p values < 0.05 were considered significant.

Results

Our major results are summarised in Table 1. In general, the studied variants were almost completely absent from chromosomes with a single SMN1 copy (1/297; 0.33%), but they were frequently detected in those carrying two copies of the gene (18/96; 18.75%) (p < 0.001).

Table 1 Screening for the presence/absence of variants c.*3 + 80 T > G and c.*211_*212del in 270 Spanish individuals
Full size table

Carriers. One classical 1/0 carrier presented the variants (1/41; 2.4%) whereas 7 of the 32 2/0 carriers were positive (7/32; 21.8%). Interestingly, the 3/0 carrier was negative for both variants (Fig. 1a). All individuals with the variants were unrelated and shared a 20-repeat allele for marker D5S1556. Further, five of them, who were from the Canary Islands, showed a 24-repeat allele for marker D5F149S1 (Supplementary Figure 1).

Fig. 1
figure 1

Identification and genetic analyses of special SMA carriers. a Familial haplotype and SMN1 quantitative analyses in a family with a 3/0 carrier. The family requested carrier studies because of the previous death of a son with clinical manifestations compatible with type I SMA. Due to the informative potential of the D5F149S1 and D5S1556 polymorphic markers and allele segregation, the more likely explanation is that the three copies of the SMN1 gene detected in the mother (I:2) were located in cis in the same chromosome. Her non-carrier daughter (II:1) harboured 4 SMN1 copies, three of them inherited from her maternal chromosome and the other one from her father. The carrier daughter (II:2) showed a single SMN1 copy inherited from her father (shared also with her sister) and the SMN1 deletion from her mother. The haplotype of SMA patient (II:3) is inferred. Black bars represent the chromosome that lacks SMN1. Each box represents a single SMN1 copy. The number of each marker allele corresponds to the CA repeats. b Association between the presence of the c.*3 + 80 T > G variant in intron 7 of the SMN1 gene and a chromosome carrier of two copies of this gene. By quantitative studies, we detected the presence of three SMN1 copies in the father (I:1) and one copy in the mother (I:2). Haplotype analyses revealed that both siblings inherited the same maternal chromosome without the SMN1 gene, while the paternal chromosome was different. The older daughter (II:1) was diagnosed as a 2/0 carrier with 2 copies of the SMN1 gene and the presence of the c.*3 + 80 T > G variant inherited from her father, while the younger son (II:2) showed one copy of the gene without the variant

Full size image

Non-carriers. All 1/1 individuals (n = 99) were negative for the studied variants, whereas 11 out of 58 subjects with three SMN1 copies (2/1) were positive (19.3%; p < 0.001). Four individuals in this last subgroup had also homozygous absence of SMN2 genes, three of whom showed the variants. A comparison between 2/0 carriers vs. 1/1 controls was also significant (p < 0.001) for the presence of the variants.

SMA patients. None of the variants were found in 16 SMA patients who lack SMN1. However, four out of 20 SMA patients with SMN2–SMN1 hybrid genes presented the c.*211_*212del variant (20%) (Supplementary Figure 2).

Discussion

To improve genetic counselling for carriers of SMA, we aimed to validate in the Spanish population two variants of the SMN1 gene, c.*3 + 80 T > G and c.*211_*212del, which had been previously associated to chromosomes with two SMN1 copies in cis in the Ashkenazi population [11].

Our results confirm that these variants are only present in SMN1. First, we did not detect them in patients with total absence of the SMN1 gene. Second, in subjects with SMN2–SMN1 hybrids we detected only the c.*211_*212del variant, which corresponds to the SMN1 half of the hybrid. Finally, we identified both variants in three individuals with three SMN1 copies each, but who lack the SMN2 gene.

We have also corroborated the utility of both variants for genetic testing of SMA carriers, as they are much more frequent in chromosomes with two SMN1 copies in cis (p < 0.001), and are also present in almost 20% of cases with three SMN1 copies (Table 1).

We report for the first time a 3/0 SMA carrier, identified in the context of a SMA family study but without the variants (Fig. 1a). Individuals with 3 SMN1 copies are usually thought to harbour two copies in one chromosome and one in the other (2/1). Our finding emphasizes the complexity of the SMA region and points to possible pitfalls in interpreting the results of non-carriers. Genetic counselling in these cases should be carefully evaluated in the context of haplotype results [13].

Chromosomes with two SMN1 copies are more frequent in the African population [2, 16] and, concomitantly, they present both SMN1 variants at higher frequencies [11]. In our cohort, six of seven cases with the variants were from the Canary Islands, a region with African genetic influence due to the territorial proximity. Most of these cases share a common haplotype (Supplementary Figure 1).

Among non-carriers, the variants were absent in 1/1 individuals but present in almost every fifth (19.3%) of the 2/1 individuals studied (p < 0.001), reinforcing their association with chromosomes harbouring 2 SMN1 copies in cis. These variants were previously reported in 116 out of 200 individuals (58%) with three SMN1 copies [11]. Ninety of these positive cases were from African origin, likely explaining the differences with our cohort.

Given that the variants are linked to SMN1 but not SMN2 (ref 10. and this work), they are not detected in SMA patients who lack SMN1. However, 20% of patients with hybrid SMN2–SMN1 genes, who usually have intron 7 of the SMN2 gene and exon 8 from SMN1 [9, 10], present the exon 8-linked c.*211_*212del variant. As expected, none of them had the intron 7-associated variant. These observations suggest that some hybrid genes originate from chromosomes with two SMN1 copies in cis. However, it is not known whether the presence of the variant makes SMN1 genes prone to rearrange in hybrid structures. It is also possible that individuals who only present the intron 7 variant may represent hybrid SMN1 intron 7 and SMN2 exon 8 structures.

In conclusion, our results indicate that SMN1 variants c.*3 + 80 T > G and c.*211_*212del are associated to chromosomes that underwent rearrangements such as those with two SMN1 copies in cis and those with hybrid SMN2–SMN1 genes (around 20% of the cases). However, absence of both variants in a subject with two SMN1 copies does not preclude the 2/0 carrier status limiting the utility of this analysis. Since most of the two-cis chromosomes and hybrids do not show these variants, their study in SMA carrier testing may have a limited geographical application, assuming a higher frequency in the Ashkenazi Jews and African population due to an increased number of chromosomes with two SMN1 copies. However, the study of the variants is useful to select individuals with increased risk of being 2/0 carriers. Indeed, the presence of one or both variants notably increases the residual risk from 1/781 to ~1 (Table 2). In these cases, testing of the parents of the individual [2] would be necessary to confirm his/her 2/0 carrier status.

Table 2 Usefulness of the c.*3 + 80 T > G and c.*211_*212del SMN1 variants for assessing the risk of being a 2/0 SMA carrier
Full size table