RIKEN Researchers Develop Predictive Score for Liver Cancer Risk

Researchers at Japan's RIKEN Center for Integrative Medical Sciences (IMS), led by Xian-Yang Qin, have made a significant breakthrough in liver cancer research. Their study, published in Proceedings of the National Academy of Sciences, identifies the protein MYCN as a driver of liver tumorigenesis, specifically in the most lethal subtype of liver cancer.

The research details the microenvironment of genes that facilitate MYCN overexpression and describes a machine-learning algorithm.

This algorithm uses the collected data to predict the likelihood of tumor development in a liver currently free of tumors.

Addressing the High Mortality of Liver Cancer

Liver cancer, or hepatocellular carcinoma, is a formidable global health challenge, contributing to over 800,000 global deaths annually. Its high mortality rate is largely attributed to late detection and a recurrence rate of 70-80%. Qin's team embarked on this investigation into the MYCN protein with the goal of developing a method for predicting at-risk livers before tumor formation.

While the MYCN gene has been recognized for its role in liver cancer development from damaged livers, its exact mechanism remained unclear.

Unraveling MYCN's Role in Tumorigenesis

The team theorized that direct overexpression of MYCN could lead to liver tumorigenesis, making it a potential biomarker. To test this, they inserted MYCN into mouse liver genomes, resulting in its overexpression.

When MYCN was overexpressed alongside always-active AKT, an astonishing 72% of the mice developed liver tumors within 50 days. These tumors exhibited characteristics consistent with human hepatocellular carcinoma. Significantly, tumors did not form when either gene was overexpressed independently, underscoring the synergistic effect.

Discovering the "MYCN Niche" with Spatial Transcriptomics

To understand the early microenvironmental cues that foster tumor development, researchers employed spatial transcriptomics, a cutting-edge technique that maps gene activity within tissues. In a mouse model of metabolic dysfunction-associated liver cancer, they monitored gene expression over time and location as tumors progressed, focusing on areas with increasing MYCN levels.

This meticulous analysis led to a pivotal discovery: a cluster of 167 genes, termed the "MYCN niche", that were differentially expressed in tumor-free liver sections exhibiting elevated MYCN levels.

A Machine-Learning Model for Early Prediction

Based on this comprehensive mouse spatial transcriptomics data, a sophisticated machine-learning model was developed. This model can analyze gene-expression patterns and generate a score indicating the presence of a MYCN niche with 93% accuracy.

Translating to Human Health: Predicting Recurrence and Outcomes

The MYCN niche score was subsequently applied to human hepatocellular carcinoma datasets. The results were compelling: patients with higher MYCN niche scores demonstrated an increased risk of tumor recurrence and poorer clinical outcomes.

Crucially, this correlation was more pronounced when the score was derived from non-tumor tissue rather than tumor tissue.

The score thus functions as a spatial biomarker for predicting prognosis based on microenvironments conducive to tumor formation.

Xian-Yang Qin affirmed the significance of their findings: "Our team developed a clinically actionable strategy to identify high-risk patients through gene expression profiling in non-tumor liver tissue. By integrating spatial transcriptomics with machine learning, we established a MYCN niche score that predicts recurrence risk and identifies precancerous microenvironments susceptible to de novo liver tumorigenesis."

Future work will aim to further explore the biological mechanisms behind these scores and how cancer-permissive environments are established and maintained.