Pre-mRNA splicing is a fundamental and highly conserved process that removes introns and joins exons in precursor mRNAs to form mature transcripts. This reaction is catalyzed by the spliceosome, large ribonucleoprotein complex composed of five small nuclear ribonucleoproteins (snRNPs)—U1, U2, U4, U5, and U6—along with numerous associated protein cofactors (Black, 2003; Kinlaw et al., 1982; Sperling et al., 1986; Vagner et al., 2000). The splicing process begins with U1 snRNP recognizing the 5′ splice site and U2 snRNP binding the branch point sequence, facilitated by U2 auxiliary factors U2AF1 and U2AF2 (Glasser et al., 2022; Kralovicova and Vorechovsky, 2017). During spliceosome activation, the subsequent recruitment of the U4/U6·U5 tri-snRNP complex facilitates the structural rearrangements required for catalysis, enabling two sequential transesterification reactions. In the first catalytic step, the 2′-OH of the branch point adenosine attacks the 5′ splice site, generating a cleaved 5′ exon and forming the intron lariat structure through a 2′–5′ phosphodiester bond (Figure 1A). The spliceosome primarily recognizes canonical GT-AG splice site motifs. Mutations within these sites can activate cryptic splice sites or cause exon skipping and intron retention, frequently resulting in aberrant or truncated protein products (Corvelo et al., 2010; Douglas and Wood, 2011; Pros et al., 2008).

Alternative splicing (AS) enables a single gene to produce multiple mRNA and protein isoforms by varying the combination of exons incorporated into the final transcript, thereby substantially increasing proteomic complexity without altering the genomic sequence (Tao et al., 2024). Common AS patterns include exon skipping, use of alternative 5′ or 3′ splice sites, intron retention, and mutually exclusive exon usage. It is estimated that approximately 75% of alternatively spliced exons encode regions that contribute to protein functional domains, often affecting surface-exposed residues important for molecular interactions (Blencowe, 2006; Romero et al., 2006).

Splice site selection is governed by the strength of core splicing signals and the presence of cis-regulatory elements, exonic and intronic splicing enhancers and silencers, which serve as binding platforms for trans-acting splicing factors (Figure 1B). Serine/arginine-rich splicing factors (SRSFs) typically enhance exon inclusion by binding to splicing enhancers, while heterogeneous nuclear ribonucleoproteins (hnRNPs) often repress splicing through silencers, though their functional roles are context-dependent and can be bidirectional (Busch and Hertel, 2012). This flexibility contributes to cell type, and signal-specific splicing programs. Additionally, DEAD-box RNA helicases such as DDX5 and DDX17 modulate RNA secondary structures and assist in spliceosome assembly (Dardenne et al., 2014; Lee et al., 2018). Post-translational modifications, particularly phosphorylation of the RS domains in SR proteins, further influence their binding affinity and protein-protein interactions, thereby fine-tuning splice site selection (Long et al., 2019).

Splicing is intimately linked with transcription and chromatin dynamics. The C-terminal domain of RNA polymerase II (Pol II) serves as a scaffold for the recruitment of spliceosomal components, facilitating co-transcriptional splicing (Giono and Kornblihtt, 2020). Transcription elongation rates can influence splice site recognition, with slower elongation favoring inclusion of alternatively spliced exons. Chromatin features, such as nucleosome positioning and histone modifications, also modulate splicing outcomes. For instance, H3K36me3 is enriched at exons and has been shown to promote exon inclusion by recruiting splicing regulatory proteins (Figure 1C) (de Almeida et al., 2011).

RNA secondary structures can further regulate splicing by masking or exposing splicing regulatory elements, as demonstrated in genes such as MAPT (tau) and FN1 (fibronectin) (Muro et al., 1999; Ong et al., 2019). The FN1 pre-mRNA adopts a specific stem-loop around the EDA exon, positioning the exon splicing enhancer within an exposed loop that facilitates proper splicing. Moreover, small nucleolar RNAs, such as HBII-52, have been implicated in modulating alternative splicing and are associated with genetic disorders like Prader-Willi syndrome (Gallagher et al., 2002). Splicing regulation is also influenced by extracellular signals through modulation of RNA-binding protein activity via phosphorylation and other post-translational modifications, adding yet another layer of dynamic control (Blaustein et al., 2007).

Genome-wide RNA sequencing (RNA-seq) analyses have revealed dynamic shifts in AS profiles during embryonic stem cell differentiation and their reversion upon cellular reprogramming, suggesting that AS contributes to the establishment and maintenance of cell type–specific gene expression programs, though it may not serve as a primary driver of lineage commitment (Laaref et al., 2020). In MSCs, differential AS patterns are observed between young and aged donors, indicating that splicing regulation plays a role in age-associated alterations in differentiation potential (Peffers et al., 2016).

The transcription factor RUNX2 is a master regulator of osteogenic differentiation, being highly expressed during early lineage commitment and subsequently downregulated in mature osteocytes (Makita et al., 2008). U2 snRNP components such as U2AF1, SF3A1, and SF3A3 are essential for proper splicing of RUNX2 transcripts. Knockdown of these factors induces exon skipping and disrupts RUNX2 function, impairing osteogenic commitment (Venables et al., 2013) (Figure 2A). In parallel, extracellular signaling pathways, including Wnt, BMP, and Notch, modulate AS by altering the expression and activity of splicing factors and RNA-binding proteins.

Osteoblasts are responsible for the production and mineralization of the extracellular matrix (ECM), which is primarily composed of type I collagen, non-collagenous proteins, and hydroxyapatite. AS plays a pivotal role in regulating the structure, composition, and mechanical properties of the ECM. For instance, COL1A1, the gene encoding the α1 chain of type I collagen, undergoes AS, and aberrant splicing events are associated with osteogenesis imperfecta (Han et al., 2020).

AS also regulates the activity of osteogenic growth factors such as VEGFA, IGF-1, and NELL1. Mechanical stimulation alters the splicing pattern of VEGFA, favoring membrane-bound isoforms under high-stress conditions and soluble isoforms under low stress. This shift affects both mineralization and angiogenesis (Faure et al., 2008). Similarly, mechanical loading induces global AS changes in osteocytes, modulating ECM composition and bone strength. A specific isoform of IGF-1 is preferentially produced by osteoblasts in response to mechanical stretch, highlighting the mechanosensitive nature of splicing regulation (Xian et al., 2006). NELL1, a potent osteoinductive factor, exists as two isoforms generated via alternative promoter usage and exon inclusion. The shorter isoform more efficiently promotes MSC proliferation, although the molecular mechanisms governing its splicing remain poorly understood (Zhang et al., 2010) (Figure 2B).

Together, these findings illustrate that RNA splicing is not merely a passive post-transcriptional process but an active regulatory mechanism essential for maintaining skeletal integrity. By enabling precise temporal and spatial control of gene expression, AS ensures that bone-forming and bone-resorbing activities are balanced in response to developmental and environmental cues.

Osteogenesis Imperfecta (OI) is a heritable connective tissue disorder primarily caused by pathogenic variants in COL1A1 and COL1A2, encoding the pro-α chains of type I collagen. Precise splicing of these genes is critical for the assembly of functional collagen triple helices. Splice site mutations can lead to exon skipping within the triple-helical domain, thereby disrupting chain alignment and compromising helix propagation—mechanisms associated with severe OI phenotypes (Chu and Prockop, 2002). In other cases, splicing alterations activate cryptic splice sites or cause intron retention, leading to premature termination codons and nonsense-mediated mRNA decay, often resulting in milder clinical presentations (Maquat, 2004). These molecular consequences underscore the importance of splicing regulation in both collagen biosynthesis and OI phenotype heterogeneity. Beyond structural defects, altered mRNA splicing impacts transcript stability and collagen stoichiometry, positioning RNA splicing as a central determinant of skeletal integrity.

Osteoporosis is a multifactorial metabolic bone disease characterized by reduced bone mineral density (BMD), microarchitectural deterioration, and increased fracture risk. While aging, hormonal imbalance, and environmental factors are key contributors, genetic and epigenetic mechanisms, including alternative splicing, play substantial roles in disease susceptibility and progression.

In monogenic forms, such as LRP5-related osteoporosis, splice site mutations disrupt exon recognition, resulting in exon skipping, in-frame deletions, or cryptic splice site activation. These mutations yield truncated or nonfunctional LRP5 proteins lacking transmembrane or intracellular domains, impairing Wnt signaling and osteoblast function (Ai et al., 2005; Laine et al., 2011). Notably, some LRP5 splicing mutations are also associated with neurological symptoms, suggesting pleiotropic effects of splicing errors beyond the skeleton.

At a polygenic level, genome-wide association studies (GWAS) and transcriptomic analyses have uncovered widespread splicing alterations correlated with skeletal fragility. A recent meta-analysis identified 32 genes exhibiting splicing events significantly associated with BMD, and an additional 10 genes harboring splicing variants linked to fracture risk (Liu et al., 2020). Among these, DBF4B is regulated by the splicing factor SRSF1, which has been identified as a central node in both protein–protein interaction and co-expression networks in postmenopausal osteoporosis (Chen et al., 2017; Qian et al., 2019). These findings underscore RNA splicing as a critical post-transcriptional mechanism shaping the genetic architecture of osteoporosis.

Osteosarcoma is the most prevalent primary bone malignancy, typically arising during adolescence and characterized by aggressive growth, metastatic potential, and resistance to conventional therapies. Increasing evidence implicates alternative splicing as a driver of osteosarcoma pathogenesis.

High-throughput transcriptome profiling has identified 63 dominant AS events in osteosarcoma tissues, many of which correlate with specific immune cell populations, including resting memory CD4⁺ T cells, dendritic cells, and mast cells. This suggests a role for aberrant splicing in remodeling the tumor immune microenvironment. A regulatory network termed the RBP–AS–immune network has been proposed, linking RNA-binding proteins (RBPs) to immune modulation and cancer progression. Key RBPs—such as NOP58, FAM120C, DYNC1H1, TRAP1, and LMNA—emerge as promising targets for therapeutic immunomodulation (Dai et al., 2023).

Among splicing regulators, SRSF3 displays oncogenic properties, and its overexpression has been shown to drive tumor formation in nude mice (Jia et al., 2010). Transcriptome-wide studies in U2OS cells demonstrate that SRSF3 modulates splicing and expression of over 200 genes involved in cell cycle control and proliferation. Its knockdown also alters microRNA expression, suggesting broader roles in gene regulatory networks (Ajiro et al., 2016).

AS also governs the isoform expression of vascular endothelial growth factor (VEGF), a key modulator of tumor angiogenesis. The proangiogenic isoform VEGF₁₆₅ is upregulated in osteosarcoma, while the antiangiogenic isoform VEGF₁₆₅b is downregulated. The splicing factor YBX1, overexpressed in osteosarcoma, promotes VEGF₁₆₅ expression and represses VEGF₁₆₅b, thereby enhancing tumor cell proliferation, migration, and invasion. High YBX1 expression correlates with poor prognosis (Bates et al., 2002; Quan et al., 2023). These findings highlight splicing regulators as potential therapeutic targets for osteosarcoma.

Several rare skeletal disorders also implicate RNA splicing in their pathogenesis.

Odontochondrodysplasia (ODCD): ODCD is a rare skeletal dysplasia characterized by short stature, joint laxity, craniofacial abnormalities, and dental defects, including dentinogenesis imperfecta (Wehrle et al., 2019). A homozygous in-frame splicing mutation in intron 9 of TRIP11 leads to the expression of an alternative transcript, likely contributing to the mild skeletal phenotype and dentinogenesis imperfecta characteristic of ODCD (Medina et al., 2020).

Marfan Syndrome (MFS): MFS is a systemic connective-tissue disorder presenting with aortic root dilation, ectopia lentis, long-bone overgrowth, and skeletal abnormalities such as scoliosis and pectus deformities. Pathogenic splicing mutations in FBN1 have been identified as causative of MFS. A recently reported donor site mutation (c.8051+1G>C) in exon 64 disrupts canonical splicing, leading to disease development (Comeglio et al., 2007; Karttunen et al., 1998; Wang et al., 2022).

X-linked Spondyloepiphyseal Dysplasia Tarda (SEDT): Caused by splice-disrupting mutations in SEDL, this disorder is characterized by short stature and joint degeneration. Splicing errors in SEDL result in abnormal protein isoforms, impairing intracellular trafficking (Shaw et al., 2003; Xiong et al., 2009).

Paget’s Disease of Bone (PDB): PDB is a chronic focal bone disorder characterized by markedly increased and disorganized bone turnover, leading to bone pain, deformity, and a heightened risk of fractures, particularly in the pelvis, spine, skull, and long bones (Hasebe and Hamasaki, 2025). RNA-seq studies identified six AS events associated with PDB in genes such as LGALS8, RHOT1, CASC4, USP4, TBC1D25, and PIDD, suggesting potential roles for splicing defects in altered bone remodeling (Klinck et al., 2014).

Antisense oligonucleotides (ASOs) are short, synthetic, single-stranded nucleic acids designed to modulate pre-mRNA splicing by binding to specific RNA sequences. They can restore normal splicing patterns by masking aberrant splice sites, blocking splicing silencers or enhancers, or promoting exon skipping or inclusion (Figure 3A). In cases where pathogenic mutations activate cryptic splice sites or disrupt exon definition, ASOs can redirect splicing to restore the correct reading frame or produce partially functional proteins.

While ASO-based therapy has not yet been applied to bone diseases, its clinical success in correcting a 3′ splice site mutation in the DMD gene for Duchenne muscular dystrophy highlights its therapeutic potential in disorders with similar molecular mechanism (Yokota et al., 2011). ASOs can also sterically block pseudo splice sites induced by mutation, thereby promoting usage of authentic splice sites (Uchikawa et al., 2007). Pharmacological agents such as kinetin have also demonstrated splicing modulation capabilities. For instance, kinetin partially restores correct splicing at a mutated 5′ splice site in the IKBKAP gene in familial dysautonomia (Axelrod et al., 2011). These findings support the plausibility of pharmacological or ASO-mediated splicing correction in bone disorders with analogous splice site mutations.

Trans-splicing, or spliceosome-mediated RNA trans-repair, offers a powerful strategy for correcting endogenous mRNA at the RNA level. This approach utilizes pre–trans-splicing molecules (PTMs) to replace mutated regions of a pre-mRNA with a wild-type sequence via splicing-mediated ligation (Figure 3B). PTMs contain an RNA binding domain specific to the target transcript, a functional splice site, and the corrective coding sequence.

This technique is particularly suited for correcting mutations at canonical splice sites and has shown efficacy in preclinical models of β-thalassemia by repairing HBB gene mutations (Berger et al., 2016; Kierlin-Duncan and Sullenger, 2007). Although trans-splicing has not yet been explored in bone-related conditions, its successful application to similarly structured genes suggests its potential translational relevance for skeletal diseases.

Engineered small nuclear RNAs (snRNAs), particularly modified U1 snRNAs, represent another promising modality for rescuing splicing defects. U1 snRNAs recognize the 5′ splice site during early spliceosome assembly. Introducing compensatory base changes in U1 snRNA can restore complementarity to mutated 5′ splice sites, thereby facilitating accurate exon recognition and inclusion (Bohnsack and Sloan, 2018) (Figure 3C). This strategy is broadly applicable to splice site mutations, particularly those that disrupt base pairing without compromising the core splicing machinery. However, mutations at the highly conserved first and second intronic nucleotides are challenging to rescue due to their essential catalytic roles (Fernandez et al., 2012; Glaus et al., 2011). Nevertheless, advancements in snRNA design may overcome these obstacles and expand their utility in bone disease treatment.

Collectively, these splicing-targeted approaches offer promising therapeutic avenues for bone disorders driven by splice site mutations. Although most of these strategies remain in preclinical stages for skeletal disorders, their proven efficacy in other genetic diseases underscores their translational potential. Further research and disease-specific validation will be essential to bring these RNA-based therapeutics into clinical application for bone pathologies.