Abstract 5288: Identification of cancer cells with metastatic potential in nasopharyngeal carcinoma through spatial multi-omics analysis

Abstract Background: Metastasis is the leading cause of nasopharyngeal carcinoma (NPC) mortality. The understanding of the process of metastasis is inadequate. Objectives: To identify the high-metastatic-potent (HMP) cluster in NPC using a set of key genes based on spatial multi-omics analysis. Method: 238 NPC samples, 110 from Hong Kong (HK) and 128 from Guangzhou (GZ), were classified into the metastatic (MET) group if the patient developed a remote metastasis within 5 years after the diagnosis of NPC, and into the non-metastasis (nonMET) group vice versa. Two HK samples (MET vs. nonMET: 1 vs. 1) were profiled by single-cell spatial transcriptome sequencing (scST) to explore the HMP cluster and its signature genes. Twenty-eight GZ samples were utilized for spatial transcriptome (GZ_ST: 14 vs. 11) and spatial proteome (GZ_SP: 17 vs. 11) profiling to streamline the signature genes into a final gene list. Validation at the transcriptome level and protein level was accomplished by bulk RNA sequencing (HK: 39 vs. 38; GZ: 41 vs. 59) and spatial proteome sequencing (HK_SP: 19 vs. 12), respectively. Independent validation was performed by multiplex immunohistochemical staining (mIHC). Results: Ten epithelial clusters from tumor cells were defined by cell segmentation and deconvolution in scST. Three clusters exhibited significant enrichment in the MET samples generating 69 signature genes for HMP. Gene set enrichment analysis (GSEA) in GZ_ST indicated a substantial activation of HMP signature in the MET group (two-sided p = 0.044). Thirty-three leading-edge genes in HMP signature were selected to evaluate the expression levels in scST, GZ_ST, and GZ_SP. Nineteen genes were included in the final gene list as their expressions were higher than average, upregulated in the MET group, and specific to HMP clusters. GSEA analysis in bulk RNA sequencing showed that the 19 HMP genes were activated in the MET group (one-sided p = 0.032). The lactate dehydrogenase A (LDHA)/enhancer of zeste homolog 2 (EZH2)/insulin-like growth factor 1 receptor (IGF1R) regulatory axis may play a key role in tumor metastasis by affecting tumor invasion. Pearson correlation analysis in bulk RNA sequencing displayed a positive correlation between HMP cluster and several lytic EBV genes transcribed from the BART regions (r > 0.4, p < 0.05), suggesting the potential role of EBV lytic reactivation in promoting metastasis. Conclusions: The high-metastatic-potent cluster was identified through this spatial multi-omics study. The selected biomarkers well define this cluster and enlighten the underlying metastatic mechanisms and its relevance of EBV lytic reactivation. Acknowledgement: This work was supported by General Research Fund (17102619) and Theme-based Research Scheme (T12-703/22-R; T123-70323-N) from Research Grant Council, and Technology Program of Guangdong Province in China (2021A0505110010). Citation Format: Songran Liu, Zhaozheng Hou, Kazi Anisha Islam, Yingpei Zhang, Larry Ka-Yue Chow, Ying Wang, Ping Feng, Michael King Yung Chung, Kiu-Wai Cheng, Dittman Lai-Shun Chung, Yunfei Xia, Wei Dai. Identification of cancer cells with metastatic potential in nasopharyngeal carcinoma through spatial multi-omics analysis [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 5288.