Real-world evidence: Risdiplam in a patient with spinal muscular atrophy type I with a novel splicing mutation and one SMN2 copy

Abstract Spinal muscular atrophy (SMA), which results from the deletion or/and mutation in the SMN1 gene, is an autosomal recessive neuromuscular disorder that leads to weakness and muscle atrophy. SMN2 is a paralogous gene of SMN1. SMN2 copy number affects the severity of SMA, but its role in patients treated with disease modifying therapies is unclear. The most appropriate individualized treatment for SMA has not yet been determined. Here, we reported a case of SMA type I with normal breathing and swallowing function. We genetically confirmed that this patient had a compound heterozygous variant: one deleted SMN1 allele and a novel splice mutation c.628-3T>G in the retained allele, with one SMN2 copy. Patient-derived sequencing of 4 SMN1 cDNA clones showed that this intronic single transversion mutation results in an alternative exon (e)5 3′ splice site, which leads to an additional 2 nucleotides (AG) at the 5′ end of e5, thereby explaining why the patient with only one copy of SMN2 had a mild clinical phenotype. Additionally, a minigene assay of wild type and mutant SMN1 in HEK293T cells also demonstrated that this transversion mutation induced e5 skipping. Considering treatment cost and goals of avoiding pain caused by injections and starting treatment as early as possible, risdiplam was prescribed for this patient. However, the patient showed remarkable clinical improvements after treatment with risdiplam for 7 months despite carrying only one copy of SMN2. This study is the first report on the treatment of risdiplam in a patient with one SMN2 copy in a real-world setting. These findings expand the mutation spectrum of SMA and provide accurate genetic counseling information, as well as clarify the molecular mechanism of careful genotype–phenotype correlation of the patient.


Introduction
Spinal muscular atrophy (SMA) is a group involving several subtypes of neuromuscular disorders characterized by the degeneration of α-motor neurons in the anterior horn of the spinal cord, leading to symmetrically progressive muscle weakness and atrophy [1,2].SMA is still the leading cause of infant mortality worldwide with an incidence of 1 in 6000-10 000 live births [3,4].More than 95% patients with SMA are caused by decreased levels of the survival motor neuron (SMN) protein resulting from the homozygous deletions of at least exon 7 of the SMN1 gene [5][6][7].The remaining (less than 5%) cases are caused by compound heterozygosity variants of SMN1, with one SMN1 deletion and one SMN1 intragenic mutation, providing additional support for SMN1 as a causative gene of SMA.The intragenic mutations in SMN1 associated with SMA include missense, nonsense, or frameshift mutations located within the exons of SMN1 and splicing mutations located in the intron regions of SMN1, which can lead to abnormal splicing of SMN1 pre-mRNAs [8].SMN2, a homologue of SMN1, can produce about 10% functional SMN protein of that of SMN1, partially compensating for the deficiency of SMN1 [9]; therefore, the SMN2 copy number in patients with SMA is inversely correlated with disease severity [10][11][12].
According to age of symptom onset and the maximum motor milestone that can be reached, SMA can be divided into five subtypes (types 0-IV) [13][14][15].Among them, SMA type 0 is the most severe form of SMA, with a prenatal or neonatal onset.These patients require assisted breathing from birth and usually survive no more than 6 months [15].SMA type I (OMIM #253300), also known as Werdag-Hoffman disease, accounts for about 50%-70% of patients with SMA, and it is a severe form of SMA, with symptoms usually occurring before 6 months of age [16].Patients with SMA type I are unable to achieve higher motor milestones without treatment, such as sitting without assistance, and they generally have shortened lifespans [17,18].The affected patients usually exhibit a decline in motor function after symptom onset, accompanied with a decline of functions in respiratory and swallowing, and feeding support or combined feeding and ventilatory support typically needs to be provided to these infants by 12 months of age [16,17].SMA type II (OMIM #253500) is characterized by onset usually before 18 months [19].The affected patients can sit independently, rarely stand, and only rely on others for support, but they cannot walk independently [20].Their lifespan is shortened, with approximately 50% of individuals surviving to the age of 40 years [21].Patients with SMA type III (OMIM #253400) have symptom onset beyond 18 months of age.They are able to walk independently with difficulty (waddling gait) but may lose this ability over time.As the condition progresses, they may require mobility support but generally with a normal life expectancy [20,21].SMA type IV (OMIM #271150), characterized by adult onset, is the mildest form of SMA, and individuals with SMA type IV have a fairly benign course of disease [19].
Currently, the U.S. Food and Drug Administration and the European Commission have approved three disease-modifying therapies for SMA.Nusinersen (SPINRAZA ® ), delivered by intrathecal injection, is an antisense oligonucleotide targeting the SMN2 gene and has been approved for the treatment of both adult and pediatric populations with SMA [22,23].Onasemnogene abeparvovec (ZOLGENSMA ® ), given by intravenous infusion, is a gene replacement therapy-based viral vector, which is approved for the treatment of patients with SMA under 2 years old (US), or patients with SMA type I or 1-3 SMN2 copies (EU) [24,25].Risdiplam (EVRYSDI ® ), administered orally, is a systemically distributed splicing modifier of SMN2, which increases the levels of functional SMN by promoting the inclusion of exon 7. It is indicated for children with SMA aged 2 months and older (USA) or aged 2 months and older with a clinical diagnosis of SMA types I, II, or III or 1-4 copies of SMN (EU) [26,27].The most appropriate individualized treatment for SMA has not yet been determined, and the best initial approach remains unclear [28].In addition, the role of the SMN2 copy number in patients treated with disease modifying therapies (DMTs) remains unclear.
In this study, we report a relatively mild case of SMA type I with a compound heterozygous variant in the SMN1 gene: one deleted SMN1 allele and a novel splicing variant c.628-3T>G in the retained allele and only with one SMN2 copy.Further mRNA sequencing in SMN1 gene showed that an insertion of two bases (AG) between exons 4 and 5 predicting will lead to a premature stop codon (p.Pro210SerfsTer4) but still retain partial normal spliced mRNA with c.628-3T>G variant, thereby explaining why the patient with only one copy of SMN2 had a mild clinical phenotype.For the patient's family to accept the treatment cost and proceed with drug treatment as early as possible, risdiplam was used for the treatment of this patient.However, the patient showed remarkable clinical improvement after treatment with risdiplam for 6 months despite carrying only one copy of SMN2.

Genetic analysis
Given the autosomal recessive genetic model and the clinical phenotype of neurological diseases, no highly suspected disease-associated subtle mutation was found by clinical exome sequencing.However, CNV analysis of NGS indicated the loss of heterozygosity of the SMN1 and SMN2 genes (Fig. 1A).MLPA was further employed to detect the copy numbers of SMN1 and SMN2, and the results showed that both exons 7 and 8 of the SMN1 and SMN2 genes were absent (Fig. 1B).The results of NGS and MLPA analyses revealed that the patient carried only one copy of SMN1 and SMN2.Given the result of loss of SMN1 in one allele, we predicted that the patient might carry a subtle mutation in the other allele.Sanger sequencing identified a novel substitution of thymine by guanine at c.628-3 sit (c.628-3T>G) in intron 4 of the SMN1/SMN2 gene in the patient (Fig. 2A and B).This mutation was predicted to change splicing of pre-mRNA in SMN1/SMN2 and had a frequency of 0 in thousands of genomes, ESP6500, gnomAD, ExAC, and in-house databases.Given that SMN1 and SMN2 are two highly similar genes, both can be amplified as templates when using the common primers for SMN.With Sanger sequencing, the specific position of this novel mutation cannot be determined.Since there was initially no clear clinical diagnosis of SMA and SMN1 and SMN2 were highly homologous, the mutation c.628-3T>G did not come to our attention when analyzing the exome sequencing data.Thus, after the Sanger sequencing test, reanalysis of NGS data at this mutation site showed that the ratio of the number of mutated base G to wild T (a total of 384 reads) was approximately 1:1 (Fig. 2C).This result was additional proof for the loss of heterozygosity of the SMN1 and SMN2 genes.

Bioinformatics analysis
The novel variant c.628-3T>G in SMN1/SMN2 was located at a splice receptor site, so in silico analysis including an online RNA Splicer tool and SpliceAI (a deep learning-based tool to identify splice variants) was employed to predict the possible effects on splicing.SpliceAI predicted a delta score of 0.87 for this variant.Two potential aberrant splice patterns were shown by the online RNA splicer tool: pattern 1 showed two bases (AG) were inserted at the 5 -end of the alternatively spliced exon 5 (Fig. 3A), which could lead to a premature termination codon (p.Pro210SerfsTer4); pattern 2 showed exon 5 (96 bp) skipping (Fig. 3A), suggesting that a new 3 splice acceptor site (YAG|(Exon), Y = pyrimidine) formed in intron 4 with pattern 1.

Splicing study of c.628-3T>G and identification of this mutation in the SMN1 gene
Given that SMN1 and SMN2 have only one copy confirmed by the above methods, SMN1 and SMN2 each can produce one form of mRNA.Therefore, c.628-3T>G could only have an abnormal splicing effect on one of the SMN1 and SMN2 genes.The splicing effect caused by c.628-3T>G in SMN1 or SMN2 was validated by Sanger sequencing of the SMN-cDNA plasmid described in the above method.A total of 14 randomly selected positive SMN-cDNA clones were used for Sanger sequencing validation, of which 10 were identified as SMN2-cDNA clones and the remaining 4 were SMN1-cDNA clones.In 4 SMN1-cDNA clones, 2 presented abnormal splicing (2 bases AG was added at the 5 -end of the alternatively spliced exon 5 predicting will lead to a premature stop codon (p.Pro210SerfsTer4)), but the other 2 showed normal splicing (Fig. 4A).All 10 SMN2-cDNA clones were identified as wild-type cDNA sequences without abnormal splicing (Fig. 4B).The retained normal-spliced SMN1 gene can produce a portion of the SMN protein with normal function; thus, the patient's phenotype did not become more severe.This aberrant splicing presented on SMN1-cDNA was consistent with splicing pattern 1 predicted by bioinformatics (Fig. 3A) and supported the patient's

Splicing study of SMN1 c.628-3T>G by minigene assay
To further characterize the altered splicing identified in vivo, we also conducted minigene analysis in the wild type and mutant type with SMN1 c.628-3T>G (Fig. 3B).A single band from the wild type (expected 451 bp) was presented by agarose gel electrophoresis of the amplified products after RT-PCR, but the mutant type showed an unexpected band that predicted less than 400 bp (Fig. 5A).Sanger sequencing showed a normal splicing isoform from the wild type and an aberrant splicing with exon 5 deletion (96 bp) for the mutant type (Fig. 5B and C).This result was inconsistent with the splicing study from peripheral blood of the proband but consistent with the second prediction (Splice pattern 2) by the online RNA Splicer tool (Fig. 3A).However, this result did not support the clinical phenotype of the patient.

Treatment
When the patient was 8 months old and had genetically acquired a definitive SMA type I diagnosis, she began to receive treatment with risdiplam.During the treatment, we evaluated the score of the patient's Children's Hospital of Philadelphia Infant Test of Neuromuscular Disorders (CHOP-INTEND), which is a reliable standard for assessing motor skills in patients with SMA and infantile neuromuscular diseases [29].This measure contains 16 items of motor function, each rated from 0 to 4, including spontaneous movements of the upper and lower limbs, trunk, hip, and shoulder and head control.The total scores ranged from 0 to 64, with higher scores indicating better motor function.At 3 months post-risdiplam treatment, the patient showed moderate improvement in hypotonia, head control, and lower limb and hand prehensile strength.The patient maintained normal breathing and swallowing function and could sit unassisted for 5 s (Fig. 6A), with the CHOP-INTEND score increasing from 8 to 26.Such a remarkable improvement was beyond our expectations, as the patient had only one copy of SMN2.After 7 months of treatment, she could sit alone and play with toys for 2 min, with the CHOP-INTEND score increasing from 26 to 40 (Fig. 6B).

Discussion
SMA is a leading monogenic cause of infant death worldwide.The estimated incidence of SMA ranges from 1:6000 to 1:1000, with a carrier frequency of 1:25 to 50 in most populations [3,4,[30][31][32].A total of more than 95% of patients with SMA (this proportion is lower in African black patients with SMA) present homozygous absence of at least exon 7 of the SMN1 gene on chromosome 5q13, and the remaining patients (less than 5%) show compound heterozygous variants with an intragenic mutation on one allele and a deletion or gene conversion on the other [33][34][35].SMN2, a nearly identical replicating gene to SMN1, has a variable copy number and can produce insufficient but essential levels of SMN protein.Given that an inverse relationship between copy numbers of the SMN2 gene and severity of the SMA phenotype has been presented [8,10,36], assessment of the SMN2 copy number in patients with SMA is essential to clearly elucidate genotype-phenotype correlations and predict the severity of the phenotype.To date, 147 variants have been recorded in the HGMD professional database, which included 111 intragenic mutations, with 56 missense/nonsense mutations, 14 splicing mutations, 40 frameshift mutations, and 1 regulatory mutation.Predicting the effects of intragenic mutations is usually not easy, which prevents a clear genotype-phenotype correlation, even though these mutations may be highly detrimental to the stability and function of SMN protein [37,38].Therefore, to establish a clear genotypephenotype correlation for a patient with SMA and an intragenic mutation, two aspects of the effect of the intragenic mutation and SMN2 copy number should be considered.Here, we identified a novel splicing variant c.628-3T>G in a patient with SMA type I and only one SMN2 copy.We reviewed all 14 splicing mutations recorded in the HGMD professional database and found that the majority of these variants were located near exon 7 (12/14) [1,33,[39][40][41][42][43][44][45][46][47][48], only one was in intron 3 [49], and the other was in intron 4 [50].By contrast, the variant in our case was located in intron 4 (Table 1).All the 14 patients had one deleted SMN1 allele (at least exon 7) in addition to splicing mutation allele.Half of all 14 cases did not have accurate splicing effect studies or SMN2 copy number(s) that made it difficult to establish clear genotypephenotype correlations.No. 1 case (Table 1), with one SMN2 copy and a complete exon 7 skipping of SMN1 transcripts caused by the splicing mutation of the single SMN1 copy, presented a very severe phenotype of type 0 SMA.By contrast, our case, despite having only one SMN2 copy, had a less severe clinical phenotype, which may be attributed to its low level of fully functional SMN protein from the SMN1 transcripts.The two other cases (cases 4 and 14 in Table 1), with two SMN2 copies, showed an unexpected mild SMA phenotype (SMA types III and IV).A similar explanation for this result was that the splicing mutations in cases 4 and 14 both induced a modest exclusion of exon 7 from the SMN1 transcripts, and part of the normal functioning SMN protein from SMN1 was retained.The above results showed that SMN1 splicing mutations may contribute more significantly to the severity of SMA phenotype than SMN2 copy numbers.
Interestingly, cases 13 and 14 both had two SMN2 copies and one mutation at the same site 3 nucleotides upstream of exon 7 in SMN1 (site c.835-3 of SMN1, case 13 with c.835-3C>A and case 14 with c.835-3C>T), but the clinical severity of the two varied greatly (case 13 with SMA type I and case 14 with SMA type IV); thus, splicing mutations caused by different base substitutions at the same site may have various effects on splicing.According to the consensus for splice acceptor sites [51], the three-nucleotide sequence YAG|(Exon) upstream of exon near the intron-exon junction (positions −3, −2, −1|(Exon) of the exon, Y = pyrimidine), as one of these sites, is recognized by U2AF 35 .Therefore, case 13 featured a severe clinical phenotype, probably because the mutation changed the pyrimidine of the nucleotide at position −3 of the exon from cytosine to adenine (C-to-A).Although the mutation in case 14 did not change the nature of the pyrimidine, the authors argued that c.835-3C>T probably induced the exclusion of exon 7 from SMN1 transcripts by reinforcing a hnRNP A1 binding site that spanned the acceptor site of the exon [47], so it had minimal effects on splicing and the case had a mild clinical phenotype.Nevertheless, a mutation at position −3 of the exon from a pyrimidine to a purine leads to unstable binding of U2AF 35 to this site that allows spanning the acceptor site of the exon, which may cause a harmful effect on the splicing of the gene's pre-mRNA.Notably, the mutation c.628-3T>G in our case was also at site −3 of the exon; only the insertion of two bases (AG) and a small amount of normal splicing but no exon skipping were found in the in vivo study.One reliable explanation is that this mutation created a new splice acceptor site (YAG) that could be stably bound by U2AF 35 , but a small portion remained bound to the original site (Fig. 7).Although the minigene splicing assay is considered a reliable, effective and relatively simple tool to functionally analyze potential splicing [52], surprisingly, in vitro minigene assay with the mutation c.628-3T>G detected a complete exon skipping that was inconsistent with our in vivo study.Exons present in heterologous environments may not necessarily characterize the exact splicing profile of the endogenous gene [53].This result suggested that the findings can be confirmed by using RNA from patients whenever possible.
In generall, the SMN2 copy number has been used to predict disease severity, but the role of the SMN2 copy number in patients treated with DMTs is unclear.Nowadays, three molecular therapies (onasemnogene abeparvovec, nusinersen, and risdiplam) with different therapeutic mechanisms and administration routes have been approved for the treatment of SMA.As a result of the lack of comparative data among these approved molecular therapies, the most appropriate individualized treatment for SMA has not yet been determined.Risdiplam (Evrysdi™), as an oral RNA splicing modifier targeting SMN2, has been approved for the treatment of patients with SMA or a clinical diagnosis of SMA type 1, 2 or 3 or 1-4 SMN2 copies (EU) [26,27].However, we reviewed four current ongoing clinical trials assessing risdiplam: the FIREFISH (NCT02913482), the JEWELFISH (NCT03032172), the RAINBOWFISH (NCT03779334), and the SUNFISH (NCT02908685) trials [54][55][56][57][58].We found that almost no patients with one SMN2 copy were included in their study, or if they did, it would not provide a detailed reference for individualized treatment.Our study is the first report on the treatment of risdiplam in a patient with one SMN2 copy in a real-world setting.On the basis of the functional skills reported by the patient's parents and CHOP-INTEND assessment, the patient showed remarkable motor improvement.We suppose that this favorable clinical response was due to the fact that the patient in our study had higher concentrations of fully functional SMN protein than those who had complete deletion of the SMN1 gene but only two copies of the SMN2 gene.Therefore, a slight increase in the concentration of fully functional SMN protein after treatment with risdiplam could lead to good clinical improvement.However, a limitation of this study is that SMN protein concentrations were not measured before and after treatment with risdiplam.No laboratory abnormalities or major side effects were detected.Assessment of the long-term efficacy and any delayed onset side effects of risdiplam is ongoing by comprehensive follow-up data collection.
Overall, first, this study is the first report on the treatment of risdiplam in a patient with one SMN2 copy in a real-world setting.Our findings expand the mutation spectrum of SMA, provide accurate genetic counseling information, and clarify the molecular mechanism of careful genotype-phenotype correlation of the patient.Second, our work highlights that the functional analysis of splicing variant is important to establish careful genotypephenotype correlations and predict disease severity, particularly to evaluate their effects on RNA splicing.Meanwhile, our findings underscore the importance of using in vivo assays to accurately determine the effect on RNA splicing of patients with SMA, as in vitro assays may not accurately ref lect splicing results in patients.Third, functional analysis of intragenic mutations in patients with SMA may provide some guidance for the selection of individualized treatment for patients.Finally, our study showed    that risdiplam could be effective in patients with SMA type I who have only one copy of SMN2 but with high levels of functional SMN protein from SMN1.

General information and clinical examination of the patient
The propositus, a female infant, was born uneventfully to nonconsanguineous parents following a normal pregnancy and delivery without family history of known genetic neurological disorders.As reported by the patient's parents, she presented a

Next-generation sequencing (NGS) and variant calling
For the consideration of a genetic neuromuscular disease, NGS was applied to screen the genetic cause of the patient.Peripheral blood from the patient and her parents was collected in an EDTA vacutainer.The QIAamp DNA Blood Midi Kit (Qiagen, Shanghai, China) was used for genomic DNA extraction of blood samples.A NanoDrop 2000 ultraviolet spectrophotometer (Thermo Fisher, USA) was used to determine the extracted DNA.GenCap MedE006 capture kit (MyGenostics, Beijing, China) was applied for clinical exome sequencing to screen for variants in the probands with a HiSeq X ten sequencer (Illumina, United States).After sequencing, the raw data were exported and stored as a FASTQ format.Low quality sequences, such as Illumina sequencing adapters and low quality reads (< 80 bp), were filtered using cutadaptor software (http://code.google.com/p/cutadapt/).The clean reads were mapped to the UCSC human reference genome (version GRCh37/hg19) with the BWA parameter in Sentieon software (https://www.sentieon.com/).A BAM file was then generated.The driver parameter of Sentieon software was used to mark and remove the duplicated reads and correct the base.The quality of corrected bases can ref lect a real mismatch probability with the reference genome.The driver parameter of sentieon software was used to locally reassemble the active region (that is, regions that are more different from the reference genome) and call the SNPs and InDels variants.The data were converted to VCF format.ANNOVAR software (http://annovar.openbioinformatics. org/en/latest/) was further used for annotation of variations.The database of normal human genomes was used to determine the frequency of mutations.Variant frequencies were determined in 1000 Genomes Project, Genome Aggregation Database (gno-mAD), Exome Aggregation Consortium (ExAC), Exome Sequencing Project 6500 (ESP6500), and in-house databases to remove common variants (sub-allelic frequency > 5%).Potential diseasecausing variants associated with the patient-standardized HPO phenotype were prioritized for this study.The pathogenicity of novel variants was predicted by SIFT, MutationTaster, PolyPhen-2, and GERP++.Segregation analysis of the suspicious variations was performed by Sanger sequencing.Whole genome copy number variation (CNV) analysis was conducted with the FASTQ format and BAM file based on the above methods.The CNV information was obtained with CNVkit software (https://cnvkit.readthedocs.io/en/stable/).

Multiplex ligation-dependent probe amplification (MLPA) analysis
Blood samples were collected from the patient, and genomic DNA was extracted as described above.MLPA analysis was performed to detect the copy numbers of exons 7 and 8 of both the SMN1 and SMN2 genes using the SALSA MLPA Probemix P060 SMA Carrier kit (P060-B2, MRC-Holland, Amsterdam, the Netherlands) according to the manufacturer's protocols.This kit contains a total of 21 MLPA probes, including two probes each for SMN1 and SMN2, and 17 reference probes for detecting sequences outside this region.
The amplified products were between 154 and 342 nt and separated by capillary electrophoresis.The results were analyzed and compared with positive control DNA samples using Coffalyser.Net software.

Bioinformatics analysis
In this study, we used the online RNA Splicer tool (https://rddc.tsinghua-gd.org/en/ai/rna-splicer)to predict the splicing patterns of potential splicing variants and evaluated the pathogenicity of novel splicing variants by SpliceAI software.

RNA extraction, RT-PCR, and mutation identification after cloning
Total RNA from the proband's peripheral blood samples was extracted using Trizol reagent (Invitrogen, USA) according to the protocol of the manufacturer.In brief, TRIzol reagent was added to the peripheral blood sample to lyse blood cells with a 3:1 volume ratio between the former and the latter.Subsequently, chloroform was added to the mixture from the previous step to separate the RNA into the upper aqueous phase.Finally, RNA was recovered by precipitation using isopropanol.The extracted RNA was determined with a NanoDrop 2000 spectrophotometer (Thermo Scientific, USA).A Takara PrimeScript™ RT reagent Kit (TaKaRa) was used to obtain cDNA from RNA reverse transcription following the manufacturer's protocol.Given the high homology of SMN1 and SMN2, the PCR products amplified from the cDNA of the SMN1/SMN2 gene in the patient with a forward primer SMN-F (5 -TTGGTACCGAGCTCGGATCCATGGCGATGAGCAGCGGC-3 ) and a reverse primer SMN-R (5 -ACGGGCCCTCTAGACTCGAGTTAATTT-AAGGAATGTGAGCACCTTC-3 ) were cloned into the pMini-CopGFP vector (Beijing Hitrobio Biotechnology Co., Ltd) with a ClonExpress II One Step Cloning Kit (Vazyme, Nanjing, China).This step was performed to confirm whether the identified splicing variant by Sanger sequencing described above has aberrant splicing effects in the mature mRNA of the SMN1 or SMN2 gene.The cloned SMN-cDNA plasmid was eventually confirmed by Sanger sequencing.

Minigene splicing assay
We carried out a minigene splicing assay in vitro to further characterize possible splicing effects resulting from the c.628-3T>G variant in SMN1.The minigene region was designed to span exon 4, intron 4, exon 5, intron 5, and exon 6 of the SMN1 gene (Fig. 3B) amplified from the normal human gDNA with a forward primer (5 -TTGGTACCGAGCTCGGATCCAATGAAAATGAAAGCCAAGTTT-CA-3 ) and a reverse primer (5 -ACGGGCCCTCTAGACTCGAGCAT-ATAATAGCCAGTATGATAGCCACTC-3 ).With a ClonExpress II One Step Cloning Kit (Vazyme, Nanjing, China), the amplified products were cloned into the pMini-CopGFP vector (Beijing Hitrobio Biotechnology Co., Ltd).The cloned wild-type recombinant SMN-plasmid was validated by Sanger sequencing.

Figure 1 .
Figure 1.Genetic analysis of candidate SMN1 and SMN2 variants.(A) Result for CNV analysis of NGS indicating a loss of heterozygosity of SMN1 and SMN2 genes; (B) Results for MLPA, indicating the copy number of exons 7 and 8 of SMN1/SMN2.

Figure 2 .
Figure 2. Identification of intragenic mutation in SMN1/SMN2.(A) Pedigree chart: Solid circle denotes the affected patient (II1).An open circle and square denotes the mother (I2) and father (I1), respectively.(B) Sanger sequencing showed that c.628-3T>G was the paternal variant.(C) Reanalysis of NGS data from the BAM file showing the mutation c.628-3T>G (only a portion of the data for all 384 reads is shown in the rectangular box).

Figure 4 .
Figure 4. Splicing study of c.628-3T>G by sanger sequencing of SMN-cDNA clones.(A) Sanger sequencing of SMN-cDNA clones showing an abnormal splicing (2 bases AG was added at the 5 -end of the alternatively spliced exon 5 predicting will lead to a premature stop codon (p.Pro210SerfsTer4)) and a normal splicing in the SMN1-cDNA clones.(B) Sanger sequencing of SMN-cDNA clones, indicating normal splicing in all the SMN2-cDNA clones.

Figure 5 .
Figure 5. Splicing study of c.628-3T>G by minigene assay.(A) Agarose gel electrophoresis results of RT-PCR for plasmid expression: The wild-type fragment is estimated to be 451 bp, and the mutant type is estimated to be less than 400 bp.WT = wild-type; MT = mutant type.(B) Sanger sequencing of RT-PCR for plasmid expression.WT = wild-type; MT = mutant type.(C) Schematic of splicing for c.628-3T>G in the minigene assay.

Figure 6 .
Figure 6.Demonstration of patient clinical improvement.(A) Display of the patient sitting for more than 5 s without assistance at 3 months postrisdiplam treatment.(B) Display of the patient sitting alone and playing with toys for 2 min at 3 months post-risdiplam treatment.

Figure 7 .
Figure 7. Schematic of the new splice acceptor site generation.

Table 1 .
Splicing mutations in SMN1 that have been recorded in the current HGMD database.