The research centres on a protein called G-protein signalling modulator 2, or GPSM2, a member of the family of proteins that regulate G-protein activity independently of receptors. GPSM2 is a 684-amino-acid protein built from eight N-terminal tetratricopeptide repeats and four C-terminal GoLoco motifs, and it is known to play an important role in mitotic spindle positioning and cell cycle regulation. It has already been implicated in several other cancers: it drives epithelial–mesenchymal transition in non-small cell lung cancer, and silencing it in breast cancer causes defective cell division and markedly slows proliferation. Yet its role in pancreatic cancer had remained largely undefined—until now.

To begin unravelling that role, the researchers performed an integrated transcriptomic analysis using data from 183 pancreatic cancer patients in The Cancer Genome Atlas, supplemented with normal pancreatic tissue samples from the GTEx database and adjacent normal samples from TCGA. After normalisation and batch correction, they identified a set of genes that were differentially expressed between tumour and normal tissue, and then screened these for survival relevance using univariate Cox regression. GPSM2 emerged as a prominent risk-associated gene, with a hazard ratio of 2.051, meaning that patients with higher GPSM2 expression faced more than double the risk of death compared with those with lower expression. Kaplan–Meier survival analysis confirmed that patients with high GPSM2 levels had significantly shorter overall survival, and immunohistochemical images from the Human Protein Atlas corroborated the protein’s elevated presence in tumour tissue.

With a statistical link established, the team turned to laboratory experiments to determine whether GPSM2 actively drives the disease or is merely a bystander. Working with two widely used pancreatic cancer cell lines, BxPC-3 and PANC-1, they engineered cells to either overexpress or silence GPSM2. The results were striking. Cells with boosted GPSM2 levels showed dramatically enhanced invasive capacity in Transwell migration assays and produced far more colonies in two-week proliferation assays. Conversely, knocking GPSM2 down suppressed both invasion and proliferation. Taken together, these findings positioned GPSM2 not as a passive marker but as a functional promoter of pancreatic cancer’s malignant behaviour.

The next question was how GPSM2 exerts these effects. The researchers’ attention turned to Yes-associated protein 1, or YAP1, a transcriptional co-activator and central node of the Hippo signalling pathway that is already recognised as a key driver of tumour initiation and progression in pancreatic cancer. When Hippo signalling is inactive, YAP1 migrates to the nucleus and partners with TEA domain transcription factors to switch on genes that promote epithelial–mesenchymal transition and a more aggressive, undifferentiated cancer state. Western blot analysis revealed that GPSM2 overexpression significantly raised YAP1 protein levels, while GPSM2 knockdown lowered them. Critically, when the researchers used the YAP1 inhibitor verteporfin, or generated YAP1 knockout cells within GPSM2-overexpressing lines, GPSM2’s ability to promote colony formation and invasion was largely abolished—demonstrating that YAP1 is the essential downstream mediator of GPSM2’s pro-tumour effects.

Digging deeper, the team discovered that GPSM2 boosts YAP1 not by increasing its production at the gene level but by stabilising its messenger RNA. Quantitative PCR showed that GPSM2 markedly increased YAP1 mRNA levels, and RNA decay assays using actinomycin D revealed that GPSM2 significantly slowed the degradation of YAP1 transcripts. The mechanism behind this stabilisation turned out to be N6-methyladenosine, or m6A, the most abundant internal chemical modification in eukaryotic messenger RNA. First identified in the 1970s, m6A influences RNA splicing, translation and stability, and it has become increasingly recognised as a powerful post-transcriptional lever that cancers pull to fuel proliferation, invasion and metastasis.

To confirm that GPSM2 acts through m6A methylation of YAP1 mRNA, the researchers employed several complementary approaches. Bioinformatic prediction tools identified putative m6A modification sites on the YAP1 transcript, and mutant reporter plasmids in which key adenosine residues were substituted with cytosine showed reduced m6A enrichment, pinpointing the modified positions. Methylated RNA immunoprecipitation followed by quantitative PCR confirmed that GPSM2 overexpression increased the m6A modification of YAP1 mRNA, and treating cells with the methylation inhibitor 3-deazaadenosine abrogated the GPSM2-driven rise in YAP1 mRNA levels. In other words, GPSM2’s stabilising grip on YAP1’s message depends on methyl marks placed directly on the RNA molecule itself.

The identity of the enzyme placing those marks proved equally important. The m6A modification is catalysed by a multicomponent methyltransferase complex whose core is the METTL3–METTL14 heterodimer, with METTL3 being the only subunit capable of binding the methyl donor S-adenosylmethionine and performing the actual catalytic transfer. When the researchers knocked down METTL3, METTL14 or WTAP individually in GPSM2-overexpressing cells, only METTL3 silencing abolished the GPSM2-induced upregulation of YAP1 protein. Conversely, overexpressing METTL3 enhanced GPSM2’s effect on YAP1. Co-immunoprecipitation experiments—both with tagged proteins and with endogenous proteins in BxPC-3 cells—showed that GPSM2 physically interacts with METTL3, suggesting that GPSM2 recruits or cooperates with the methyltransferase machinery to install m6A marks on YAP1 transcripts.

But methyl marks alone do not stabilise RNA; they must be recognised by so-called reader proteins. The IGF2BP family of readers, which possess two RNA recognition motif domains and four K-homology domains, is known to enhance mRNA stability upon binding m6A sites. Analysing TCGA data, the team found that high expression of IGF2BP2 and IGF2BP3 was associated with poorer overall survival in pancreatic cancer patients, and both readers correlated positively with GPSM2 expression. Overexpressing either reader in GPSM2-high cells further increased YAP1 protein and mRNA levels. Using a series of HA-tagged domain truncation constructs, the researchers demonstrated through RNA immunoprecipitation that it is specifically the KH3-4 domains of IGF2BP2 and IGF2BP3 that recognise the m6A-modified YAP1 transcripts and anchor them against degradation. The full circuit—GPSM2, METTL3, m6A, IGF2BP2/3, YAP1—now formed a coherent mechanistic chain linking a poorly understood signalling modulator to one of cancer’s most potent growth drivers.

Importantly, the story did not end in the petri dish. In xenograft experiments, pancreatic cancer cells overexpressing GPSM2 were injected subcutaneously into athymic nude mice, and the resulting tumours grew significantly larger and heavier than controls. Immunohistochemical staining of the excised tumours confirmed that GPSM2 drove up the expression of METTL3, IGF2BP2, IGF2BP3 and YAP1 within the tumour tissue itself, validating the in vitro mechanism in a living system. This in vivo confirmation strengthens the case that the GPSM2–METTL3–YAP1 axis is not a laboratory artefact but a genuine feature of pancreatic tumour biology.

The implications of the work are twofold. First, GPSM2 may serve as a biomarker: its strong association with unfavourable overall survival suggests it could help stratify patients by risk, potentially guiding treatment intensity and follow-up. Second, and perhaps more excitingly, each node in the pathway represents a potential therapeutic target. Drugs that block the GPSM2–METTL3 interaction, inhibit METTL3’s catalytic activity, disrupt the binding of IGF2BP readers to m6A sites, or suppress YAP1 itself—verteporfin being an existing example of the latter—could, in principle, collapse the entire growth-promoting circuit. The researchers caution that several questions remain open, including the precise regulatory mechanism between GPSM2 and METTL3 and the full dependency network linking GPSM2 to YAP1, and larger clinical cohorts will be needed to correlate GPSM2 protein expression with tumour staging and grading. Nevertheless, in a disease where effective molecular targets are desperately scarce, the delineation of a complete GPSM2-to-YAP1 signalling axis—woven together by RNA methylation—offers a genuinely new roadmap for therapeutic development and a reminder that some of cancer’s most important vulnerabilities may lie not in DNA, but in the chemical decoration of its messenger molecules.