Blueprint of the Human Spliceosome Could Help Find Cancer’s Achilles’ Heel

By Deborah Borfitz 

November 13, 2024 | The spliceosome, the cellular machinery that catalyzes the process of splicing in genes, was discovered nearly five decades ago and has been successfully leveraged in the development of therapies for rare diseases such as spinal muscular dystrophy where concerns about side effects tend to be secondary to correcting the underlying pathology. The question now is how to move the strategy to more complex diseases involving more than a single gene and splicing mutation, according to Malgorzata Rogalska, Ph.D., researcher with the Center for Genomic Regulation (CRG) in Barcelona. 

It is an almost unfathomably difficult undertaking. Although the human genome is relatively small compared to that of a plant or even some simple bacteria, those 20,000 genes produce over 100,000 proteins, she points out. 

“For humans, the complexity does not really lie in the DNA component but how those genes are getting transcribed and processed,” says Rogalska. In fact, scientists explain the mismatch between genes and biological intricacy through RNA processing, notably the splicing step when non-coding segments of RNA are removed, and the remaining coding sequences get stitched together to form a recipe for protein production. 

The complex of RNA and proteins that make up the spliceosome has 150 core components plus another 150 that can come into play at different points, she continues. They each have a role to play, which might include selecting an RNA segment for removal, ensuring cuts are made at the right place in the RNA sequence, or controlling the time-to-action of other components. It is thought to be possibly the largest and most confusingly interrelated molecular machine found within a human cell. 

This machinery interprets pieces of DNA and “splices” them together, a sailing term meaning to join two ropes together. Spliced transcripts differ widely across tissue types and probably the most remarkable example is the brain where alternative splicing occurs at a particularly high frequency, says Rogalska.  

It is relatively easy to tease out a single causal mutation by looking at what’s happening in diseased versus healthy tissue, she says. “Correct it and in theory everything is okay.” 

With a more complex disease like cancer, “probably a correcting of one gene won’t ever be sufficient to correct the phenotype,” Rogalska says. “We have to understand the molecular machinery itself.” 

To that end, researchers at the CRG created the first blueprint of the human spliceosome that was published recently in Science (DOI: 10.1126/science.adn8105). It was a scientific feat that represents the culmination of more than a decade of work and the insights of more than a dozen experts. 

‘Cascade of Events’

Dysregulation of splicing has been recognized as a common feature of cancer for at least the last 15 years, and some of the more prominent culprit genes (e.g., SF3B1, U2AF1, PRMTs, and MYC) are core components of the spliceosome or play roles in regulating its expression, says Rogalska. But there is yet no approved therapies that take advantage of the information because, as the new blueprint reveals, “affecting one component of the spliceosome activates a cascade of events.”  

Specifically, the research team manipulated spliceosome component SF3B1, which is known to be mutated in many cancers and a target for anti-cancer drugs although the mechanisms of action were previously unclear. The study found that altering the expression of SF3B1 in cancer cells affected a third of the cell’s entire splicing network causing a succession of failures that overwhelmed the cell's ability to fuel growth. 

Understanding those chain reactions will likely be key to designing effective therapies, including its therapeutic window and how to minimize resistance, she adds. Some previously tested therapies aimed at spliceosome components turned out to be toxic enough to halt clinical trials.  

The blueprint represents a transcriptome-wide splicing network and indicates the order and density of the connections between RNA splicing factors, Rogalska explains. It can be used as a dynamic resource to explore the mechanisms of splicing regulation by dissecting the indirect effects of different perturbations. 

As it turns out, proteins at the core of the spliceosome are quite specialized and allow cells to send out precise genetic messages that might be exploited for drug development purposes, she continues. Up to now, knowledge about the spliceosome has come largely from the cryo-structures created under ideal conditions in the laboratory that suggested its components always appeared in the same order and in the same composition on the RNA.  

In the environment of the cell, this is not the case. In the latest study, researchers show that U1 small nuclear ribonucleoprotein (U1 snRNP) complex—what Rogalska describes as “beautiful” particles she has long studied that were thought to be inflexibly associated with three specific binding proteins—only really need a copy of two of the three and the third component is only important for a specific function. If this information is leveraged in the development of snRNP-targeting cancer drugs, it might be possible to fine-tune the core components to avoid dysregulating RNA splicing and some of the toxic side effects that have been hindering progress, she says. 

The other reality is that the spliceosome is highly interconnected such that disrupting one component can have widespread ripple effects throughout the entire network. For instance, the SF3B1 modulator RVT-2001 (formerly H3B-8800), developed to target splicing in myelodysplastic syndromes, was recently discontinued following disappointing interim results in a phase I/II trial, Rogalska cites as an example.  

Early trials of SF3B1 inhibitors were similarly halted due to on-target, off-tumor toxicity in the retina, and RVT-2001 was later found to induce cardiac toxicities. Understanding that some components of the spliceosome are like “master regulators of the machinery itself” means researchers can either try to contradict some of the toxicity that has been seen in the clinic or find a new way forward that more precisely targets the problem.  

Developers of the blueprint are currently working on a web interface so researchers anywhere can go in and play with the data, says Rogalska, which is highly homogenous and well curated. Many scientists have been anticipating release of the dataset, she adds, some of whom want to incorporate their own data. 

The raw RNA-seq data is currently available via the public database ArrayExpress, and tables included in the published paper list the strength and confidence value of each of the identified interactions. The code for the blueprint was written in R, the go-to programming language for researchers, data scientists, and analysts worldwide. 

Moving Forward Faster

Rogalska says she and her colleagues are currently looking at mutations of splicing RNA binding factors, which is seen in many patients with colorectal, blood, and prostate cancers. Using data in the spliceosome blueprint, they’ll try to “decode” the footprint of those mutations and determine if it causes a vulnerability that could be targeted. 

“We also want to explore synthetic lethality approaches where you combine the targeting of a drug transcription factor with splicing regulators,” she adds, since many of those factors were included in the blueprint. The idea is to use information about the interconnectedness between transcription and splicing to design a more precise synthetic lethality sequence with less toxic effects.   

A third area for further study is the generation of neoantigens, which is currently a hot topic in the RNA field. “In theory, we have perturbed everything there is to perturb in the spliceosome, so we should have produced all of the possible [aberrant RNA] isoforms that there are to be produced,” says Rogalska. This could then be used as a basis for testing tumor-specific neoantigens that can trigger an immune response. 

Overall, the blueprint of the spliceosome is expected to hold great appeal to scientists looking to precisely map splicing dysregulation. “Finding mutations in DNA is now relatively easy but sometimes you have transcriptomic effects that are not trivial, and the blueprint helps un-convolute what is happening,” she says. This might variably be done in the lab and in patients. 

“We did it in HeLa cells [the first immortal human cell line] because it is the most widely used and very easily accessible for... any [conceivable lab] experiment,” says Rogalska. Data from the sequencing of those cells could be incorporated into the blueprint to arrive at an idea of the inhibitory effects of a drug, based on dysregulation of the isoforms, “to at least have a hypothesis if any splicing perturbations that you observed were caused by which of the factors.” 

These days, seemingly everyone is using the new generation of proteomics to model transcripts, she says. But for this to apply to more complex diseases, it is imperative to understand what happens when the RNA processing machinery is perturbed. “Sometimes, taking a step back will allow you to move forward faster because you will understand better what you are trying to change.”