Frequent somatic mutations in components of the RNA processing machinery in chronic lymphocytic leukemia

Chronic lymphocytic leukemia (CLL) is a common neoplasia of B lymphocytes with a heterogeneous clinical course ranging from an indolent disorder, with a normal lifespan, to an aggressive disease and short survival.1 The genetic mutations and associated molecular alterations that are able to accurately recapitulate the molecular pathogenesis of CLL are only now coming to light.2, 3, 4, 5 Foremost amongst the mutated genes is SF3B1, a core component of the RNA splicing machinery. SF3B1 mutations were found to be somatically acquired in 10% of CLL patients, and correlate with a more aggressive behavior of the disease.4, 5, 6, 7 Consistently, the recent identification in myeloid neoplasms of recurrent mutations in SF3B1 and other spliceosome components in a mutually exclusive manner suggests a common functional consequence in the regulation of RNA splicing.8, 9, 10

Virtually all of the eukaryotic mRNAs are synthesized as precursor molecules that need to be extensively processed in order to serve as a blueprint for proteins. The most prevalent steps are the capping reaction at the 5′-end, the removal of intervening sequences by splicing, formation of 3′-end poly (A)-tails and the nuclear export of the generated mature mRNA. Other RNA molecules undergo some of these maturation steps. Thus, 5′-end capping and transport to the cytosol are necessary to produce functional small nuclear ribonucleoproteins (snRNP), which in turn catalyze mRNA splicing.11 Although these maturation steps affect most transcripts in the cell, deregulation of splicing has recently been recognized as a driver of oncogenic processes.12 Moreover, it has been shown that the anti-tumoral mechanism of the drug spliceostatin A involves altered fidelity of splice-site recognition and changes in alternative splicing, particularly of cell-cycle-control genes.13 With these antecedents, we observed that the list of the 60 genes with the highest corrected mutational frequencies in CLL (RM-CLL genes class 1) contained four genes encoding factors, which has known roles in RNA splicing (SF3B1, SFRS1/SRSF1, U2AF65/U2AF2 and BRUNOL4/CELF4).4 In addition, we found three genes encoding RNA-export-machinery factors (NXF1, XPO1 and DDX3X).4 This prompted us to investigate the role of mutations affecting the RNA processing machinery, not limited to splicing, in the development of CLL.

To assess the involvement of RNA-processing factors in the development of CLL, we extended our previous high-throughput sequencing experiment in 105 patients to a cohort of 140 patients. Thus, we compared the whole exomes of tumor and normal samples from each patient with the Sidrón algorithm (Supplementary Methods). With this method, we identified 52 somatic mutations in 44 patients affecting 29 genes (Figure 1 and Supplementary Table 1). This suggests that mutations on these factors may contribute significantly to CLL development. Consistent with this, only 3 out of the 43 missense mutations identified were predicted as neutral for the function of the protein (Supplementary Table 1). We compared this frequency with the background proportion of neutral mutations, as estimated with the same method on a set of randomly chosen mutations in genes not related to RNA processing (Supplementary Methods). The proportion of predicted neutral mutations in RNA-processing-related genes was significantly lower than in the background (3/43 versus 39/122, P=0.009). The remaining nine mutations are predicted to cause severe effects in six genes through frameshifts, premature stop codons or missed splicing sites. Notably, two of the mutated genes, EIF4A3 and MAGOH, belong to the exon junction complex (EJC). These genes, together with RBM8A and CASC3, form the core of this complex,14 which suggests that the EJC core might constitute a third target for somatic mutations in CLL.

Figure 1

Somatically acquired mutations in components of the RNA processing pathway in CLL. Presented is an overview illustrating the individual components of the RNA-processing pathway, with those components identified as being somatically mutated highlighted (*) and the mutated protein listed in red. Initially, nascent pre-mRNA transcripts undergo 5′ capping and binding of the cap-binding complex (CBC), followed by the formation of the major spliceosome, the machinery responsible for the removal of pre-mRNA introns via a stepwise mechanism. Initial assembly steps include formation of pre-spliceosome complex A (top left nuclear complex) involving recognition of the 5′ splice site by U1 snRNP (an interaction stabilized by members of the serine-arginine-rich (SR) protein family) and recognition of the 3′ SS region by the U2 Auxiliary factor U2AF and by U2snRNP. U2AF binds to the intronic polypyrimidine tract and 3′SS, and facilitates binding of U2 snRNP to the branch-point sequence. Stable U2 snRNP binding requires stabilizing interactions of the U2 snRNP heteromeric protein complexes SF3A and SF3B with the pre-mRNA and/or the 65 kDa subunit of U2AF. Transition to spliceosome complex B (bottom left nuclear complex) involves recruitment of the pre-assembled U4/U6/U5 tri-SNP and more than 35 non-RNP proteins, including protein components of the PRP19 and retention and splicing (RES) complexes. Activation of the spliceosome, a requirement for the first catalytic step of splicing and conversion to complex C (bottom right complex), requires the release of U1 and U4 snRNPs, followed by the second catalytic splicing step and dissociation of the spliceosome, and the simultaneous binding of the exon-junction complex (EJC) a few nucleotides upstream of the exon junction. Processed mRNA is exported from the nucleus mainly via interactions with the non-karyopherin heterodimer of NXF1 and NXT1, although a subset of mRNAs are exported via the karyopherin XPO1 and its binding partner DDX3X.

In spite of this evidence, several of the studied genes show low somatic-mutation frequencies, suggesting that most of the encoded factors may not be driving tumoral progression. Also, the complementation pattern of these mutations does not suggest a clear underlying biochemical mechanism (Figure 2). Therefore, we decided to focus on U2 snRNP and RNA-transport factors, whose genes feature frequent somatic mutations in a large number of CLL patients. As the U2 snRNP complex is involved in splicing of mRNA, we then studied the consequences of somatic mutations affecting these factors in mRNA splicing and expression patterns. In this analysis, we used AceView to classify the splicing sites predicted by RNA-Seq of 12 tumor samples with mutations in U2 snRNP factors and 42 control tumor samples (Supplementary Table 2) as known (described in AceView) or novel (otherwise), and calculated the coverage for each of those sites (Supplementary Methods). As previously reported for SF3B1, the somatic mutation of RNA-processing factors does not affect overall splicing accuracy as assessed by RNA-Seq experiments. Thus, the ratio of reads covering the known splicing sites versus reads covering novel splicing sites is similar in U2 snRNP-mutated tumors (9.42±0.24) and control samples (9.46±0.36). Therefore, even if the effect of mutations in splicing factors is mediated by splicing aberrations, these changes must be localized in a few genes, as has been shown for FOXP1.4 This is not surprising, as splicing is required for the synthesis of most proteins and RNAs, and widespread changes in this process are likely to result in cell death. In addition, mutations in factors involved in RNA metabolism may affect the ratio between RNA isoforms, their export to the cytoplasm or the decay of the mRNA, leading to changes in the expression of proteins important for tumor progression.

Figure 2

Distribution of somatic mutations in genes involved in RNA processing in CLL patients. Each red box indicates the somatic mutation of a gene (row) in a tumor sample (column). Mutations in other frequently mutated genes and the mutational status of IGHV are shown at the bottom.

To advance in the characterization of somatic mutations in U2 snRNP and RNA-transport factors, we performed a gene set expression analysis on RNA-Seq experiments from 12 tumor samples with U2 snRNP complex mutations compared to a control set of 42 tumor samples without mutations in known RNA-processing factors (Supplementary Table 2). Likewise, we compared six tumor samples with RNA transport mutations with 11 IGHV-unmutated tumor samples without mutations in known RNA-processing factors (Supplementary Table 2). A direct comparison of the coverage of all transcripts in each sample identified those genes whose expression is most likely to be affected by mutations in those factors (Supplementary Table 3). To extract information on the putative effects of those changes, we performed gene set enrichment analyses. These analyses identified several significant sets, whose genes are differentially expressed in one of the groups (Supplementary Table 4). Interestingly, the FRS2/PLC cascade is enriched in genes with differential expression in samples with somatic mutations in RNA-transport factors. The FRS2/PLC cascade regulates multiple downstream processes, including Akt phosphorylation and phosphatidyl inositol hydrolysis, some of which have been related to RNA export.15 Therefore, mutations in the RNA-export machinery might act in concert with deregulated FRS2/PLC signaling to facilitate the development of CLL. Further experiments will be necessary to validate this hypothesis and ascertain the nature of this putative link.

Finally, we investigated the clinical consequences of somatic mutations in RNA processing factors. To this end, we reviewed the clinical evolution of patients carrying somatic mutations in U2 snRNP or RNA-transport factors compared to patients with no mutations in each of those pathways (Supplementary Table 5). This analysis showed that somatic mutations in the RNA-processing machinery are heterogeneous in clinical and biological terms. As previously reported,4, 5, 6, 7 somatic mutations in SF3B1 are correlated with unfavorable clinical and biological parameters. However, a multivariate Cox analysis showed that only SF3B1 mutation, and not mutation in other factors of the U2 snRNP spliceosome, has independent prognostic value. This suggests that, in spite of the biochemical relationship between these genes, somatic mutations affecting different factors of the U2 snRNP spliceosome entail different clinical consequences. Therefore, the first approach to pharmacological targeting of this complex in CLL should focus on drugs that bind to SF3B1,13 which might be useful against aggressive forms of the disease. On the other hand, mutations affecting RNA-transport factors are associated with biological features related to adverse prognosis, notably lack of somatic hypermutation (Supplementary Table 5). This result singles out RNA transport as a new factor in the development and evolution of this aggressive form of CLL, and suggests that the pharmacological targeting of this pathway might provide new drugs with important clinical implications for the treatment of these recalcitrant cases. Further studies will be necessary to gain statistical power and to ascertain the putative mechanisms underlying this relationship, which might prove valuable for the prediction of the clinical outcome of CLL from early stages of the disease.

Taken together, these data suggest that somatic mutations affecting some factors in the RNA processing pathway contribute to the development of CLL. Furthermore, the preferential targets of these mutations are components of the U2 snRNP spliceosome and the RNA-transport machinery. However, we have provided evidence of clinical disparities between patients with mutations in SF3B1 and patients with mutations in other components of the U2 snRNP spliceosome. Finally, we have shown for the first time the putative clinical relevance of somatic mutations in RNA-transport factors. Therefore, the RNA-maturation pathway provides attractive targets for the development of specific drugs that may be useful in the treatment of a significant percentage of CLL patients.


  1. 1

    Rozman C, Montserrat E . Chronic lymphocytic leukemia. N Engl J Med 1995; 333: 1052–1057.

  2. 2

    Zenz T, Mertens D, Kuppers R, Dohner H, Stilgenbauer S . From pathogenesis to treatment of chronic lymphocytic leukaemia. Nat Rev Cancer 2010; 10: 37–50.

  3. 3

    Puente XS, Pinyol M, Quesada V, Conde L, Ordonez GR, Villamor N et al. Whole-genome sequencing identifies recurrent mutations in chronic lymphocytic leukaemia. Nature 2011; 475: 101–105.

  4. 4

    Quesada V, Conde L, Villamor N, Ordonez GR, Jares P, Bassaganyas L et al. Exome sequencing identifies recurrent mutations of the splicing factor SF3B1 gene in chronic lymphocytic leukemia. Nat Genet 2012; 44: 47–52.

  5. 5

    Wang L, Lawrence MS, Wan Y, Stojanov P, Sougnez C, Stevenson K et al. SF3B1 and other novel cancer genes in chronic lymphocytic leukemia. N Engl J Med 2011; 365: 2497–2506.

  6. 6

    Rossi D, Bruscaggin A, Spina V, Rasi S, Khiabanian H, Messina M et al. Mutations of the SF3B1 splicing factor in chronic lymphocytic leukemia: association with progression and fludarabine-refractoriness. Blood 2011; 118: 6904–6908.

  7. 7

    Quesada V, Ramsay AJ, Lopez-Otin C . Chronic lymphocytic leukemia with SF3B1 mutation. N Engl J Med 2012; 366: 2530.

  8. 8

    Yoshida K, Sanada M, Shiraishi Y, Nowak D, Nagata Y, Yamamoto R et al. Frequent pathway mutations of splicing machinery in myelodysplasia. Nature 2011; 478: 64–69.

  9. 9

    Papaemmanuil E, Cazzola M, Boultwood J, Malcovati L, Vyas P, Bowen D et al. Somatic SF3B1 mutation in myelodysplasia with ring sideroblasts. N Engl J Med 2011; 365: 1384–1395.

  10. 10

    Damm F, Nguyen-Khac F, Fontenay M, Bernard OA . Spliceosome and other novel mutations in chronic lymphocytic leukemia and myeloid malignancies. Leukemia 2012; 26: 2027–2031.

  11. 11

    Wahl MC, Will CL, Luhrmann R . The spliceosome: design principles of a dynamic RNP machine. Cell 2009; 136: 701–718.

  12. 12

    David CJ, Manley JL . Alternative pre-mRNA splicing regulation in cancer: pathways and programs unhinged. Genes Dev 2010; 24: 2343–2364.

  13. 13

    Corrionero A, Minana B, Valcarcel J . Reduced fidelity of branch point recognition and alternative splicing induced by the anti-tumor drug spliceostatin A. Genes Dev 2011; 25: 445–459.

  14. 14

    Le Hir H, Andersen GR . Structural insights into the exon junction complex. Curr Opin Struct Biol 2008; 18: 112–119.

  15. 15

    Okada M, Ye K . Nuclear phosphoinositide signaling regulates messenger RNA export. RNA Biol 2009; 6: 12–16.

Download references


This work was funded by the Spanish Ministry of Science and Innovation (MICINN) through the Instituto de Salud Carlos III (ISCIII) and Red Temática de Investigación del Cáncer (RTICC) del ISCIII. CL-O is an Investigator of the Botín Foundation. We are grateful to all members of the CLL Spanish Consortium for their continuous support to this project. We are also very grateful to all patients with CLL who have participated in this study.

Author information

Correspondence to V Quesada.

Ethics declarations

Competing interests

The authors declare no conflict of interest.

Additional information

Supplementary Information accompanies this paper on the Leukemia website

Supplementary information

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Ramsay, A., Rodríguez, D., Villamor, N. et al. Frequent somatic mutations in components of the RNA processing machinery in chronic lymphocytic leukemia. Leukemia 27, 1600–1603 (2013).

Download citation

Further reading