LUC7L2 Is a Novel RNA-Splicing Regulatory Factor Mutated in Myelodysplastic Syndromes

Myelodysplastic syndromes (MDS) are unique among cancers because of the frequent occurrence of somatic mutations impacting spliceosome machinery. At least 65% of MDS patients harbor a mutation in one of several splicing factors including U2AF1, SF3B1 and SRSF2. Whole exome sequencing of MDS bone marrow uncovered somatic frameshift mutations in LUC7L2, the mammalian ortholog of a yeast splicing factor. LUC7L2 is located in the most commonly deleted region of chromosome 7. Deletions and frameshifts lead to haploinsufficient expression and therefore it can be approximated that a combined 14% of MDS patients have low expression of LUC7L2. Restoring expression of LUC7L2 in del(7q)-iPSCs partially rescues the differentiation of iPSCs into CD45⁺ myeloid progenitors. Although perhaps partly due to associated losses of other genes on chromosome 7, low expression of LUC7L2 correlates with a poorer patient prognosis, so its haploinsufficiency may play an important role in bone marrow failure. While U2AF1, SF3B1, and SRSF2 are well-characterized splicing factors, the function of LUC7L2 in pre-mRNA splicing is unexamined and its role in the MDS pathogenesis is undefined. We hypothesize that low expression of LUC7L2 results in the aberrant splicing of oncogenes and tumor suppressor gene transcripts thus reducing expression or altering function and contributing to the pathogenesis of MDS.

We have characterized LUC7L2 as an alternative splicing regulatory protein that plays a repressive role in the regulation of alternative RNA splicing. We generated HEK-293 cells overexpressing V5-tagged LUC7L2 for immunoprecipitation-mass spectrometry, to ascertain protein interactions with LUC7L2. LUC7L2 co-immunoprecipitated with splicing regulators which are involved in splice site recognition. We performed cross-linking-IP-high-throughput-sequencing (CLIP-seq) to identify LUC7L2 binding sites on RNA. We identified 301 LUC7L2 RNA-binding sites as well as binding sites on U1 and U2 which is common for splicing regulatory proteins. Metagene analysis of these binding sites showed that LUC7L2 bound near splice sites in exonic sequences.

We knocked down LUC7L2 expression in HEK293 and K562 cells to phenocopy the frameshifts and deletions observed in MDS patients. We used a PCR-based assay to measure the splicing efficiency of introns near LUC7L2-binding sites. Knockdown of LUC7L2 increased the splicing efficiency of 8/13 selected introns; this suggests that LUC7L2 represses selective splice site usage. We also performed RNA-seq to characterize global mis-splicing events. Analysis of RNA transcripts revealed a multitude of splicing changes, including enhanced exclusion of alternative introns. Knockdown LUC7L2 cells exhibited-altered expression of other splicing factors; this could have further contributed to the vast number of splicing changes observed.

To identify specific splicing changes that could contribute to the pathogenesis of MDS, we compared the splicing profiles of LUC7L2-knockdown in K562 cells with RNA-seq data from K562 cells expressing U2AF1^S34F, SRSF2^P95H or SF3B1^K700E. This analysis yielded several exon-skipping splicing patterns in cancer-relevant transcripts, such as oncogene PRC1, splicing factor PTBP1 and MRPL33. Additionally, we noticed commonly mis-spliced transcripts among the four datasets in which the missplicing events occurred in the functional domain, potentially conferring a functional change. Surprisingly, we observed missplicing of U2AF1 in LUC7L2-knockdown, SRSF2^P95H, and SF3B1^K700E K562 cells, which altered the length of the RNA-recognition UHM domain by inclusion of a mutually exclusive exon or retention of an intron. In this way, low expression of LUC7L2, or point mutants U2AF1^S34F, SRSF2^P95H, and SF3B1^K700E,could alter U2AF1 function as a distal convergence point.

In summary, we identified a novel splicing factor implicated in the pathogenesis of MDS. We characterized LUC7L2 as a splicing repressor and discovered many splicing changes caused by low expression of LUC7L2. Several genes were also mis-spliced in U2AF1^S34F, SRSF2^P95H and SF3B1^K700E K562 cells targeting these for further study. Commonly mis-spliced targets such as U2AF1 may indicate that some of the novel therapeutics may have spliceosome mutation agnostic effects. If this applies to the LUC7L2 mutations, then they may also be effective in del7/del7q cases.