MYCN gene expression as an indicator for predicting the risk of liver cancer development

A research group led by Senior Research Scientist Xian-Yang Qin and Team Director Harukazu Suzuki from RIKEN Center for Integrative Medical Sciences (IMS), together with researchers from the Chiba Cancer Center Research Institute, Gifu University, and the Jikei University School of Medicine, has announced that the transcription factor MYCN protein drives the development of primary liver cancer. The group also developed a spatial feature score reflecting the spatial distribution of MYCN gene expression, which can predict the risk of liver cancer development. These results are expected to improve the prognosis and quality of life (QoL) of liver cancer patients through early intervention. The results were published in PNAS on February 18.

Global deaths from liver cancer exceeded 830,000 in 2020, roughly double the figure from two decades earlier. With the recurrence rate within five years of liver cancer treatment reported at 79%, the development of technologies for early detection and prevention has become an urgent priority.

The liver is composed of parenchymal cells (hepatocytes) and non-parenchymal cells such as immune cells, fibroblasts, and vascular cells, which interact with one another in a complex manner to form a "tumor microenvironment." In chronic liver disease, repeated cycles of hepatocyte necrosis and abnormal regeneration transform the entire liver into a microenvironment that promotes cancer development, leading to multicentric recurrence-the emergence of tumors at multiple sites.

The research group had previously searched for candidate factors governing responsiveness to acyclic retinoid, a compound that suppresses liver cancer recurrence, and identified the oncogene MYCN as a target of acyclic retinoid. The researchers also developed a method to quantify MYCN levels in blood and demonstrated that blood MYCN levels serve as a useful biomarker for selecting patients who respond to acyclic retinoid. MYCN is specifically expressed in liver cancer stem cells, and while the MYCN gene itself carries no mutations, its expression level is associated with the malignancy of liver cancer. At the same time, MYCN gene expression is low in normal liver tissue but shows a transient peak during liver regeneration, particularly during the active hepatocyte division phase 48 hours after partial hepatectomy.

The group therefore investigated whether elucidating the gene expression dynamics of MYCN and its regulatory mechanisms within the liver cancer microenvironment could enable a distinction between physiological liver regeneration and pathological liver tumorigenesis.

First, the MYCN gene was introduced into mouse livers, confirming that this promoted liver tumor formation.

Next, using spatial transcriptomics, the group tracked spatiotemporal changes in MYCN gene expression in liver tissue from a mouse model of liver tumorigenesis and identified regions of high MYCN expression displaying signals characteristic of liver cancer stem cells.

The frequency of these regions increased as tumorigenesis progressed, and they were found to be spatially adjacent to areas with high expression of AFP, a liver cancer tumor marker. Marked activation of pathways involved in maintaining cancer stem cell properties was confirmed within the MYCN niche. Accumulation of macrophages was also observed in the surrounding area.

Furthermore, culture supernatant from inflammatory macrophages was found to induce MYCN expression in hepatocytes, while MYCN overexpression in hepatocytes reduced the responsiveness of those cells to macrophage stimulation. This revealed a feedback loop control mechanism in which MYCN expression and inflammatory responses mutually influence each other. The group also developed a "MYCN score" using a machine learning algorithm to quantify the spatial features of the MYCN microenvironment.

When this formula was applied to human hepatocellular carcinoma gene expression databases, the MYCN score in non-tumor tissue was found to correlate strongly with the risk of liver cancer recurrence. The findings showed that when a MYCN-like tumor-promoting microenvironment is present in non-tumor tissue, liver cancer is more likely to recur in the future.

Qin commented: "The development of epithelial tumors, including liver cancer, results from complex interactions among diverse factors in the microenvironment, and fully understanding that process has been a major challenge. In this study, we succeeded in quantitatively assessing the overall picture of the liver tumor microenvironment by combining spatial omics with machine learning. Going forward, we aim to advance the clinical application of the spatial feature score as a biomarker and to uncover the true nature of liver tumorigenesis as captured by the machine learning-derived formula."

Journal Information
Publication: PNAS
Title: Oncogenic function and transcriptional dynamics of MYCN in liver tumorigenesis
DOI: 10.1073/pnas.2521923123