Abstract

   SF3B1 gene mutations are prevalent in myelodysplastic syndrome (MDS)
   and define a distinct disease subtype. These mutations are associated
   with dysregulated genes and pathways, offering potential for novel
   therapeutic approaches. However, the aberrant mRNA alternative splicing
   landscape in SF3B1-deficient MDS cells remains underexplored. In this
   study, we investigated the influence of SF3B1 gene alterations on the
   pre-mRNA splicing landscape in MDS cells using transcriptomic data from
   two independent MDS cohorts. we identified over 5000 significant
   differential alternative splicing events associated with SF3B1
   mutation. This work corroborates previous studies, showing significant
   enrichment of MYC activity and heme metabolism in SF3B1 mutant cells. A
   key novel finding of this study is the identification of a gene
   expression signature driven by SF3B1 mutations, centered on protein
   post-translational modifications. Notably, we discovered aberrant
   alternative splicing of the tumor suppressor gene UBA7, leading to
   significantly reduced gene expression. This dysregulation implicates
   UBA7 as a critical player in MDS pathogenesis. Importantly, the
   clinical relevance of this finding is underscored by the observation
   that low UBA7 gene expression was associated with poor overall survival
   in chronic lymphocytic leukemia (CLL), another hematological malignancy
   with frequent SF3B1 mutations. Furthermore, a similar association
   between low UBA7 gene expression and poor survival outcomes was
   observed across multiple tumor types in the TCGA database, highlighting
   the broader implications of UBA7 dysregulation in cancer biology. These
   findings provide new insights into the mechanisms by which SF3B1
   mutations reshape the pre-mRNA splicing landscape and drive disease
   pathogenesis in MDS. Furthermore, they underscore the potential of UBA7
   as a biomarker to stratify SF3B1-mutant MDS and CLL patients, offering
   a refined approach for risk assessment and highlighting opportunities
   for targeted therapeutic interventions.

   Subject terms: Myelodysplastic syndrome, Cancer genomics

Introduction

   Myelodysplastic syndrome (MDS) is defined as a hematopoietic stem cells
   defect characterized by inefficient blood cell production and abnormal
   cell morphology. Clinically, it presents as peripheral blood low cell
   counts and carries a risk of progressing to acute myelogenous leukemia
   (AML). MDS affects 3 to 4 people per 100,000 in the United States, with
   its occurrence rising as individuals age^[34]1. Research over the past
   twenty years has shown that MDS is a diverse group of malignancies
   resulting from hematopoietic stem cell abnormal function, immune system
   dysregulation, altered apoptosis, and various aberrant genomic
   alternations^[35]1–[36]4.

   In recent years, significant progress has been made in characterizing
   the genetics landscape of MDS. Research has revealed that somatic
   mutations affecting genes with roles in the spliceosome occur in about
   50% of MDS patients and that there is a distinct correlation between
   the type of spliceosome mutations and MDS disease
   stratification^[37]5–[38]7. SF3B1 gene alterations represent the most
   prevalent pre-mRNA splicing changes in MDS. The status of SF3B1 gene is
   a key indicator for the presence of ring sideroblasts (RS), occurring
   in the majority of MDS patients who have ring sideroblasts
   (MDS-RS)^[39]8,[40]9.

   Genetic alterations in splicing factors are known to be early events in
   tumor development, and some researchers suggest that these mutations
   are not the primary drivers of the disease but rather add more fitness
   to tumor cells by facilitating pro-cancer changes^[41]10–[42]12.
   Functional studies have shown that these gene changes are prevalent in
   signaling pathways and cellular processes that are dysregulated to
   trigger the acceleration of the disease, significantly influencing the
   clinical characteristics of MDS^[43]13–[44]15. Moreover, research has
   identified specific genes and biological pathways that are prone to
   dysregulation in association with SF3B1 mutations, presenting promising
   new therapeutic strategies^[45]15.

   The pre-mRNA splicing landscape changes in SF3B1 mutant MDS cells are
   not fully characterized due to the low number of reported cases of
   mutant and wild-type SF3B1 MDS samples and/or the use of a population
   of cells that do not necessarily represent SF3B1 deficient
   cells^[46]16–[47]18.

   Here, we performed a thorough and systematic analysis to assess the
   consequence of SF3B1 alterations on the splicing landscape in MDS cells
   using larger cohorts. Key cellular processes and dysregulated pathways
   that supported prior findings were identified in CD34 + MDS cells.
   Genes that were prone to aberrant alternative splicing events were
   identified, and their gene expression and clinical relevance were
   assessed. Findings in this work were also evaluated in other
   independent cohorts with SF3B1 mutation occurrences, such as chronic
   lymphocytic leukemia (CLL). The clinical relevance was also evaluated
   in other tumor contexts in the cancer genome atlas project (TCGA).

Methods

Patient cohorts

   This research utilized open-access transcriptomic data from multiple
   independent MDS studies. The inclusion criteria were consistent initial
   cellular samples (CD34), samples from untreated MDS patients, and a
   paired-end NGS sequencing configuration to reduce potential technical
   noise. Two MDS studies met these criteria, and their transcriptomic
   data was found in public databases (GEO). Raw transcriptomic data for
   all two studies (n = 94) was retrieved from the GSE database. These
   studies are [48]GSE114922 and [49]GSE63569 collected in 2015, 2018,
   respectively^[50]19,[51]20. Detailed RNA-seq sample information is
   provided in supplementary table S1 (Supplement S1). The data was
   accessed from the NCBI GEO database on January 19, 2024, using SRA
   tools. CLL (CLL-ES cohort N = 304) gene expression count matrix, and
   CLL patient’s clinical information was obtained from the International
   Cancer Genome Consortium (ICGC) open-access public repository at
   ([52]https://dcc.icgc.org). The author did not have access to
   information that could recognize individual patients during or after
   data collection, as the data was already anonymized in the GEO and ICGC
   databases.

Raw data processing and mapping

   Raw sequencing data underwent quality control evaluation before being
   aligned to the hg38 reference genome using the STAR using 2-pass
   configuration to reliably and precisely identify novel splice junctions
   for differential splicing analysis^[53]21. The resulting BAM files were
   filtered, and duplicates were removed using Picard. Subsequently, read
   groups (RG) tags were added to sequencing reads using the tool Genome
   Analysis Toolkit (GATK)^[54]22. Sequencing reads within bam files
   containing (Ns) in their cigar string were split as they span splicing
   events in the data using the GATK tool.

Percent-spliced-in (PSI) quantification and differential ASE analysis

   Alternative splicing events (ASEs), such as inclusion/skipping of
   cassette exons (skipped exons, SE), alternative 5′/3′ splice site usage
   (A5SS/A3SS), intron retention (IR), are quantified using the
   percent-spliced-in (PSI) parameter, calculated by equation^[55]23. The
   formula for PSI calculation is as follows:
   graphic file with name 41598_2025_95738_Article_Equa.gif

   ASE isoforms are identified based on the region size that is
   alternatively spliced, with the included isoform having a larger region
   than the excluded isoform. For mutually exclusive events such as the
   inclusion of mutually exclusive exons (MXE) and alternative first or
   last exons (AFE/ALE), the included isoform is defined by the shorter
   first intron. Isoform abundances are estimated using junction read
   counts, which do not require length normalization like exon
   body-aligned reads. For intron retention (IR) events, the included
   isoform’s abundance is determined using the mean depth of coverage
   across the intron, trimmed by 30%, as per IRFinder’s method^[56]24.

   We identified two independent MDS cohorts with SF3B1 mutations (N = 36)
   and those with wild-type SF3B1 (N = 58). Differential splicing analysis
   was conducted by comparing SF3B1-mutant MDS samples to their wild-type
   counterparts. Differential ASE analysis was applied, utilizing a
   generalized linear model (GLM) based approach using the edgeR
   package^[57]25. In this analysis, estimated included and excluded
   isoform coverage matrices where columns representing samples and rows
   representing AS events were used as input for differential ASE
   analysis. The SF3B1 mutation status and the study cohort were described
   as factors in the GLM equation. AS events were defined as significant
   if they passed the significance cut off ( FDR < 0.05 and
   LogFC > −/ + 1.5).

Differential gene expression

   Gene expression count matrix was produced using the featureCounts
   tool^[58]26 with the GTF file from Gencode (GENCODE 46 (May
   2024))^[59]27. The resulting count matrix was then used for
   differential gene expression analysis. DESeq2^[60]28 was used to find
   differentially expressed genes (DEGs) by contrasting SF3B1-mutant
   against SF3B1-wt MDS samples, considering a gene significant if it had
   an adjusted p-value below 0.05 and more than 50% fold change. This
   stringent cut-off produced lists of DEGs that were either upregulated
   or downregulated in MDS patients. Subsequently, Gene Set Enrichment
   Analysis (GSEA)^[61]29 was applied to the DEGs, identifying significant
   hallmarks. All plots were produced using R (Version 4.4.1).

Survival analysis for TCGA and ICGA data

   To evaluate the impact of UBA7 gene expression in various tumors, we
   analyzed the TCGA cohort’s transcriptomic data and related clinical
   information. Normalized gene expression metrics for each cohort were
   obtained from the UCSC Xena project^[62]30. Cancer samples were
   categorized as High or Low based on the median expression value of UBA7
   within each cohort. Survival curves were constructed using the
   survminer package and deemed significant if the log-rank test p-value
   was below 0.05. For CLL-ES gene expression data, the count data were
   normalized using the DESeq2 package and generated survival plots
   following the same methodology as with the TCGA data.

Results

   MDS cohorts raw data that met the inclusion criteria (methods) were
   identified and downloaded from the GEO database. The raw data were then
   supplied to the analysis pipeline. MDS samples (N = 94) with confirmed
   SF3B1 genotype status were then identified and used for splicing
   analysis. The alternative splicing (AS) events were characterized, and
   splicing profiles were generated in SF3B1-mutant vs. SF3B1-wt in the
   transcriptome of MDS samples. Alternative splicing count estimation at
   the event level utilized both exon body and junction reads. A
   percentage spliced-in (PSI) was calculated for each splicing event,
   representing the proportion of reads supporting exon inclusion compared
   to the total number of reads^[63]31. Using a threshold of FDR < 0.05
   and a log2 fold change above 1.5, a total of 5679 significantly
   differential AS events were identified (Supplement S2).

   These significant AS events were distinct and showed a differential
   pattern in the SF3B1-mutant vs. SF3B1-wt comparison (Fig. [64]1A). To
   have a holistic understanding of the distribution of these significant
   AS events, a volcano plot was generated, which shows the majority of AS
   events were increased in SF3B1-mutant MDS samples as expected since the
   splicing machinery is defective in these samples (Fig. [65]1B). The
   majority of these significant AS events were classified as alternative
   3′ splice site usage (A3SS), which represents more than 60 percent,
   followed by skipped exons (SE), alternative 5′ splice site
   usage (A5SS), alternative first exons (AFE), intron retention (IR),
   alternative last exons (ALE), and inclusion of mutually exclusive exons
   (MXE) (Fig. [66]2A). Gene set enrichment analysis (GSEA) on
   differentially expressed genes identified an increased activity of MYC
   and heme metabolism pathways (Fig. [67]2B), which is consistent with
   previous reports^[68]32–[69]34 (Supplement 3).

Fig. 1.

   [70]Fig. 1
   [71]Open in a new tab

   SF3B1 mutant MDS cells show a distinctive spliceosome landscape. (A)
   Unsupervised hierarchical clustering of significant AS events in
   SF3B1-mutant and SF3B1-WT MDS samples. AS events, PSI normalized values
   (Z-score) across samples are represented in rows. As events were
   considered significant if pass the predefined threshold (FDR < 0.05 and
   LogFC > −/ + 1.5). (B) Volcano plots showing the significance and
   effect size of identified AS events in SF3B1-mutant and SF3B1-WT MDS
   samples. Colors represent the type of AS event. (A3SS) alternative 3′
   splice site usage, (SE) skipped exons, (A5SS) alternative 5′ splice
   site usage, (AFE) alternative first exons, (IR) intron retention, (ALE)
   alternative last exons, (NS) not significant.

Fig. 2.

   [72]Fig. 2
   [73]Open in a new tab

   Splicing event distribution and pathway enrichment analysis in
   SF3B1-mutant MDS. (A) Bar plot depicting the distribution of
   alternative splicing (AS) event types identified in SF3B1-mutant MDS
   samples. The majority of events were annotated as alternative 3’ splice
   site (A3SS), followed by skipped exons (SE), alternative 5’ splice site
   (A5SS), alternative first exon (AFE), intron retention (IR),
   alternative last exon (ALE), and mutually exclusive exons (MXE).
   Percentages represent the proportion of each splicing event type
   relative to the total number of identified events. (B) Bar plot showing
   the pathway enrichment analysis based on Hallmark gene sets in
   SF3B1-mutant MDS samples. Pathways with positive normalized enrichment
   scores (NES) include MYC targets, oxidative phosphorylation, and heme
   metabolism, indicating significant upregulation. Pathways with negative
   NES, such as apoptosis, TGF-beta signaling, and TNF-alpha signaling,
   were downregulated. Results highlight the dysregulation of key pathways
   associated with SF3B1 mutations.

   To gain biological insights into the impact of genes exhibiting
   differential alternative splicing events (ASE) with an FDR < 0.05, we
   conducted a gene pathway enrichment analysis using the Reactome
   database^[74]35. As anticipated, genes associated with the mRNA
   splicing machinery and translation were predominantly affected by SF3B1
   mutations in MDS samples (Fig. [75]3A). Intriguingly, pathways related
   to protein post-translational modifications, including
   ubiquitin-mediated protein degradation, SUMOylation of ubiquitinylation
   proteins, and deubiquitination, were also significantly enriched.
   Additionally, we observed enrichment in pathways involved in interferon
   signaling and ISG15 antiviral signaling, highlighting the broader
   regulatory disruptions caused by defective splicing machinery.

Fig. 3.

   [76]Fig. 3
   [77]Open in a new tab

   SF3B1 mutations alter splicing and gene expression, highlighting
   pathways in protein modification and immune response. (A) Reactome
   pathway enrichment analysis of differentially spliced events
   (ΔPSI > 0.1, FDR < 0.05). The results highlight significant enrichment
   in pathways such as mRNA splicing, eukaryotic translation initiation,
   and post-translational protein modifications, underscoring the broad
   impact of SF3B1 mutations on splicing and cellular processes. (B) Venn
   diagram illustrating the intersection of differentially expressed genes
   (DGE) (adjusted p-value < 0.05 and > 50% fold change) and
   differentially spliced events (DSE), identifying 83 genes that exhibit
   both significant changes in expression and splicing, further
   implicating their potential roles in SF3B1-driven pathogenesis. (C)
   Pathway analysis of the intersecting genes (DGE and DSE) reveals their
   involvement in key processes, including post-translational protein
   modifications, antiviral mechanisms by interferon-stimulated genes, and
   protein metabolism. Sankey plot highlights representative genes such as
   UBA7, SNCA, and TUBA1A, mapping their contributions to the associated
   pathways. The dot plot on the right shows the gene ratio, statistical
   significance (−log10 p-value), and enrichment magnitude for the top
   pathways.

   To further refine the analysis and identify high-confidence genes, we
   intersected differential ASE genes (ΔPSI > 0.1, FDR < 0.05) with genes
   showing significant differential expression (adjusted p-value < 0.05
   and > 50% fold change) (Supplement 4). This filtering process yielded
   611 significant alternative splicing events (DSE) and 3,390
   differentially expressed genes (DGE). The overlap between these two
   categories identified 83 genes that were both differentially spliced
   and differentially expressed (Fig. [78]3B). Reactome pathway enrichment
   analysis of this subset revealed predominant involvement in protein
   post-translational modifications, antiviral mechanisms mediated by
   interferon-stimulated genes, protein metabolism, and interferon
   signaling pathways (Fig. [79]3C).

   Notably, these 83 genes exhibited a distinct expression profile that
   differentiated SF3B1-mutant MDS from SF3B1-wildtype MDS, suggesting a
   unique gene expression signature associated with SF3B1 mutation status
   (Fig. [80]4). Specifically, the gene UBA7 was involved in more than two
   pathways, underscoring its potentially profound impact and paving the
   way for a more focused investigation into its specific role in these
   processes.

Fig. 4.

   [81]Fig. 4
   [82]Open in a new tab

   Transcriptomic landscape of SF3B1-mutant samples reveals distinct gene
   expression signatures. Hierarchical clustering heatmap showing the
   Z-score normalized expression of differentially expressed genes (DEGs)
   between SF3B1-mutant and wild-type (WT) samples. Each row represents a
   gene, and each column corresponds to a sample. Red indicates higher
   expression levels, while blue indicates lower expression levels. The
   SF3B1 mutation status is denoted by the color bar above the heatmap
   (red for mutated and blue for WT). Genes are clustered based on their
   expression profiles, revealing distinct patterns of upregulated and
   downregulated genes associated with SF3B1 mutations. These expression
   differences highlight the molecular impact of SF3B1 mutations on the
   transcriptomic landscape of the samples analyzed.

   Functional alterations associated with SF3B1 mutations are likely due
   to abnormal splicing events in critical genes involved in
   tumorigenesis^[83]36,[84]37. Approximately half of the genes linked
   with aberrant splicing events were subjected to nonsense-mediated
   decay, reducing gene and protein expression^[85]32,[86]38. Therefore,
   this observation may affect key genes involved in cancer-protective
   mechanisms. To find genes subjected to an increased level of aberrant
   AS events, the number of AS events was counted for each gene and
   plotted as a horizontal bar plot (Fig. [87]5A). The results showed high
   AS events in genes such as UBA7, HLA, RNF10, PSMB4, HNRNPA1, ERGIC3,
   SUGP1, and RPL13A. To have a holistic overview of UBA7 significant
   splicing events among other events a volcano plot was generated
   highlighting these events (Supplement 5A). Consistent with our previous
   findings (Fig. [88]3C), genes exhibiting a high frequency of
   alternative splicing events were predominantly associated with pathways
   related to protein post-translational modifications and interferon
   signaling. Among these genes, UBA7 was significantly downregulated
   (Fig. [89]5B), which may have resulted from the observed alterations in
   splicing events that occurred during the maturation of mRNA produced
   from this gene.

Fig. 5.

   [90]Fig. 5
   [91]Open in a new tab

   Spliceosome-mediated disruption of UBA7 in SF3B1-mutant samples. (A)
   Bar plot showing the top genes with the highest number of alternative
   splicing (AS) events in SF3B1-mutant samples. UBA7 demonstrates the
   greatest number of AS events, highlighting its susceptibility to
   splicing dysregulation in the context of SF3B1 mutations. (B) Boxplot
   illustrating the downregulation of UBA7 expression in SF3B1-mutant
   (red) compared to wild-type (WT) samples (blue). Gene expression is
   presented as log2-normalized counts, with a significant difference
   observed (adjusted p-value < 0.0001). This result suggests a functional
   impact of SF3B1 mutations on UBA7 expression, potentially contributing
   to disease pathogenesis.

   To validate our findings, we analyzed the splicing landscape and
   expression profiles of key splicing machinery genes, including SRSF2,
   U2AF1, and ZRSR2, which are known to be frequently mutated in
   hematological malignancies. While we observed significant alternative
   splicing events in these genes, their frequency was notably low and did
   not substantially impact their gene expression profiles, suggesting
   that deregulated gene expression driven by aberrant alternative
   splicing is specific to UBA7 (Supplement 5D–F, Supplements 8, 9, and
   10). Furthermore, we identified significant alternative splicing events
   in previously established gene associated with SF3B1-mutant tumors,
   BRD9, which were accompanied by a marked reduction in gene expression
   (Supplement 5B and Supplement 6). These results are consistent with
   previous studies^[92]32,[93]36,[94]39, supporting a similar mechanism
   underlying the disruption of UBA7 expression.

   Ubiquitin-like modifier-activating enzyme 7 (UBA7) is a specialized
   E1-like enzyme that plays a pivotal role in the conjugation of
   interferon-stimulated gene 15 (ISG15). Most prior research on the
   UBA7/ISG15 signaling pathway in cancer has concentrated on lung and
   breast cancer and reported that UBA7 gene expression was
   downregulated^[95]40–[96]43. Combined with its association with various
   diseases, it underscores its biological importance. The high frequency
   of splicing events in UBA7 observed in the MDS data highlights its
   potential impact on the regulation of protein post-translational
   modifications and interferon signaling. Additionally, the limited
   existing research on UBA7 in the context of hematological malignancies
   and MDS in particular compared to other genes presents a unique
   opportunity to explore novel regulatory mechanisms and therapeutic
   targets. Therefore, the focus will be on UBA7 for the remainder of this
   paper to elucidate its functional implications and potential
   contributions to disease pathogenesis.

   The aberrant AS events changes associated with the UBA7 gene were A3SS,
   MXE, and IR. In all these events, higher PSI values were associated
   with SF3B1 mutations, selected AS events changes were plotted in
   Fig. [97]6A, B. In addition, the sashimi plots illustrate the RNA-seq
   read coverage and splicing patterns for UBA7 in SF3B1-mutant and
   SF3B1-WT samples. Compared to WT samples, SF3B1-mutant samples exhibit
   increased intron retention and alternative 3’ splice site usage,
   highlighting significant splicing alterations in UBA7 transcripts
   (Fig. [98]6C, D). These findings suggest that SF3B1 mutations
   significantly impact the splicing and the expression of UBA7. Together,
   these findings suggest that SF3B1 mutations contribute to MDS
   pathogenesis by disrupting the balance of protein post-translational
   modifications through aberrant alternative splicing events,
   consequently affecting key cancer barrier pathways.

Fig. 6.

   [99]Fig. 6
   [100]Open in a new tab

   UBA7 alternative splicing events and gene expression are dysregulated
   in SF3B1 mutant MDS cells. (A) Boxplot showing the Percent-Spliced-In
   (PSI) values for intron retention (IR) events in UBA7, comparing
   wild-type (WT, blue) and SF3B1-mutant (red) samples. A significant
   increase in IR is observed in SF3B1-mutant samples (****p < 0.0001).
   (B) Boxplot displaying the PSI values for alternative 3’ splice site
   (A3SS) events in UBA7, with SF3B1-mutant samples (red) exhibiting
   significantly higher PSI values compared to WT (blue) (****p < 0.0001).
   (****FDR < 0.0001, ***FDR < 0.001, **FDR < 0.01, *FDR < 0.05). (C)
   Sashimi plot comparing RNA-seq read coverage and splicing patterns for
   UBA7 between SF3B1-mutant (red) and WT (blue) samples, focusing on a
   specific genomic region. Numbers on the curved lines represent the
   average junction-spanning read counts across biological replicates,
   normalized using RPKM. The average number of junction-spanning reads is
   clearly higher in WT samples compared to SF3B1-mutant samples,
   indicating reduced splicing efficiency in the mutant group. (D) Sashimi
   plot comparing another genomic region of UBA7 between SF3B1-mutant
   (red) and WT (blue) samples, showing similar trends with higher
   junction-spanning read counts in WT samples. The bottom black tracks
   represent the genomic localization of splicing events, providing a
   structural overview of exon–intron junctions. This figure highlights
   the spliceosome-mediated dysregulation of UBA7 in SF3B1-mutant samples,
   characterized by increased intron retention, alternative 3’ splice site
   usage, and reduced splicing efficiency compared to wild-type samples.

   Mutations in SF3B1 are highly prevalent and significant in
   hematological malignancies^[101]44,[102]45. In patients with chronic
   lymphocytic leukemia (CLL), SF3B1 gene alterations occur in
   approximately 10–14% of cases. These mutations are linked to reduced
   overall survival rates in CLL patients^[103]46. To determine if the
   findings observed in MDS also apply to CLL, public gene expression data
   (n = 304) from the International Cancer Genome Consortium (ICGC) for
   CLL was analyzed.

   To determine if SF3B1 mutations in CLL are correlated with a similar
   reduction in UBA7 gene expression as observed in MDS, we analyzed the
   ICGC CLL dataset. Although we observed a similar trend of UBA7
   downregulation in SF3B1-mutant CLL samples, the change was not
   statistically significant (Supplement 11A). Next, given that SF3B1
   mutations are known to co-occur with 11q deletions in CLL^[104]47, we
   investigated whether UBA7 expression is associated with 11q deletions.
   Interestingly, we observed a slight upregulation of UBA7 in CLL samples
   with 11q deletion, but this change was also not statistically
   significant (Supplement 11B). This observation suggests that UBA7 gene
   expression changes may be specific to SF3B1 mutations, emphasizing the
   need for future functional studies to validate this finding and
   elucidate the underlying mechanisms. These observations may stem from
   the presence of non-functional SF3B1 mutations, variations in allele
   frequencies, and the well-documented heterogeneity of the 11q deletion
   region in CLL, further underscoring the complex and multifaceted nature
   of CLL as a disease^[105]48,[106]49.

   Using the CLL ICGC associated clinical information, including survival
   data for CLL patients, allowed the investigation into the impact of
   UBA7 gene expression on overall survival. CLL patients with SF3B1
   mutations were identified, and their gene expression data were
   imported. After normalizing this data, the median UBA7 expression was
   used as a threshold to categorize patients into high and low UBA7
   expression groups. Kaplan–Meier survival plots were then generated
   using the gene expression and clinical data. The analysis showed that
   CLL patients with low UBA7 expression had significantly shorter
   survival times compared to those with high UBA7 expression
   (Fig. [107]7), with the difference being statistically significant
   (log-rank p-value = 0.0027). SF3B1 mutations significantly co-occur
   with 11q deletions, both of which are associated with aggressive
   disease and poor overall survival in CLL^[108]47. Our findings
   demonstrated that UBA7 downregulation in SF3B1-mutant CLL is strongly
   linked to poor survival, suggesting its potential integration into
   prognostic models to enhance the prediction of disease progression.
   Furthermore, we were able to stratify CLL patients harboring SF3B1
   mutations based on the UBA7 gene expression profile, providing a
   refined approach for risk assessment.

Fig. 7.

   [109]Fig. 7
   [110]Open in a new tab

   UBA7 downregulation correlates with worse survival outcomes in chronic
   lymphocytic leukemia (CLL). Kaplan–Meier survival curves comparing
   overall survival probabilities between patients with high (red) and low
   (blue) UBA7 gene expression levels. The analysis reveals that patients
   with low UBA7 expression have significantly poorer survival outcomes
   compared to those with high expression levels (p = 0.0027). The shaded
   areas represent the 95% confidence intervals for each group. Below the
   survival curves, the "Number at risk" table indicates the number of
   patients remaining in each group at different time points.

   The Cancer Genome Atlas (TCGA) database represents an invaluable
   resource for studying cancer. To further examine the influence of UBA7
   gene expression on cancer patient survival, the integration of TCGA
   gene expression data with clinical information was utilized to
   construct Kaplan–Meier survival plots for each TCGA cohort (Supplement
   S12). Notably, six TCGA cancer cohorts exhibited significant results,
   showing that patients classified with low UBA7 gene expression profiles
   have reduced overall survival rates, and this observation was
   statistically significant. These cohorts include Esophageal Carcinoma
   (ESCA) with a log-rank p-value of 0.024, Skin Cutaneous Melanoma (SKCM)
   with a log-rank p-value of less than 0.0001, Liver Hepatocellular
   Carcinoma (LIHC) with a log-rank p-value of 0.035, Breast Invasive
   Carcinoma (BRCA) with a log-rank p-value of 0.00029, Bladder Urothelial
   Carcinoma (BLCA) with a log-rank p-value of 0.0032, and Adrenocortical
   Carcinoma (ACC) with a log-rank p-value of 0.017.

   These findings are consistent with previous observations in MDS and
   CLL, further supporting the role of UBA7 as a potential tumor
   suppressor. Our analysis identified a distinct gene expression
   signature associated with SF3B1 mutations, centered on protein
   post-translational modification (PTM) and immune response pathways.
   Notably, UBA7 exhibited a high frequency of alternative splicing events
   and low gene expression, implicating its dual role in PTM and
   interferon signaling. In conclusion, the consistent association between
   low UBA7 expression and reduced survival across diverse cancer cohorts
   underscores the importance of UBA7 as a key contributor to cancer
   progression and patient outcomes.

Discussion

   This work evaluated the impact of SF3B1 gene alterations on the
   splicing landscape in the context of MDS. The results demonstrated that
   the splicing landscape in SF3B1-mutant MDS samples was significantly
   deregulated, impacting genes involved in protein post-translational
   modifications, as reflected in the identified gene expression
   signature. Among these affected genes, UBA7 stood out due to its
   involvement in multiple biological pathways and its established role in
   carcinogenesis. These findings provide novel insights into how SF3B1
   mutant cells gain more fitness and participate in MDS pathogenesis and
   extend work done by Pellagatti et al.^[111]19.

   The findings in this work with an increased sample size showed results
   that further support work done by Pellagatti et al. in terms of
   transcriptomic signatures associated with mutant SF3B1 gene in pathways
   such as heme metabolism and mitochondrial metabolism (Fig. [112]2B).
   This work demonstrated that protein post-translational modification is
   disrupted in SF3B1-mutant MDS, as evidenced by the deregulation of the
   UBA7 gene. This downregulation was driven by aberrant alternative
   splicing events, consistent with mechanisms observed in previous
   studies^[113]36,[114]39 where BRD9, another gene subjected to
   spliceosome dysregulation, exhibited impaired gene expression.

   SF3B1 mutations play a multifaceted role in tumor pathogenesis. In
   numerous cancer cells, SF3B1 mutations have been shown to disrupt
   various cellular processes, such as heme metabolism, immune
   surveillance, DNA damage maintenance, and Notch
   signaling^[115]50–[116]52. Additionally, these mutations affect several
   cellular pathways, notably the Notch, NF-κB, and mitochondrial
   pathways^[117]51.

   The splicing machinery has been reported to play diverse roles in
   cellular processes beyond its canonical function in splicing^[118]53.
   Notably, alternative splicing (AS) and protein post-translational
   modifications (PTMs) are known to work in concert, enabling precise,
   time- and cell-specific protein modifications essential for numerous
   cellular functions^[119]54. In support of this, studies have shown that
   the splicing machinery actively enhances the regulation of PTMs, and
   inefficient pre-mRNA splicing can disrupt multiple post-transcriptional
   regulatory processes, as demonstrated in Caenorhabditis
   elegans^[120]55. In our work, we observed significant gene expression
   deregulation in genes associated with protein post-translational
   modifications, supporting the notion that SF3B1 mutations have roles
   extending beyond splicing. Specifically, UBA7, a critical member of the
   protein post-translational modification pathway, was notably impacted
   by this deregulation, providing novel insights into how SF3B1-mutant
   cells gain a survival advantage by exploiting these disrupted pathways.
   These findings underscore the intricate interplay between splicing and
   post-translational modifications in maintaining cellular homeostasis.

   In this work, the UBA7 gene was found to exhibit increased levels of
   alternative splicing (AS) events, which significantly impacted its gene
   expression and potentially its biological role. UBA7 plays a critical
   role in protein post-translational modification by catalyzing the
   attachment of ISG15 to cellular proteins through a process known as
   ISGylation. ISGylation is a reversible modification wherein the
   ubiquitin-like protein ISG15 is processed into a mature 15-kDa form,
   exposing its carboxyl-terminal LRLRGG motif. This motif facilitates the
   recognition and binding of ISG15 to lysine residues on target proteins,
   thereby influencing their function, stability, and
   localization^[121]56. The ISGylation process occurs via a cascade of
   three enzymatic steps: the E1 group of enzymes, such as UBA7 activates
   ISG15 by forming a high-energy thioester bond, the E2 group of enzymes
   transfers the activated ISG15, and E3 ligases mediate its covalent
   attachment to specific lysine residues on target proteins. Importantly,
   this modification is dynamically regulated by the deubiquitinase USP18,
   which removes ISG15 from proteins, ensuring precise control and
   cellular adaptability under varying conditions^[122]57

   The role of ISGylation in cancer remains a topic of debate, with
   studies suggesting both tumor-suppressing and cancer-promoting effects.
   Elevated ISGylation has been shown to inhibit the
   camptothecin-dependent proteasome-mediated degradation of topoisomerase
   I in breast cancer cells, contributing to the efficacy of camptothecin
   as an anticancer chemotherapeutic agent^[123]58. On the other hand,
   several studies have demonstrated that ISGylation inhibits
   tumorigenesis by destabilizing proteins involved in growth regulation.
   For example, UBA7 targets key regulatory proteins such as cyclin D1 and
   PML-RARα for proteasomal degradation, suppressing the growth of human
   lung cancer cells^[124]59,[125]60. Furthermore, free ISG15 has been
   shown to inhibit tumor growth when administered extracellularly,
   inducing the infiltration of NK cells into tumors in nude mice.
   Intracellular free ISG15 also enhances the 26S proteasome-dependent
   surface expression of MHC Class I complexes on breast cancer cells,
   further contributing to its tumor-suppressive role^[126]61. Recent
   findings by Wei Y. et al. reported that UBA1, a member of the E1
   ubiquitin-activating enzyme family, is downregulated at both the mRNA
   and protein levels in myelodysplastic syndrome. Notably, their study
   revealed a tendency for SF3B1 mutations to co-occur with UBA1
   downregulation, further highlighting the potential role of E1 enzyme
   family dysregulation in the pathogenesis of SF3B1-mutant MDS^[127]62.

   Consistent with this notion, this study identifies a novel role for
   UBA7 in MDS, and potentially CLL, as a tumor suppressor subjected to
   downregulation through disruptions in alternative splicing. Low UBA7
   gene expression was associated with poor survival as seen in CLL
   patients harboring SF3B1 mutations, suggesting its potential use as a
   biomarker for adverse prognostic outcomes. These findings broaden our
   understanding of MDS pathogenesis, highlighting the critical role of
   the interferon-stimulated gene ISGylation network in tumor suppression.

   The findings presented in this work are derived from comprehensive
   computational analyses, providing valuable insights into the
   dysregulation of UBA7 in MDS and its potential roles in tumor biology.
   However, the absence of experimental validation represents a limitation
   that should be addressed in future research. Functional studies,
   including in vitro assays and in vivo models, are essential to confirm
   the involvement of UBA7 in ISGylation processes, immune evasion
   mechanisms, and tumor progression. These experiments could elucidate
   the molecular pathways regulated by UBA7 and its interactions with
   other key players in tumor biology. Moreover, evaluating UBA7’s
   functional relevance across diverse tumor types could reveal its
   broader implications as a pan-cancer target. Such studies would not
   only provide mechanistic validation but also enhance the translational
   significance of UBA7 as a potential therapeutic target and prognostic
   biomarker, paving the way for the development of targeted therapies and
   personalized treatment strategies in hematologic malignancies and
   beyond.

Supplementary Information

   [128]Supplementary Information 1.^ (2.3MB, zip)
   [129]Supplementary Information 2.^ (1.8MB, pdf)

Author contributions

   Conceptualization: SA Data retrival and analysis: SA Writing – original
   draft: SA & IA, and FA Review & editing: FA & OA, and AA All authors
   reviewed and agreed on the manuscript.

Funding

   This study did not receive any external funding.

Data availability

   RNA-seq data used in this study are publicly available at GEO database
   and can be accessed on the following links:
   [130]https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE63569,
   [131]https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE114922.

Declarations

Competing interests

   The authors declare no competing interests.

Ethics approval

   Due to the nature of this design, no additional ethical approval was
   needed. Consequently, the Local Ethics Committee waived the ethical
   approval. This study analyzed publicly available data from multiple
   independent cohorts ([132]GSE114922 and [133]GSE63569 for MDS, CLL-ES
   for CLL from ICGC, and cancer cohorts from the TCGA database).

Footnotes

   Publisher’s note

   Springer Nature remains neutral with regard to jurisdictional claims in
   published maps and institutional affiliations.

Supplementary Information

   The online version contains supplementary material available at
   10.1038/s41598-025-95738-9.

References