Deconvoluting the multi-faceted roles of alternative splicing events in cancer: From underlying mechanisms to innovative therapeutics

Abstract

Alternative splicing (AS) is well-known to be a critical mechanism to enhance proteomic diversity from a limited genome. Dysregulation of AS may represent a key tumor-specific molecular feature; thereby driving tumorigenesis and progression. Elucidating the mechanisms underlying AS dysregulation in oncogenesis is paramount for developing more novel therapeutic targets and agents. Thus, in this review, we comprehensively demonstrate the pivotal roles of AS dysregulation across the 14 established cancer hallmarks. We systematically summarize current therapeutic targets and related strategies for targeting AS in cancer, including small molecules and antisense oligonucleotides (ASOs). Moreover, we further discuss the progress, challenges, and potential solutions in drug development by providing both preclinical data and clinical trial outcomes for AS-targeted therapies. In summary, this review aims to deconvolute the multi-faceted roles of alternative splicing events in cancer from underlying mechanisms to innovative therapeutics.

Graphical abstract

Alternative splicing dysregulation permeates all 14 hallmarks of tumorigenesis and progression; targeting aberrant splicing in tumors represents a pivotal anti-cancer therapeutic strategy.

1. Introduction

Since the discovery in 1977 that introns in eukaryotic “split” genes can be removed by spliceosome¹, alternative splicing (AS), a regulatory mechanism that processes a single pre-mRNA into multiple isoforms through differential splice site selection, has been recognized as a critical step in the expression of most human genes^2,3. Almost all transcripts of human protein-coding genes undergo one or more forms of AS from pre-mRNA to mature mRNA⁴, such as exon skipping, intron retention, selective 5′- and 3′-splice sites, and neighboring exons being mutually exclusive, which allows a single gene to generate many different mature mRNAs encoding a wide range of proteins⁵. These splicing forms are regulated by multiple RNA-binding proteins (RBPs) and are dependent on cis-acting elements and trans-acting factors⁶. AS plays a critical role in various physiological functions and holds significant implications. Regarding physiological functions, it regulates gene expression levels, influencing mRNA stability and translation efficiency. Furthermore, it alters the function, activity, and subcellular location of proteins. Regarding its importance, AS plays a role in regulating specific tissues and developmental stages, aiding cellular stress responses^5,6. The comprehensive examination of AS has emphasized its crucial function in both healthy and diseased conditions, thus garnering significant attention from the scientific community.

The role of AS in cancer has been extensively studied7, 8, 9, 10. Current studies have shown that AS not only plays an important role in the occurrence and development of cancer, but also may serve as a potential therapeutic target¹¹. Aberrant AS events, such as exon skipping, intron retention, altered expression of isoforms, and splicing errors that result in the failure of tumor suppressors or promote oncogene expression, are considered to be key markers of cancer progression¹². Since the identification of the 14 hallmarks of cancer, substantial evidence has demonstrated that aberrant AS events play a pivotal role in these hallmark characteristics¹³. For instance, the dysregulation of AS events mediated by polypyrimidine tract-binding protein 1 (PTBP1) and serine/arginine-rich splicing factor 1 (SRSF1) is intricately associated with tumor proliferation, invasion, metastasis, and immune escape mechanisms^14,15. In recent years, the cancer-promoting phenotype of transcripts generated by aberrant splicing can be altered by targeting the molecular and cellular processes disrupted during aberrant splicing events, which is of great clinical value. Therefore, small molecules targeting oncogenic splicing factors or components of the splicing machinery are being developed as anticancer therapies¹⁶.

In this review, we first systematically summarize the major types and underlying mechanisms of AS. Subsequently, an in-depth elaboration on AS events and their associated proteins linked to the 14 hallmarks of cancer are provided, with the aim of identifying potential drug targets for targeted intervention of AS in tumor therapy. In addition, we highlight that dysregulation of AS in non-coding RNAs (including lncRNAs and circRNAs) contributes to tumor progression, thereby expanding the therapeutic potential of targeting this mechanism. Finally, we summarize the current research landscape for drugs targeting AS in cancer therapy—such as small molecule inhibitors and antisense oligonucleotides (ASOs)—covering both preclinical findings and clinical trial outcomes, along with associated challenges and potential solutions. In conclusion, this review provides a comprehensive overview of the multifaceted roles of AS in tumorigenesis and development, along with potential therapeutic targets for tumor treatment, emphasizing that targeting AS represents a critical strategy for effective tumor therapy.

2. Dysregulated alternative splicing events in tumor pathogenesis

AS is a fundamental step in eukaryotic gene expression, during which spliceosomes—comprising five small nuclear RNAs (U1, U2, U4, U5, and U6) and numerous proteins—precisely remove introns and ligate exons to form mature mRNA¹⁷ (Fig. 1). This process relies on the recognition of conserved splicing elements, including the 5′ and 3′ splice sites, branch point, and polypyrimidine tract^18,19. Splicing occurs through a dynamic five-step cycle involving spliceosome assembly, activation, catalysis, and disassembly. AS outcomes are modulated by cis-regulatory elements (ESE, ESS, ISE, ISS) and trans-acting splicing factors, which together guide splice site selection²⁰. To date, seven major AS types have been identified, such as exon skipping (ES), intron retention (RI), and alternative splice site usage²¹. While AS greatly enhances transcriptomic diversity, its dysregulation can lead to aberrant isoforms and is now recognized as a driver of tumorigenesis. In this section, we will examine how AS contributes to different cancer hallmarks.

Figure 1.

The regulatory mechanisms and classification of AS. (A) The regulatory mechanism of AS. Initially, the U1 snRNA within the U1 snRNP recognizes the 5′ SS of the pre-mRNA through complementary base pairing. Subsequently, SF1 binds to the BPS, while the U2 snRNP subunits U2AF2 and U2AF1 bind to the PPT and the 3′ SS, respectively, forming the E complex. Next, U2 snRNP replaces SF1 and binds to the BPS, leading to the formation of the A complex. This is followed by the recruitment of the U4, U5, and U6 trimeric snRNP complexes, which facilitates the re-arrangement of the B complex into a catalytically active form. Finally, two consecutive transesterification reactions occur, resulting in the formation of mature mRNA and an intron lariat. (B) Cis-regulatory elements within pre-mRNA molecules interact with splicing factors to modulate the splicing process. Specifically, SRSF proteins function as splicing activators by binding to (ESEs and ISEs, thereby promoting efficient splicing. In contrast, hnRNPs act as inhibitors by interacting with ESSs and ISSs, which prevents the binding of splicing factors to these sites. (C) The types of AS. RNA splicing encompasses both constitutive splicing and AS (retention intron, exon skipping, alternative 3′UTR, alternative 3′ splicing, alternative 5′ splicing, alternative 5′UTR and mutually exclusive exon). The purple boxes denote constitutive exons, the pink boxes indicate alternative exons, the lines or gray boxes represent introns, and polyadenylation sites are marked by AAA.

2.1. Dysregulated mRNA alternative splicing: A foundational driver of tumor pathogenesis

Dysregulation of AS serves as a pivotal driver in tumorigenesis and malignant progression²². Genetic mutations or aberrant expression profiles of spliceosomal machinery components and splicing regulatory factors can disrupt cancer-relevant gene networks and signaling pathways, primarily by reprogramming the splicing patterns of downstream target genes¹². Here, we will conduct an innovative and in-depth analysis of the role and mechanism of abnormal AS from the perspective of the canonical hallmarks of cancer. Cancer is characterized by a series of defining biological hallmarks, such as unbridled cell division, defiance of antiproliferative signals, and insusceptibility to programmed cell death, which collectively encapsulate its essence²³. Meanwhile, we comprehensively summarize the potential targets for tumor treatment by targeting alternative splicing events (Table 124, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91).

Table 1.

Dysregulation of alternative splicing affects tumor hallmarks.

Hallmark of tumor	Mechanism	Splicing factor/RNA isoform	Ref.
Sustaining proliferative signaling	Recurrent mutations in splicing factors	RBM10	24
	Splicing factor expression alterations	HNRNPA1, HNRNPA2, PTB, SF3A2, PUF60, SF3B3, SRSF1, SRSF3, SRSF9	25, 26, 27, 28, 29
	Aberrantly spliced RNA isoforms	EIF4H, MKRN1, TMEM, ILF3, HER2, EGFR, HRAS, MDM2, BIRC5, MKNK2, GAP17, CCND1, PPS6KB1, PKM, NUMB	25,26,29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39
Evading growth suppressors	Splicing factor expression alterations	HNRNPK	37
	Aberrantly spliced RNA isoforms	P53, GAP, KLF6, PTEN	37,40, 41, 42, 43
Avoiding immune destruction	Splicing factor expression alterations	PTBP3	44
	Aberrantly spliced RNA isoforms	IL-18, HLA-G, MHC-I, IL-1RAP, CD19, CD44	12,44, 45, 46
Enabling replicative immortality	Splicing factor expression alterations	NOVA1, hnRNPK, hnRNPD, SRSF11, hnRNPH2, hnRNPL, NOVA1, PTBP1	47, 48, 49, 50, 51
	Aberrantly spliced RNA isoforms	TERT	47,49
Tumor-promoting inflammation	Splicing factor expression alterations	PTBP1	52
	Aberrantly spliced RNA isoforms	EXOC7, CD45, IL7	12,52
Activating invasion & metastasis	Splicing factor expression alterations	ESRPs, SRSFs, RBFOXs, hnRNP family, MBNLs, QKI, AKAP8, RBM47, NSrp70	53, 54, 55, 56, 57
	Aberrantly spliced RNA isoforms	ARHGEF11, CCDC50, CDK2, CLSTN1, CTNND1, CTTN, DNM2, ENAH, EXOC7, FGFR2, FMNL3, FN1, ITGA6, MAP3K7, MPRIP, MYL6, PARD3, PLOD2, PKM2, RAC1, RON, SCRIB, SLK, TCF7L2, TPM1, TPM2, TSC2	58
Inducing or accessing vasculature	Splicing factor expression alterations	SRSF6, SRSF1, SRSF2	59
	Aberrantly spliced RNA isoforms	VEGFA, VEGFR, NRR, UNC5B	59, 60, 61, 62, 63, 64
Genome instability & mutation	Splicing factor expression alterations	SNRPB, RBM39	65,66
	Aberrantly spliced RNA isoforms	BRCA1, BRCA2, Chk1	65,67
Resisting cell death	Splicing factor expression alterations	SAM68, RBM4, RBM10, PTBP1, RBM25, SRSF1, hnRNPF, hnRNPH, hnRNPK, SRSF9, RBM5, SRSF2	68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78
	Aberrantly spliced RNA isoforms	BCL-X, BCL-2, MCL-1, Fas, CASP9, CASP8, BIM, BIN1, CASP2	77, 78, 79, 80, 81, 82, 83
Deregulating cellular metabolism	Splicing factor expression alterations	SRSF3, PTBP1, hnRNPA1, hnRNPA2	84, 85, 86
	Aberrantly spliced RNA isoforms	PKM, LDHA, TNFSF13, ARHGAP4, MARCH7, PCBP2, LRCH3	84, 85, 86, 87
Unlocking phenotypic plasticity	Splicing factor expression alterations	MBNL1	88
	Aberrantly spliced RNA isoforms	MAP2K7, p63	88,89
Nonmutational epigenetic reprogramming	Splicing factor expression alterations	SRSF3, SHNRNP	90,91
	Aberrantly spliced RNA isoforms	Flower, PRSS3	90,91
Senescent cells	Splicing factor expression alterations	PTBP1	52
	Aberrantly spliced RNA isoforms	EXOC7	52

2.1.1. Sustaining proliferative signaling

The capacity for indefinite proliferation is a significant characteristic of tumors¹³. This is because dysregulation of AS allows tumor cells to remain unaffected by normal growth-inhibitory signals, whereas in normal cells, their proliferation is regulated^12,92. For example, in lung cancer cells, mutations in RBM10 have been identified, which affect the AS of EIF4H exon 5 and stimulate tumor proliferation^24,30. In addition, changes in the expression levels of splicing factors may cause alterations in the AS of their downstream target genes. For example, the oncogenic transcription factor c-Myc is known to upregulate the transcription of splicing factors PTB, hnRNPA1, and hnRNPA2. This resulting overexpression of these splicing factors, in turn, alters the splicing of pyruvate kinase isoenzyme (PKM), favoring the production of the PKM2 isoform, which is known to promote tumor cell proliferation²⁵. Similarly, the overexpression of SF3A2 specifically regulates the AS of MKRN1 and promotes the expression of the dominant oncogenic isoform MKRN1-T1, accelerating the proliferation of triple-negative breast cancer (TNBC) cells²⁶. As a key oncogene, SRSF1 has been demonstrated to regulate splicing alternation in breast cancer by directly binding to a motif within exon 3 of the PTPMT1 gene, thereby inducing abnormal proliferation of breast cancer cells⁹³ (Fig. 2). Moreover, the splicing of the RPS6KB1 gene, which encodes the protein ribosomal S6 kinase 1 (S6K1), is associated with the continuous proliferation of breast and lung cancer cells¹². The splicing of RPS6KB1 is regulated by SRSF1⁹⁴. In breast cancer, the splice variant RPS6KB1-2 isoform of S6K1, a shorter isoform produced by the inclusion of the cassette exons 6a, 6b, and 6c, promotes tumor proliferation⁹⁵.

Figure 2.

SRSF1 modulates multiple tumor characteristics via AS. (A) SRSF1 facilitates tumor progression by regulating the AS of genes associated with tumor immunity, apoptosis, and proliferation. (B) SRSF1 modulates the AS of VEGFA pre-mRNA to enhance tumor angiogenesis. (C) SRSF1 promotes tumor invasion and metastasis by regulating the inclusion of exon 3 in SRA1.

EGFR, belonging to the receptor tyrosine kinase family, is important for cell division. EGFRvIII, the alternative splice isoform of EGFR, is a mutant commonly found in a variety of cancers (e.g., glioblastoma and breast sarcoma-like tumors) that activates downstream signaling pathways independently of ligands, thereby consistently stimulating cell proliferation⁹⁶. In head and neck squamous cell carcinoma, isoform A of EGFR exon 18‒21 mutations promote cell proliferation, differentiation, migration and inhibits apoptosis⁹⁷. Other genes that produce isoforms that maintain proliferative signaling in cancer include TMEM134³¹, ILF3²⁹, HER2³², HRAS³³, MDM2³⁴, BIRC5³⁵, MKNK2³⁶, GAP17³⁷, CCND1³⁸, NUMB³⁹, they may be regulated by splicing factors SF3A2, PUF60, SF3B3, SRSF1, SRSF3 and SRSF9^27,28.

2.1.2. Evading growth suppressors

Tumor cells have the ability to evade growth suppressors, which relies on the function of tumor suppressor genes¹³. Numerous tumor suppressors have been discovered that restrict cell growth and proliferation through different mechanisms, with the p53 protein and retinoblastoma-associated protein (RB) being the more traditional examples⁹⁸. As a transcription factor, p53 orchestrates cell cycle arrest and apoptosis by activating downstream genes in response to cellular stress and injury⁹⁹. The N-terminal truncated p53 isoform (Δ40 p53/ΔNp53) is still able to activate gene expression⁴⁰. Moreover, mutations in p53 specific to cancer trigger the translation of Δ160p53, which promotes tumor development⁴¹. Mutant p53 increases the expression of the splicing regulator hnRNPK, which causes the loss of cell membrane binding of the cytosine-rich exon-containing GTPase-activating protein isoform, leading to the development of pancreatic ductal adenocarcinoma (PDAC)³⁷. A comprehensive analysis of RB mutations in cancer revealed that more than 15% of these mutations are located at splice sites¹⁰⁰. Other tumor suppressor genes such as PTEN⁴³ and kruppel-like factor 6⁴² also produce oncogenic isoforms that promote tumor development.

2.1.3. Avoiding immune destruction

The body recognizes and removes abnormal tumor cells through immune surveillance mechanisms to inhibit cancer development, but tumor cells are able to escape recognition and attack by the immune system through a variety of mechanisms, thereby promoting their own proliferation and ultimately the formation of cancer^92,101. One of these escape mechanisms is the expression of splicing factors, such as polypyrimidine region-binding protein 3 (PTBP3), an oncogene that drives the progression of hepatocellular, gastric, and colorectal carcinomas. PTBP3 was found to be highly expressed in gallbladder cancer (GBC) by bioinformatics, and mechanistic studies revealed that PTBP3 promotes exon skipping of IL-18, generating the heterodimer ΔIL-18 that is specifically expressed in tumors, thus promoting immune escape from GBC⁴⁴. Beyond PTBP3, other splicing factors also play pivotal roles in immune evasion. For instance, SRSF1 regulates PD-1 splicing via alternative splicing, generating a PD-1 splice variant lacking exon 3 (ΔEx3PD1). This variant is endogenously expressed in PBMCs and T cells and prevents tumor cell-mediated T cell exhaustion¹⁰² (Fig. 2). Conversely, PTBP1 is essential for regulating T cell homeostasis and antitumor immunity. Studies indicate that specific PTBP1 deficiency in dendritic cells (DCs) increases MHC II expression and enhances antitumor immunity by modulating alternative splicing of PKM and several interferon-responsive genes¹⁰³ (Fig. 3). Moreover, HLA-G, a class I human leukocyte antigen, is extensively expressed across various tumor types. The conversion of the splicing isoform HLA-G1 to HLA-G2 can mediate immune evasion of melanoma cells⁴⁵. Other splice isoforms reported to be associated with tumor immune escape include MHC-I, IL-1RAP, CD19 and CD44^12,46.

Figure 3.

PTBP1 modulates multiple tumor characteristics via AS. (A) PTBP1 modulates the exon 10 skipping of AXL pre-mRNA, consequently impacting tumor proliferation, invasion, and metastasis. (B) PTBP1 controls the AS of PKM, thereby regulating tumor immunity. (C) PTBP1 governs the AS of genes including MBNL1, ZNF207, APP, and CD44, which in turn affects genomic stability and apoptosis in tumor cells。.

2.1.4. Enabling replicative immortality

85%‒90% of cancers maintain their malignant growth by lengthening telomeres via telomerase, which gives them unlimited replication capacity¹⁰⁴. Multiple splicing isoforms of hTERT, a crucial part of telomerase activity, have been identified, yet the splicing factors that promote or suppress these isoforms remain largely unknown. Ludlow et al.⁴⁷ found by genetic engineering experiments that in non-small cell lung cancer the splicing factor NOVA1 promotes exon inclusion in the structural domain of the hTERT reverse transcriptase, resulting in the generation of full-length TERT transcripts. In breast cancer cells, beta deletion is the most highly expressed hTERT transcript, regulated by the splicing regulators hnRNPK, hnRNPD, SRSF11, hnRNPH2, hnRNPL, NOVA1 and PTBP147, 48, 49, 50, 51.

2.1.5. Tumor-promoting inflammation

Inflammation, especially chronic inflammation, plays a decisive role in the process of tumor development, progression, malignant transformation, infiltration and metastasis by activating the survival and proliferation programs of cancer cells, promoting gene mutation, influencing immune surveillance, establishing an immune microenvironment, and providing growth factors, making it one of the culprits in tumor development and progression¹⁰⁵. Expression of the splicing factor PTBP1 is positively correlated with growth and poor prognosis of various cancers, and it has been found that PTBP1-regulated EXOC7 variable splicing regulates inflammation in vivo⁵². In addition, splice isoforms defined by Weinberg and Hanahan as being associated with inflammatory markers are CD45 and IL7, which are closely associated with tumor development¹².

2.1.6. Activating invasion & metastasis

The invasion-metastasis cascade is one of the complex functions of cancer cells and involves local invasion of primary tumor cells, entry and survival in the circulatory system, metastasis to distant tissues and colonization. Micrometastatic clusters then proliferate to form detectable metastatic foci, a process that is the primary cause of more than 90 per cent of cancer-related deaths¹⁰⁶. To overcome selective pressure during tumor metastasis, cellular plasticity enables tumor cells to switch phenotypic and functional characteristics through the reversal processes of epithelial mesenchymal transition (EMT) and mesenchymal to epithelial transition (MET), which enhances migratory and invasive capabilities¹⁰⁷. This complex process is accompanied by the regulation of a large number of gene expression, and AS is an important way to regulate gene expression, which occupies a certain position in tumor metastasis^53,54.

A variety of splicing factors, including ESRPs, SRSFs, RBFOXs, heterogeneous nuclear ribonucleoproteins (hnRNPs), and muscleblind-like proteins, have been identified as regulators of EMT-related alternative splicing^53,55. ESRPs (ESRP1 and ESRP2) were originally discovered in the regulation of FGFR2 splicing^108,109. FGFR2 comprises two mutually exclusive exons, IIIb and IIIc, with the FGFR2-IIIb isoform present in epithelial cells and the FGFR2-IIIc isoform found in mesenchymal cells^110,111. In addition, ESRPs promote the EMT in breast cancer cells by influencing the AS of CD44 to produce the isoform CD44s¹¹². Interestingly, the SR-associated protein NSrp70 regulates CD44 exon 6‒13 skipping, generating the variant CD44v10, which promotes TNBC cell migration⁵⁷. ESRPs directly influence the alternative splicing of numerous other EMT-related genes, such as CTNND1, Exo70, ITGA6, SLK, ARHGEF11, CLSTN1, ENAH, MAP3K7, TCF7L2, Fibronectin, RAC1, RON and TSC2^56,113, 114, 115, 116. As an EMT inducer, TGF-β suppresses the expression of ESRP1 and ESRP2, thereby antagonizing their mediated splicing events^58,117. Among the RBFOX protein family, RBFOX2 exhibits the broadest expression. The increase in RBFOX2 expression induced by TGF-β further promotes the expression of the MAP3K7 isoform¹¹⁸. Moreover, RBFOX2 collaborates with ESRPs, MBNL1, hnRNP F/H/M, PTBP1, and SRSF1 to determine splicing choices that affect whether certain exons are included or excluded119, 120, 121, 122, 123. These interactions can be synergistic or antagonistic. For example, in hepatocellular carcinoma cells, PTBP1 inhibits AS of exon 10 by competing with U2AF2 for binding to the polypyrimidine tract (TCCTCTCTGTCCTTTCTTC) on intron 9 of Axl, a receptor tyrosine kinase involved in cancer metastasis¹²⁴ (Fig. 3). This promotes the production of Axl-S, thereby influencing cellular invasion and metastasis. Similarly, SRSF1 has been shown to promote exon 3 inclusion in SRA1 precursor mRNA by interacting with its third exon, thereby enhancing the invasive capacity of hepatocellular carcinoma cells¹²⁵ (Fig. 2).

2.1.7. Inducing or accessing vasculature

Tumor development and progression is dependent on angiogenesis to meet its needs for oxygen and nutrients¹²⁶. During angiogenesis, pro-angiogenic factors play a key activating role, such as VEGF-A, vascular endothelial growth factor receptor (VEGFR) and neurofibrillary protein (NRP)¹²⁷. The AS events of these pro-angiogenic factors play an integral role in angiogenesis.

VEGF-A is the most potent member of the VEGF family, with splicing proteins like SRSF6, SRSF1, and SRSF2 playing a role in its splicing regulation^59,128. There are six major isoforms of VEGF-A, namely VEGF-A111, VEGFA121, VEGF-A145, VEGF-A165, VEGF-A189 and VEGF-A206. And different splicing modes can produce pro-angiogenic VEGF-A_xxxa isoforms and anti-angiogenic VEGF-A_xxxb isoforms^129,130. Two mutually exclusive AS sites VEGF₁₆₅ and VEGF₁₆₅b found in terminal exon 8 of the VEGF-A gene¹³¹. VEGF-A₁₆₅, a highly potent pro-angiogenic factor, is commonly found to be upregulated in tumors⁶⁰. For example, extracellular matrix (ECM) promotes SRSF1 expression via the YAP/RUNX2/SRSF1 signaling axis, whereas VEGF-A undergoes specific shearing by SRSF1 to produce large amounts of VEGF-A₁₆₅ and promote neuroblastoma (NB) angiogenesis¹³². A similar mechanism has been identified in HCC, where LINC01446 binds to SRPK2, activating SRSF1. This enhances the splicing and expression of the VEGF-A₁₆₅ isoform, ultimately promoting HCC progression¹³³ (Fig. 2). VEGF-A₁₆₅b is endogenously expressed in healthy tissues and contributes to the maintenance of normal vascular function and homeostasis. However, the expression level of VEGF-A₁₆₅b tends to be downregulated in a variety of cancers (e.g., colon, kidney, prostate, and metastatic melanoma)¹³⁴. Interestingly, in lung adenocarcinoma, the VEGF₁₆₅b isoform is elevated and a high VEGF₁₆₅b/VEGF₁₆₅ ratio is associated with lymph node metastasis¹³⁵. Down-regulation of VEGF-A₁₆₅b is usually accompanied by up-regulation of pro-angiogenic VEGF-A₁₆₅, allowing tumors to access a greater blood supply, thus supporting their growth and spread. For example, in colon cancer, variable splicing of the T-cell Intracellular Antigen (TIA-1) may regulate the balance between VEGF-A₁₆₅ and VEGF-A₁₆₅b¹³⁶. In addition, ubiquitin-specific peptidase 39 (USP39) binds specifically to serine-arginine protein kinase 1 (SRPK1), which inhibits the phosphorylation of SRSF1 by SRPK1, resulting in the SRSF1 fails to shear VEGF-A, which blocks the production of VEGF-A₁₆₅b, thereby promoting malignant proliferation and angiogenesis in renal cell carcinoma^137,138. Due to these properties, VEGF-A₁₆₅b and its regulatory mechanisms have become an important topic in cancer research, and researchers are investigating ways to inhibit tumor growth and angiogenesis by regulating VEGF-A₁₆₅b expression, which may provide new ideas for anticancer therapy.

VEGFR, NRP and Unc-5 reticulin receptor B (UNC5B) can also be variably spliced and potentially promote angiogenesis. For example, in squamous cell lung cancer, SOX2 inhibits the NOTCH signaling pathway and contributes to increased expression levels of sVEGFR1-i13 in endothelial cells⁵⁹. sVEGFR1-i13 promotes the growth of squamous cell lung cancer cells and enhances their resistance to anti-angiogenic therapy by regulating the β1 integrin/VEGFR autocrine loop^61,62. NRP1 splice variants NRP1-Δ4 and NRP1-Δ5 secreted by cellular exosomes can activate the FAK/p130Cas signaling pathway through interaction with Met and β1-integrin and promote tumor angiogenesis, thereby facilitating migration, invasion and metastasis of colorectal cancer⁶³. UNC5B is a netrin-1-dependent receptor that plays an important role in mediating angiogenesis and tumorigenesis, and its spliceosome isoform, UNC5B-Δ8, can regulate angiogenesis by inducing apoptosis through its cytoplasmic death domain⁶⁴.

2.1.8. Genome instability & mutation

Abnormal proliferation of tumor cells increases the tendency for genomic alterations and mutations, resulting in impairment of multiple genes that regulate cell division and tumor suppression. Persistent genomic changes underlie multiple features of tumors, and mutant genotypes provide selective advantages for cellular subcloning and drive tumor progression⁹². Mutations are alterations in an organism’s DNA sequence; key proteins and genes crucial for DNA damage repair and genome stability encompass BRCA1, BRCA2, and Chk1^92,139,140.

Alterations in the BRCA1 and BRCA2 genes, which are crucial oncogenes linked to breast cancer risk, cause faulty homologous recombination, leading to a marked rise in genomic instability¹⁴¹. Mutation screening has now identified 20,000 unique germline BRCA1 and BRCA2 variants¹⁴². BRCA1 is a highly variable splicing gene. Its inactivation is mainly seen in hereditary breast and ovarian cancers, as well as in basal-like breast cancers. Inactivation of the BRCA1 gene is also seen in other cancers such as prostate, pancreatic and colon cancers, but occurs less frequently¹⁴³. The splice variants of BRCA1, BRCA1a and BRCA1b, lose certain protein-binding domains, including RB, p53, MYC, RAD50, and TUBG, due to the lack of a nuclear localization signal located in exon 11⁶⁷. And its subcellular localization is closely related to its function, with nuclear BRCA1a promoting apoptosis and cytoplasmic BRCA1a promoting breast cancer cell proliferation¹⁴⁴. In conclusion, the complex functions of BRCA1a and BRCA1b in cancer are closely related to their splice variant status and cellular localization. Mutations in the BRCA2 gene are predominantly found in hereditary breast, ovarian, pancreatic and prostate cancers. For example, deletion of BRCA2 exon 3 results in deletion of the PALB2 binding domain necessary for homologous recombination and promotes ovarian cancer progression, and this process is regulated by the splicing factor SNRPB⁶⁵. In addition, the splicing factor RBM39 plays an important role in ovarian cancer patients with mutations in the BRCA1/2 gene⁶⁶. In the DNA damage response, the long noncoding RNA LUCAT1 interacts with PTBP1 in CRC cells, promoting PTBP1 binding to certain DNA damage-related genes and altering their alternative splicing patterns, thereby regulating CRC cell resistance to DNA-damaging agents¹⁴ (Fig. 3).

Chk1 is a serine/threonine protein kinase that responds rapidly to DNA damage processes. Jumps in exon 3 of CHK1 produce the shorter isoform Chk1-S (highly expressed in ovarian, testicular, and hepatocellular carcinomas) that lacks the amino-terminal structural domain for ATP binding, thereby inhibiting full-length CHK1^145,146.

2.1.9. Resisting cell death

Cancer cells acquire resistance to cell death; aberrant regulation of apoptosis is a major mechanism driving malignant tumorigenesis^92,147. BCL-2 family members produce different isoforms, such as BCL-X/BCL2L1, BCL-2, or MCL-1, thereby contributing significantly to cancer therapy⁷⁹. BCL-X produces the anti-apoptotic BCL-XL and the pro-apoptotic BCL-XS around its 5′ splice site, respectively⁷⁹. And it is regulated by splicing factors SAM68, RBM4, RBM10, PTBP1, RBM25, SRSF1, hnRNPF, hnRNPH, hnRNPK and SRSF968, 69, 70, 71, 72, 73, 74, 75, 76. Among these, PTBP1 has been demonstrated to directly regulate BCL-X splicing. Bielli et al.⁷⁰ showed that PTBP1 overexpression promotes selection of the distal 5′ splice site of BCL-X exon 2, leading to the generation of the pro-apoptotic BCL-Xs isoform. Conversely, PTBP1 depletion enhances splicing of the anti-apoptotic BCL-XL variant (Fig. 3). SRSF1 is another key regulator of apoptosis. Its overexpression has been shown to promote the transformation of mammary epithelial cells by regulating AS of BIM (BCL2L11) and BIN1, generating isoforms lacking pro-apoptotic functions⁹³ (Fig. 2). Three MCL-1 isoforms have been identified, MCL-1L, MCL-1S, and MCL-1ES. Among them, MCL-1L has an anti-apoptotic effect, while MCL-1S and MCL-1ES have a pro-apoptotic effect¹⁴⁸. Notably, studies on the splicing regulators of MCL-1 have mostly been conducted in parallel with studies on BCL-X. BCL-2 exists as BCL-2α and BCL-2β isoforms. Among these, BCL-2α, similar to MCL-1L and BCL-XL, is the most extensively studied for its significant anti-apoptotic effects^149,150.

In addition, RBM5, which is aberrantly expressed in lung and breast tumors, regulates splicing of Fas receptor exon 6 to generate pro-apoptotic membrane-bound or anti-apoptotic soluble Fas receptors⁷⁷. The CASP9 gene produces pro-apoptotic Casp9a and pro-survival Casp9b through inclusion/exclusion of exon 3, 4, 5, and 6 cassettes and is regulated by SRSF1 and SRSF2⁷⁸. There are also genes that produce isoforms involved in resistance to cell death, such as CASP8, BIM, BIN1 and CASP280, 81, 82, 83.

2.1.10. Deregulating cellular metabolism

The Warburg effect reveals that tumor cells preferentially obtain energy through glycolysis, even when oxygen is sufficiently available, producing lactate as the end product¹⁵¹. Although glycolysis is not an efficient way to supply energy, tumor cells use it to meet the demands of rapid proliferation and to gain the ability to reprogram their energy metabolism. This metabolic reprogramming is an adaptation to the environment during tumor formation or metastasis¹⁵². The genes of some key metabolic enzymes in the aerobic glycolysis pathway are prone to AS, resulting in abnormal splice variants that provide a metabolic advantage to tumor cells. In the final step of glycolysis, phosphoenolpyruvate is converted into pyruvate through the catalytic action of pyruvate kinase (PKM)¹⁵³. The splicing of the PKM gene can cause imbalances in cellular metabolism. This gene generates the isoforms PKM1 and PKM2 by splicing two exons that are mutually exclusive, exon 8 and exon 9. PKM1 is normally expressed routinely in adult normal cells and stimulates oxidative phosphorylation, whereas PKM2 emerges during embryonic development, only to be switched off in adult cells, but in many cancers is re-expressed as a promoter of aerobic glycolysis¹⁵⁴. PKM2 appeared upregulated in colorectal, hepatocellular, breast, and pancreatic ductal adenocarcinoma and was associated with survival time, advanced stage, and poor prognosis^84,86. For example, in colorectal cancer, AS increases the PKM2/PKM1 ratio and promotes aerobic glycolysis in response to metabolic stress. This process is regulated by SRSF3, PTBP1, hnRNPA1 or hnRNPA2 splicing factors^84,85.

Furthermore, in an RNA-Seq study, it was shown that splicing events in breast cancer cells under hypoxic conditions may be related to cellular metabolism. For instance, this includes intron retention affecting targets LDHA, TNFSF13, and ARHGAP4, and exon skipping affecting targets MARCH7, PCBP2, and LRCH3⁸⁷.

2.1.11. Unlocking phenotypic plasticity

Under normal conditions, cell development is usually accompanied by terminal differentiation. In contrast, cancer cells escape or escape from the state of terminal differentiation by unlocking the ability of phenotypic plasticity, which is a key component of the mechanism of carcinogenesis¹³. This plasticity can manifest itself in several different forms, including dedifferentiation, resistance to differentiation and trans differentiation, and AS plays a key role in these processes. MBNL1 is a suppressor of tumor dedifferentiation, and its downregulation increases the organism’s tumorigenicity and progenitor-like properties. Downregulation of MBNL1 in breast, lung and gastric adenocarcinomas suggests lower overall survival, and in triple-negative breast cancers suggests recurrence and distant metastasis. Mechanistic studies revealed that MAP2K7 exon 2 deletion in MBNL1 low-expressing cancer cells caused tumor dedifferentiation⁸⁸. p63 is a master regulator of basal cells. In human eNEC cells, ectopic expression or inhibition of EZH2 upregulates the major p63 isoform ΔNp63α, thereby promoting squamous trans differentiation of eNEC cells⁸⁹.

2.1.12. Non-mutational epigenetic reprogramming

Genomic instability and mutations are key factors in cancer formation, however, there exists an independent mode of genomic reprogramming known as “non-mutational epigenetic reprogramming”. This primarily alters gene expression through epigenetic regulation, which mainly involves DNA methylation, histone modifications, and non-coding RNAs¹⁵⁵. Notably, increasing evidence indicates that aberrant physical properties of the tumor microenvironment trigger extensive epigenetic reprogramming during tumor development and progression, where AS plays a significant role. For example, in highly differentiated plasmacytoid ovarian cancer (HGSC), tumor cells promote cancer growth, metastasis and invasion by actively reducing microenvironmental cell viability. Tu-Stroma, a long non-coding RNA predominantly synthesized by tumor cells and subsequently delivered to stromal tissues via exosomes, compels microenvironmental cells to express Flower Lose isoforms, thereby inducing a hypo-adaptive state. Mechanistic studies revealed that Tu-Stroma forms DNA–RNA adducts with exon 3 of the Flower gene and induces DNA methylation, histone methylation, HP1 binding, and recruitment of the splicing factor SRSF3, which leads to exon 3 exclusion and generation of the Flower Lose isoform⁹⁰.

Epigenetic modifications like DNA methylation, RNA methylation, histone modification, and protein phosphorylation, mediated by AS, are crucial in the progression of hepatocellular carcinoma (HCC). For example, site-specific CpG methylation leading to abnormal expression of PRSS3 splice variants might synergistically exert the oncogenic effects of PRSS3-V2 with the tumor-suppressor functions of PRSS3-V1, ultimately promoting molecular diversity and functional adaptability of PRSS3 in HCC. Divergent long-stranded non-coding RNA transcribed from the GATA3 locus generates a regulatory element through AS, and this element directs VIRMA to m6A methylation at the 3′ end of the GATA3 pre-mRNA, and VIRMA promotes degradation of the GATA3 pre-mRNA through this m6A modification, thereby promoting hepatocellular carcinoma development. In addition, phosphorylation of the splicing factors SRSF3 and HNRNP A18 may be a major mechanism for the progression of non-alcoholic fatty liver disease (NAFLD) to HCC⁹¹.

2.1.13. Polymorphic microbiomes and senescent cells

The polymorphic microbiome and senescent cells represent two emerging frontiers in cancer biology, both harboring unexplored connections to AS mechanisms in tumor progression. The polymorphic microbiome, in symbiosis with body barrier tissues exposed to the external environment (especially epidermis and internal mucous membranes), the gastrointestinal tract, the lungs, the breasts, and the genitourinary system, broadly influences the acquisition of hallmark capabilities for cancer¹³. The microbiome, comprising bacteria, fungi, viruses, and mycoplasmas, interacts dynamically with barrier tissues and has been implicated in over 20 malignancies¹⁵⁶. While its role in DNA mutagenesis is well-documented, its influence on RNA splicing remains enigmatic. Future studies could investigate whether microbial metabolites, pathogen-associated molecular patterns, or dysbiosis-induced inflammation directly modulate spliceosome activity or splicing factor expression (such as hnRNPs and SR proteins), thereby generating oncogenic mRNA isoforms that drive tumor heterogeneity.

Cellular senescence was historically viewed as a tumor suppressor, in certain situations, aging cells promote the growth and advancement of tumors^157,158. The main mechanism is thought to be senescence-associated secretory phenotype, which transmits signaling molecules in a paracrine manner to neighboring surviving cancer cells as well as to other cells in the tumor microenvironment, thereby transmitting signature capabilities such as proliferative signaling, avoiding apoptosis, inducing angiogenesis, stimulating invasion and metastasis, and suppressing tumor immunity¹³. Recent RNAi screens identified PTBP1—a splicing regulator—as a SASP controller via EXOC7 isoform switching, linking AS directly to tumor-promoting secretory programs⁵². This discovery suggests broader possibilities: other splicing factors may govern SASP component diversity (such as IL-6/IL-8 variants), and senescent cell-derived extracellular vesicles might deliver splicing-regulatory RNAs to neighboring cells. Notably, microbiome-senescence crosstalk remains unexplored—for instance, whether microbial communities’ prime senescent cells to adopt AS-dependent pro-tumorigenic phenotypes or vice versa.

2.2. Non-coding RNA splicing misregulation: Expanding the regulatory repertoire of tumorigenesis

With the advancement of technologies like sequencing and high-resolution splicing structure analysis, the role of AS of non-coding RNAs in tumor progression has been gradually recognized. This also offers more opportunities for targeting AS to treat tumors, as the quantity of non-coding RNAs in the human body is significantly higher than that of mRNAs. Taking the AS regulated by long non-coding RNAs (lncRNAs) as an example, numerous studies have demonstrated that the dysregulation of AS mediated by lncRNAs is closely associated with the malignant progression of various tumors. For example, Bian et al.⁷ discovered that LINC01852 can regulate the AS of PKM by facilitating the ubiquitination and degradation of SRSF5, reducing the PKM2 splicing subtype, and consequently inhibiting tumorigenesis and chemotherapy resistance in colorectal cancer. In liver cancer, LINC01089 mediates exon 3 skipping of DIAPH3 through interaction with hnRNPM¹⁵⁹. Mechanistically, knocking down LINC01089 enhances the inclusion of DIAPH3 exon 3, which harbors an important m6A modification site. IGF2BP3 can recognize this site to enhance the stability of DIAPH3 mRNA, thus suppressing the ERK/Elk1/Snail axis to prevent liver cancer metastasis¹⁵⁹. Additionally, researchers have found that the dysregulation of AS mediated by lncRNAs is also closely linked to tumor stemness¹⁶⁰. For instance, after being transcriptionally activated by CREB1, lncRNA RAB30-DT can directly bind to the splicing kinase SRPK1, thereby reshaping the AS landscape in liver cancer cells, such as the splicing of the cell cycle regulator CDCA7, and subsequently driving tumor stemness and malignancy¹⁶⁰.

Another representative example highlighting the crucial role of non-coding RNA-mediated AS in tumor progression is circRNA, exemplified by circFLNB¹⁶¹. The biogenesis of circFLNB is regulated by QKI-and CELF4-dependent AS triggered by chemotherapy. CircFLNB can bind to miR-3127-3p to modulate the MOB1B/Hippo pathway, thus suppressing the onset of colorectal cancer¹⁶¹. circ_0007386, a novel circular RNA identified to be upregulated in non-small cell lung cancer tissues and cell lines, can regulate the proliferation and apoptosis of non-small cell lung cancer both in vitro and in vivo¹⁶². Specifically, under hypoxic conditions, the interaction between YAP1 and EIF4A3 is strengthened, which facilitates the reverse splicing of CRIM1 pre-mRNA, enhances the formation of circ_0007386, and further accelerates the progression of non-small cell lung cancer¹⁶². Additionally, Zhang et al.¹⁶³ discovered that CircRAPGEF5 can interact with RBFOX2, thereby promoting the skipping of exon 4 of TFRC and resulting in the resistance of endometrial cancer cells to ferroptosis. In conclusion, these findings emphasize the great potential of non-coding RNA-mediated AS in the development of new anti-tumor targets and drugs.

In this chapter, we systematically synthesize the crucial role of aberrant AS in tumorigenesis and progression, examining its impact from two perspectives: coding RNAs and non-coding RNAs. Anchored in the 14 canonical hallmarks of cancer, we examined the oncogenic effects of aberrant AS across multiple facets, delineated the underlying molecular mechanisms, and identified potential therapeutic targets for tumor therapy via targeting this process.

3. Targeting alternative splicing for cancer therapy: Strategies and clinical translation progress

3.1. Small molecules targeting the core spliceosome

The process of RNA splicing fundamentally entails the sequential assembly, activation, catalysis, and disassembly of the spliceosome¹². Consequently, directly targeting the spliceosome in tumors characterized by aberrant splicing dysregulation is considered a promising therapeutic strategy. SF3B1, as an essential component of the spliceosome, plays a pivotal role in the selection of the BPS and the 3′SS. Inhibition of its function can disrupt splicing during the early stages of spliceosome assembly¹⁶⁴. On this basis, researchers have developed a series of spliceosome-targeting drugs (Table 2165, 166, 167, 168, 169, 170, 171, 172, 173, 174, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 188, 189, 190, 191, 192, 193, 194, 195, 196, 197, 198, 199, 200, 201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212, 213, 214, 215, 216) (Fig. 4).

Table 2.

Small molecule compounds targeting alternative splicing in tumors.

Drug	Structure	Target	Therapy	Tumor type	Ref.
Pladienolide B		SF3B1	Inhibit cellular proliferation; Induces cell cycle arrest and apoptosis	Cervical cancer; Colon cancer; Gastric cancer; Lung cancer; Breast cancer	165, 166, 167, 168
E7107		SF3B1	Inhibit cellular proliferation	Colon cancer; DMPM; CRPC	165,169,170
Meayamycin B		SF3B1	Induces cell cycle arrest and apoptosis	Head and neck cancer; DMPM	169,171
H3B 8800		SF3B1	Inhibit cellular proliferation; Induce cell apoptosis	AML; CML; T-ALL	172,173
GEX1A		SF3B1	Inhibits the pre-mRNA splicing process by binding to SF3B1; Induces both G1 and G2/M arrest; Induce cell apoptosis	Human epidermoid carcinoma; Lung cancer; Colon cancer; Leukaemia	174,175
Jerantinine A		SF3B1	Induces cell cycle arrest and apoptosis	Nasopharynx carcinoma; Breast carcinoma; Colon carcinoma; Lung carcinoma	176, 177, 178
Isoginkgetin		Spliceosome	Induces apoptosis and activates autophagy	MM; Fibrosarcoma; Breast carcinoma	179, 180, 181
Madrasin		Spliceosome	Induce cell cycle arrest	Cervical cancer	182,183
FR901464		SF3B1 and PHF5A	Induces both G1 and G2/M arrest; Induces DNA fragmentation and cell shrinkage	Breast cancer; NSCLC; Colon cancer; Leukaemia	184, 185, 186
Spliceostatin A		SF3B1	Induce cell apoptosis	CLL; Prostate cancer	187,188
Sudemycin E		SF3B1	Induces cell cycle arrest and apoptosis	Mantle cell lymphoma; Prostate adenocarcinoma; Colon cancer; Rhabdomyosarcoma; Cervical cancer	189,190
NSC 194308		U2AF2-RNA	Selective induction of cell death in leukemia cell lines harboring spliceosome mutations	Leukemia	191
NB-506		Spliceosome	Induce cell apoptosis;Phosphorylate SF2/ASF	Leukaemia; Gastric carcinoma; Colon carcinoma; Lung carcinoma	192,193
E7070		RBM39	Induces cell cycle arrest and apoptosis	Nasopharynx carcinoma; Colon cancer; Murine leukemia	194, 195, 196
E7820		RBM39	Induces cell cycle arrest and apoptosis	Colon cancer; Lung cancer; Breast cancer	197,198
Tasisulam		RBM39	Induces cell cycle arrest and apoptosis	Melanoma; Lung cancer	199
Chloroquinoxaline		RBM39	Induces cell cycle arrest and apoptosis	CML; Colon cancer	200
GSK3368715		PRMT1	Inhibit tumor growth	Diffuse large B cell lymphoma; solid tumors	201
JNJ-64619178		PRMT5	Inhibit tumor growth	Non-Hodgkin’s lymphoma; myelodysplastic syndrome; solid tumors	202
GSK3326595		PRMT5	Inhibit tumor growth	AML; Myelodysplastic syndrome; non-Hodgkin’s lymphoma	203
PF06939999		PRMT5	Induce cell cycle arrest; Inhibit tumor growth	NSCLC	204
TG003		CLK1	Inhibits cell proliferation and invasion	Gastric cancer; Prostate cancer	205, 206, 207
T025		CLK2	Inhibit cell proliferation	Lung carcinoma; Colon carcinoma; Breast carcinoma; Kidney cancer	208
SM04690		CLK2	Inhibit cell proliferation	Colon carcinoma	209,210
SM08502		CLK	Inhibit cell proliferation; Induces apoptosis and DNA fragmentation	Gastrointestinal cancer; Colon carcinoma	210
SRPIN340		SRPK1	Induces apoptosis and autophagy	Leukemia	211
SRPKIN-1		SRPK1/2	Inhibits phosphorylation of serine/arginine (SR)-rich splicing factor proteins	Cervical cancer	212
SPHINX31		SRPK1	Inhibits phosphorylation of serine/arginine-rich splicing factor 1	Cholangiocarcinoma	213
Amiloride		APAF1	Induce cell apoptosis	Glioblastoma	214
Caffeine		SRSF3	Induces cell cycle arrest and apoptosis	Cervical cancer	215
Digoxin		SRSF3	Induces cell cycle arrest, DNA damage and apoptosis	Cervical cancer	216

DMPM, diffuse malignant peritoneal mesothelioma; CRPC, castration-resistant prostate cancer; AML, acute myeloid leukemia; CML, chronic monocytic leukemia; T-ALL, T-cell acute lymphoblastic leukemia; NSCLC, non-small cell lung cancer; CLL, chronic lymphocytic leukemia.

Figure 4.

Small molecule compounds targeting AS in tumors. Different colored boxes represent distinct mechanisms by which small molecule compounds modulate AS. Specifically, Isoginkgetin, NB-506, and Madrasin inhibit the recruitment of the U4/U5/U6 tri-snRNP, resulting in stalling at the pre-spliceosomal A complex. Pladienolide B, E7107, H3B 8800, and others interfere with AS by targeting SF3B1. NSC 194308 prevents pre-mRNA splicing by stabilizing the early U2AF2‒RNA complex. Additionally, other small molecules inhibit AS by targeting post-translational modifications of splicing factors, such as methylation (Me), phosphorylation (P), or ubiquitination (Ub), exemplified by GSK3368715, E7070, SRPKIN-1, and others.

Pladienolide B is a macrolide natural product isolated by researchers from Streptomyces platensis. Simultaneously, six analogs were also obtained from the same source, named pladienolide A and pladienolide C–G²¹⁷. Subsequent studies revealed that pladienolide B exhibited superior in vitro and in vivo anti-tumor activity compared to the other six homologs. In further research into its anti-tumor mechanisms, Kotake and colleagues synthesized chemical probes of pladienolide B and identified its potential target as SF3B1¹⁶⁵. Additionally, they reported that the IC₅₀ values of pladienolide B in WiDr and HeLa cells were 5.6 ± 0.5 and 0.86 nmol/L, respectively¹⁶⁵. Subsequently, an increasing number of studies demonstrated that pladienolide B possesses broad-spectrum anti-tumor activity, such as against gastric cancer, colorectal cancer, and breast cancer166, 167, 168. Notably, Kotake’s research team^169,170 discovered that the carbamate derivative of pladienolide B, E7107, exhibited superior physicochemical properties and stronger anti-tumor activity. For instance, the IC₅₀ of E7107 in the STO cell line for diffuse malignant peritoneal mesothelioma (DMPM) was 0.34 ± 0.08 nmol/L, and in PC3 cells for castration-resistant prostate cancer (CRPC), it was 0.5089 nmol/L. Importantly, E7107 entered Phase I clinical trials as a first-in-class compound in 2009; however, the clinical trial was unfortunately terminated due to vision loss observed in some patients, posing a potential risk of blindness²¹⁸. Driven by interest in targeting spliceosome drugs, researchers developed another derivative of pladienolide, H3B 8800 (now renamed RVT-2001), in 2017. This compound demonstrated promising results in clinical trials, showing significantly enhanced anti-tumor activity compared to E7107 without any adverse ocular side effects. Since there are no spliceosome-targeting drugs currently approved for clinical use, RVT-2001 was granted orphan drug designation by the FDA in 2017 for the treatment of acute myeloid leukemia (AML) and chronic monocytic leukemia (CMML). Additionally, RVT-2001 is currently undergoing Phase I clinical trials targeting patients with myelodysplastic syndromes carrying SF3B1 mutations¹⁷². Furthermore, it was also observed that RVT-2001 exhibits significant pharmacological activity against T-cell acute lymphoblastic leukemia (T-ALL). The IC₅₀ values in three T-ALL cell lines, namely JURKAT, CUTLL1, and DND41, range from 21 to 36 nmol/L¹⁷³.

Herboxidiene (GEX1A) is an effective plant-toxic polyketide compound isolated from the culture broth of Streptomyces sp. Researchers also isolated five novel compounds related to herboxidiene, namely GEX1Q1-Q5, from the same culture. All GEX1 compounds except GEX1Q3 and GEX1Q exhibited cytotoxicity against human tumor cell lines (A431, A549, and DLD-1) in vitro, with IC₅₀ values ranging from 0.0037 to 0.99 μmol/L. Among these, GEX1A showed the best activity, with an IC₅₀ of 3.7 nmol/L against the human epidermoid carcinoma A431 cell line¹⁷⁴. Hasegawa et al.²¹⁹ used a photoaffinity labeling method to identify SF3B1 as the target of GEX1A. By binding to SF3B1, GEX1A specifically impairs the function of SF3b, making it a novel splicing inhibitor. Recent studies have also revealed that GEX1A has inhibitory effects on leukemia initiation and progression. It induces a conformational change in MCL-1 protein isoforms in all types of leukemia cells, converting the anti-apoptotic MCL-1L subtype into the pro-apoptotic MCL-1S subtype, thereby promoting leukemia cell apoptosis¹⁷⁵.

Jerantinine A was isolated by Lim and colleagues^176,177 from Tabernaemontana corymbosa in 2008. During this study, a total of seven compounds were identified, named jerantinines A‒G (1‒7). Among these, jerantinines A‒E exhibited significant cytotoxic activity against human KB cells with IC₅₀ values < 1 μg/mL. Chung and colleagues¹⁷⁸ further investigated the anti-tumor mechanisms of Jerantinine A. Their research revealed that Jerantinine A exerts its anti-tumor activity by inhibiting tubulin polymerization and inducing G2/M cell cycle arrest. Through genome-wide shRNA library screening and proteomics analysis, the authors found that Jerantinine A targets the spliceosome by upregulating SF3B1 and SF3B3 in breast cancer cells, leading to a significant increase in unspliced precursor mRNAs. Further analyses demonstrated that Jerantinine A stabilizes endogenous SF3B1 protein in breast cancer cells and induces its dissociation from nuclear speckle-associated complexes. In conclusion, this study suggests that Jerantinine A may target SF3B1 and SF3B3 to inhibit the spliceosome and exert its anti-tumor activity.

Isoginkgetin is a dibenzopyrone compound extracted from Ginkgo biloba leaves and exhibits anti-tumor activity against multiple cancers. O’Brien et al.179, 180, 181 found that this compound is a general inhibitor that suppresses both major and minor spliceosome activity, inhibiting splicing in vitro and in vivo at similar micromolar concentrations. The mechanism likely involves preventing the stable recruitment of the U4/U5/U6 tri-snRNP complex, resulting in the accumulation of pre-spliceosome A complexes. However, other studies have suggested that Isoginkgetin actually inhibits transcriptional elongation by causing RNA polymerase to accumulate in promoter-proximal regions²²⁰.

Madrasin is a splicing regulator identified by the Pawellek group from 71,504 drug-like small molecules. It can inhibit pre-mRNA splicing both in vitro and in vivo and regulate the splicing of multiple pre-mRNAs in HeLa cells and HEK293 cells. Further mechanistic studies reveal that Madrasin modulates AS by interfering with the early stages of spliceosome assembly, causing the process to stall at the A complex stage. In vitro experiments show that Madrasin exhibits cytotoxicity at higher concentrations, while at lower concentrations, it induces cell cycle arrest and promotes specific reorganization of subnuclear protein localization^182,183.

Nakajima and colleagues identified three natural products, FR901463, FR901464, and FR901465, during their search for transcriptional regulators in soil microbes. These compounds were isolated from Pseudomonas strain 2663. Further studies revealed that these natural products exhibited antitumor activity. Specifically, the IC₅₀ values of FR901463 against tumor cell lines MCF-7, A549, HCT116, SW480, and P388 were 0.46, 0.35, 0.22, 0.4, and 0.82 ng/mL, respectively. For FR901464, the IC₅₀ values against these five tumor cell lines were 0.91, 0.66, 0.31, 0.51, and 1.69 ng/mL. Similarly, FR901465 exhibited IC₅₀ values of 0.59, 0.44, 0.34, 0.53, and 0.48 ng/mL against the same cell lines. Furthermore, FR901464 was discovered to boost transcriptional activity from the SV40 promoter and cause cell cycle arrest at the G₁ and G₂/M phases. Furthermore, FR901464 demonstrated in vivo efficacy by extending the lifespan of tumor-bearing mice^184,185,221. Owing to the distinctive pharmacological properties of FR901464, it has garnered significant attention from researchers such as Kaida et al. ^186,222, leading to an in-depth investigation into its mechanism of action. As a result, in 2007, they presented strong evidence that both FR901464 and its methylated form, Spliceostatin A, block pre-mRNA splicing by non-covalently attaching to the SF3b complex within the U2 snRNP. Furthermore, subsequent studies have revealed additional anti-tumor activities of SSA, including its efficacy against chronic lymphocytic leukemia (CLL) and prostate cancer^187,188. Notably, inspired by SSA’s potent anti-tumor activity and its ability to modulate AS, researchers have synthesized a series of SSA derivatives, including Spliceostatin B‒G. These derivatives have demonstrated significant anti-tumor activities in various tumor cell lines, such as H1975, N87, BT474, MDA-MB-DYT2, and MDA-MB-468²²³. In 2009, Lagisetti and colleagues¹⁸⁹ developed a series of FR901464 analogs. Among these, Sudemycin E (compound 5) demonstrated significant cytotoxicity in multiple tumor cell lines, such as JeKo-1 and JVM-2, PC-3, and WiDr, with IC₅₀ values of 0.12, 0.19, 2.2, and 0.91 μmol/L, respectively. Further studies revealed that, similar to SSA, Sudemycin E binds to the SF3B1 component of the U2 snRNP. This interaction interferes with U2 snRNPs’ capacity to preserve the H3K36me3 modification at genes being actively transcribed, causing alterations in chromatin structure that lead to chromatin condensation and the death of cancer cells¹⁹⁰. Additionally, Sudemycin E analogs, such as Sudemycin D6 and Sudemycin K, were also shown to exert antitumor effects through the regulation of AS^224,225. Similarly, in 2013, Gao and colleagues²²⁶ reported a FR901464 analog, Meayamycin B, which also targets the SF3B complex within the spliceosome. This compound exhibited multiple antitumor activities, including against head and neck cancer and DMPM^169,171.

NSC 194308 is a U2AF2‒RNA complex enhancer that inhibits pre-mRNA splicing by disrupting the assembly of the spliceosome. This disruption occurs through the recruitment of U2 snRNP to the branch point, facilitated by U2AF, ultimately leading to cell death in leukemia cell lines harboring spliceosome mutations¹⁹¹. NB-506, a glycosylated indolocarbazole derivative, effectively inhibits both the kinase and relaxing functions of topoisomerase I and has been recognized as a promising AS regulator with multiple anti-cancer properties. NB-506 regulates pre-mRNA splicing by inhibiting SR protein phosphorylation, thereby affecting gene expression, including Bcl-X, CD44, SC35, and Sty^192,193.

Clinical translation. Inspired by the promising pre-clinical data regarding tumor suppression through targeting the spliceosome, multiple relevant clinical trials have been initiated (Table 3). Among these trials, the most representative ones are E7107 and H3B-8800. As the first SF3b spliceosome complex inhibitor to enter clinical trials, E7107’s performance in clinical trials has drawn extensive attention. The findings of an open-label, single-arm, dose-escalating phase I clinical study (NCT00499499) indicated that the maximum tolerated dose of E7107 was 4.3 mg/m², and the most prevalent drug-related adverse events were nausea, vomiting, and diarrhea. Nevertheless, in this clinical trial, 2 patients suffered from vision loss, which led to the termination of the trial²²⁰. The results of another clinical trial (NCT00459823) assessing the safety, tolerability, pharmacokinetics, and pharmacodynamics of E7107 revealed that the maximum tolerated dose (MTD) was 4.0 mg/m^{2 227}. At this dose, the pharmacokinetic profile of E7107 was marked by a large volume of distribution, a plasma elimination half-life ranging from 5.3 to 15.1 h, and the overall drug exposure increased in a dose-dependent manner. The common side effects were predominantly gastrointestinal reactions. Still, 1 patient experienced grade 4 blurred vision. Overall, as a distinctive spliceosome inhibitor, E7107 has exhibited remarkable pharmacokinetic properties and dose-dependent pharmacodynamic responses in clinical trials. However, its potential toxic reactions, particularly the risk of vision loss, imply that more meticulous evaluation of its safety and effectiveness is required in future research²²⁷. As a derivative of E7107, H3B-8800 demonstrated more promising performance in clinical trials compared to E7107²²⁸. In a Phase I clinical trial for myeloid tumors (NCT02841540), H3B-8800 was employed to treat patients suffering from myelodysplastic syndromes, chronic myelomonocytic leukemia, and acute myeloid leukemia. Although neither complete nor partial remission was observed, it exhibited a certain degree of efficacy in patients with low-risk myelodysplastic syndromes. However, the red blood cell transfusion independence rate was relatively low, indicating that the dosage regimen requires further investigation²²⁸.

Table 3.

Clinical trials investigating small molecules that target alternative splicing in tumor.

Drug	Target	Phase	NCT number	Tumor type
E7107	SF3B	I	NCT00459823	Solid tumors
			NCT00499499
H3B-8800	SF3B	I	NCT02841540	AML, myelodysplastic syndrome, chronic myelomonocytic leukemia
E7070	RBM39	II	NCT01692197	AML, high-risk myelodysplastic syndrome, melanoma, breast cancer, colorectal cancer, gastric cancer
			NCT00014625
			NCT00165880
			NCT00165867
			NCT00165594
E7820		II	NCT05024994	Myeloid malignancies associated with splicing factor mutations
Tasisulam		I	NCT01284335	Solid tumors
		II	NCT00490451	Advanced unresectable or metastatic soft tissue sarcoma
		III	NCT01006252	Metastatic melanoma
GSK3326595	PRMT5	I	NCT03614728	AML, myelodysplastic syndrome, non-Hodgkin’s lymphoma
			NCT02783300
JNJ-64619178	PRMT5	I	NCT03573310	Solid tumors, non-Hodgkin’s lymphoma, myelodysplastic syndrome
PF-06939999	PRMT5	I	NCT03854227	Solid tumors
PRT543	PRMT5	I	NCT03886831	Solid tumors, hematologic malignancies
MRTX1719	PRMT5	I/II	NCT05245500	Solid tumors
AMG 193	PRMT5	I/II	NCT05094336	Solid tumors
GSK3368715	PRMT1	I	NCT03666988	Solid tumors
SM08502	CLK	I	NCT03355066	Solid tumors

Although both E7107 and H3B-8800 showed a certain level of efficacy and toxic reactions in clinical trials, the selective killing effect of H3B-8800 in specific cancer types and its potential synergistic therapeutic effect render it a promising treatment alternative.

3.2. Small molecules targeting splicing factors

As previously discussed, splicing factors, also referred to as trans-acting factors, play a pivotal role in the regulation of AS by interacting with sequence elements on pre-mRNA. Accumulating evidence suggests that splicing factors are intimately associated with tumorigenesis, development, and drug resistance. Consequently, researchers have developed a range of drugs targeting splicing factors to correct aberrant AS in tumors. One of the most notable splicing factors is RNA-binding motif protein 39 (RBM39). RBM39 not only binds to key splicing factors within specific complexes via UHM/ULM interactions, thereby promoting the recruitment of U2 snRNP, but it is also considered crucial for recognizing 3′ splice sites²²⁹. Additionally, studies have demonstrated that up to 20% of variable exons are regulated by RBM39 in conjunction with U2AF65²³⁰. Ina word, RBM39 is an essential splicing factor involved in regulating AS. Therefore, several compounds targeting RBM39 have been developed to address abnormal alternative splicing in tumors. For instance, Indisulam (E7070), a sulfonamide-based cell cycle inhibitor with potent anti-tumor activity (IC₅₀ of 0.11 μg/mL in HCT116 cells), was discovered by Yoshino et al.¹⁹⁵ in 1992. Given the outstanding anti-tumor activity of E7070, researchers were eager to identify its molecular target¹⁹⁴. Through relentless efforts, Han and colleagues finally revealed in 2017 using CRISPR/Cas9 technology that Indisulam promotes the recruitment of RBM39, a splicing factor belonging to the RNA-binding protein family involved in transcriptional co-regulation and AS, to the CUL4-DCAF15 E3 ubiquitin ligase complex. This leads to the polyubiquitination and proteasomal degradation of RBM39, causing disordered RNA splicing and ultimately killing cancer cells¹⁹⁶. Therefore, Indisulam is also considered a degrader of the splicing factor RBM39. It is noteworthy that E7820, Tasisulam, and chloroquinoxaline sulfonamide (CQS), which belong to the aryl-sulfonamide class and are structural analogs of E7070, have been demonstrated to exert their antitumor effects via RBM39 degradation197, 198, 199, 200.

Clinical translation. The strategy of treating tumors by targeting splicing factors has been proven feasible through numerous positive clinical trials (Table 3). Among these trials, the most representative drugs are E7070, E7820, and Tasisulam. A common characteristic of these drugs is their ability to degrade the splicing factor RBM39. Although in a randomized, open-label phase III clinical trial (NCT01006252) for patients with metastatic melanoma, the efficacy of Tasisulam was not significantly superior to that of paclitaxel, leading to the early termination of the trial²³¹. Nevertheless, the unique anti-cancer mechanism and potential clinical application value of Tasisulam should not be overlooked. In future research, dose management and toxicity control need to be optimized to guarantee its safety and effectiveness in clinical applications²³¹. In addition, E7070 yielded positive results in a phase II clinical trial (NCT01692197). This study assessed the efficacy of indisulam in combination with idarubicin and cytarabine in patients with relapsed/refractory acute myeloid leukemia (AML) and high-risk myelodysplastic syndrome (MDS)²³². The findings indicated that among the 31 patients who received the indisulam-combined chemotherapy regimen, 11 (35%) achieved remission, and the median duration of remission was 5.3 months. The estimated 1-year overall survival rate of the remitters was 51%, significantly higher than the 8% of the non-remitters²³². In a phase II clinical trial (NCT05024994) evaluating the efficacy of E7820 in relapsed and/or refractory (R/R) myelodysplastic syndrome (MDS) and acute myeloid leukemia (AML), despite its limited clinical efficacy, the favorable safety profile and oral convenience of E7820 showcased its unique advantages in combination therapy²³³.

3.3. Small molecules targeting the post-translational modifications of splicing factors

Given the complexity of RNA splicing as a multifaceted biological process regulated at various levels, several novel drugs targeting the splicing regulatory mechanisms have been developed successively. For example, inhibiting the post-transcriptional modification of splicing factors has been shown to indirectly interfere with RNA splicing by disrupting spliceosome assembly and reducing splicing catalytic efficiency. Small molecule inhibitors of PRMT1 and PRMT5 is prime examples²³⁴. These drugs can significantly impair RNA splicing by inhibiting arginine dimethylation mediated by type I or type II protein arginine N-methyltransferases (PRMTs)²³⁵. GSK3368715 is a reversible SAM non-competitive inhibitor of PRMT1, which directly binds to a peptide site adjacent to the SAM-binding pocket. In cellular experiments, GSK3368715 alone induces changes in the methylation state of intracellular protein substrates, transitioning from ADMA to MMA and SDMA. When used in combination with a PRMT5 inhibitor, it reduces the accumulation of MMA and SDMA caused by inhibition of type I PRMTs and has significant effects on AS, leading to decreased activity of splicing regulatory factors, transcriptional activation, and cell growth suppression. GSK3368715 was tested on 249 cancer cell lines belonging to 12 tumor types and primary samples obtained from patients with diffuse large B cell lymphoma (DLBCL). GSK3368715 inhibited over 50% of the cancer cells, and the inhibition rate in samples from DLBCL patients exceeded 80%. Moreover, GSK3368715 exhibited good tolerability. In xenograft mouse models of DLBCL, clear cell renal cell carcinoma, triple-negative breast cancer, and pancreatic adenocarcinoma, it was able to significantly inhibit tumor growth²⁰¹.

PRMT5 has the ability to methylate key proteins of the spliceosome, and its dysregulation is linked to the clinical characteristics of various cancers, such as lymphoma, lung cancer, and breast cancer. JNJ-64619178, GSK3326595, and PF06939999 are all PRMT5 inhibitors, which exhibit significant anti-tumor activities against a wide range of tumors. Mechanistically, the anti-tumor activities of these three inhibitors are all associated with the inhibition of spliceosome activity202, 203, 204.

Of course, apart from the methylation modification of splicing factors, many splicing factors, particularly SR proteins, can also undergo phosphorylation modification. The activity and localization of these phosphorylated splicing factors will change correspondingly. Importantly, these changes are crucial for their splicing activity. Therefore, inhibiting the kinases that regulate these phosphorylation modifications might be a viable strategy to reduce the activity of oncogenic SR proteins. Serine-rich protein kinase (SRPK) and CDC-like kinase (CLK) are currently the two most well-reported kinases for phosphorylating splicing factors in the regulation of AS, and corresponding inhibitors have been discovered one after another. For example, inhibitors of CLKs, such as TG003, T025, SM04690, and SM08502, have obvious tumor-inhibitory effects in multiple tumors205, 206, 207, 208, 209, 210^,236. In addition, inhibitors of SRPKs, including SRPIN340, SRPKIN-1, and SPHINX31, have been demonstrated to decrease the phosphorylation levels of diverse SR proteins and exert anti-angiogenic effects via SRSF1-mediated AS of VEGF211, 212, 213.

Clinical translation. As mentioned above, disrupting RNA splicing by inhibiting the post-translational modification of splicing factors, which subsequently affects spliceosome assembly and splicing catalytic efficiency, is also a tumor treatment strategy that has been proven effective in clinical trials (Table 3). For instance, specific drugs can inhibit PRMTs-mediated methylation, thus suppressing RNA splicing. Among these drugs, the representative ones are inhibitors of PRMT1 and PRMT5. Regarding small molecule inhibitors of PRMT5, five compounds have advanced to Phase I clinical trials, specifically JNJ-64619178 (NCT03573310), PF-06939999 (NCT03854227), PRT543 (NCT03886831), MRTX1719 (NCT05245500), and AMG 193 (NCT05094336). Among these, the clinical trial results of JNJ-64619178 demonstrate that it exhibits preliminary toxicity and efficacy against adenoid cystic carcinoma (ACC) and other solid tumors, and thrombocytopenia is the sole dose-limiting toxicity identified thus far²³⁷. MRTX1719, a synthetic lethal inhibitor of the PRMT5 complex, is currently undergoing Phase I/II clinical trials for MTAP-deficient solid tumors. This study has shown anti-tumor activity in lung cancer, pancreatic cancer, and mesothelioma^238,239. In contrast, the PRMT1 inhibitor GSK3368715 entered phase I clinical trials for advanced solid tumors but was prematurely terminated due to an unexpectedly high incidence of treatment-related adverse events, limited target binding at lower doses, and lack of observed clinical efficacy²⁴⁰. It is noteworthy that the significant roles of certain sulfonamide compounds, which function as molecular glues between RBM39 and CUL4-DCAF15 E3 ligase, in regulating AS have also been gradually clarified. Significantly, these compounds have demonstrated excellent safety in clinical trials and exhibit certain anti-tumor effects, such as E7070 (NCT00014625, NCT01692197), E7820 (NCT05024994), and Tasisulam (NCT01006252, NCT00490451, NCT01284335). Among them, Tasisulam has advanced to Phase III clinical trials (for the treatment of metastatic melanoma), yet the study was prematurely halted due to safety concerns. In addition, SM08502, a novel CLK inhibitor, has been demonstrated to possess potent antitumor effects in gastrointestinal cancer models. Mechanistically, SM08502 inhibits the Wnt pathway by inducing intron retention of DVL2, ERBB2, LPR5, and TCF7, along with exon skipping of TCF7L2 and LEF1. At present, SM08502 has advanced into Phase I clinical trials. One study is appraising its efficacy when combined with hormones or chemotherapy (NCT05084859), and another is evaluating its safety and pharmacokinetics as a monotherapy (NCT03355066).

3.4. Other small molecules

Amiloride, a potassium-sparing diuretic approved by the FDA for treating edema and hypokalemia associated with hypertension, heart failure, and cirrhosis, has been shown to enhance the radiosensitivity of glioblastoma multiforme (GBM) 8401 cells. This effect is mediated through the regulation of Akt phosphorylation and alternative splicing of APAF1 (APAF1-LN, -S), particularly in cells subjected to amiloride treatment followed by ionizing radiation (IR)²¹⁴. These findings suggest that Amiloride may serve as a potential regulator of AS. Research has also demonstrated that Caffeine can modulate the splicing of p53 in HeLa cells. Specifically, Caffeine reduces the expression of p53α via the splicing factor SRSF3 and promotes the expression of p53β, which contains an alternatively spliced C-terminal region. This modulation regulates cell cycle arrest and apoptosis²¹⁵. Furthermore, Digoxin has been identified to influence the transformation of p53α to p53β through SRSF3, thereby regulating G2/M phase arrest, DNA damage, and apoptosis in HeLa cells. These results indicate that downregulating splicing factors such as SRSF3 to alter cell cycle progression and induce cell death may represent a promising target for drug repositioning of Digoxin²¹⁶.

Clinical translation. Although drugs such as Amiloride and Digoxin have not been evaluated in clinical trials for their impact on AS in tumors, these preclinical data nonetheless indicate that they represent potential agents for targeting AS. Future clinical evaluation of their AS-modulatory mechanisms holds promise for expanding new clinical indications of these approved drugs or optimizing combination regimens.

3.5. Antisense oligonucleotides

ASO can specifically bind to certain regions of pre-mRNA, which influences the recognition and binding of splicing factors and subsequently modifies the splicing pattern of pre-mRNA²⁴¹. Researchers can adopt this approach to transform anti-apoptotic splicing subtypes into pro-apoptotic splicing subtypes, thereby attaining the anti-tumor objective. For instance, they can change the anti-apoptotic BCL-XL protein subtype into the pro-apoptotic BCL-XS subtype and convert the tumor-associated STAT3α protein subtype into the pro-apoptotic STAT3β subtype242, 243, 244, 245, 246, 247. In addition, ASO is also capable of transforming carcinogenic splicing subtypes into tumor-suppressive splicing subtypes, consequently achieving anti-tumor effects. For instance, ASO can induce the conversion of the carcinogenic Mnk2b into the tumor-suppressive Mnk2a, and trigger the shift from the expression of the carcinogenic subtype of the HER4 protein to the expression of the tumor-suppressive subtype^248,249. Similarly, researchers have utilized ASOs to alter the expression of the key PKM2 protein subtype involved in aerobic glycolysis and tumor growth to the expression of the PKM1 protein subtype associated with cancer cell apoptosis, thereby achieving anti-hepatocellular carcinoma effects²⁵⁰. Furthermore, an increasing number of targets for ASO therapy have been reported, including regulators of p53 (such as MDM2 and MDM4), DNA damage response proteins (such as BRCA2 and ATM), and chromatin remodeling and transcription factors (such as BRD9 and ERG)^234,251, 252, 253, 254, 255, 256, 257, 258, 259, 260, 261, 262, 263, 264, 265, 266, 267, 268 (Table 4). Despite the challenges associated with in vivo delivery of ASOs, an increasing body of evidence demonstrates that ASOs capable of correcting cancer-related AS events exhibit promising anti-cancer effects in both cell lines and animal models. In conclusion, these studies underscore the potential of targeting alternative splicing through ASOs as a significant direction for tumor therapy.

Table 4.

Antisense oligonucleotides designed to target alternative splicing events in tumors.

Pre-mRNA target	The splicing outcomes following ASO intervention	Tumor type	In vitro experimental model	In vivo experimental model	Ref.
EZH2	Poison exon skipping	Leukaemia	K562 cell	/	191
STAT3	Induce exon 23 skipping	Breast carcinoma	MDA-MB-435s cell	Xenograft mouse model	245
BCLx	Induce the conversion of the anti-apoptotic Bcl-xL isoform to the pro-apoptotic Bcl-xS isoform	Breast carcinoma Glioblastoma Lung adenocarcinoma Prostate adenocarcinoma	MCF-7 cell PC-3 cell A549 cell MelJuSo cell	/	246,247
MKNK2	Switches Mnk2b to Mnk2a	Glioblastoma	U87MG cell	Xenograft mouse model	248
ERBB4	Induce Exon 26 skipping	Breast carcinoma	MCF7 and T47D cells	Xenograft mouse model	249
PKM2	Switches PKM2 to PKM1	Hepatocellular Carcinoma	Huh7 cell	Xenograft mouse model	250
ATM	Blocks ATM exon 28 inclusion	Leukaemia	VAVY cell	/	251
BIM	Promote inclusion of BIM exon 4	Leukaemia	K562 and KCL22 cell lines	/	252
BRCA2	Block the cryptic exon inclusion	Ovarian cancer	HeLa cell	/	253
BCL2L12	Promotes exon 3 skipping	Ovarian cancer	A2780 cell	Xenograft mouse model	254
BRD9	Induce Exon 14a skipping	Uveal melanoma	MEL202 cell	Xenograft mouse model	255
ERG	Induce exon 4 skipping	Prostate cancer	MG63 and VCaP cells	Xenograft mouse model	256
FGFR1	Promote Exon α inclusion	Glioblastoma	SNB-19 cell	/	257
GLDC	Induce exon 7 skipping	Non-small-cell lung carcinoma	A549 cell	Xenograft mouse model	258
IL5R	Induce exon 5 skipping	Leukaemia	TF-1 cell	/	259
MCL1	Induce exon 2 skipping	Basal Cell Carcinoma	BCC and AGS cells	/	260
MDM2	Induce exon 4 skipping	Uterine corpus endometrial carcinoma	JAR cell	/	261
MDM4	Induce exon 6 skipping	Melanoma	A375 cell	Patient-derived xenograft mouse model	262
MSTR1	Induce exon 11 skipping	Breast carcinoma Stomach adenocarcinoma	MDA-MB-435S and TOIII cells	/	263
PD-L1	Induce exon 3 skipping	Oral cancer	CAL 27 and SCC-9 cells	/	264
SLAMF6	Enhance the Expression of SLAMF6Δ17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42,43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65	Melanoma	Jurkat cell	/	265
SRSF3	Induce poison exon inclusion	Oral squamous cell carcinoma	CAL 27 and FaDu cells	/	266
TRA2Β	Induce poison exon inclusion	Breast carcinoma	MDA-MB231 and SUM159 cells	/	267
BAX	Promote exon 2 inclusion	Ovarian cancer	HEY and OV90 cells	/	268

Clinical translation. Targeting AS dysregulation events in tumors via ASOs is regarded as a highly promising therapeutic strategy. Presently, clinical trials of ASO therapy are being vigorously conducted and have achieved varying degrees of progress. The findings of a clinical trial assessing the efficacy of the STAT3 ASO AZD9150 in patients with relapsed or refractory lymphoma indicated that the drug exhibited good tolerability in some patients, and certain efficacy was noted in patients with diffuse large B-cell lymphoma, including 2 cases of complete remission (median duration of remission 10.7 months) and 2 cases of partial remission. Nevertheless, there were also some drug-related adverse reactions, such as elevated transaminases, fatigue, and thrombocytopenia²⁶⁹. Overall, ASOs offer new hope for tumor treatment, yet there are also some challenges. For instance, the delivery efficiency of ASOs remains low. In the future, novel delivery strategies need to be explored.

4. Conclusions

Alternative splicing (AS) is a post-transcriptional regulatory mechanism by which introns are removed from pre-mRNA and exons are spliced together under the action of the spliceosome³. Owing to the selection of different splicing sites, multiple mRNA isoforms can be generated via this mechanism, thereby enriching the diversity of the proteome. In tumors, this mechanism is frequently found to be abnormal, and the generated mRNA isoforms contribute to tumor progression¹². As previously mentioned, dysregulation of RNA splicing can enable tumors to generate self-sufficient signals, become insensitive to growth inhibitory signals, achieve immune escape, and evade apoptosis, thus promoting tumor occurrence, development, and drug resistance to therapeutic agents^12,21. However, this precisely demonstrates that targeting AS is a highly promising strategy for tumor treatment. Significantly, drugs targeting AS, such as small molecule inhibitors and antisense oligonucleotides (ASOs), have also yielded positive results in clinical studies.

Although there are currently no drugs targeting AS for tumor treatment that have been approved for marketing, the drug with the fastest research progress is still in Phase III clinical trials²³¹. The reasons for this situation are as follows: 1) The biological complexity and heterogeneity of AS in tumors. On one hand, AS can generate multiple mRNA isoforms, which may have the same or opposite functions. There is also a possibility of mutual transformation between two splicing subtypes. Therefore, this poses a huge challenge to the development of drugs targeting AS. Moreover, the regulation of AS involves the interaction of multiple factors, such as splicing factors. When one splicing factor is inhibited, other splicing factors may compensatorily increase their expression or change their functions to maintain the abnormal splicing pattern. On the other hand, there are also significant differences in the splicing profiles among different tumor patients. 2) It is difficult to reproduce the results of pre-clinical studies in clinical trials. This is mainly because pre-clinical research models, such as cell lines and patient-derived xenograft models, have fundamental defects in simulating AS abnormalities in human tumors. Additionally, the effective targets verified in pre-clinical studies may show off-target effects in clinical trials, such as ASOs drugs. 3) There are technical bottlenecks in drug delivery, and this dilemma is particularly prominent in the delivery of ASO drugs. 4) The drugs have serious side effects. For instance, E7107 has severe visual toxicity.

Therefore, the clinical translation issues of drugs targeting AS for tumor treatment can be addressed from the following aspects. Firstly, in the basic research of AS, by integrating single-cell sequencing, spatial transcriptomics, etc., with artificial intelligence, machine learning, and deep learning technologies, the splicing profiles and key mechanisms in tumor patients can be comprehensively analyzed from multiple dimensions. Secondly, in pre-clinical studies, more clinically relevant models for tumor patients, such as organoids, should be employed. Concerning the off-target problem of ASO drugs, sequence-specific targeting of tumor cryptic splice sites can be utilized to selectively degrade only the pro-cancer isoforms without affecting normal transcripts. Moreover, the innovation of the ASO delivery system is crucial for promoting its application in the treatment of solid tumors. Extracellular vesicles like exosomes and migrasomes can be harnessed for specific delivery^270,271. Finally, in terms of mitigating the toxic and side effects of drugs, it can be accomplished through drug design. Take H3B-8800, the first oral small-molecule SF3B1 (SF3B1 is a highly promising target for tumor treatment) modulator to enter clinical trials, as an example²²⁸. Its main adverse reaction is bone marrow suppression. The primary cause of this severe adverse reaction is the lack of tissue specificity of H3B-8800. Subsequent studies have demonstrated that this drug offers obvious benefits and milder toxic and side effects to patients carrying the SF3B1 K700E mutation^228,272. Therefore, based on the characteristic of SF3B1 mutation, AI-assisted drug design can be applied to explore allosteric pockets and develop highly selective small-molecule inhibitors, thus reducing the impact on the wild-type SF3B1 in normal cells. In general, as people’s understanding of the cancer-promoting mechanism of AS in tumors continues to improve and technology develops rapidly, the difficulties currently encountered in the clinical translation of targeted AS for tumor treatment will gradually be overcome.

This review systematically explores the multifaceted roles of AS in tumorigenesis and development, clarifies its underlying mechanisms, and highlights the latest progress in the preclinical and clinical development of AS-targeted therapies. Notably, based on the existing evidence, this review prospectively elaborates on the potential future research directions of AS in the field of oncology, aiming to promote further breakthroughs in precision oncology.

Author contributions

Lei Hu, Xiaofang Wang and Bo Fan wrote the manuscript, created the figures and tables. Minru Liao, Yingying Lu, Hongyao Li, Qian Zhao and Yajun Zhong polished the language. Yingying Lu, Xiangyu Fu, Huiping Wang, Yajun Zhong and Qian Zhao proofread and revised the manuscript. Cheng Du: Conceptualization, Supervision. Bo Han: Conceptualization, Funding acquisition. Leilei Fu: Conceptualization, Supervision, Funding acquisition. All authors approved the final manuscript.

Conflicts of interest

The authors have declared no conflict of interest.