Dissecting the Diagnosis and Prognosis of Hepatocellular Carcinoma by Circulating and Liver Transcriptomes

Abstract

Hepatocellular carcinoma (HCC) is the most prevalent form of primary liver cancer. Despite its aggressive nature and poor prognosis, patients diagnosed at an early stage can achieve significantly improved 5-year survival rates through curative treatments such as surgical resection or liver transplantation. Unfortunately, most HCC patients are diagnosed at advanced stages due to the lack of reliable and effective screening tools for early diagnosis, rendering curative treatments unsuitable for these patients. Alpha-fetoprotein (AFP) is the most widely used serum biomarker for HCC. Elevated AFP levels in adults can be a sign of liver regeneration due to malignancy. However, the sensitivity and specificity of AFP are suboptimal, especially for the detection of early-stage HCC. Recent researches have highlighted the potential of circulating microRNAs (miRNAs) as biomarkers for early cancer diagnosis. Combining circulating miRNAs with AFP has demonstrated promising results in improving the diagnosis of HCC. However, most findings from earlier studies cannot be reproduced in subsequent studies, suggesting the need for future studies to adopt a more standardized study design.

Tumor heterogeneity is a well-documented phenomenon in HCC, characterized by both intra- and inter-tumor heterogeneity. Different subtypes of HCC can exhibit diverse phenotypes and molecular profiles both within and across HCC tumors, complicating the disease’s management. This heterogeneity in HCC poses significant challenges for diagnosis, prognosis and treatment, since a single tumor biopsy may not accurately represent the entire tumor population within a patient or among different patients. Despite these challenges, studying heterogeneity in HCC offers substantial clinical value by paving the way for the development of effective targeted therapies and personalized treatment options. Advances in spatial technologies, such as spatial transcriptomics (ST) and metabolomics (SM), enable researchers to study multiple distinct regions from a single tumor independently. By examining these regions, researchers can identify subclonal populations and uncover spatial patterns of heterogeneity. The integration of findings from ST and SM facilitates a comprehensive examination of the interactions between spatial gene expressions and metabolite levels. This multi-layered approach provides a more nuanced understanding of tumor heterogeneity and insights into the mechanisms driving HCC progression and treatment resistance.

In this thesis, we aim to study both circulating and liver transcriptomes to establish the use of molecular biomarkers for the diagnosis and prognosis of HCC. The main findings of our works are summarized as follows:

(1)Plasma miRNA profiling of HCC patients and LC patients were performed in a multi-stage study. Differential expression analysis identified 15 upregulated miRNAs in HCC patients, most of which were associated with aggressive clinical characteristics such as advanced tumor stage and large tumor size. A 6-miRNA panel (miR-361-5p+ miR-130a-3p+ miR-27a-3p+ miR-148a-3p+ miR-30d-5p+ miR-193a-5p) combined with AFP was established as a minimal-invasive diagnostic tool to classify patients into high-risk and low-risk groups. Our established miRNA panel demonstrated superior diagnostic performance over using AFP alone, suggesting its potential to improve the current population screening for HCC.

(2)A systematic meta-analysis was conducted to compare the results of our first study with previous studies on HCC, aiming to identify circulating miRNAs that showed consistent differential expression across multiple studies. We evaluated the combined differential expressions of 13 circulating miRNAs reported in at least three studies with available expression data using the measure of effect size. In subgroup analyses based on different study designs, we found that the use of normalization control demonstrated the most significant impact on result heterogeneity. Overall, miR-122-5p, miR-21-3p, miR-192-3p and miR-29a-3p were identified as the 4 circulating miRNAs showing consistent upregulations across studies using miR-16 or spike-in as the normalization controls.

(3)The integration of ST and SM was employed to investigate the intra- and intertumoral heterogeneity in HCC. Targeted re-clustering of malignant region leads to the identification of a unique malignant cluster, referred as the malignant cluster 2, which exhibited a distinct molecular profile compared to other malignant clusters. Notably, the malignant cluster 2 demonstrated a refined architecture pattern across different samples and showed unique activation of regulatory pathways. This cluster was characterized by enriched copy number variations and was associated with poorer prognosis according to survival data from the TCGA-LIHA cohort. The integrated multi-omics data revealed significant upregulations in arginine and proline metabolism as well as glutathione metabolism within malignant cluster 2. These findings highlighted the distinct metabolic pathways activated in this cluster, distinguishing it apart from other identified malignant clusters.

Our findings provided valuable insights into using circulating miRNAs for diagnosing HCC and applying region-specific transcriptomes and metabolite levels to determine the prognosis of HCC patients. Future research focusing on dysregulated circulating miRNAs in primary liver, particularly within the refined architecture of HCC, could provide additional functional insights from these targeted miRNAs.

Date of Award	8 Apr 2025
Original language	English
Awarding Institution	City University of Hong Kong
Supervisor	M YANG (Supervisor)