The discovery centers on what happens after a gene has been transcribed. A newly made messenger RNA contains a protein-coding sequence as well as untranslated regions that help determine how long the molecule survives, where it travels inside the cell and how efficiently it is converted into protein. At the molecule’s three-prime end, cellular machinery cuts the RNA at a selected site and adds a tail of adenine nucleotides, known as a poly(A) tail. This process, called cleavage and polyadenylation, can occur at more than one location. When a cell chooses an upstream polyadenylation signal, the resulting messenger RNA has a shorter 3′UTR; when it uses a downstream signal, the 3′UTR is longer. These alternative transcripts encode the same protein, but their regulatory behavior can be dramatically different. Shortening may remove binding sites for regulatory proteins and other factors that normally restrain gene expression, allowing cancer cells to amplify oncogenic programs without changing the protein-coding DNA itself.

The research team, led by investigators at the University of Milan and collaborating institutions in Italy, Switzerland and the United Kingdom, combined several kinds of sequencing data to reconstruct the NFYA 3′UTR landscape. They examined bulk RNA sequencing, single-cell RNA sequencing and specialized three-prime-end sequencing from prostate cancer cell lines and patient-derived material. This approach identified four functional NFYA 3′UTR isoforms, each terminating at a different polyadenylation site, although one was predominantly used across the cell lines and tissues examined. By measuring the relative use of proximal and distal polyadenylation sites, the researchers could determine whether cancer cells favored shortened or lengthened transcripts. The analysis showed a broad shift toward NFYA 3′UTR shortening in prostate cancer, rather than an isolated change in a small subgroup of tumors. The pattern was associated with higher tumor grade and metastatic disease, suggesting that RNA-end selection tracks with clinically aggressive biology.

The consequences of this shortening were substantial. Tumor samples and prostate cancer models using shorter NFYA transcripts contained more NF-YA protein, while cells with longer 3′UTRs produced less. NF-Y is a transcription factor complex that binds specific DNA elements and regulates genes involved in cell-cycle control, proliferation and other growth-related processes. NF-YA acts as a regulatory component that helps determine which genes the complex can control, so changing its abundance can reshape a large downstream transcriptional network. The investigators found that the short NFYA 3′UTR was linked to increased proliferation and other traits associated with aggressive disease. In this model, cancer progression was not driven simply by producing more NFYA messenger RNA. Instead, the tumor appeared to gain an advantage by selecting an RNA architecture that made the message more effective at generating protein.

The team also investigated how the long 3′UTR reduced NF-YA output. A longer untranslated region can contain additional docking sites for microRNAs, RNA-binding proteins and cellular transport machinery, but the experiments did not support increased microRNA-mediated repression as the main explanation. Instead, lengthening the NFYA 3′UTR reduced messenger RNA stability, impaired translation and increased retention of the transcript inside the nucleus. Messenger RNA stability determines how long a transcript remains available before degradation, while translation is the process by which ribosomes read the coding sequence and build a protein. Nuclear retention creates another bottleneck: even a transcript that has been produced may be less useful if it cannot efficiently reach the cytoplasm, where most translation occurs. Together, these effects sharply reduced the amount of NF-YA protein without eliminating the NFYA gene.

The RNA pattern also changed with the state of the cancer cell. When prostate cancer cells entered quiescence, a relatively inactive state in which proliferation pauses, they shifted toward longer NFYA 3′UTRs. A similar lengthening occurred after treatment with enzalutamide, an androgen-receptor inhibitor used in prostate cancer therapy. The observation connects NFYA RNA processing to both cellular dormancy and drug response. Prostate tumors often adapt to androgen-deprivation strategies, and treatment-resistant disease can eventually progress despite continued therapy. The study does not establish that NFYA 3′UTR lengthening explains enzalutamide’s clinical effects or that manipulating the RNA switch will overcome resistance in patients. It does, however, suggest that the choice of polyadenylation site is dynamic rather than permanently fixed and may reflect the balance between a proliferating, treatment-adapted state and a more restrained cellular condition.

To test whether the RNA-processing switch was merely associated with malignancy or could be manipulated therapeutically, the researchers used two different strategies. In one, CRISPR/Cas9-mediated deletion removed a polyadenylation signal, forcing cells away from the site that generates the shorter transcript and toward production of longer NFYA 3′UTRs. In the other, antisense oligonucleotides were designed to bind and mask polyadenylation signals. These short synthetic nucleic-acid molecules can be engineered to recognize a chosen RNA sequence and physically obstruct the proteins that assemble at a polyadenylation site. Redirecting cleavage in this way offers a potentially gene-specific intervention: rather than degrading every NFYA transcript or blocking NF-YA protein after it is made, the treatment changes which version of the transcript the cell produces. In cultured prostate cancer cells, both approaches lowered NF-YA protein and reduced phenotypes associated with tumor aggressiveness, including enhanced growth.

The strongest test came in vivo, where enforced NFYA 3′UTR lengthening also suppressed aggressive tumor traits and reduced tumor progression in experimental models. The results provide proof of concept, not a ready-made treatment. Antisense drugs must reach the relevant tumor cells, remain stable in the body, enter the correct cellular compartment and avoid unintended effects on other RNAs. Prostate tumors are biologically diverse, and the balance of polyadenylation signals and RNA-binding proteins may differ between patients, treatment histories and metastatic sites. Future studies will need to establish how reliably NFYA 3′UTR patterns predict outcome, whether they can be measured in clinical samples such as biopsies or circulating tumor material, and whether antisense-mediated remodeling is safe and durable in more representative models. Even so, the study expands the therapeutic map of cancer genetics. It shows that an oncogenic protein can be controlled not only by mutations, transcription or protein degradation, but also by the precise way its messenger RNA is finished. For prostate cancer, that overlooked decision at the end of an RNA molecule could become an important new target for precision therapy.

ASOs are short, chemically modified strands of DNA designed to recognize specific messenger RNA molecules. Messenger RNA carries genetic instructions from DNA to cellular protein-making machinery. When an ASO binds to its matching messenger RNA, the hybrid molecule can be recognized and destroyed by cellular enzymes, preventing production of the corresponding protein. This strategy has already shown promise in several inherited and neurodegenerative diseases, and researchers are investigating whether it can be used to silence mutant proteins that drive cancer growth.

The study focused on cET-ASO^KRas, an antisense drug developed by Ionis Pharmaceuticals to target messenger RNAs encoding mutant KRAS. KRAS proteins act as molecular switches that regulate cell growth and division. Mutations that lock KRAS in an active state can fuel uncontrolled proliferation and are common in pancreatic, colorectal, and lung cancers. In laboratory models, treatment with cET-ASO^KRas reduced mutant KRAS levels and inhibited the growth of pancreatic cancer cells and tumor-like spheroids.

Although ASOs are known to enter cells through endocytosis, the detailed trafficking process has remained poorly understood. During endocytosis, the cell surface folds inward and encloses external material in a membrane-bound compartment called an endosome. For an ASO to work, it must avoid being permanently trapped or degraded inside these compartments. It must instead reach the cell’s interior, escape from the endosome, and encounter its target messenger RNA in the cytoplasm or near the nucleus.

The researchers found that cET-ASO^KRas first binds to CD44, a receptor protein located on the surface of pancreatic cancer cells. CD44 is frequently abundant in aggressive tumors and has been associated with cancer stem-cell properties, nutrient uptake, invasion, and resistance to treatment. Binding to CD44 initiated a signaling cascade that activated a second receptor, EPHA2. This receptor then promoted the internalization of the ASO into endosomes and helped direct those endosomes toward the nucleus.

This positioning step proved critical. The nucleus is the compartment where most messenger RNA is produced and processed before being transported into the cytoplasm. By anchoring ASO-containing endosomes close to the nucleus, EPHA2 appears to place the therapeutic cargo near a major source of potential molecular targets. The study suggests that EPHA2 does more than participate in uptake: it acts as a trafficking organizer that determines where the ASO-containing compartments travel and how effectively the drug can reach the cell’s RNA.

Once near the nucleus, the endosomal membranes became damaged and leaky. This allowed cET-ASO^KRas to escape into the cytoplasm, where it could bind to mutant KRAS messenger RNA. The researchers demonstrated the importance of this pathway by removing CD44 or EPHA2 from pancreatic cancer cells. They also tested altered forms of EPHA2 that could no longer anchor endosomes near the nucleus. In each case, the ASO became less effective at lowering KRAS production and suppressing tumor growth.

The study also uncovered a cellular defense mechanism that limits antisense therapy. When endosomal membranes become leaky, the cell responds by forming stress granules, temporary assemblies of RNA and proteins that help protect and reorganize genetic messages during cellular stress. The researchers found that these structures can contribute to the repair or sealing of damaged endosomes, reducing the amount of ASO that escapes into the cytoplasm. In effect, the cell’s protective response can restrict the drug’s access to its target.

Blocking this response with ISRIB, a compound that inhibits stress-granule formation, enhanced the ability of cET-ASO^KRas to suppress KRAS production in laboratory experiments. The finding raises the possibility that drugs designed to improve endosomal escape could be combined with ASOs to increase their potency. However, because stress responses are also important for normal cell survival and tissue protection, such combinations would require careful testing to determine their safety and therapeutic window.

The researchers propose that the CD44–EPHA2 pathway could be exploited to improve the delivery of ASOs and other nucleic-acid medicines to aggressive tumors. Since both receptors are highly expressed in subsets of pancreatic cancer, they may provide a natural entry route that can be used to direct therapeutic molecules into particularly difficult-to-treat cells. The work does not yet establish an effective treatment for patients, but it identifies a series of molecular checkpoints that could guide the design of more efficient antisense therapies. By understanding how cancer cells import, transport, and restrict these drugs, scientists may be able to turn their own trafficking systems into a delivery advantage.

Subject of Research: Cells

Article Title: EPHA2/CD44-directed trafficking enhances endosomal leakiness and antisense therapy delivery

News Publication Date: 11 August 2026

Web References: Journal of Cell Biology: https://rupress.org/jcb ; DOI: https://doi.org/10.1083/jcb.202507217

References: Marco et al., 2026, Journal of Cell Biology, DOI: 10.1083/jcb.202507217

Image Credits: © 2026 Marco et al. Originally published in Journal of Cell Biology.

Keywords: Antisense oligonucleotides, ASO therapy, KRAS, pancreatic cancer, CD44, EPHA2, endocytosis, endosomal trafficking, endosomal escape, stress granules, ISRIB, cancer treatment, molecular biology