Background

Pancreatic cancer is classified as a fatal disease with a 5-year survival rate of approximately 10%. This is due not only to the diverse genetic mutations observed in patients but also to the rapid development of resistance by cancer cells even when targeted therapies are administered. Previous studies have primarily attempted to characterize cancer by analyzing messenger RNA (mRNA) expression levels in cancer cells. However, genomic-level gene expression data do not directly reflect the active states of proteins, which are critical in determining cellular behavior. As a result, it has been difficult to elucidate the precise mechanisms by which cancer cells evade surveillance and survive during drug treatment.

The plasticity of cancer cells to switch survival states in response to drug attacks is a key factor that hinders pancreatic cancer treatment. This plasticity is often cited as the reason why monotherapies with single drugs are easily rendered ineffective. Therefore, a new approach was needed to identify common survival states across all pancreatic cancer cells, regardless of the type of genetic mutation, and to uncover the core protein networks that regulate these states.

Key Findings

A research team led by Professor Andrea Califano at Columbia University applied the VIPER (Virtual Inference of Protein activity by Enriched Regulon) algorithm, a single-cell protein activity analysis method, to precisely analyze the states of pancreatic cancer cells. The team discovered that nearly all pancreatic ductal adenocarcinoma (PDAC) cells, regardless of their genetic mutation profiles, consist of six distinct molecular states. These six cellular states are driven by master regulator (MR) proteins, which form intracellular protein interaction networks and serve as the source of plasticity that enables rapid adaptation of cancer cells to therapeutic environments.

The six states are categorized into three major developmental lineages: gastrointestinal-like state (GLS), morphogenesis and epithelial-mesenchymal transition state (MOS), and alveolar-like state (ALS). Each lineage is further divided into high- and low-activity substates based on the activation levels of the MAPK signaling pathway, resulting in a total of six states. The research team demonstrated that these cellular states are not fixed but can flexibly transition in response to drug stimuli or environmental changes. When a single-drug therapy is initiated, some cells may die, but the remaining cells rapidly switch to alternative survival states to acquire resistance.

This discovery is considered a breakthrough that overcomes the limitations of conventional genomic analysis, which has focused solely on individual gene mutations. While traditional gene expression comparison methods could not detect the subtle changes in the six cellular states, the network analysis centered on protein activity enabled the precise mapping of the complex transformation pathways of cancer cells.

Significance and Prospects

This study suggests that the mechanism of cellular state transitions in cancer cells is highly conserved, going beyond individual patient mutation profiles. This insight lays the foundation for developing universal combination therapies that can simultaneously suppress the six common cellular states, overcoming the limitations of personalized gene-targeted therapies. By designing drug combinations that regulate specific MR proteins, it may become possible to establish therapeutic strategies that block the plastic escape routes of cancer cells.

However, challenges remain before clinical application. The safety and in vivo delivery efficiency of multi-drug combinations that organically regulate the six cellular states must be validated. Technological advancements are needed to minimize toxic responses in patients while delivering multiple drugs simultaneously to the target site. The research team plans to validate optimal drug combinations using patient-derived organoid models and animal experiments to assess clinical feasibility.