Pancreatic cancer, attacked from three sides at once
Three preprints take on pancreatic cancer from different directions: a protein that organises the tumour's surroundings, a pig model that reproduces the human disease, and a drug pair that hits it in combination. A fourth preprint, on brain cancer, underlines the same lesson about place.
- In a preprint, B7-H3 acts as a spatial organiser of the pancreatic cancer microenvironment, with its expression linked to a shift toward fibro-inflammatory immune states. Self-reported m:…7821
- In a preprint, genetically engineered KPC pigs show accelerated pancreatic cancer progression, including precursor lesions, invasive carcinoma, perineural invasion and lymphatic spread. Self-reported m:…0595
- In a preprint, TR-107 reduces viability in pancreatic cancer organoids and cell lines at low-nanomolar concentrations and synergises with the AKT inhibitor capivasertib. Self-reported m:…3748
- In a preprint on IDH-wildtype glioblastoma, proximity to the lateral ventricles is associated with shorter survival. Self-reported m:…0596
The neighbourhood runs the tumour
A preprint on bioRxiv argues that a single protein, B7-H3, acts as a spatial organiser of the pancreatic cancer microenvironment. Where B7-H3 is expressed, the tumour's architecture shifts from an epithelial-intrinsic pattern to a fibro-inflammatory immune state. The fibro-inflammatory state linked to B7-H3 was also conserved within the neoadjuvant landscape, the setting in which tumours are studied around treatment given before surgery. m:…7821
The finding matters because it treats the tumour's surroundings as something organised rather than accidental. B7-H3, in this account, is not merely a marker found at high levels in pancreatic cancer but the organiser of the fibro-inflammatory state itself: the arrangement of fibroblasts and immune cells around the tumour, and the immune suppression that comes with it. m:…7821
A pig that gets the human disease
The second front is the model itself. A preprint describes a genetically engineered pig model of pancreatic ductal adenocarcinoma, built with KRASG12D and TP53R167H mutations, that shows accelerated progression: precursor lesions known as PanINs, invasive carcinoma, invasion along nerves and spread through the lymphatic system. The tumours in these pigs carry an organised immune landscape, with infiltrating effector and regulatory T cells. m:…0595
The model also gives drug researchers something to work with. Cell lines derived from the pig tumours, called PoPaCa lines, showed sensitivity to the GPX4 inhibitor RSL3 and partial responsiveness to trametinib. That is the test of a model: not whether it looks like the human disease under the microscope, but whether it behaves like it when a drug is applied. m:…0595
Two drugs, one combination
The third front is the drug combination. A preprint reports that TR-107, a compound that activates the mitochondrial protein ClpP, reduces the viability of pancreatic cancer organoids and cell lines at low-nanomolar concentrations, and lowers expression of ClpX in a dose-dependent manner. Crucially for a cancer whose tumours are starved of oxygen, TR-107 kept its effect and its mechanism of action under hypoxic conditions. m:…3748
Paired with capivasertib, an AKT inhibitor, TR-107 synergised to suppress the viability of pancreatic cancer cells: two drugs hitting mitochondrial and survival signalling at once. It is the kind of combination the pig model exists to test, and the kind of result the B7-H3 work suggests will also have to contend with the tumour's neighbourhood. m:…3748 m:…0595
Location matters beyond the pancreas
A fourth preprint, on a different cancer, underlines the same lesson about place. In IDH-wildtype glioblastoma, tumours sitting close to the brain's lateral ventricles were associated with shorter survival. Those ventricular-proximal tumours carried an inflammatory yet immunoregulatory profile, with antigen-presentation and checkpoint transcripts, and proteomic analysis found matching complement and myeloid signals. m:…0596
They also showed more proliferating Ki67-positive and SOX2-positive stem-like cells, without a corresponding rise in CD4-positive or CD8a-positive T cells. The geography of the tumour, in other words, tracks with both its biology and its outcome. m:…0596
None of this is a treatment. All four studies are preprints, and the pancreatic findings come from tissue studies, animal models and organoids in dishes, not from patients. What they share is a direction: the next attack on pancreatic cancer will have to deal with the tumour's neighbourhood, test its ideas in models that reproduce the human disease, and combine drugs that strike different machinery, remembering that where a tumour sits can decide how it behaves. m:…7821 m:…0595 m:…3748 m:…0596
What is still unresolved
- Whether blocking or modulating B7-H3 changes the fibro-inflammatory state, or only marks it.
- Whether the pig model predicts drug responses in human pancreatic cancer better than existing mouse models.
- Whether the TR-107 and capivasertib combination holds up in animals, and at doses patients could tolerate.
Provenance by Synorb. Every claim links to its source. This story was written by an agent from the Manifests listed below. Read the methodology and limitations, or send a correction.
Free to reuse, including commercially. Credit “Hanging Context, with provenance by Synorb” and link to this story. Keep the cited Manifest ids. Reuse terms v1.