Functional impact of a deep intronic variant in the RPS19 gene detected in a case of Diamond-Blackfan anemia syndrome

Diamond-Blackfan anemia syndrome (DBAS) is a rare inherited disorder characterized by pure red cell aplasia, usually presenting in infancy.¹ It can also be associated with a variety of malformations and a high risk of developing malignancy. To date, causative mutations for DBAS have been found in more than 20 ribosomal protein (RP) genes,^1-3 TSR2,⁴ HEATR3,⁵ G ATA1,⁶ and TP53,⁷ but remain unclear in approximately 20% of DBAS cases.³ Here, we report the identification of a deep intronic RPS19 variant causing aberrant splicing in a patient with DBAS that was revealed by whole genome sequencing (WGS) analysis.

A boy was born at full term with bilateral polydactyly of the thumbs. He had no family history of anemia. At one month old, he presented with pallor and tachycardia, and peripheral blood analysis showed red blood cell (RBC) count 0.64x10¹²/L, hemoglobin (Hb) 2.4 g/dL, mean corpuscular volume 107.8 fL, reticulocytes 0.2%, white blood cell count 4.10x10⁹/L, platelets 258x10⁹/L, and fetal Hb 16.4%. Bone marrow aspiration revealed hypercellularity, with 0% erythroid cells. His erythrocyte adenosine deaminase activity (eADA) and glutathione (GSH) levels were 1.11 IU/g Hb (normal range 0.46-1.50 IU/g Hb) and 122 mg/dL RBC (normal range 62.4-98.9 mg/dL RBC), respectively, suggesting a DBA pattern using a previously reported formula.⁸ After diagnosis, he became transfusion dependent.

This study was conducted with the approval of the Ethics Committees of Hirosaki University and the National Institute of Global Health and Medicine, Japan, and peripheral blood samples were collected with informed consent. All experiments were performed in accordance with the principles of the Declaration of Helsinki.

Whole genome sequencing was performed on the Illumina NovaSeq 6000 platform (Illumina) with 150-base pair (bp) paired-end reads. The target sequencing depth was approximately 30× for each individual. Quality-filtered reads were aligned to the human reference genome (GRCh38/ hg38) using BWA-MEM (version 0.7.17). Single-nucleotide variants (SNV) and small insertions/deletions (indels) were identified using GATK HaplotypeCaller (version 4.1.0.0), following the GATK Best Practices workflow. Structural variants, including deletions, duplications, inversions, and translocations, were detected using Manta (version 1.6.0). Variant annotation was performed using ANNOVAR and InterVar for SNV and small indels, and the Ensembl Variant Effect Predictor for structural variants.

Functional impacts were assessed using LOFTEE (version 1.0) for loss-of-function prediction, SpliceAI (version 1.3.1) for splicing prediction, and CADD (version 1.6) for deleterious-ness scoring. We did not detect any mutations that could cause DBAS in coding regions, but SpliceAI predicted a splice abnormality to a de novo variant in the third intron of RPS19 (hg38:chr19:41862119A>G, NM_001022:c.172+907A>G) (Figure 1A, B). This variant was not present in the 1000 Genomes Project, gnomAD or the Japanese Multi Omics Reference Panel 61KJPN database. Given the phenotype of the patient, it was considered potentially pathogenic. Barrio et al. had previously reported the variant as novel-20 (rs61761229) in a paper on RPS19 targeted resequencing using DBA and transient erythroblastopenia in childhood (TEC) samples,⁹ but it has not yet been identified as a disease-causing mutation or characterized functionally. Upon reviewing the details in the web version of SpliceAI (https://spliceailookup.broadinstitute.org), we found that the variant would create a novel donor splice site one base upstream of the variant (Donor Gain D score 0.83; values above 0.8 are high precision), and that the AG sequence of the original cryptic splicing site upstream of the variant (Acceptor Gain D score: 0.81) would result in an acceptor splice site 152 bp upstream (Figure 1A).

Reverse transcription polymerase chain reaction (RT-PCR) detected a transcript approximately 150 bp longer than the normal product exclusively in the patient, and only in trace amounts; no such transcript was observed in either parent (Figure 1C). Subsequent sequencing of the RT-PCR product identified a 152-bp cryptic exon inserted between the third and fourth exons, consistent with the prediction by SpliceAI (Figure 1D). The extremely low abundance of this aberrant transcript is likely due to nonsense-mediated mRNA decay (NMD),¹⁰ triggered by a premature termination codon (PTC) within the cryptic exon (Figure 1A, E).

To evaluate the impact of this variant on splicing, we performed genome editing using the erythroid cell line HUDEP-2.¹¹ In practice, we first created HUDEP-Cas9 cells, which constitutively express Cas9 to enhance genome editing efficiency, and performed the experiment. The methods used have been described previously.¹² A>C corresponds to the control of this experiment and A>G is a single-base substitution identical to the variant in the patient. Both were efficiently incorporated into the cells (Figure 2A). Subsequent RT-PCR confirmed that the RPS19 transcript pattern in the patient was recapitulated in HUDEP-2 Cas9 cells (Figure 2B). In the quantitative RT-PCR (qRT-PCR), a decrease in RPS19 mRNA was observed in the A>G genome-edited cells (fold change: 0.58 ± 0.09), while the transcript levels of RPL5 and RPL11 were little changed (Figure 2C). This suggests that the reduction in RPS19 mRNA caused by this A>G genome editing is specific to this gene.

We hypothesized that if the reduction in RPS19 mRNA levels observed in A>G genome-edited cells was due to aberrant splicing induced by the A>G variant, blocking the splice site of the cryptic exon could restore normal expression of RPS19 mRNA. Therefore, we transfected a S19 int3 morpholino antisense oligonucleotide (MO)¹³ targeting the splicing donor site of the cryptic exon containing the A>G substitution into genome-edited cells. As a result, qRT-PCR showed that RPS19 mRNA levels were restored according to the concentration of S19 int3 MO in A>G genome-edited cells, while no change was observed in A>C genome-edited cells (Figure 2D, left). Conversely, transcripts containing the cryptic exon in A>G genome-edited cells were reduced according to the concentration of S19 int3 MO (Figure 2D, right). Thus, the reduction in RPS19 mRNA levels in the RPS19 A>G genome-edited cells was confirmed to be due to the A>G variant causing aberrant splicing. These results suggest that this deep intronic variant causes RPS19 haploinsufficiency, which would lead to DBA.

Figure 1.Sequence of the RPS19 intronic variant and splice aberrations in the patient. (A) Schematic diagram of RPS19 gene structure and location of the variant. Black and white boxes indicate coding and non-coding exons, respectively. Arrow indicates location of a single nucleotide substitution A>G in this patient. Red box shows cryptic exon (CE) predicted to be formed by presence of this variant; arrowhead indicates a premature termination codon (PTC) is indicated. (B) Pedigree of the patient and Sanger sequencing results of the variant region using genomic DNA. Red arrows indicate the position of nucleotide substitution c.172+907A>G. (C) Reverse transcription polymerase chain reaction (RT-PCR) analysis using the primer set located on the second and fourth exons of the RPS19 gene. Arrow indicates fragments corresponding to normal transcripts; arrowhead indicates fragments corresponding to transcripts containing the cryptic exon. (D) Sanger sequencing of RPS19 RT-PCR products. Electropherograms of the boundary between the third exon and the adjacent exon are shown. (Top) Transcript containing the cryptic exon. (Bottom) Normal transcript. (E) RPS19 cDNA coding sequence with the cryptic exon. First line shows the RPS19 cDNA sequence; second line shows the corresponding amino acid. cDNA sequence of the cryptic exon region in red. The protein from this transcript is p.Ala58Glyfs*25.

Figure 2.Functional analysis of the variant by genome editing. (A) Confirmation of genome editing of RPS19 intronic variant in HUDEP-2 Cas9 cells. Genomic DNA extracted from samples four days after genome editing was subjected to Sanger sequencing. (B) Real-time polymerase chain reaction (RT-PCR) analysis using RNA extracted from samples four days after genome editing. (Top) The same primer pair used in Figure 1C was used here. Arrow indicates fragments corresponding to normal transcripts; arrowhead indicates fragments corresponding to transcripts containing the cryptic exon. (Bottom) To specifically amplify transcripts containing the cryptic exon, nested PCR was performed using samples from the above panel as templates. The primer pairs used were in the third exon and the cryptic exon. (C) Quantitative RT-PCR (qRT-PCR) analysis of RPS19 genome-edited cells at day 14. The total RPS19 transcript was quantified using the same primer pair located on the fifth and sixth exons of the RPS19 gene. The vertical axis represents a relative value with each value of A>C cells set at 1. Results are averages of 3 independent experiments. Error bars represent standard deviation. P value by two-sided t test. (D) Splicing inhibition of the cryptic exon by Morpholino antisense oligonucleotide (MO). RPS19 genome-edited day 14 cells that had been frozen were thawed and transfected with MO two days later. Samples were collected 48 hours after transfection and qRT-PCR was performed. + and – symbols at the bottom of the graphs indicate whether control MO or S19 int3 MO was used. The target sequence of S19 int3 MO is 5’-GCTTC-CCACCTACTTCTTACCTGGA-3’, the underlined nucleotide indicates the antisense base of the patient’s variant. + indicates addition of MO at 10 μM. In lanes 6 to 9 of the left graph and lanes 3 to 6 of the right graph, MO was increased to 0.05, 0.5, 5 and 10 μM. (Left) The total RPS19 transcript was quantified using the same primer pair as in Figure 2C. The vertical axis represents a relative value of A>C cells with control MO 10 μM (lane 2) set at 1. (Right) Transcripts containing the cryptic exon were specifically quantified and the primer pair used was the same as in Figure 2B, bottom. The vertical axis represents a relative value of A>G cells with control MO 10 μM (lane 2) set at 1. Results are averages of 3 independent experiments. Error bars represent standard deviation. Two-sided t tests compared to lane 2 values for both graphs, respectively. *P<0.05, **P<0.01, ***P<0.001.

Identification of responsible genes for inherited disease is crucial for appropriate diagnosis, and for selecting and developing treatment. In recent years, WGS has come into use, but there are still cases where the causative gene of DBAS cannot be identified. This indicates that the genetic abnormality that affects DBAS may not be restricted to the coding regions but may exist in deep intron regions or expression regulatory regions. Deeper insight is needed to interpret the biological significance of mutations in these regions. In this study, we utilized in silico analysis tools to pick up a variant in the deep intron region that causes splice abnormalities.

Recently, Wen et al. reported 2 cases of DBAS in which WGS was initially negative, but further analysis had revealed a splice mutation at the end of the first non-coding exon of RPS7 and a deep intronic de novo mutation in RPS19.¹⁴ In their paper, they also made use of splice prediction tools. They showed that analysis of variants in non-coding regions is useful for the diagnosis and treatment of DBAS, and our current study supports their findings. Meanwhile, they reported that their RPS19 mutation (NM_001022.4:c.172+350C>T) is a splicing mutation that creates a novel exon in intron 3, resulting in the appearance of a PTC. This is similar to our variant but differs in that their data show that the transcripts containing the novel exon account for 50.9% of the total RPS19 mRNA, which is considerably more than ours. To investigate the reason for this difference, we analyzed their transcript sequences. Contrary to their interpretation, the c.172+350C>T variant appeared to cause an 81-bp insertion without a PTC, resulting in the in-frame addition of 27 amino acid. We, therefore, consider our transcript with a PTC and their transcript without a PTC as different types. It is important to note that, in the case of a deep intronic variant resulting in PTC, such as in our case, transcripts could be readily degraded by NMD, preventing mutation function analysis.

MO can induce exon skipping by inhibiting mRNA splicing, and has already been applied in studies of neuromuscular diseases such as Duchenne muscular dystrophy.¹⁵ In this study, MO was also effective in suppressing abnormal splicing of RPS19 mRNA, so it may be possible to use MO in the future to treat DBAS with a splicing abnormality that causes extra exons to appear in the transcripts.

In conclusion, by extending the scope of analysis to deep introns of WGS data, we identified a variant that causes splicing abnormalities in RPS19 in a case of DBAS. This study demonstrates the importance of analyzing deep intronic regions in the search for mutations causing DBAS when no mutations are detected in coding regions.

Footnotes

• Received October 29, 2025
• Accepted December 30, 2025

Correspondence

E. ITO - eturou@hirosaki-u.ac.jp

Disclosures

No conflicts of interest to disclose.

Contributions

Funding

This work was supported by the research grant (Research on Measures for Intractable Disease: 25FC1010) from the Ministry of Health, Labour and Welfare of Japan and Practical Research Project for Rare/Intractable Disease from Japan Agency for Medical Research and Development (AMED).

Acknowledgments

The authors would like to thank H. Kudo and Y. Kudo for their technical assistance. We would also like to thank Dr. Atsushi Miyawaki for use of the HUDEP-2 cell line.