RNA Sequencing Addresses a 5’ UTR Variant Leading to X-Linked Agammaglobulinemia and Broader Immune Dysregulation Authors: Kexin Xu, Bingqing Zhang, Sen Zhao, Yun Zhang, Nan Wu, Min Shen, Xuejun Zeng Published: 2025 ABSTRACT AbstractPurposePathogenic variants in BTK cause X-linked agammaglobulinemia (XLA), yet some cases remain undiagnosed despite extensive genetic testing. Additionally, autoimmune complications have been increasingly reported, but the underlying etiology is still unclear. This study aimed to use RNA sequencing (RNA-seq) to resolve an undiagnosed case and characterize transcriptomic features associated with disease mechanisms of XLA.MethodsWe performed genome sequencing and RNA-seq, detected outliers against 89 controls to identify disease-causing variants, and validated functional consequences through RT-PCR. We further conducted gene set enrichment analysis (GSEA) to delineate the transcriptomic features of XLA.ResultsWe recruited an 18-year-old male who experienced recurrent infections since infancy but developed idiopathic portal hypertension recently. He had received four noninformative genetic test results. RNA-seq enabled the identification of significantly decreased BTK expression caused by a de novo hemizygous variant (NM_000061.3:c.-32A > G) located in the 5’ UTR. Outlier analysis revealed reduced expression of genes critical for not only immunoglobulin synthesis, but also T-cell development and inflammation intensification, such as IGHM, TRBV24-1, and PLVAP. GSEA additionally revealed significant enrichment in pathways associated with immune response and inflammatory response, suggesting broader immune dysregulation. These findings align with the growing recognition of autoimmune complications in XLA.ConclusionThis study emphasizes the clinical utility of RNA-seq and provides a comprehensive expression profile of XLA. These findings can guide future molecular analyses in undiagnosed cases and enhance understanding of autoimmune complications in XLA, suggesting a Multidisciplinary Team approach in both diagnostic and treatment processes.Supplementary InformationThe online version contains supplementary material available at 10.1007/s10875-025-01971-2.Keywords: X-linked agammaglobulinemia, BTK, 5’ UTR, RNA sequencing, Transcriptomics, Autoimmune disorders Introduction Agammaglobulinemia is a predominantly antibody deficiency characterized by early-onset recurrent bacterial infections, with a reduction in all classes of serum immunoglobulins and a decreased number of B cells [1]. X-linked agammaglobulinemia (XLA, MIM: 300755) accounts for approximately 85% of all cases of agammaglobulinemia and mainly affects males [2]. XLA is caused by pathogenic variants in BTK (HGNC:1133), which is located at Xq22.1 and encodes the Bruton tyrosine kinase (BTK). BTK is a non-receptor tyrosine kinase that plays a crucial role in the development and functioning of B cells. Defects in BTK block early B-cell development, resulting in decreased or absent immunoglobulin production, ultimately leading to XLA. Current mainstay treatments, including intravenous immunoglobulin (IVIG) and antibiotics, have effectively increased life expectancy [3]. However, timely diagnosis and effective management can be challenging [4]. Although typical clinical features observed in males usually indicate XLA, there is considerable phenotypic overlap between XLA and 11 other forms of agammaglobulinemia, with late-onset and atypical manifestations being more frequently reported [2, 5–7]. The increasingly recognized phenotypic and genetic heterogeneity presents challenges to establishing a diagnosis based solely on clinical evaluation. Advances in next-generation sequencing and variant interpretation, including sequencing methods ranging from targeted gene panels to genome sequencing (GS), the development of increasingly comprehensive variant databases, and updated bioinformatics pipelines, have improved diagnostic yield in inborn errors of immunity [8]. For example, in a cohort in which 58% of males remained molecularly undiagnosed, exome sequencing (ES) identified 16% of potential causal variants in previously tested X-linked genes, including intronic single nucleotide variants (SNVs) and copy number variants (CNVs) that were missed by prior diagnostic Sanger sequencing [9]. Nevertheless, undiagnosed cases still persist after intensive testing. One key contributing factor is the overlooking of non-coding variants, which may not be captured or lack functional data for accurate interpretation. In disease management, timely immunoglobulin replacement therapy has effectively reduced early mortality. However, patients surviving into adulthood are at high risk of developing life-threatening autoimmune complications that affect multiple organ systems [3, 10]. These emerging complications lack effective therapies, and their underlying mechanisms remain unknown. RNA sequencing (RNA-seq) is a powerful tool in clinical diagnostics and has been increasingly used to uncover underlying disease mechanisms. Integrating RNA-seq into family-based GS has significantly improved diagnostic outcomes in previously unsolved cases by as much as 35%, largely attributed to the functional consequences and transcriptome phenotypes [11–14]. RNA-seq can detect aberrant expression, splicing errors, and mono-allelic expression, therefore providing functional evidence for the reclassification of variants of uncertain significance (VUSes). RNA-Seq may also identify candidate variants in cases where prior DNA-based analysis has failed to provide informative results. Further, RNA-seq provides global transcriptomic features, enhancing the understanding of underlying mechanisms and paving the way for precision medicine [15]. Gene expression is highly tissue-specific, and a large proportion of immunodeficiency-related genes are highly expressed in whole blood [13]. However, the global transcriptomic profile of XLA has never been reported. In the present study, we used RNA-seq to address the pathogenicity of a previously overlooked 5’ UTR variant in BTK. We further performed gene set enrichment analysis (GSEA) to investigate the transcriptomic profile of XLA and to provide insights into the mechanism underlying autoimmune complications in XLA. The ultimate goal was to leverage these molecular findings to facilitate timely and accurate diagnosis and improve patient outcomes. Methods Clinical Data The patient was evaluated at Peking Union Medical College Hospital (PUMCH). Clinical data were collected, including demographic information, current and past medical history, findings on physical examination, laboratory test results, radiological assessments, and family history. DNA Sequencing and Interpretation Raw data of duo-exome sequencing (proband and mother) were obtained from the external diagnostic laboratory that performed the primary analysis. Trio-GS was performed using DNA extracted from the peripheral blood of the patient and both parents, followed by the preparation of PCR-free sequencing libraries. Paired-end, 150-bp read sequencing was performed on the MGI DNBSEQ-T7 platform (MGI Tech Co., Ltd., Shenzhen, China), with a mean depth of 40X. Raw data of the above-mentioned ES and GS were processed using the previously reported in-house developed Peking Union Medical College Hospital Pipeline (PUMP) [16, 17]. Data were visualized using the Integrative Genomics Viewer (IGV) [18, 19]. SNVs, small insertions/deletions, and single-gene CNVs were interpreted and classified in accordance with the ACMG/AMP guidelines [20]. Multigenic CNVs were interpreted and classified in accordance with the 2019 ACMG-ClinGen guidelines [21]. Structural variants were interpreted according to our laboratory’s internal protocol. Mitochondrial DNA variants were interpreted following the ACMG/AMP specifications for mitochondrial variant interpretation [22]. Variant density was defined as the number of observed nonpathogenic variants in the gnomAD database divided by the length of the region and was used to estimate regional tolerance in the 5’ UTR [23]. RNA Sequencing and Outlier Detection Whole-blood samples of the patient and both parents were collected in PAXgene Blood RNA Tube (QIAGEN). RNA was extracted and processed following the manufacturer’s recommendation. RNA purity and concentration were examined using NanoDrop 2000. RNA integrity and quantity were measured using the Agilent 4200 system. mRNA was isolated from total RNA using polyT and then fragmented into fragments of 300–350 bp. First-strand cDNA was synthesized through reverse transcription using fragmented RNA and dNTPs (dATP, dTTP, dCTP, and dGTP). Subsequently, second-strand cDNA synthesis was performed. Remaining overhangs of double-strand cDNA were converted into blunt ends via exonuclease/polymerase activities. After adenylation of the 3’ ends of DNA fragments, sequencing adaptors were ligated to the cDNA and the library fragments were purified. The template was enriched by PCR, and the PCR product was purified to obtain the final library. The concentration of library was measured by the Qubit fluorometer (Thermo Fisher Scientific). Then, paired-end 150nt sequencing was performed on Illumina NovaSeq 6000 (Illumina, San Diego, CA, USA). The obtained data were aligned to the GRCh38/hg38 reference sequence using STAR v2.7.10a with default parameters using GeneCode v39 annotation [24, 25]. RNA-seq data were visualized using IGV [18, 19]. Raw read counts were normalized to reads per million mapped reads to account for differences in sequencing depth across samples. Outliers were called using OUTRIDER and FRASER (FDR < 0.05), integrated into the detection of RNA outliers pipeline [26–28]. Control samples were obtained from 89 unrelated patients and healthy family members who were recruited as part of Deciphering disorders Involving Scoliosis and COmorbidities (DISCO) study (http://www.discostudy.org/). Among patients with diagnosed disorders, a diverse range of conditions was included to prevent bias from disease-specific expression variations. All individuals underwent genome or exome sequencing, and none of them had a personal or family history of recurrent infections, autoimmune disorders, or inflammatory disorders. Transcription Polymerase Chain Reaction (RT-PCR) The amplification was performed using primers in exon 1 of BTK (5’- AGTGTCTGCTGCGATCGAG-3’) and exon 5 of BTK (5’- TTGAGCTGGTGAATCCACC-3’). For all samples, the annealing temperature was 59 °C, and amplification was carried out for 40 cycles. ABL1 was amplified with forward primer 5’-GCCAGTAGCATCTGACTTTGAGC-3’ and reverse primer 5’-TTGACTGGCGTGATGTAGTTGC-3’ as an endogenous reference gene. Gene Set Enrichment Analysis Results from expression outlier analysis, filtered by a threshold of TPM >1 and p < 0.05, were used for Gene Set Enrichment Analysis (GSEA) using the clusterProfiler package in R with pre-defined sets of genes in Gene Ontology (GO) and custom gene sets [29]. To balance inclusivity and interpretability, gene sets containing fewer than three genes were excluded from the analysis. P-value was used as the ranking metric. The custom gene set was generated by extracting genes associated with GO terms containing the key term in the Description, including all GO annotations across evidence types. Ethics Statement, Informed Consent and Consent to Publish This study was approved by the Institutional Review Board of Peking Union Medical College Hospital (No. I-24PJ0761, approved on April 8, 2024; No. I-25PJ1020, approved on April 28, 2025). Written informed consent and consent to publish were obtained from all participating individuals or their legal guardians. Results Clinical Presentation An 18-year-old male was admitted to our hospital with a chief complaint of progressive lower limb pitting edema and ascites persisting for one year (Fig. 1a). He was the second child of non-consanguineous parents without an obvious family history (Fig. 1b). His medical history was remarkable for recurrent fever and infections since he was three months old, including recurrent otitis media, pneumonia, and septic hip arthritis. XLA was suspected at the age of 10 years based on low serum immunoglobulin levels (Table S1). However, results from four genetic tests, including two immunodeficiency panel tests, one BTK single gene test, and one duo-exome sequencing, were either negative or non-informative (Table S2). Nevertheless, he received regular IVIG supplements based on clinical suspicion, and his fever generally resolved, with only occasional low-grade fever that responded to oral antibiotics. The Clinical Findings (a) Lower limb edema; (b) The pedigree. Filled-in symbol indicates the affected individual. X represents the X chromosome. Y represents the Y chromosome. WT represents wide type. NT represents not being tested. (c) Computed tomography image showing splenomegaly and splenic vein dilation; (d) Endoscopic image showing esophageal varices Laboratory tests showed mild pancytopenia and hypoalbuminemia (Table S1). Abdominal CT and vascular ultrasound showed dilatation of the inferior vena cava, left hepatic vein, portal vein, superior mesenteric vein, and splenic vein, as well as splenomegaly and ascites (Fig. 1c). Liver stiffness measurement was 11.2 kPa on liver ultrasound elastography. Gastroduodenoscopy detected esophageal varices (Fig. 1d). Liver biopsy reported mild chronic nonspecific liver changes. He was diagnosed with portal hypertension and then treated with spironolactone, diuretics, and propranolol. His lower limb edema was generally controlled during follow-up. However, the etiology remained unclear. Genetic Data Reanalysis and Multidisciplinary Team Consultation We reanalyzed the patient’s duo-ES data and identified a hemizygous variant (NM_000061.3(BTK):c.−32A >G) located in the 5’ UTR in the proband, but not in the mother (Table S3). The c.−32 A >G variant was absent from gnomAD and had not been reported in individuals with XLA. SpliceAI predicted a donor loss at −1 bp with a score of 0.16 and a donor gain at + 11 bp with a score of 0.47 [30]. Variants in the 5’ UTR of BTK were infrequently observed in the gnomAD database, with a variant density of 0.09, indicating that variants in this region were less tolerable. No functional studies had been reported. Consequently, the c.−32 A >G variant was interpreted as reportable but classified as a VUS (VUS = PM2_P + PP3 + PS1_P + PM6_P). Due to the unclear clinical picture and nondiagnostic results from multiple genetic tests, the patient was referred to the Multidisciplinary Team (MDT) Consultation Program for Rare Diseases at PUMCH [31]. The patient was suspected of having agammaglobulinemia, with portal hypertension and hypersplenism considered uncommon hepatic comorbidities associated with autoimmune dysregulation. In order to achieve an accurate diagnosis, combined trio-GS and RNA-seq were requested to reclassify the c.−32 A >G variant and to exclude other variants more likely responsible for the patient’s complete clinical presentation. Molecular Findings from Integrated RNA-seq and GS Trio-GS revealed the same c.−32A >G variant and confirmed it as de novo (Fig. 2a). We performed both phenotype- and genotype-driven analyses on variants in genes with definitive gene-disease associations in the OMIM database [32]. Particular attention was given to genes associated with various forms of agammaglobulinemia, other immunodeficiencies, and disorders known to present with portal hypertension and edema, as documented in the OMIM database (e.g., IGLL1, TCF3, and GATA2). Sequencing reads of these genes were then manually inspected using IGV. No additional pathogenic or likely pathogenic variants were identified, and no other variants were interpreted as more likely to be disease-causing (rare variant in BTK is shown in Table S4). The sashimi plot of trio RNA-seq data showed a complete absence of junction reads spanning the 5’ UTR in the biologically relevant transcript of the proband (Fig. 2b, loss of junction reads highlighted by arrow). Statistical comparison with the control cohort identified a significant decrease in expression of BTK (z score = −7.53, log2(fold change) = −2.23, adjusted p-value = 2.72 × 10− 6, Table S5), whereas both parents exhibited expression levels comparable to controls (Fig. 2c). The aberrant expression of BTK represented the most significant deviation among the proband’s transcriptome (Table S5). RT-PCR analysis confirmed the complete absence of BTK expression in the proband, while both parents exhibited normal expression levels (Fig. 2d, results of endogenous reference gene are shown in Figure S1). The Molecular testing results (a) Results of genome sequencing. IGV of the trio-genome sequencing data showing the de novo hemizygous c.−32A > G variant. For clarity, only the MANE Select transcript is displayed. (b) Sashimi Plot of RNA Sequencing. Sashimi plot of the trio-RNA sequencing data showing a complete lack of junction reads from the 5’ UTR in the proband (indicated by a red arrow), indicating loss of canonical transcript usage. Broken axes were used for the parents to better visualize the significantly reduced expression in the proband. For clarity, only the MANE Select transcript is displayed. (c) Expression of BTK compared with 89 controls. The proband exhibits significantly decreased BTK expression (log2(fold change) = −2.23, adjusted p-value = 2.72 × 10⁻6), while both parents display expression levels comparable to controls (log2(fold change) ≈ 0, adjusted p-value = 1). (d) RT-PCR results of the BTK MANE transcript. A clear amplification product was observed in both parents, whereas no detectable product was found in the proband, indicating absence of the biologically relevant BTK transcript Transcriptomic Features Revealed by Gene Set Enrichment Analysis We further analyzed the expression outliers to provide functional evidence that would assist future interpretation of other VUSes in BTK and enhance the understanding of the biological processes underlying XLA. A total of 290 genes met the filter criteria (TPM >1 and p < 0.05) and were included in the analysis (Table S5). The results revealed significantly decreased expression of six other genes, IGHM, IGKC, TRBV24-1, HES4, ADGRE1, and PLVAP (Fig. 3a). These genes are essential for not only immunoglobulin synthesis (BTK, IGHM, IGKC) but also T cell development (TRBV24-1, HES4, ADGRE1), and may also intensify inflammation (PLVAP) [33–37]. The significant downregulation of genes, including BTK and HES4, is consistent with previous reports [38]. The decreased expression of other genes, such as IGHM and TRBV24-1, has not been previously described. We next analyzed enriched pathways through GSEA (top 20 results in the GO Biological Process category are shown in Fig. 3b, all pathways passing a p-value cutoff of 0.05 are listed in Table S6). Significant enrichments were observed in B cell-related biological processes, such as B cell differentiation, humoral immune response, and immunoglobulin mediated immune response, consistent with our patient’s clinical presentation and known pathophysiology of XLA. We also performed the same analyses on control samples, but did not identify notable enrichment of pathways involving immune function as observed in the patient (Figure S2). Further analysis of differential gene expression and enrichment pathways (a) Volcano plot of expression outlier. The results of expression outlier analysis (p-value < 0.05, TPM > 1) were visualized using a volcano plot. Genes with a p-value < 0.05 and |log2 fold change| >1.5 were labeled. The expression level of BTK was among the lowest and exhibited the most statistically significant change. (b) Gene Set Enrichment Analysis (GSEA) against Gene Ontology (GO) Database. The dot plot shows the top 20 results in the GO Biological Process category. (c) GSEA focused on immune response. (d) GSEA focused on inflammatory response. Note: In panels b–d, the Gene Ratio refers to the proportion of input genes that contribute most to an enriched pathway relative to the total number of input genes annotated to that pathway Among all significantly enriched pathways in the proband, many were related to immune response and inflammatory response (Table S6, highlighted in red and orange, respectively). Considering the portal hypertension of our patient was attributed to autoimmune dysregulation, we further performed two enrichment analyses focused on “immune response” and “inflammatory response” using related pathways from the GO database (genes and their corresponding pathways are listed in Table S7). Multiple pathways under both “immune response” and “inflammatory response” were enriched, as shown in Fig. 3c and d. We also performed GSEA using the same pathways on control samples. No enrichment was observed for immune response pathways in the control samples. For inflammatory response-associated pathways, only a single control sample showed enrichment of one pathway, with lower statistical significance compared with the patient’s results (Figure S3). These analyses indicate that the dysregulation of the pathways in the patient reflects disease-associated pathophysiology. We further analyzed the expression patterns of genes in pathways associated with immune response and inflammatory response, as shown in Figs. 3c–d. The results are presented in Table 1. Most of the enrichment in immune response pathways was attributable to the downregulation of genes such as IGHM, IGKC, BTK, CMKLR1, C3AR1, CCR7, BCL2, and PRKCQ, and the upregulation of genes including IRF4, TRIM32, DUSP3, CLPB, WDFY1, KMT5C, and CD40. A relatively smaller number of genes were significantly up- or downregulated in inflammatory response pathways, with enrichment primarily involving the downregulation of BTK, CCR7, and PIK3CG. Definitive relationship has been established between BTK and IGHM with agammaglobulinemia, which is characterized by defective B-cell development (MIM: 300755 and MIM: 601495). IGKC encodes the immunoglobulin kappa chain, which is important for B-cell development and is also a marker of B-cell activity [33, 39]. IRF4 is critical for the activation and differentiation of both B and T cells, while BCL2 regulates apoptosis in immune cells such as naïve T cells, naïve B cells, and natural killer cells [40, 41]. CCR7, PRKCQ, and CD40 encode proteins involved in the activation and function of both CD4+ and CD8+ T cells [42–44]. In addition to adaptive immune response, TRIM32, DUSP3 and CLPB play key roles in regulating innate immune response through various mechanisms [45–47]. CMKLR1, C3AR1, WDFY1, and PIK3CG are directly involved in NF-κB and Toll-like receptor signaling [48–51]. Beyond transcriptional regulation, KMT5C encodes a histone methyltransferase that catalyzes H4K20 dimethylation, a key epigenetic regulator of chromatin structure that contributes to oncological immune responses [52]. Collectively, these aberrant gene expression patterns and pathway enrichment features provide evidence of broader immune dysregulations in XLA and may inform future interpretation of other VUSes in BTK. Table 1Significantly upregulated and downregulated genes in enriched pathways related to immune response and inflammatory responseIDDescriptionDownregulated GenesUpregulated GenesImmune responseGO:0006959humoral immune responseIGHM, JCHAIN, BCL2, CCR7/GO:0002253activation of immune responseIGHM, IGKC, BTK, CMKLR1, C3AR1, BCL2, CD79B, CCR7, HRAS, FGR, CSKIRF4, TRIM32, DUSP3, CLPB, WDFY1GO:0002757immune response-activating signaling pathwayIGHM, IGKC, BTK, CMKLR1, C3AR1, BCL2, CD79B, CCR7, HRAS, FGR, CSKIRF4, TRIM32, DUSP3, CLPB, WDFY1GO:0002429immune response-activating cell surface receptor signaling pathwayIGHM, IGKC, BTK, CMKLR1, C3AR1, BCL2, CD79B, CCR7, HRAS, FGR, CSKTRIM32, DUSP3, WDFY1GO:0002764immune response-regulating signaling pathwayIGHM, IGKC, BTK, CMKLR1, C3AR1, BCL2, CD79B, CCR7, HRAS, FGR, PRKCQ, CSKIRF4, CD40, TRIM32, DUSP3, CLPB, WDFY1GO:0002768immune response-regulating cell surface receptor signaling pathwayIGHM, IGKC, BTK, CMKLR1, C3AR1, BCL2, CD79B, CCR7, HRAS, FGR, PRKCQ, CSKCD40, TRIM32, DUSP3, WDFY1GO:0016064immunoglobulin mediated immune responseIGHM, IGKC, BTK, IL2RB, TBX21KMT5C, CD74GO:0002460adaptive immune response based on somatic recombination of immune receptors built from immunoglobulin superfamily domainsIGHM, IGKC, BTK, HRAS, LEF1, IL2RB, TBX21, PRKCQ, SASH3IRF4, CD40, KMT5C, CD74GO:0050778positive regulation of immune responseIGHM, IGKC, BTK, CMKLR1, C3AR1, BCL2, CD79B, CCR7, HRAS, TBX21, FGR, PRKCQ, SASH3, POLR3C, CSK, STX4IRF4, KMT5C, HLA-DMA, TRIM32, DUSP3, CD74, RBM47, WDFY1, CLPB, AKIRIN2GO:0050776regulation of immune responseIGHM, IGKC, BTK, CMKLR1, C3AR1, BCL2, CD79B, CCR7, GPX1, HRAS, SAMSN1, TBX21, FGR, PRKCQ, SASH3, POLR3C, CSKIRF4, CD40, KMT5C, HLA-DMA, TRIM32, DUSP3, CD74, RBM47, WDFY1, CLPB, AKIRIN2Inflammatory responseGO:0002675positive regulation of acute inflammatory responseBTK, CCR7, PIK3CG/GO:0050729positive regulation of inflammatory responseBTK, CCR7, PIK3CG/GO:0002673regulation of acute inflammatory responseBTK, CCR7, PIK3CG/GO:0002437inflammatory response to antigenic stimulusBTK, CCR7, GPX1, FGR/GO:0002861regulation of inflammatory response to antigenic stimulusBTK, CCR7, GPX1, FGR/GO:0002526acute inflammatory responseBTK, ITIH4, CCR7, PIK3CG/GO:0050727regulation of inflammatory responseBTK, CCR7, GPX1, FGR, PIK3CG/ Significantly upregulated and downregulated genes in enriched pathways related to immune response and inflammatory response Management Approach Informed by Molecular Findings Based on the updated clinical information from the MDT consultation and the molecular findings, the c.−32 A > G variant was reclassified from VUS to LP (LP = PM2_P + PS2_M + PS3_M + PP4, Table S3). The patient was therefore diagnosed with XLA and initiated on IVIG therapy through a nation-sponsored program, alleviating both emotional and financial burdens. Informed by the broader immune dysregulation revealed through RNA-seq, the patient has been managed by our MDT, not only addressing portal hypertension but also implementing long-term monitoring for autoimmune complications. Discussion In this study, we integrated RNA-seq with GS to investigate a previously overlooked de novo variant in BTK, thereby ending an 18-year diagnostic odyssey. We further provided a global transcriptomic profile of XLA, demonstrating broader immune dysregulation that had not been fully characterized at the transcriptomic level. These findings explain the emerging clinical observation of increased incidence of autoimmune complications in patients with XLA and indicate a disease mechanism beyond the well-recognized B-cell defects. RNA-Seq guides the identification and reclassification of variants, particularly non-coding variants. Compared with conventional methods such as RT-PCR, RNA-seq does not rely on well-designed specific primers, allowing for the simultaneous analysis of both known and unknown isoforms at a lower cost [53–55]. Cummings et al. used RNA-seq to successfully assign pathogenicity to variants challenging to interpret, including those in TTN [14]. Murdock et al. used decreased PQBP1 expression to guide the identification of a causative PQBP1 deep intronic variant [13]. This PQBP1 variant was missed during the initial analysis, likely due to its intronic location and a benign prediction provided by SpliceAI. Variants in the 5’ UTR of BTK have been identified in two patients with previously negative ES results and labeled as “pitfalls.” [56] The variants, c.−31G >T and c.−31G >A, led to an 11-nucleotide insertion elongating the 5’ UTR. However, the underlying mechanism was unclear, as only the c.−31G >T variant created a new upstream open reading frame (uORF). Additionally, it was only investigated through RT-PCR, which was not capable of detecting functional consequences beyond the specific region of BTK or revealing any downstream effects on protein function or cellular pathways. The assessment of other variants in the 5’ UTR of BTK remains challenging. In this study, RNA-seq revealed a complete loss of the biologically relevant transcript in the proband, and the outlier analysis revealed a significantly decreased expression of BTK, providing critical functional evidence that enabled the reclassification of the c.−32A > G variant from VUS to LP (Fig. 2, Table S3). This variant was missed by all four previous tests, including two immunodeficiency panel tests, one BTK single gene test, and one duo-ES. The variant was detected by the duo-ES but was not reported by the institute during their initial analysis. This was likely due to the 5’ UTR location, together with a SpliceAI score of 0.47, which is below the recommended threshold of 0.5. Data from the other three tests were not available, but this variant was probably either undetected or filtered out due to its location or uncertain predicted impacts. In this study, RNA-seq enabled the reclassification of the c.−32 A > G variant, leading to an accurate and definitive diagnosis that concluded an 18-year diagnostic odyssey. However, the exact mechanisms remain unclear but may involve splicing errors caused by the c.−32 A > G variant. Aberrant splicing can lead to intron elongation or the creation of a novel uORF that triggers nonsense-mediated mRNA decay. Nonetheless, these results from integrated DNA and RNA analyses informed tailored treatment and long-term MDT management, improving patient outcomes and alleviating both emotional and financial burdens. Although XLA has been recognized for more than half a century, comprehensive transcriptomic features of XLA remain underexplored. Previous studies on gene expression in XLA primarily focused on genes known to interact with BTK, such as Toll-like receptors (TLRs) that are involved in innate immunity [57, 58]. Transcriptomic analyses are limited and often performed on selected cell types, for example, classical monocytes [38]. While these studies provide valuable insights, they may overlook broader gene expression dysregulation across diverse cell types, which could reveal novel mechanisms involved in the pathophysiology of XLA. The RNA-seq we performed in this study addressed this limitation, and revealed aberrant gene expression that has not been previously reported. The lack of prior recognition is likely due to the underexplored relationship between these genes and XLA, as well as their low expression in classical monocytes. For example, the normalized transcript per million (nTPM) of IGHM is 3.7, compared with 141.2 for BTK (nTPM data from The Human Protein Atlas, www.proteinatlas.org, accessed August 11, 2025) [59]. Subsequent GSEA demonstrated enrichment of immune and inflammatory pathways, revealing a previously underexplored transcriptomic profile of XLA (Figs. 3b-3d). These transcriptomic features can serve as a valuable resource, aiding in the further interpretation of functional consequences of other variants in BTK, especially those located in the 5’ UTR region. Besides, it presents an altered immune network, indicating broader immune dysregulation that extends beyond the well-recognized B-cell defects. Our transcriptomic findings are consistent with the recent clinical observations that patients with XLA are susceptible to developing non-infectious comorbidities later in life, especially autoimmune and inflammatory disorders. Traditionally, XLA treatment has focused on managing infections, with standard treatments leading to improved life expectancy. However, recent long-term follow-up studies involving patients from multiple ethnic groups have reported an unprecedented rate of non-infectious comorbidities. For example, 5.2% of patients exhibit immune dysregulation, and approximately half of patients are affected by respiratory and gastrointestinal conditions [3, 10, 60]. Among them, liver disorders were infrequently identified, but have led to liver transplantation and death. In other types of predominantly antibody deficiency, liver comorbidities, particularly nodular regenerative hyperplasia (NRH), have been associated with increased morbidity and mortality, possibly attributed to autoimmune dysregulation [61]. A recent study has suggested that, in XLA, NRH is an underestimated but severe comorbidity, occurring at an incidence of 29% and resulting in portal hypertension in all affected patients [62]. The pathophysiology remains unclear. However, based on reports of infiltrating CD8+ cytotoxic T cells, a source of increased IFN-γ, observed in NRH liver sinusoids from patients with common variable immune deficiency, the onset of NRH in XLA has been proposed to be associated with a disturbed autoimmune process [61–63]. Autoimmune and inflammatory disorders affecting multiple organ systems have been reported in approximately 15% of patients with XLA, with inflammatory bowel disease being the most commonly identified [3, 64]. The underlying mechanisms were not fully understood. B cells bidirectionally interact with T cells. With defects in B cells, a spectrum of abnormalities in the T cell compartments has been associated with XLA, including defects of regulatory T cells and T helper 17 cells, but increased CD8+ T cells and skewed T helper 1 response [65]. Further, the overproduction of autoantibodies has been implicated in XLA, likely due to excessive stimulation of TLRs and their secondary messengers, such as NF-κB [64, 66, 67]. Additionally, dysregulated innate immune function has also been suggested [38]. All these dysfunctions are well-established contributors to autoimmune disorders, and they have been proposed to predispose patients with XLA to autoimmune and inflammatory diseases [64, 65, 68]. In this study, the portal hypertension with possible NRH observed in our patient has also been associated with autoimmune abnormalities. The transcriptomic features revealed through RNA-seq indicate broadly disrupted immune response and inflammatory response. These findings are consistent with clinical manifestations of autoimmune disorders in patients with XLA and further support the proposed underlying mechanisms. For example, downregulation of PRKCQ can impair T-cell activation, while upregulation of WDFY1 can lead to TLR3- and TLR4-mediated activation of NF-κB [43, 50]. Moreover, aberrant expression of TRIM32, DUSP3, and CLPB may contribute to dysregulation of innate immune responses through multiple mechanisms [45–47]. Together, these genes may act synergistically, disrupting the complex and interconnected immune network. Overall, these findings provide a systemic transcriptomic profile that may explain the development of autoimmune complications in XLA. The limitations of this study include the following factors. Although the BTK expression in the proband decreased significantly and no other variant is more likely to be disease-causing, we cannot entirely rule out the possibility that other variants, within or outside BTK, could also influence gene expression and contribute to the aberrant expression. Additionally, because no amplification product was detected in the RT-PCR of BTK mRNA corresponding to the MANE Select transcript, we were unable to identify potential splicing errors caused by the c.−32 A > G variant. Furthermore, there was only one case in the GSEA, which may introduce biases or noise related to the specific individual, rather than reflecting generalizable findings for XLA. Finally, it is possible that the decreased expression of B cell-related genes reflected the markedly reduced number of B cells. Future studies clarifying the respective contributions of intrinsic transcriptional dysregulation and B-cell deficiency will further enhance the understanding of the disease mechanisms of XLA. Nevertheless, the RNA-seq in our study provides the first comprehensive view of the gene expression profile associated with XLA, offering valuable insights for assessing the pathogenicity of other variants in undiagnosed cases and advancing the understanding of disease mechanisms that inform disease management strategies. Conclusions In this study, we used RNA-seq to provide functional evidence for a previously overlooked disease-causing variant in BTK and to generate a comprehensive transcriptomic profile of XLA. Our findings highlight the importance of RNA-seq in both clinical diagnostics and the potential to advance transcriptome-wide profiling of genetic disorders. Further, the results from this study suggest an MDT approach for XLA in both diagnostic and treatment processes, ultimately contributing to improved patient outcomes. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 (PDF 692 KB) Supplementary Material 2 (DOCX 22.8 KB) Supplementary Material 3 (DOCX 18.2 KB) Supplementary Material 4 (DOCX 19.6 KB) Supplementary Material 5 (DOCX 17.3 KB) Supplementary Material 6 (XLSX 52.2 KB) Supplementary Material 7 (XLSX 43.7 KB) Supplementary Material 8 (XLSX 4.20 MB) Acknowledgements We thank the patient and his family, as well as all individuals in the control cohort, for participating in this study. We thank our clinical care, genetics and laboratory teams at Peking Union Medical College Hospital for their contributions to this study. We thank GeneSeeq Technology Inc. and Ekitech Technology Inc. for providing sequencing and technical support. We acknowledge Zhejiang Biosan Biochemical Technologies for performing the RT-PCR experiments. We thank the Clinical Biobank (ISO 20387) at Peking Union Medical College Hospital, Chinese Academy of Medical Sciences, for their assistance with sample deposition and management. Author Contributions X.Z., M.S., and N.W. concepted and designed the study. X.Z. and N.W. acquired data. K.X. and B.Z. analyzed and interpreted data. K.X., B.Z., S.Z., and Y.Z. validated the data. K.X. and B.Z. drafted the article. K.X., B.Z., S.Z., Y.Z., N.W., M.S. and X.Z. critically revised the article. X.Z., M.S., and N.W. supervised the study. X.Z., N.W. and Y.Z. acquired the fundings. The first authorship is shared by K.X. and B.Z. All authors have read and approved the manuscript and its submission to Journal of Clinical Immunology. Funding This research was funded in part by the National Key Research and Development Program of China (2022YFC2703102 to N.W.), CAMS Innovation Fund for Medical Sciences (CIFMS, 2021-I2M-1-051 to N.W.), National High Level Hospital Clinical Research Funding (No. 2022-PUMCH-B-044 to X.Z., 2022-PUMCH-D-004 to N.W., 2022-PUMCH-C-033 to N.W.), 2019 Discipline Development Project of Peking Union Medical College faculty development program in family medicine (No.201920200106 to X.Z.), the National Key Clinical Specialty Construction Project and the Non-profit Central Research Institute Fund of Chinese Academy of Medical Sciences (No. 2019PT320025 to N.W.), and the Peking Union Medical College Hospital Outstanding Young Talent Development Program (No. UBJ10806 to Y.Z.). Data Availability Data are not openly available due to reasons of privacy and are available from the corresponding authors upon reasonable request. Declarations Ethics Statement, Informed Consent and Consent To Publish This study was approved by the Institutional Review Board of Peking Union Medical College Hospital (No. I-24PJ0761, approved on April 8, 2024; No. I-25PJ1020, approved on April 28, 2025). Written informed consent and consent to publish were obtained from all participating individuals or their legal guardians. Competing Interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Kexin Xu and Bingqing Zhang contributed equally and are designated to have co-first authorship. Contributor Information Nan Wu, Email: dr.wunan@pumch.cn. Min Shen, Email: shenmpumch@163.com. Xuejun Zeng, Email: zxjpumch@126.com. References Associated Data Supplementary Materials Supplementary Material 1 (PDF 692 KB) Supplementary Material 2 (DOCX 22.8 KB) Supplementary Material 3 (DOCX 18.2 KB) Supplementary Material 4 (DOCX 19.6 KB) Supplementary Material 5 (DOCX 17.3 KB) Supplementary Material 6 (XLSX 52.2 KB) Supplementary Material 7 (XLSX 43.7 KB) Supplementary Material 8 (XLSX 4.20 MB) Data Availability Statement Data are not openly available due to reasons of privacy and are available from the corresponding authors upon reasonable request.