| Author: | Huang, Mohan |
| Title: | Hypoxia-induced modulation of immunotherapy efficacy in hepatocellular carcinoma |
| Advisors: | Chan, Wing-chi Lawrence (HTI) Tai, William (ABCT) Zou, Xiang (HTI) |
| Degree: | Ph.D. |
| Year: | 2025 |
| Subject: | Liver -- Cancer -- Immunotherapy Liver -- Cancer -- Genetic aspects Drug resistance in cancer cells Hong Kong Polytechnic University -- Dissertations |
| Department: | Department of Health Technology and Informatics |
| Pages: | 179 pages : color illustrations |
| Language: | English |
| Abstract: | Background Hepatocellular carcinoma (HCC) is a leading cause of cancer-related mortality, with most cases diagnosed at an advanced stage, making immunotherapy a key treatment strategy. However, the response rate to PD-L1 inhibitors remains low, necessitating further exploration of resistance mechanisms and predictive biomarkers. Hypoxia is a major contributor to immunotherapy resistance, as HIF-1α upregulates PD-L1 expression and activates genes that help tumor cells adapt to hypoxia, ultimately reducing immunotherapy efficacy. This study integrated bioinformatics, machine learning, and deep learning to identify key hypoxia-associated genes and pathways contributing to PD-L1 expression. A hypoxia risk score model was developed to stratify cases by risk, and a Kolmogorov-Arnold Network (KAN) deep learning model was constructed to predict immunotherapy response. Additionally, an in vitro hypoxia-induced drug-resistant HepG2 cell model was established, and the role of NOXA in apoptosis regulation was examined through flow cytometry and AI-based image analysis. Results and Conclusion 52 HCC-Hypoxia Overlap genes (HHOs) were identified, with 14 PD-L1 regulatory genes and 10 hub genes influencing immunotherapy response. PMAIP1 (NOXA) was significantly associated with immunotherapy response (p < 0.001). A hypoxia risk score model integrating PMAIP1 and 9 hypoxia risk-associated genes demonstrated high predictive accuracy (AUC = 0.815, 0.774, 0.771 for 1-, 2-, and 3-year survival, respectively). The KAN deep learning model incorporating 11 key genes achieved high predictive accuracy (AUC = 0.936 training, 0.7 test). SVM-based integration of hypoxia risk score and KAN model improved prediction performance (AUC = 0.725 test set). Experimental validation demonstrated that hypoxia enhances drug resistance in HepG2 cells, while NOXA knockdown alters apoptosis patterns, potentially modulating treatment response. These findings highlight NOXA as a potential therapeutic target and establish a robust model for predicting immunotherapy response, advancing precision medicine in HCC treatment. |
| Rights: | All rights reserved |
| Access: | open access |
Copyright Undertaking
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item:
https://theses.lib.polyu.edu.hk/handle/200/13858