A new spatial map of pancreatic precancer reveals a striking disconnect: PanIN cells can look increasingly cancer-like, yet their surrounding tissue may lack the stromal and immune changes needed to support malignant transformation.
Study: Asynchronous evolution of epithelium and stroma differentiates precursor lesions from pancreatic cancer. Image Credit: Nemes Laszlo / Shutterstock
In a recent study published in the journal Cancer Discovery, researchers found that epithelial cells in pancreatic intraepithelial neoplasia (PanIN) gradually resembled cancer cells. However, the surrounding tissue environment markedly differed until late-stage transformation. The lack of full stromal reprogramming may therefore help explain why most PanIN lesions do not progress to pancreatic cancer.
Pancreatic Ductal Adenocarcinoma (PDAC) remains a lethal condition globally since it is often diagnosed at late stages, when early disease has gone undetected, and cancer cells may already have begun to spread.
Scientists have been focusing on PanIN, the microscopic precursor lesion responsible for most pancreatic cancers. But these lesions are difficult to study in humans since they require invasive procedures and are rarely obtained from healthy individuals. This is why researchers mainly rely on samples obtained from tumor-adjacent regions or genetically engineered mouse models, which differ in molecular and anatomical characteristics.
About the study
In the present study, researchers analyzed healthy donor pancreata obtained through a collaboration with the Gift of Life Michigan. The broader donor program accrued more than 150 pancreata from individuals aged 20 to 70 years. They also obtained pancreatic cancer samples from the University of Michigan. Advanced spatial transcriptomics and single-cell RNA sequencing (scRNA-seq) mapped cellular and molecular changes underlying disease progression.
The researchers used advanced spatial gene-mapping technologies to examine 11 PanIN-containing donor samples, seven PDAC samples, and two adjacent-normal pancreas samples from cancer patients. Board-certified gastrointestinal pathologists reviewed all samples. They improved resolution by combining spatial transcriptomics with a comprehensive scRNA-seq dataset comprising over 200,000 cells from healthy, adjacent normal, and pancreatic cancer tissues.
The researchers used computational tools to identify epithelial, immune, and fibroblast cells in the tissues. These tools performed tasks such as spot deconvolution, spatially informed clustering, pseudotime analysis, and ligand-receptor interaction analysis to analyze the samples.
The team also studied protein expression using immunofluorescence. They analyzed publicly available Xenium and CosMx datasets to confirm their findings. The team also developed three-dimensional organoid-fibroblast mini models in the laboratory using patient-derived pancreatic cancer organoids and fibroblasts. These models simulated tumor cell interactions with surrounding stromal cells.
They then assessed macrophage responses to tumor-derived signaling molecules. These approaches allowed researchers to compare cells and their surrounding environments in healthy tissues, precancerous lesions, and invasive pancreatic cancer.
Results
The epithelial cells within PanIN lesions showed molecular similarities with pancreatic cancer, despite remaining non-invasive. Spatial transcriptomic analyses revealed that PanIN epithelial cells exhibited gradual, progressive changes that increasingly resembled those of pancreatic cancer. In particular, as the disease advanced, these cells showed increases in RAS/MAPK or KRAS pathway activity, inflammatory pathways, hypoxia, and specific metabolic changes, including reduced oxidative phosphorylation and fatty acid metabolism.
Markers indicative of aggressive disease also became elevated in poorly differentiated tissues. This included Janus kinase (JAK) and signal transducer and activator of transcription 3 (STAT3) signaling molecules. Tumor genes such as keratin 17 (KRT17) and Wnt family member 7a (WNT7A) were also elevated.
Although the epithelial cells showed tumor-like changes, the surrounding microenvironment of PanIN lesions behaved more like healthy pancreatic tissue than a tumor. The area adjacent to PanIN lesions closely resembled normal pancreatic ducts. There were many plasma cells, or immune cells that produce antibodies, around PanIN lesions. In addition, macrophages, which can promote inflammation and tumor growth, did not cluster close to the PanIN lesions and remained distant from the epithelium.
These findings suggest a potentially protective, controlled immune environment surrounding PanIN lesions, which may serve as an important barrier to their progression to cancer. Tumor cells showed the opposite pattern, with macrophages clustered near cancer cells and plasma cells dispersed throughout the tissue.
The team found that a tumor-enriched fibroblast population expressed specific proteins in high amounts. Examples include smooth muscle actin (SMA) and leucine-rich repeat-containing 15 (LRRC15). Such fibroblasts were associated with a poor prognosis.
The lab-grown organoids demonstrated that cancer cells triggered these fibroblasts to switch on lymphoid enhancer-binding factor 1 (LEF1), a gene involved in the wingless-related integration site (Wnt) signaling pathway, which is linked to cancer growth and tissue remodeling. Such actions were associated with cancer-associated molecular changes and a shift toward a more aggressive basal tumor-like phenotype.
Researchers also found that over 60% of healthy donor pancreata contained PanIN lesions. This suggests that the precursor lesions are surprisingly common but rarely progress to cancer.
Conclusions
The findings suggest that although PanIN lesions show some tumor-like changes, they have not yet developed the cancer-promoting microenvironment seen in pancreatic tumors. This may help explain why most PanIN lesions do not progress to cancer.
Future studies will require more advanced technologies to study pancreatic tissue at single-cell resolution. This will help understand how molecular pathways, such as WNT signaling, influence cells surrounding PanIN epithelial lesions. Such efforts could improve understanding of the natural barriers that may restrain the progression of precursor pancreatic lesions to cancer, which could be targeted to inform cancer interception strategies and lower cancer burden.
Download your PDF copy by clicking here.
ANU researchers map hidden cellular networks to better understand diseases