The study protocol was approved by the Humanities and Ethics Committee of Fujian Children’s Hospital, Fujian Branch of Shanghai Children’s Medical Center Affiliated to Shanghai Jiaotong University School of Medicine (2022ETKLR07003). The parents and five healthy volunteers provided written informed consent before peripheral blood samples were collected in accordance with institutional policies and the Declaration of Helsinki.

The Shanghai Children’s Medical Center maintained Ccryopreserved K562 and 293T cells. The 293T cells were maintained in DMEM supplemented with 10% fetal bovine serum (FBS) (Sigma), and the K562 cells were maintained in RPMI 1640 medium supplemented with 10% FBS. All the cells were grown in a 37°C incubator with 5% CO₂.

Genomic DNA was isolated from the patient’s and his parent’s peripheral blood, while an oral scraping sample was collected as a germline control. Whole-exome sequencing (WES) was performed by inputting 500 ng of genomic DNA from the patient. Briefly, the WES library was prepared by KindStar, Inc. Exome capture using the MGIEasy Exome Capture V4 Probe (MGI) was followed by paired-end read sequencing (2 × 100 bp read length) on the MGISEQ-2000 platform, with an average depth of ≥100-fold. Exome sequencing data were analyzed as described previously (Ulirsch et al., 2018).

The GATA1 genotype of samples from the patient’s blood and oral scrapings, and parents’ blood was examined using Sanger sequencing. Polymerase chain reaction (PCR) was performed to amplify the relevant DNA fragments using the primers specified in Supplementary Table S1.

A 320-bp region encompassing GATA1 ChrX: 48,649,737 was amplified using CloneAmp HiFi PCR Premix (Clontech) (primer sequences listed in Supplementary Table S1) and cloned into the pGL4.23-mini/P vector with a minimal SV40 promoter upstream of the firefly luciferase gene sequence. Mutagenesis was performed using the QuikChange II Site-Directed Mutagenesis Kit (Agilent Technologies), with the primers listed in Supplementary Table S1. For reporter assays, 2 × 10⁶ K562 cells were resuspended in 100 μL of Nucleofector Solution Kit V (Lonza) with the addition of 1.9 μg of pGL4.23 constructs and 100 ng of renilla pTK plasmid. The cells were electroporated and incubated for 24 h at 37°C with 5% CO₂. Similarly, HEK293T cells (6 × 10⁴) were plated in 96-well plates (flat bottom), co-transduced with 95 ng pGL4.23 constructs and 5 ng renilla pTK, and incubated for 24 h. The experiments were performed in triplicate. Firefly luciferase activity was normalized to the Renilla luciferase activity to control for cell number and transfection efficiency. Measurements are presented as a ratio relative to the activity of the pGL4.23-mini/P vector with the wild-type G allele.

Total RNA was extracted using the RNeasy Micro Kit (Qiagen, Hilden, Germany) according to the manufacturer’s protocol. We reverse transcribed 500 ng of RNA into cDNA, and the qRT-PCR was performed using an ABI Prism 7900HT detection system (Applied Biosystems) with FastStart SYBR Green master mix (Roche). GAPDH was used as an internal control. The primer sequences used in this study are listed in Supplementary Table S1.

EDTA-treated blood was stained with monoclonal PE-conjugated CD33, PE-conjugated CD55, and FITC-conjugated CD59 and analyzed by flow cytometry as previously described (Abdulhay et al., 2019).

Genomic DNA was isolated using the QIAamp DNA Blood Mini Kit (Qiagen, Hilden, Germany). A chromosomal microarray assay was performed using a high-resolution genotyping single nucleotide polymorphism microarray (Affymetrix CytoScan 750 K Array; Affymetrix, Santa Clara, CA, United States). CNVs were identified based on records associated with the human reference genome 37 (hg19) of the National Center for Biotechnology Information.

SPSS version 26.0 for Windows (Armonk, NY, IBM Corp) was used for data analysis. p < 0.05 was considered to be statistically significant.

A 5-year-old child was diagnosed with severe anemia for over 2 years ago. Physical examination revealed pallor without obvious deformities, palpable lymphadenopathy, or hepatosplenomegaly. Laboratory tests revealed mild macrocytic anemia with 94.5–99.6 fL mean corpuscular volume, as indicated by a 44 g/L hemoglobin level and an absolute reticulocyte count of 13.6 × 10⁹/L. Bone marrow aspiration (BMA) revealed erythroid aplasia with normal granulopoiesis and thrombocytopoiesis (Figures 1A, B). Autoimmune and infectious diseases were excluded from systemic laboratory tests. Therefore, the pure red cell aplasia (PRCA) diagnosis was made. A series of systemic hematopoietic tests were performed to determine the cause further. A bone marrow biopsy revealed 30% bone marrow cellularity and increased adipose tissue proliferation. Abnormal localization of immature precursor cells (ALIP) has not yet been identified. The results of CD55, CD59, and FLAER assays evaluated by flow cytometry were normal (Supplementary Figures S1A, B). No copy number variation was identified among chromosomes 5 and 7 by chromosomal microarray assay (Supplementary Figures S2A, B), suggesting that the possibility of myelodysplastic syndromes (MDS) and paroxysmal nocturnal hemoglobinuria (PNH) was excluded. Whole-exome sequencing (WES) was performed to investigate the genetic alterations underlying this PRCA case. The GATA1 c.220 + 1G>C mutation found by WES was further confirmed (Figure 1C) by Sanger sequencing, which was not identified in the Genome Aggregation Database (gnomAD) SV database (125,748 genomes) or the 15,708 Genomes database (Karczewski et al., 2020). We performed Sanger sequencing of the patient’s oral scraping sample to confirm the origin of this mutation. Therefore, a GATA1 germline mutation was identified (Figure 1C). Sanger sequencing of his parents’ blood samples revealed that the GATA1 c.220 + 1G>C mutation was a de novo mutation rather than an inherited variation (Figure 1C). A diagnosis of DBS with non-RP gene mutations was made.

We performed a qPCR assay to examine GATA1 transcription to confirm the pathological role of this de novo mutation and found that the normal GATA1 (GATA1-l) transcription was significantly suppressed in comparison to healthy controls (N = 5) (p < 0.0001) (Figure 2A), indicating the pathogenic role of the mutation. By examining the chromatin state annotations of this genomic region across 42 cells and tissues from the Roadmap Epigenomics Project (Moore et al., 2020), we observed that GATA1 c.220 + 1G>C mutation was aligned with a putative weak enhancer with H3K27ac enrichment (Figure 2B). To probe the enhancer function of this regulatory DNA element and investigate how its activity is influenced by GATA1 c.220 + 1G>C, we first tested the 320-bp fragment surrounding GATA1 c.220 + 1G using a reporter gene assay in 293T cells. The fragment identical to the mutant C allele revealed a transcription effect similar to that of the wild-type allele (Figure 2C). We performed the same reporter assay using K562 cells to exclude the impact of the cellular matrix, an erythroid cell line. Figure 2C shows that GATA1 c.220 + 1G>C mutation did not impair transcription activity in this erythroid context.

Because such intronic mutations did not impair GATA1 transcriptional activity (Figures 2B, C), we hypothesized that they might alter the splicing process. To preliminarily probe this possibility, we used the RNA Splicer analytical tool developed by the Research Institute of Tsinghua (https://rddc.tsinghua-gd.org/). GATA1 c.220 + 1G>C mutation potentially disturbed the GATA1 normal transcription via three alternative splicing models (Figure 3A). We designed two primer pairs to determine the splicing effect on GATA1 transcription and quantify the two GATA1 isoforms (Figure 3B). The GATA1 short isoform (GATA1-s) was significantly higher than that in healthy controls (N = 5) (p < 0.0001) (Figure 3C). We found that the GATA1 short isoform was expressed preferentially after relating these data to the readings of the GATA1 long isoform (Figures 2A, 3C).

Before this diagnosis, red blood cell transfusion was occasionally prescribed to relieve the symptoms. Building upon DBS diagnosis, we treated the patient with prednisone at a dose of 2 mg/kg/d, which was tampered with after 2 weeks of treatment. We evaluated the treatment response through routine total cell counts and reticulocyte percentage measurements over the first month. Figure 4A shows that the hemoglobin level gradually increased from 44 to 91 g/L, accompanied by a concurrent reticulocyte increase, indicating a response effect. After 2 months of therapy, the hemoglobin level remained steady over the follow-up period, between 91 and 98 g/L (Figure 4A). We performed qPCR to investigate the effect of prednisone on GATA1 transcription and evaluate GATA1 transcription alteration after prednisone exposure. Figures 2A, 3C, 4B show that either GATA1-l or GATA1-s transcription was affected by prednisone treatment, further confirming the splicing error due to this de novo intronic mutation. We selected several erythropoiesis-related genes to primarily investigate the glucocorticosteroid responding program and perform the qPCR assay. We observed that TAL1 and MPL were up-regulated in the assay, while other genes (GATA2, RUNX1, GFIB, GP1BA, and NFE2) were not altered (Figure 4B), suggesting a treatment-induced transcriptional reprogramming induced by glucocortisteroids.