Overview
CLDN18.2 is a tight-junction protein that has attracted major attention as a therapeutic target in several gastrointestinal cancers, especially gastric and gastro-oesophageal junction cancers. In pancreatic ductal adenocarcinoma (PDAC), however, treatments aimed at CLDN18.2 have shown only limited benefit so far. This study helps explain why: the tumour’s genetic background and metabolic environment can change where CLDN18.2 sits in the cell and how well it can be targeted.
The researchers found that KRAS mutations, which are present in most PDAC cases, together with high blood sugar, promote a chemical modification called O-GlcNAcylation on CLDN18.2. This modification appears to trap CLDN18.2 inside the cell rather than allowing it to remain on the cell membrane, where CLDN18.2-targeted therapies are designed to bind. As a result, the cancer becomes more aggressive and less responsive to treatment.
Why CLDN18.2 Matters
CLDN18.2 is normally involved in maintaining tight junctions, which help cells stick together and preserve tissue structure. In cancer, abnormal CLDN18.2 expression can create an opportunity for targeted therapy, including antibody-based approaches that recognize CLDN18.2 on the tumour cell surface.
For these therapies to work well, CLDN18.2 must be present on the outer surface of the cancer cell membrane. If the protein is moved into the cytoplasm, its accessibility to therapeutic antibodies decreases, weakening the treatment effect. The study shows that this mislocalisation is not random, but driven by a specific molecular mechanism involving KRAS signalling and O-GlcNAcylation.
What the Researchers Studied
The investigation combined patient samples with multiple preclinical models, including humanised patient-derived xenografts, patient-derived organoids, orthotopic xenografts, KPC mice, and KPC-Cldn18.2 knockout mice. This multi-model approach allowed the team to examine how CLDN18.2 behaves in real human tumours and in experimental systems that closely resemble pancreatic cancer biology.
The study specifically asked three questions: first, whether O-GlcNAcylation changes CLDN18.2 localisation; second, whether this affects tumour progression and metastasis; and third, whether blocking this modification can restore sensitivity to CLDN18.2-targeted therapy.
Key Findings
The central discovery was that KRAS mutation and hyperglycaemia act together to drive O-GlcNAcylation of CLDN18.2 at threonine 204 (T204), a site near the C-terminal region of the protein. This modification promotes accumulation of CLDN18.2 in the cytoplasm instead of the cell membrane.
Once CLDN18.2 becomes O-GlcNAcylated, it supports cancer progression in several ways. It increases migration and invasion of pancreatic cancer cells and contributes to metastasis. At the same time, it reduces the effectiveness of anti-CLDN18.2-targeted therapies, because the target is less available at the cell surface.
The study also identified a mechanism behind this effect. Normally, the phosphatase PTP1B can bind to CLDN18.2 and help regulate its phosphorylation status. O-GlcNAcylation weakens this interaction, leading to increased tyrosine phosphorylation of CLDN18.2. The modified protein then recruits Src through Src’s SH2 domain, which triggers Src activation. Since Src is a well-known driver of cancer cell movement and invasion, this helps explain the aggressive tumour behaviour observed in the models.
How KRAS and High Glucose Interact
KRAS is one of the most common oncogenic drivers in pancreatic cancer, and mutations such as KRAS G12D help maintain tumour growth and survival. In this study, KRAS mutation did not act alone. Hyperglycaemia, a hallmark of diabetes and abnormal glucose metabolism, further enhanced CLDN18.2 O-GlcNAcylation.
This is important because many patients with pancreatic cancer also have diabetes or glucose intolerance. The findings suggest that metabolic conditions may influence how well targeted therapies perform by altering cancer cell surface proteins and signalling pathways. In other words, the tumour microenvironment and the patient’s systemic metabolism may both shape drug response.
Restoring Membrane CLDN18.2
A major practical question was whether the harmful effects of O-GlcNAcylation could be reversed. The answer was yes. The researchers showed that either genetic blockade, using a T204A mutation that prevents O-GlcNAcylation at that site, or pharmacological inhibition of O-GlcNAcylation restored CLDN18.2 localisation to the cell membrane.
When membrane localisation was recovered, tumour progression was suppressed. This suggests that the modification is not merely a marker of aggressive disease, but a functional driver of treatment resistance and metastasis.
Therapeutic Implications
The study has several clinically relevant implications. First, it provides a biological explanation for why CLDN18.2-targeted therapy may be less effective in some PDAC patients than in gastric cancer. Second, it suggests that evaluating tumour metabolism and KRAS status may help identify patients most likely to benefit from CLDN18.2-directed treatment.
Most importantly, the researchers found that low-dose MRTX1133, a KRAS G12D inhibitor, reduced CLDN18.2 O-GlcNAcylation and worked synergistically with CLDN18.2-targeted therapy in KRAS-mutant PDAC models. This combination improved treatment response while causing minimal side effects in the experimental setting.
MRTX1133 is not yet a routine standard treatment, but the concept is highly promising: instead of targeting CLDN18.2 alone, suppress the upstream KRAS-driven process that makes CLDN18.2 harder to reach. This kind of combination strategy may be especially relevant for cancers driven by strong oncogenic signalling and metabolic reprogramming.
What This Means for Patients
For patients, the study underscores that pancreatic cancer is not driven by one alteration alone. Genetic mutations, such as KRAS, and metabolic factors, such as elevated glucose, can work together to change how cancer behaves and how it responds to therapy.
If future clinical trials confirm these findings, testing for KRAS mutation status, CLDN18.2 expression pattern, and metabolic context may become more important in treatment planning. In particular, patients whose tumours express CLDN18.2 but show poor response to CLDN18.2-targeted therapy might benefit from a strategy that also addresses KRAS signalling or O-GlcNAcylation.
Limitations and Future Directions
Although the results are compelling, they come from a combination of patient-derived materials and animal models rather than a completed large-scale clinical trial. That means the findings need further validation in human studies before they can directly guide routine care.
Future research will need to clarify how common this modification is across different PDAC subtypes, whether blood glucose control can influence CLDN18.2 targeting, and which patients are most likely to benefit from KRAS inhibition plus CLDN18.2-directed therapy. It will also be important to identify the safest and most effective way to block O-GlcNAcylation in patients.
Conclusion
This study reveals a new mechanism by which KRAS-mutant pancreatic cancer becomes more aggressive and more difficult to treat. KRAS mutation and hyperglycaemia promote O-GlcNAcylation of CLDN18.2 at T204, causing the protein to accumulate in the cytoplasm, enhancing Src signalling, and reducing the effectiveness of CLDN18.2-targeted therapy.
Encouragingly, low-dose MRTX1133 can reduce this modification, restore CLDN18.2 to the cell membrane, and improve therapeutic response in preclinical models. These findings open the door to a more precise combination approach for KRAS-mutant PDAC, especially in patients whose tumours express CLDN18.2 but respond poorly to current targeted strategies.