Article Text
PDF
Abstract
Background Numerous variants of uncertain significance (VUSs) have been identified by whole exome sequencing in clinical practice. However, VUSs are not currently considered medically actionable.
Objective To assess the splicing patterns of 49 VUSs in 48 families identified clinically to improve genetic counselling and family planning.
Methods Forty-nine participants with 49 VUSs were recruited from the Reproductive and Genetic Hospital of CITIC-Xiangya. Bioinformatic analysis was performed to preliminarily predict the splicing effects of these VUSs. RT-PCR and minigene analysis were used to assess the splicing patterns of the VUSs. According to the results obtained, couples opted for different methods of reproductive interventions to conceive a child, including prenatal diagnosis and preimplantation genetic testing (PGT).
Results Eleven variants were found to alter pre-mRNA splicing and one variant caused nonsense-mediated mRNA decay, which resulted in the reclassification of these VUSs as likely pathogenic. One couple chose to undergo in vitro fertilisation with PGT treatment; a healthy embryo was transferred and the pregnancy is ongoing. Three couples opted for natural pregnancy with prenatal diagnosis. One couple terminated the pregnancy because the fetus was affected by short-rib thoracic dysplasia and harboured the related variant. The infants of the other two couples were born and were healthy at their last recorded follow-up.
Conclusion RNA splicing analysis is an important method to assess the impact of sequence variants on splicing in clinical practice and can contribute to the reclassification of a significant proportion of VUSs. RNA splicing analysis should be considered for genetic disease diagnostics.
- genetic testing
- mutation
- sequence analysis
- RNA
Data availability statement
Data are available upon reasonable request.
Statistics from Altmetric.com
Introduction
Monogenic disorders are caused by pathogenic variations in a single gene. To date, more than 7000 monogenic disorders have been identified in humans, accounting for 5.3% of the newborns and 10% of paediatric hospitalisations, thereby posing a considerable threat to human health.1 Identification of genetic aetiologies to perform accurate diagnoses is essential for the treatment of monogenic disorders. Whole exome sequencing (WES) is now routinely used in clinical practice to identify the underlying genetic cause (pathogenic or likely pathogenic variation) of monogenic disorders.2 3 However, numerous variants of uncertain significance (VUSs) have been identified in a substantial proportion of cases. Furthermore, 48% of the variants are of uncertain significance in the ClinVar database.4
The identification of an increasing number of VUSs in genetic testing, such as missense, silent and intronic variants, poses considerable problems for genetic counselling and clinical management.5–7 Therefore, additional evidence is necessary for the reclassification of VUS. Functional assays may help provide new evidence to support reclassification, which can be used to assess the impact of genetic variants on the gene or protein product, including splicing pattern and protein activity.8 9 Many variants causing diseases have been reported to actually cause aberrant gene splicing.10 However, functional testing of splicing patterns is challenging and is currently not broadly used in clinical practice.
In the present study, we conducted RNA analysis (RT-PCR and minigene analysis) to assess the splicing patterns of 49 VUSs identified in the clinical setting. Eleven variants were found to disrupt splicing patterns and one variant caused nonsense-mediated mRNA decay (NMD), which provided additional evidence for the reclassification of their pathogenicity and lays a foundation for genetic counselling and family planning.
Materials and methods
Patients and variants
Forty-nine participants from 48 families with 49 VUSs located across 47 genes were recruited from the Reproductive and Genetic Hospital of CITIC-Xiangya. The participants were divided into the following two categories: (1) participants or their family members who were diagnosed with monogenic disorders based on parameters such as clinical manifestations and neuroimaging features (families 1–16, 19–36 and 38–48); and (2) participants who underwent preconception carrier screening (families 17–18 and 37). All variants were identified by WES, which was performed by the Beijing Genome Institute at Shenzhen, and data analysis was performed using the Genome Analysis Toolkit, as our previous study.11 Two of these variants have been previously reported, but they are also classified as VUSs.12 13 The details of the variants are described in table 1 and online supplemental table S1.
Supplemental material
Splicing analysis results for the first 12 variants with altering pre-mRNA splicing
Peripheral blood from 49 participants and their family members was collected in EDTAK2 vacuum tubes. Muscle tissue was sampled from the member of family 3. All participants provided written informed consent to participate in the study.
In silico splicing predictions
The effects on mRNA splicing of variants detected by DNA sequencing were first predicted using Alamut Batch Plus and SpliceAI.14 15 All analyses were conducted using the default threshold values of SpliceSiteFinder-like (15%),16 MaxEntScan (5%),16 17 NNSPLICE (0.4)17 and SpliceAI (0.2).15
Transcript analysis
Depending on the strategies used for different mutant genes, the following protocols were used: (1) RT-PCR based on blood or tissues: RNA was extracted from lymphocytes for the genes expressed in blood or using muscle tissues for WD repeat domain 35 (WDR35) variants; and (2) genes that cannot be amplified by peripheral blood cells were examined using the minigene system.
Minigene system
The minigene system was constructed according to the methods described by a previous study.18 Briefly, a minigene vector containing a fragment of the C1 inhibitor gene (SERPING1/C1NH), with at least two exonic regions separated by an intron, was cloned into the mammalian expression vector pcDNA3.1(−). A genomic segment encompassing two exons upstream and downstream of the variant site, as well as the complete bridging intronic sequence, was amplified by PCR using genomic DNA of the patient and cloned into the minigene expression plasmid by insertion into the SERPING1/C1NH intron. The resulting minigene construct was transiently transfected into cultured Human Embryonic Kidney 293 cells (HEK293) using Lipofectamine 3000 (Invitrogen).
RT-PCR analysis
RNA from fresh whole blood was extracted using the QIAamp RNA Blood Mini Kit (Germany) within 2 hours after venepuncture, and RNA was extracted from tissues and HEK293 cells using the TRIzol reagent (Ambion).
cDNA was synthesised using the Promega Reverse Transcription System (Promega, USA). PCR was then performed with cDNA samples using specific primers. The products were then subjected to either direct Sanger sequencing or thymine-adenine (TA) cloning combined with Sanger sequencing to determine the changes in pre-mRNA splicing. The primers used are listed in online supplemental table S2.
Supplemental material
Reproductive interventions
The couples who underwent intracytoplasmic sperm injection-in vitro fertilisation (IVF) treatment to conceive children opted for preimplantation genetic testing (PGT). On day 5 after fertilisation, trophectoderm cells that had herniated out of the zona pellucida were selected for biopsy from blastocysts for subsequent genetic testing following previously described methods.19 20 Whole-genome amplification of trophectoderm cells was performed using a commercial kit (REPLI-gMidi Kit, QIAGEN, Germany) according to the manufacturer’s instructions. Next-generation sequencing was then used to detect the target variations and to screen for the adjacent SNPs within 1 Mb upstream and downstream. The embryo genotype was also determined by haplotype analysis based on the SNP information.
Prenatal diagnosis (PND) was performed on women who conceived naturally and was performed as a confirmatory test after PGT. Amniocentesis was conducted on pregnant women at 18–20 gestational weeks. Detection of the corresponding variant in the fetus was performed using DNA samples extracted from the amniocytes.
Results
Bioinformatics-based splicing predictions
To preliminarily predict splicing effects, the genomic variant exploration software (Alamut Visual Plus and SpliceAI) was used. A total of 22 variants were predicted to be likely deleterious or deleterious by at least two splicing prediction tools. The details are presented in online supplemental table S1.
RNA splicing analysis
Expression analysis revealed that 2 of the 47 genes, namely TATA box-binding protein-like 2 (TBPL2) and MPDZ, were not or very poorly expressed in blood cells, which made them undetectable by RT-PCR. RT-PCR of blood RNA samples revealed that 10 variants caused abnormal splicing and were reclassified as likely pathogenic (c.575G>T of SLC16A2 (family 1), c.140+5G>T of CTNS (family 2), c.215-8T>G of WDR35 (family 3), c.238+6C>G of TRAPPC2 (family 4), c.765G>A of COL4A3 (family 5), c.1141-6G>T of OPA1 (family 6), c.558+4_558+7del of COL4A4 (family 7), c.908+5G>A of GALC (family 8), c.909-10A>G of GALC (family 8) and c.1573+5G>A of STAG3 (family 9)). In family 3, the splicing pattern of mutant WDR35 from the muscular tissue of the aborted fetus was the same as that derived from blood.
As TBPL2 and MPDZ cannot be amplified using blood, minigene analysis revealed that c.788+3A>G of TBPL2 (family 10) led to exon 4 skipping (r.697_788del), and c.3841-3C>G of MPDZ (family 11) resulted in inserting 2 bp of intron 27 (r.3840_3841ins3841-2_3841–1). The two variants were both predicted to generate a truncated protein (p.Arg233Ter of TBPL2 and p.Ala1281Argfs*65 of MPDZ). This finding of TBPL2 has been reported previously.21
Particularly, c.575G>T of solute carrier family 16 member 2 (SLC16A2) (from family 1) is located in the last base of exon 2. Sanger sequencing demonstrated heterozygous 575 G/T at the genomic DNA level, whereas cDNA derived from the patient’s blood cells only showed the wild-type allele (G). This indicated that the variant probably caused NMD.
In total, 11 variants altered pre-mRNA splicing and 1 variant caused NMD. Supplemented with RNA test results, the pathogenicity of variants was re-evaluated according to the American College of Medical Genetics and Genomics (ACMG) standards and guidelines for interpretation of variations. Eleven variants causing aberrant RNA splicing and 1 variant resulting in NMD were reclassified into likely pathogenic variants, while the remaining 38 variants were still of unknown significance. Wimmer et al 22 reported that splicing variants could be divided into five categories. Five and four variants were classified as type I (exon skipping; families 2 and 7–10) and type IV (activation of cryptic splice sites; families 3, 4, 6 and 8), respectively. The details are presented in figure 1, online supplemental figure 1 and table 1.
Supplemental material
Illustrative positive examples of splicing analysis. The splicing analysis of variations from the first eight families was based on RNA extracted from blood cells. The splicing analysis of variations from families 9 and 10 were reported in our earlier studies (He et al 21 29). Minigene analysis of family 11 revealed that c.3841-3C>G of MPDZ leads to the inserting 2 bp of intron 27(r.3840_3841ins3841-2_3841–1). The results showed that such variations caused abnormal splicing, including NMD, exon skipping and activation of cryptic splice sites. L indicates ladder; P, Mo, FA, S and C indicate patient, patient’s mother, patient’s father, patient’s sister and the normal control, respectively; MT, WT and EV indecate Mutant Type, WildType and Empty vector. The position of variants is in the brackets. The asterisk indicates non-specific band.
Reproductive interventions and follow-up
One couple (family 1) chose to undergo IVF with PGT. A total of five blastocysts were biopsied and tested in two IVF cycles, which revealed that four out of five embryos were free of the tested monogenic disorders. Based on the results obtained, embryo 1 (free of SLC16A2 variant) was transferred; however, this transfer yielded unsuccessful results. Subsequently, embryo 4 (free of SLC16A2 variant) was transferred in the second cycle and the patient reported pregnancy, showing a positive human chorionic gonadotropin pregnancy test. PND showed that the fetal genotype was in agreement with that of the original embryo 4.
Three couples (families 2–4) opted for natural pregnancies with PND. Two out of the three fetuses were free of the tested monogenic disorders and the couples chose to continue the pregnancies. The other couple (family 3) chose to terminate the pregnancy because the fetus was affected by short-rib thoracic dysplasia caused by homozygous c.215-8T>G in the WDR35 gene. The results of PGT and PND are summarised in table 1.
By May 2021, two infants were born (families 2 and 4) and one pregnancy was ongoing (family 1). Further follow-up revealed that the two babies were healthy.
Discussion
In the present study, RT-PCR and minigene analysis were used to investigate whether VUS affected the splicing patterns of select, mutated genes, revealing that 24.49% (12 of 49) of these variants influenced pre-mRNA processing conditions or led to NMD. Additionally, two couples delivered two healthy babies after PND, according to the results.
The ACMG and the Association for Molecular Pathology (AMP) have recommended a standard terminology to describe gene variants identified in Mendelian disorders, including pathogenic, likely pathogenic, uncertain significance, likely benign and benign.7 Variants in likely pathogenic or pathogenic categories could be used to guide medical decision-making processes for patients, whereas VUS should be considered as clinically non-actionable.23 24 However, variants can be reclassified over time on discovery of their pathogenicity. Functional studies, such as enzymatic assays, subcellular localisation and pre-mRNA splicing pattern assays, are powerful tools for understanding the pathogenicity of variants. Splicing pattern analysis can reduce the uncertainty in clinical interpretation.25 In our study, 49 variants were identified by genetic testing and were initially classified as VUSs, which could not be used to formulate medical decisions. To reassess their pathogenicity, functional assays, including RT-PCR and minigene analysis, were used to investigate the impact of these variants on splicing. The results showed that 24.49% (12 of 49) of these variants influenced pre-mRNA processing conditions that was predicted to affect protein functions. Based on the results, these VUSs were reclassified as likely pathogenic. This can help physicians formulate medical decisions, which can increase the clinical diagnostic rate and have profound implications for reproductive interventions. Therefore, we suggest that it is necessary to obtain functional evidence of VUSs and to reclassify them in the clinical setting.
It has been reported that the in silico tools might have yielded 16%–25% incorrect predicting rate for a splicing impact based on the results of RNA analysis.26 In our study, RNA analysis was also performed on a significant proportion of non-canonical splicing variants that are predicted to be benign, including the TRAPPC2 variant ultimately determined to cause aberrant splicing. It is because that the phenotypes of patients are similar with that caused by the mutant genes. In addition, the experimental evidence of abnormal splicing is more reliable than the evidence from in silico predictors according to the guidelines from the ACMG/AMP. Therefore, we suggest that it is necessary to obtain functional evidence of VUS when clinicians highly suspect that the variant is associated with the phenotype.
It has been reported that the pathogenic mechanisms of a substantial number of pathogenic variants involve splicing patterns of the corresponding genes, which is estimated to account for approximately 35% of disease variants.27 28 Splicing analysis of blood cell RNA can help clarify the functional effects of a significant proportion of VUSs and aid the reclassification of their pathogenicity.4 25 Although two variants (c.909-10A>G in GALC, c.2869A>G in SEC24D) have been reported in earlier studies, the functional impacts of each variant are unclear and thus they are classified as VUSs, according to the ACMG standards and guidelines for the interpretation of variations.12 13 In the present study, 10 variants, including the 1 reported variant (c.909-10A>G in GALC), were observed to influence splicing based on RNA samples extracted from blood cells and were reclassified as likely pathogenic (STAG3 in c.1573+5G>A reported in our earlier study).29 They accounted for ~20% of the total variants. Thus, we suggest that splicing analysis based on blood cell-derived RNA samples is a crucial method for assessing the impact of sequence variants on splicing in clinical practice.
It has been reported that there are more than 3000 genes that have been identified as tissue-specific genes and a large number of genes are not or poorly expressed in blood.30 We first investigated the expression level of the gene of interest in databases. However, low values in RNAseq-based databases are indicative of all expressed genes in the tissue, and even when the expression of the gene of interest is low compared with all other genes, they may be sufficient to be amplified by RT-PCR, such as STAG3 and SLC16A2. Thus, irrespective of the expression level in databases, we should finally perform PCR with fresh blood RNA to explore whether the transcript can be amplified in blood. In addition, when only one transcript is amplifiable, we also should assay for the presence or absence of the other heterozygous coding SNPs in the corresponding gene to confirm whether this variant leads to a much longer transcript not readily amplifiable or leads to NMD.
The prerequisite for splicing analysis based on RNA samples derived from blood or other tissues is that the corresponding genes are expressed in these samples. Unfortunately, patient RNA extracted from relevant samples is often not available in clinical practice, such as from the brain and ovary tissues. Minigene analysis can compensate for this lack of availability and can replicate the splicing pattern in vitro.18 31 In the present study, minigene analysis was used to investigate the splicing patterns of genes in vitro. The minigene analysis results showed that c.788+3A>G in TBPL2, which is highly expressed in developing oocytes, led to the skipping of exon 4 and was predicted to generate truncated TBPL2 protein, which has already been reported in our previous study.32 Minigene analysis results also demonstrated that c.3841-3C>G in MPDZ caused inserting 2 bp of intron 27 (r.3840_3841ins3841-2_3841-1). Using this method, the splicing patterns of two additional variants were assessed. Therefore, we suggest that not only patient RNA analysis based on blood but also in vitro minigene analysis should be considered in genetic disease diagnostics.
There are numerous benefits of a definitive diagnosis of monogenic disorders, such as selection of the most suitable therapy and early interventions. Another important benefit of accurate diagnosis based on genetic testing is reproductive interventions for family members, including PND and PGT.33 34 In the present study, according to the results of RNA analysis, PGT and PND for monogenic disorders were performed for one family and three families, respectively. Subsequently, the two delivered infants free of target variants were healthy, which indicated the necessity that VUSs were reclassified based on RNA detection.
Our study had two main limitations. First, the NMD of the mutant allele in SLC16A2 was reached by detecting no signal of the mutant base ‘T’ on r.575 in family 1. However, there were no other heterozygous SNP to support the result. Under the circumstance, the undetected ‘T’ might be caused by two possibilities: (1) pseudoexon activation leading to a considerably longer transcript that is not readily amplifiable; or (2) a premature termination codon introducing and resulting in NMD. These two conditions could not be distinguished here. Second, allelic expression quantification of aberrant transcripts was not performed in our study. For those variants that may affect mRNA expression but not splicing patterns, the available results could not offer functional evidence. Some relevant information can be obtained from gels and direct Sanger sequencing results, but it will not be available in a real context from variants assayed via minigene analysis.
In summary, we used two methods to assess whether the 49 VUSs affected the splicing patterns of the mutated genes. Twelve variants were found to cause abnormal splicing or lead to NMD and were reclassified as likely pathogenic. According to these results, a definitive diagnosis was formulated for these patients, followed by the birth of healthy infants after PGT and PND. These findings shed light on the importance of RNA splicing analysis in clinical practice, especially in blood cell-derived RNA analysis. Further studies are necessary to confirm these findings.
Data availability statement
Data are available upon reasonable request.
Ethics statements
Patient consent for publication
Ethics approval
This study involves human participants and was approved by the institutional ethics committees of the Reproductive and Genetic Hospital of CITIC-Xiangya (LL-SC-2019-025). Participants gave informed consent to participate in the study before taking part.
Acknowledgments
We would like to thank all individuals who participated in this study.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
W-BH, W-JX and C-LD contributed equally.
HH and JD contributed equally.
Contributors HH and JD designed the study. W-BH, WX, CD and Y-Q T analysed the data and wrote the paper. Y-RW and SZ performed the bioinformatic analysis. W-BH, WX, CD, CT and LM performed the study. XL, FG, G-XL and GL performed the clinical work. JD is responsible for the overall content of the manuscript acting as guarantor.
Funding This work was supported by grants from the National Key Research and Development Program of China (2018YFC1004901), the National Natural Science Foundation of China (81771645 and 81971447), the Hunan Provincial Natural Science Foundation of China (2019JJ51006, 2019JJ50397), the Hunan Provincial Grant for Innovative Province Construction (2019SK4012), the Key Grant of Prevention and Treatment of Birth Defect from Hunan Province (2019SK1012), and the Research Grant of CITIC-Xiangya (YNXM-201915, YNXM-201913, YNXM-201912, YNXM-202002).
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.