Spinal muscular atrophy caused by a novel Alu‐mediated deletion of exons 2a‐5 in SMN1 undetectable with routine genetic testing

Abstract Background Spinal muscular atrophy (SMA) is an inherited neuromuscular disease affecting 1 in 8,000 newborns. The majority of patients carry bi‐allelic variants in the survival of motor neuron 1 gene (SMN1). SMN1 is located in a duplicated region on chromosome 5q13 that contains Alu elements and is predisposed to genomic rearrangements. Due to the genomic complexity of the SMN region and genetic heterogeneity, approximately 50% of SMA patients remain without genetic diagnosis that is a prerequisite for genetic treatments. In this work we describe the diagnostic odyssey of one SMA patient in whom routine diagnostics identified only a maternal heterozygous SMN1Δ(7–8) deletion. Methods We characterized SMN transcripts, assessed SMN protein content in peripheral blood mononuclear cells (PBMC), estimated SMN genes dosage, and mapped genomic rearrangement in the SMN region. Results We identified an Alu‐mediated deletion encompassing exons 2a‐5 of SMN1 on the paternal allele and a complete deletion of SMN1 on the maternal allele as the cause of SMA in this patient. Conclusion Alu‐mediated rearrangements in SMN1 can escape routine diagnostic testing. Parallel analysis of SMN gene dosage, SMN transcripts, and total SMN protein levels in PBMC can identify genomic rearrangements and should be considered in genetically undefined SMA cases.


| INTRODUCTION
Spinal muscular atrophy (SMA) is an inherited neuromuscular disease characterized by progressive degeneration of alpha motor neurons in the spinal cord leading to muscle weakness and paralysis. SMA is the most prevalent monogenic cause of death in infancy (Glascock et al., 2018) with an incidence of ~1:8,000 in Caucasians and Asians and carrier frequency of ~1:50 (Verhaart et al., 2017). The severity of SMA can vary from early postnatal onset and muscular weakness with respiratory insufficiency to milder forms presenting during infancy or adolescence (Schorling et al., 2019). The majority of SMA patients carry bi-allelic variants in the survival of motor neuron 1 gene (SMN1, OMIM 600354) that localizes to a duplicated region on chromosome 5q13.
The survival of motor neuron 2 gene (SMN2, OMIM 601627) is a homologue of SMN1. Genetic investigations have revealed zero to six copies of SMN2 that are located next to SMN1 on 5q13 (Crawford et al., 2012). SMN1 and SMN2 ("SMN genes") encode for the same SMN protein. However, expression of the SMN protein from SMN2 is substantially lower than from SMN1 due to a single nucleotide sequence difference at the sixth position of exon 7 in SMN2 (hg38, chr5:70076526, T). This sequence difference alters splicing and results in the predominant production of an SMN2 transcript that skips exon 7 (SMN2Δ7) and encodes for an unstable and less functional protein. SMN2 is therefore unable to fully compensate the deficit of SMN1. However, SMN expression from SMN2 may be increased in a gene dose-dependent manner (Crawford et al., 2012). This SMN2 dose variance was suggested to have a compensatory effect on SMN expression and ameliorate the severity of SMA (Butchbach, 2016).
The SMN region on chromosome 5q13 is enriched for primate-specific nonautonomous retrotransposons belonging to a class of short interspersed elements (SINE)-repetitive DNA sequences called Alu elements. The Alu elements are about 280 base pairs long and are formed by two diverged dimers (Deininger, 2011). They are divided into subfamilies based on single nucleotide differences. The main Alu subfamilies are AluJ, AluS, and AluY (Kim, Cho, Han, & Lee, 2016) with the AluY being the evolutionarily youngest and AluS the most numerous. The youngest Alu subfamilies AluS and AluY increase the likelihood of genomic rearrangements that result in the formation of a new chimeric Alu-Alu element at the breakpoint junction.
It is important to note that the term "deletion of SMN1 exons 7 and/or 8" is commonly used to describe the results of routinely performed multiplex ligation-dependent probe amplification (MLPA) assays, which are the current gold standard of SMA diagnostics (Mercuri et al., 2018). Deletions of these two particular exons can be detected by the MLPA assay, which is designed to target exclusively exons 7 and 8 and distinguishes SMN1 and SMN2 based on single nucleotide sequence differences. In reality, these deletions can extend beyond exons 7 and 8 and include the entire SMN1, then even extending further to include multigene deletions of the 5q13 region.
Other Alu-mediated genomic rearrangements in the SMN region identified in SMA patients lead to formation of SMN1-SMN2 hybrid genes (van der Steege et al., 1996). Interestingly, Alu-mediated deletion of exons 4 to 6 with intact exons 7 and 8 has thus far been reported in just a single case (Wirth et al., 1999). Other SMN1 variants identified in SMA patients include small intragenic deletions and missense variants. A full list of these variants can be found in the Human Gene Mutation Database records (Stenson et al., 2014).
Due to the complexity of the SMN 5q13 genomic region, approximately 50% of all SMA patients remain without a genetic diagnosis after routine genetic testing (Karakaya et al., 2018). The ability to identify the genetic cause of SMA is critically important for patients because only patients with bi-allelic SMN1 variants are eligible for genetic therapies (Michelson et al., 2018). These potential treatments include Zolgensma and the antisense oligonucleotide treatment Spinraza. The precise identification of the causal variants in SMA patients is also important for genetic counseling in affected families.
In this work we describe the diagnostic odyssey for one SMA patient and her parents from Slovakia in whom the routine MLPA assay and subsequent direct sequencing of SMN1 coding regions identified only a heterozygous, maternally inherited deletion of exons 7 and 8 in SMN1.

| Ethical compliance
The study was approved by the appropriate institutional review boards and the investigations were performed according to the Declaration of Helsinki principles. Parents provided informed consent. | 3 of 10 JEDLIČKOVÁ Et aL.

| Clinical report
The patient was clinically diagnosed at the Children Teaching Hospital Košice, Slovakia. The infant was born by normal spontaneous delivery to a 31-year-old mother following a full-term pregnancy from the first gravidity and with no reported abortions; the postnatal adaptation of the infant was standard. The patient presented with global muscle hypotonia and hyporeflexia suggestive of SMA at the age of 1 month and showed markedly decreased mobility at the age of 3 months. Global respiratory failure required tracheostomy and mechanical ventilation from the age of 7 months. At the time of investigation, the patient was 2 years and 8 months old and ventilator dependent. She suffered from severe global weakness and hypotonia. Muscle atrophy predominantly affected the lower limbs. The patient could respond only with eye contact. There were no sensory deficits. Both parents were neurologically intact.

| MLPA, panel sequencing, and cytogenetic analyses
Genomic DNA (gDNA) was isolated from peripheral venous blood in the patient and her parents using standard protocol. The patient's karyotype was assessed by G-banding and comparative genomic hybridization (aCGH) was performed using Agilent SurePrint HD 4x44 platform at the Medirex Group, Slovakia. Genes associated with a set of neuromuscular disorders were analyzed using a custom SeqCap EZ kit (Roche) and Illumina sequencing at CMBGT in Brno, Czech Republic. The TruSight One Sequencing Panel (Illumina) was used for analysis of more than 4,800 genes associated with human diseases at Medirex. The presence of the deletion of exons 7 and 8 in SMN genes was assessed using the MLPA assay; SALSA MLPA P060 SMA Carrier probemix (MRC-Holland) at Genexpress. The coding regions of SMN genes were analyzed using paired-end sequencing of PCR amplicons on Illumina MiSeq (Illumina) at Alpha Medical.

| SMN1 and SMN2 mRNA/ cDNA analysis
Total RNA and cDNA were isolated and prepared from peripheral blood mononuclear cells (PBMC) using ProtoScript® II Reverse Transcriptase (NEB). Full-length SMN1 and SMN2 cDNAs were PCR amplified using primers SMN575_F (Sun et al., 2005) and SMN_541C1120_R (Lefebvre et al., 1995) (Table 1) amplifying together both SMN genes derived transcripts from the first coding exon to the last untranslated exon 8 (NM_000344.3). PCR products were analyzed using the agarose gel electrophoresis, and Sanger sequenced using the version 3.1 Dye Terminator cycle sequencing kit (ThermoFisher Scientific) with electrophoresis on an ABI 3500XL Avant Genetic Analyzer (ThermoFisher Scientific).

| Western blot analysis of SMN protein
The quality and amount of the SMN protein were assessed in lysates of PBMC using Western blot analysis. PBMC were isolated using the Histopaque-1077 reagent (Sigma-Aldrich). The cell pellet was resuspended in 50 mM Tris pH 6,8, 50 mM DTT, 2% SDS, and Complete Protease Inhibitor Cocktail (Roche), sonicated using the Covaris S2 Ultrasonicator (Covaris), and denaturated at 100°C for 10 min. The protein content in the supernatant was determined using an infrared spectrometer Direct Detect (Millipore) according to the manufacturer's protocol. Protein lysates equivalent to 22 μg of protein were reduced and denatured at 100°C for 10 min in a sample buffer with 1% beta-mercaptoethanol before SDS-PAGE electrophoresis. After protein transfer to the polyvinylidene fluoride (PVDF) membrane, the membrane was blocked by 5% skimmed milk and 0.1% Tween 20 in PBS for 1 hr at room temperature (RT). The SMN protein was visualized by incubation with mouse monoclonal SMN antibody (610646, BD Transductions) at 1:5,000 in 5% BSA and 0.1% Tween 20 in PBS over night at 4°C, followed by incubation with goat anti-mouse HRP (Pierce) at 1:10,000 in 0.1% Tween 20 in PBS for 60 min, and detection was performed by Clarity Western ECL Substrate (Bio-Rad). The actin protein was visualized by incubation with rabbit Actin antibody (A2103; Sigma-Aldrich) at 1:1,000 in 0.1% BSA and 0.1% Tween 20 in PBS for 1 hr at RT, followed by incubation with goat anti-rabbit HRP (Pierce) under conditions and using detection as described above. Relative quantification of the SMN protein was performed using GeneTools software (4.03.03.0, Syngene). SMN protein levels were normalized to actin; the experiment was performed in three technical replicates. The statistical significance was determined using one-way ANOVA test.

| Copy number analysis of SMN genes
The total number (sum) of SMN genes dosage was determined in the gDNA of the patient and her parents using the quantitative PCR (qPCR) analysis of four retrotransposonfree SMN genomic regions: (a) in intron 1 (I1), (b) in the exon 3-intron 3 junction (E3I3), (c) in the intron 5-exon 6 junction (I5E6), and (d) ~1 kb downstream from exon 8 (+1 kb) ( Table 1). The qPCR was performed using the LightCycler® 480 System (Roche Applied Science). All of the qPCR reactions were performed in triplicates in 10 μl reactions with the following final concentrations of the reagents: 80 nM UPL probe, 300 nM primers, 1x Roche Probe Master Mix, and total 15 ng of genomic DNA per reaction. Sample quantitation cycle (cq) values were determined using the Second Derivative Maximum Method and normalized using the RNAse P and albumin genes as a references. Relative quantification using the 2−ΔΔCT method (Livak & Schmittgen, 2001) was performed to determine the sum copy number. Unrelated healthy controls were used as control samples.

| Mapping of SMN1 deletion breakpoint/ junction by Alu PCR
The Alu PCR was performed as described previously by Majer et al. (2020). Briefly, based on the SMN1 transcript analysis that identified deletion of exons 2a to exon 5 in SMN1 cDNA, we anticipated that the deletion breakpoints must be located in adjacent introns 1 and 5. We identified in these regions sequences of Alu retrotransposons ( Figure S1) and designed (a) a set of universal reverse primers targeting the terminal parts of the Alus in intron 1 (Alu_259_ wt_R, Alu_259_4A_R, Alu_259_3C_R, Alu_259_1A_R) (Table 1, Figure S1), and (b) one SMN1-specific forward primer targeting from intron 5 toward intron 1 (SMN_ i5_979_R, Table 1). We performed four separate PCR reactions ( Figure S2) with the intron 5 SMN1-specific primer and one of the four Alu universal primers. Resulting PCR products of the four reactions were column isolated and Sanger sequenced using a gene-specific primer located in intron 5 Alu preceding region (SMN_i5_821_R) ( Table 1). The Sanger sequencing was performed as described above.
The deletion-spanning PCR method for testing of family members for the Δ2a-5 deletion was performed using primers annealing to the intron 1 (SMN_i1_10748_F) and exon 6 of SMN1 (SMN_i5e6_1R, Table 1). Resulting PCR products were column isolated and Sanger sequenced as described above.

| Routine genetic testing in the patient identified only heterozygous, maternally inherited deletion of exons 7 and 8 in SMN1
To establish the diagnosis, MLPA analysis was performed and revealed a deletion of SMN1 exons 7 and 8 and two copies of SMN2 exons 7 and 8. Heterozygous deletion of SMN1 exons 7 and 8 was also found in the patient's mother, who also carried only one copy of SMN2 exons 7 and 8. The father carried two copies of exons 7 and 8 of both SMN genes (Table 2). Subsequent targeted sequencing of the coding regions of SMN genes in the patient and parents did not reveal any further pathogenic variants that would explain the SMA phenotype. To search for other potential disease-causing variants, karyotype assessment and array-based comparative genomic hybridization assay (aCGH) were performed but no gross chromosomal abnormalities were detected. The panel sequencing of genes associated with a set of neuromuscular disorders and the TruSight One Sequencing Panel did not reveal any definitive or likely pathogenic variants related to the phenotype. Due to inconclusive results of genetic testing, the patient was referred to the Research Unit for Rare Diseases of the First Faculty of Medicine, Charles University in Prague, and included in the "Undiagnosed Disease Program" that aims to identify genetic diagnosis in cases of rare genetic diseases with negative results of genetic and genomic analyses.
Before looking for other genetic causes, the primary focus was to further analyze the SMN genomic region in this patient whose clinical phenotype was highly suggestive of SMA.

| Direct sequencing of SMN cDNAs revealed an absence of a full-length SMN1 transcript in the patient
To study the SMN1 transcript we performed reverse transcription polymerase chain reaction (RT-PCR) and amplified the full-length SMN cDNAs from PBMC of the patient, both parents and controls. Using the agarose gel electrophoresis, we observed in the patient and her father an abnormal PCR product of ~600 bp that was not present in mother and controls (Figure 1a). Subsequent Sanger sequencing of the gel isolated PCR products revealed the presence of the fulllength SMN1 and SMN2 cDNA in the mother, father, and control. In the patient, only the full-length SMN2 cDNA was present and the full-length SMN1 cDNA was lacking. The abnormal shorter cDNA of ~600 base pairs that was detected in the patient and father was, based on the presence of two SMN1-specific sequence variants in exons 7 and 8, identified as a SMN1 that was lacking exon 2a to exon 5 (Δ2a-5) (Figure 1a).

| Western blot analysis confirmed decreased amount of SMN protein in PBMC
To analyze the effect of identified mRNA changes on the protein quality and abundance, we immunodetected and quantified SMN protein in PBMC lysates from the patient, her parents, and a control. In all samples we detected only one immune-reactive protein at ~40 kDa, corresponding to the predicted molecular weight of full-length SMN (Figure 1b). The protein at the length of ~10 kDa corresponding to a predicted molecular weight of the deleted SMN(Δ2a-5) was not detected probably due to its altered immunogenicity or reduced protein stability. Compared with controls, the amount of SMN normalized to actin was reduced to 10% in the patient, 27% in the mother, and 41% in the father (Figure 1c). F I G U R E 2 Identification of SMN1 variants. (a) qPCR analysis of four retrotransposon-free SMN genomic regions in the intron 1 (I1), in the exon 3-intron 3 junction (E3I3), in the intron 5-exon 6 junction (I5E6), and ~1 kb downstream from exon 8 (+1 kb). Compared with four copies of SMN genes that are present in controls, we found that the patient has three copies at the I1, I5E6, and +1 kb loci and two copies at the E3I3 loci. The mother has three copies through the I1 to I5E6 loci and two copies at the +1 kb loci. The father has three copies only at the I5E6 loci. 3.4 | SMN1 and SMN2 gDNA copy number analyses suggested the presence of the intragenic SMN1 deletion in the patient and father and whole-gene SMN1 deletion in the patient and her mother To assess SMN gene dosage, we performed qPCR analysis of four retrotransposon-free SMN genomic regions: (a) in intron 1 (I1), (b) in the exon 3-intron 3 junction (E3I3), (c) in the intron 5-exon 6 junction (I5E6), and (d) ~1 kb downstream from exon 8 (+1 kb). Compared with four copies of SMN genes that are normally present in each of four tested loci in controls, we found that the patient has three copies at the I1, I5E6, and +1 kb loci and two copies at the E3I3 loci. The mother has three copies through the I1 to I5E6 loci and two copies at the +1 kb loci. The father has three copies only at the I5E6 loci ( Figure 2a). Together with the MLPA assay findings (Table 2), this analysis suggested that the patient had deletion of the whole SMN1 on the maternal allele (NC_000005.9:g.(?_70221078)_ (70249850_?)del), where the genomic coordinates denote the central position of the qPCR probes targeting the I1 and +1 kb genomic loci, and a deletion encompassing Δ2a-5 exons of SMN1 on the paternal allele.

| Sequencing of long-range gDNA PCR products spanning SMN genomic sequence indicated bi-allelic deletion of the SMN1 in the patient
To identify the variant causing the deletion of SMN1 exons 2a-5 in the patient's and father's cDNA, we PCR amplified the genomic DNA and Illumina sequenced four overlapping long-range amplicons (PCR1-PCR4) spanning exons 1-8 of SMN1 and SMN2. Using specific single nucleotide sequence differences that distinguish the SMN genes, we found that in the patient amplicon PCR4, spanning intron 6, exon 7, intron 7, and exon 8 (Figure 2b), had been amplified exclusively from SMN2. These findings indicated that SMN1 gene must be disrupted on both alleles in the patient.

| SMN1 deletion breakpoint/ junction mapping revealed paternal Alumediated deletion
To identify the exact nature of the variant on the paternal allele, we considered the Alu-mediated rearrangement of SMN1 as the most likely mechanism. Using computer analysis we identified in the candidate breakpoint region of intron 1 a set of 21 Alus (Figure 2b); in the candidate region of intron 5 we identified only one AluSq (Figure 2b). We considered AluSq to be a candidate breakpoint start site and anticipated recombination between the AluSq and one of the Alus located in intron 1 resulting in the formation of a new chimeric Alu. With this assumption we performed four PCR reactions using always one of the universal Alu primers and the SMN1 intron 5-specific primer flanking the AluSq. Sanger sequencing of obtained PCR products revealed that two of them contained both the wild-type (wt) AluSq sequence of intron 5 and the sequence originating from a new chimeric Alu resulting from the rearrangement with AluSp element originating from the intron 1 (Figure 2c; Figure S2).
Using this sequence we designed and performed a deletion-spanning PCR allowing for genotyping of the Δ2a-5 deletion. The deletion-spanning PCR product was obtained from DNA of the patient and her father, but not from the mother and control (Figure 2d). Sanger sequencing of this PCR product (Figure 2e) defined the new Alu-mediated SMN1 deletion ranging 8,978 base pairs as NC_000005.9:g.70232118-70241095del; NM_000344.3:c.82-2548_723+515del.

| Genetic analysis of the SMN region correlates with the variance in SMN expression in PBMC
In summary, the genetic analysis (Figure 2f) established that the patient is a compound heterozygote for the ~9 kbp deletion in SMN1 that she inherited from her father and the deletion of the entire SMN1 that she inherited from her mother. Maternal and paternal SMN2 alleles were intact. This genotype correlates with the very low (10% of controls) SMN content in PBMC that must be expressed exclusively from SMN2 (Figure 2b,c). In addition to the deletion of the entire SMN1, the mother also has a deletion of the terminal part of SMN2. This correlates with reduced (27% of controls) SMN content in PBMC compared with the father (41% of controls), who has only the ~9 kbp deletion on one of the SMN1 alleles and both SMN2 alleles intact.

| DISCUSSION
Spinal muscular atrophy is devastating inherited neuromuscular disease resulting from variants of SMN1 and deficiency of the survival motor neuron protein (SMN). Several clinical or experimental therapies for SMA augmenting levels of SMN are currently available or in development (Groen, Talbot, & Gillingwater, 2018). The treatment is provided only to individuals with an established genetic diagnosis. Despite great progress in genetic diagnostic methods, approximately half of the patients suspected to have SMA still remain without a genetic diagnosis after routine genetic testing (Karakaya et al., 2018). This unsatisfactory situation is due to the complexity of the SMN genomic region and lack of methods allowing routine identification of individual genomic rearrangements.
In this work, we describe the diagnostic odyssey for one SMA patient in whom routine diagnostic procedures identified only a heterozygous, maternally inherited deletion of exons 7 and 8 in SMN1. SMN is ubiquitously expressed and detectable in PBMC (Sumner et al., 2006). To identify the other SMN1 variant in this case, we obtained PBMC from the patient and his parents. In this material we successively assessed SMN transcripts, SMN protein content, SMN genes dosage, and SMN genomic sequence. Using this approach we found in this case that SMA was caused by a novel Alu-mediated deletion encompassing exons 2a to exon 5 (Δ2a-5) of SMN1 on the paternal allele and by deletion of whole SMN1 on the maternal allele.
The (Δ2a-5) variant of SMN1 escaped detection by the routine MLPA assay that targets only exons 7 and 8. To the best of our knowledge, this is only the second reported Alumediated deletion in SMN1 that does not encompass the exons 7 and 8 (Wirth et al., 1999).
Our work suggests that other Alu-mediated rearrangements in the SMN region that escape detection with routine genetic testing may be more common and should be considered in SMA cases who remain without a genetic diagnosis after standard genetic testing.
We demonstrate that in these cases measurement of SMA protein in PBMC may successfully identify patients with SMA in whom a genetic diagnosis cannot be made. These individuals can then undergo further testing, including SMN gene dosage, full characterization of SMN transcripts, and precise characterization of the eventual genomic rearrangements. All these parameters are critical for targeted genotyping and eventually for prenatal or preconception analysis in affected families and therapeutic eligibility for affected individuals. We are interested to study other similar cases and can provide the genetic testing and biochemical analyses described here. Please contact ivana.jedlickova@lf1. cuni.cz.

ACKNOWLEDGMENTS
This work was supported by Ministry of Health of the Czech Republic, grant nr. NV19-08-137. Institutional support was provided by Charles University institutional programs PRVOUK-P24/LF1/3, UNCE 204064, and SVV2016/260148, by the project LQ1604 NPU II from the Ministry of Education, Youth and Sports of the Czech Republic. We thank The National Center for Medical Genomics (LM2018132) for their instrumental and technical support with the WES analyses. Instrumental support was provided by grant CZ.1.05/2.1.00/19.0400 from the Research and Development for Innovations Operational Programme (RDIOP) cofinanced by European regional development fund and the state budget of the Czech Republic.