Genomic analysis uncovers prevalence of RNA splicing mutations in 33 tumor types

Illustration of three DNA helices
In addition to opening up new indications for H3 Biomedicine, the findings also improve the field's understanding of RNA splicing mutations in cancer. (Darwin Laganzon)

Eisai spinoff H3 Biomedicine has published an analysis of data from The Cancer Genome Atlas (TCGA) showing that genetic mutations in RNA splicing factors are much more common and appear in more cancers than previously thought. The findings open up new indications for H3, which has a clinical-stage candidate for blood cancers and highlights RNA splicing factors as a target for others developing cancer treatments. 

Genes make proteins when instructions from DNA are transcribed to mRNA, which then translates these instructions into proteins. RNA splicing happens between these transcription and translation steps, when pre-mRNA turns into RNA. 

Somatic mutations in core splicing factors have been observed in some blood cancers before, but the breadth of them wasn't completely understood.

"The findings published today give us for the first time a comprehensive understanding of just how common splicing factor mutations are across many types of cancer and of the role they may play in promoting tumor formation and progression,” said H3 CEO Markus Warmuth, M.D., in a statement. 

The company used whole exome and RNA sequencing to analyze 10,000 patient samples across 33 tumor types. The analysis looked at more than 400 splicing factors, turning up 199 genes with putative driver mutations, including loss-of-function mutations and hotspot, or oncogenelike, mutations. They also found splicing factor mutations in all 33 tumor types, with mutations occurring in as many as 60% of samples in some tumor types. The research was published in the journal Cell Reports.

"Prior to this publication, there have been reports of somatic mutations in the spliceosome, both in hemotological malignancies and solid tumors," said H3 Chief Scientific Officer Peter Smith in an interview. It's the first comprehensive analysis of its kind, he said. 

H3's clinical candidate, H3B-8800, is an orally available, small-molecule inhibitor of wild-type and mutant SF3b complex, a splicing factor gene. The drug is in a phase 1 study evaluating its safety and efficacy in patients with myelodysplastic syndromes, acute myeloid leukemia and chronic myelomonocytic leukemia who carry mutations in splicing factor genes. Preclinical results show it has antitumor effects in models of various spliceosome-mutant cancers. 

“We’re particularly proud of this work because, beyond informing our own drug discovery efforts, it will help advance the entire field’s understanding of the role dysregulated RNA splicing plays in cancer," said Lihua Yu, Ph.D., H3's chief data science officer.

TCGA, which was founded in 2005, has been criticized by scientists who thought its funding would be better directed elsewhere. But the dataset has helped answer some questions, including how to develop targeted treatments for cervical cancers that are not caused by human papillomavirus (HPV). Last year, the TCGA Research Network discovered a group of eight cervical cancers with specific genetic characteristics that could lead to the development of personalized therapies.

Smith said that his company's analysis will allow scientists there to envision different scenarios for H3B-8800.

"It allows us to think about additional indications or genotypes where that compound should have therapeutic use in patients and it has inspired a number of translational studies we've done to follow up," he said.