Functional splicing analysis in an infantile case of atypical hemolytic uremic syndrome caused by digenic mutations in C3 and MCP genes

An accurate prediction model of digenic interaction for estimating pathogenic gene pairs of human diseases

•

Numerous pathogenic interactions are yet to be revealed efficiently due to dimension burden.

•

Existing methods are underpowered and inaccurate to estimate pathogenic interactions.

•

We developed an accurate bioinformatic method to predict digenic interaction effects based on gene-level features.

•

The new method created a valuable resource of genome-wide pathogenic digenic interactions.

•

We found that known causal genes in Mendelian and Oligogenic diseases may be enriched with interactive effects for the first time.

Increasing evidence shows that genetic interaction across the entire genome may explain a non-trivial fraction of genetic diseases. Digenic interaction is the simplest manifestation of genetic interaction among genes. However, systematic exploration of digenic interactive effects on the whole genome is often discouraged by the high dimension burden. Thus, numerous digenic interactions are yet to be identified for many diseases. Here, we propose a Digenic Interaction Effect Predictor (DIEP), an accurate machine-learning approach to identify the genome-wide pathogenic coding gene pairs with digenic interaction effects. This approach achieved high accuracy and sensitivity in independent testing datasets, outperforming another gene-level digenic predictor (DiGePred). DIEP was also able to discriminate digenic interaction effect from bi-locus effects dual molecular diagnosis (pseudo-digenic). Using DIEP, we provided a valuable resource of genome-wide digenic interactions and demonstrated the enrichment of the digenic interaction effect in Mendelian and Oligogenic diseases. Therefore, DIEP will play a useful role in facilitating the genomic mapping of interactive causal genes for human diseases.

The Contribution of COL4A5 Splicing Variants to the Pathogenesis of X-Linked Alport Syndrome

X-linked Alport syndrome (XLAS) is caused by pathogenic variants in COL4A5 and is characterized by progressive kidney disease, hearing loss, and ocular abnormalities. Recent advances in genetic analysis and further understanding of genotype-phenotype correlations in affected male patients raises the importance of detecting splicing variants in COL4A5. Aberrant splicing of COL4A5 is caused not only by canonical splice site variants but also non-canonical splice site variants such as deep intronic changes or even substitutions in exons. Patients with splicing variants account for ~15% of all cases in XLAS. In addition, it has been shown that there is a significant difference in kidney survival depending on the aberrant splicing patterns of transcripts- in particular in-frame or out-of-frame nucleotide changes in transcripts. Therefore, cDNA analysis of patient mRNA is necessary to determine the impact of splice site variants and to confirm a diagnosis of XLAS and to predict the kidney prognosis. However, it is usually difficult to amplify COL4A5 transcripts extracted from peripheral blood leukocytes. For these cases, in vitro minigene assays or RNA sequence extracted from urine derived cells can confirm aberrant splicing patterns. Moreover, controlling aberrant splicing by nucleic acids or small molecular compounds in genetic diseases are attracting attention as a potential therapeutic strategy. Here, we review the frequency of splicing variants in COL4A5, the latest diagnostic strategies, and the prospects for new therapeutic approaches.

Usefulness of functional splicing analysis to confirm precise disease pathogenesis in Diamond-Blackfan anemia caused by intronic variants in RPS19

Diamond-Blackfan anemia (DBA) is mainly caused by pathogenic variants in ribosomal proteins and 22 responsible genes have been identified to date. The most common causative gene of DBA is RPS19 [NM_001022.4]. Nearly 180 RPS19 variants have been reported, including three deep intronic variants outside the splicing consensus sequence (c.72-92A > G, c.356 + 18G > C, and c.411 + 6G > C). We also identified one case with a c.412-3C > G intronic variant. Without conducting transcript analysis, the pathogenicity of these variants is unknown. However, it is difficult to assess transcripts because of their fragility. In such cases, in vitro functional splicing assays can be used to assess pathogenicity. Here, we report functional splicing analysis results of four RPS19 deep intronic variants identified in our case and in previously reported cases. One splicing consensus variant (c.411 + 1G > A) was also examined as a positive control. Aberrant splicing with a 2-bp insertion between exons 5 and 6 was identified in the patient samples and minigene assay results also identified exon 6 skipping in our case. The exon 6 skipping transcript was confirmed by further evaluation using quantitative RT-PCR. Additionally, minigene assay analysis of three reported deep intronic variants revealed that none of them showed aberrant splicing and that these variants were not considered to be pathogenic. In conclusion, the minigene assay is a useful method for functional splicing analysis of inherited disease.

Exon-Trapping Assay Improves Clinical Interpretation of COL11A1 and COL11A2 Intronic Variants in Stickler Syndrome Type 2 and Otospondylomegaepiphyseal Dysplasia

Stickler syndrome (SS) is a hereditary connective tissue disorder affecting bones, eyes, and hearing. Type 2 SS and the SS variant otospondylomegaepiphyseal dysplasia (OSMED) are caused by deleterious variants in COL11A1 and COL11A2, respectively. In both genes, available database information indicates a high rate of potentially deleterious intronic variants, but published evidence of their biological effect is usually insufficient for a definite clinical interpretation. We report four previously unpublished intronic variants in COL11A1 (c.2241 + 5G>T, c.2809 − 2A>G, c.3168 + 5G>C) and COL11A2 (c.4392 + 1G>A) identified in type 2 SS/OSMED individuals. The pathogenic effect of these variants was first predicted in silico and then investigated by an exon-trapping assay. We demonstrated that all variants can induce exon in-frame deletions, which lead to the synthesis of shorter collagen XI α1 or 2 chains. Lacking residues are located in the α-triple helical region, which has a crucial role in regulating collagen fibrillogenesis. In conclusion, this study suggests that these alternative COL11A1 and COL11A2 transcripts might result in aberrant triple helix collagen. Our approach may help to improve the diagnostic molecular pathway of COL11-related disorders.

Dual Fluorescence Splicing Reporter Minigene Identifies an Antisense Oligonucleotide to Skip Exon v8 of the CD44 Gene

Splicing reporter minigenes are used in cell-based in vitro splicing studies. Exon skippable antisense oligonucleotide (ASO) has been identified using minigene splicing assays, but these assays include a time- and cost-consuming step of reverse transcription PCR amplification. To make in vitro splicing assay easier, a ready-made minigene (FMv2) amenable to quantitative splicing analysis by fluorescence microscopy was constructed. FMv2 was designed to encode two fluorescence proteins namely, mCherry, a transfection marker and split eGFP, a marker of splicing reaction. The split eGFP was intervened by an artificial intron containing a multicloning site sequence. Expectedly, FMv2 transfected HeLa cells produced not only red mCherry but also green eGFP signals. Transfection of FMv2CD44v8, a modified clone of FMv2 carrying an insertion of CD44 exon v8 in the multicloning site, that was applied to screen exon v8 skippable ASO, produced only red signals. Among seven different ASOs tested against exon v8, ASO#14 produced the highest index of green signal positive cells. Hence, ASO#14 was the most efficient exon v8 skippable ASO. Notably, the well containing ASO#14 was clearly identified among the 96 wells containing randomly added ASOs, enabling high throughput screening. A ready-made FMv2 is expected to contribute to identify exon skippable ASOs.

Functional analysis of suspected splicing variants in CLCN5 gene in Dent disease 1

Background

In recent years, the elucidation of splicing abnormalities as a cause of hereditary diseases has progressed. However, there are no comprehensive reports of suspected splicing variants in the CLCN5 gene in Dent disease cases. We reproduced gene mutations by mutagenesis, inserted the mutated genes into minigene vectors, and investigated the pathogenicity and onset mechanisms of these variants.

Methods

We conducted functional splicing assays using a hybrid minigene for six suspected splicing variants (c.105G>A, c.105+5G>C, c.106−17T>G, c.393+4A>G, c.517−8A>G, c.517−3C>A) in CLCN5. We extracted information on these variants from the Human Gene Mutation Database. We reproduced minigene vectors with the insertion of relevant exons with suspected splicing variants. We then transfected these minigene vectors into cultured cells and extracted and analyzed the mRNA. In addition, we conducted in silico analysis to confirm our minigene assay results.

Results

We successfully determined that five of these six variants are pathogenic via the production of splicing abnormalities. One showed only normal transcript production and was thus suspected of not being pathogenic (c.106−17T>G).

Conclusion

We found that five CLCN5 variants disrupted the original splice site, resulting in aberrant splicing. It is sometimes difficult to obtain mRNA from patient samples because of the fragility of mRNA or its low expression level in peripheral leukocytes. Our in vitro system can be used as an alternative to in vivo assays to determine the pathogenicity of suspected splicing variants.

Electronic supplementary material

The online version of this article (10.1007/s10157-020-01876-x) contains supplementary material, which is available to authorized users.

Determination of the pathogenicity of known COL4A5 intronic variants by in vitro splicing assay

X-linked Alport syndrome (XLAS) is a congenital renal disease caused by mutations in COL4A5. In XLAS cases suspected of being caused by aberrant splicing, transcript analysis needs to be conducted to determine splicing patterns and assess the pathogenicity. However, such analysis is not always available. We conducted a functional splicing assay using a hybrid minigene for seven COL4A5 intronic mutations: one was identified by us and six were found in the Human Gene Mutation Database. The minigene assay revealed exon skipping in four variants, exon skipping and a 10-bp insertion in one variant, and no change in one variant, which appeared not to be pathogenic. For one variant, our assay did not work. The results of all three cases for which transcript data were available were consistent with our assay results. Our findings may help to increase the accuracy of genetic test results and clarify the mechanisms causing aberrant splicing.