Single-cell RNA-sequencing and spatial transcriptomic analysis reveal a distinct population of APOE- cells yielding pathological lymph node metastasis in papillary thyroid cancer

Thyroid cancer is one of the most common endocrine tumors worldwide, especially among women and the metastatic mechanism of papillary thyroid carcinoma remains poorly understood.

We identified a distinct APOE- tumor cell population associated with cervical metastasis and poor prognosis. Our results and models have potential clinical, prognostic, and therapeutic implications for advanced PTC. Key points: A subpopulation of tumor cells with lower expression levels of APOE was strongly associated with more advanced stages and metastasis of PTC. APOE-negative (APOE-) cellsoverall exhibited weaker interactions with immune cells. A machine-learning bioinformatics model based on scRNA-seq data of in-situ thyroid cancer tissue was established to predict lymph node metastasis.

Thyroid cancer tissue samples were obtained for single-cell RNA-sequencing and spatial transcriptomics, aiming to intratumoral and antimetastatic heterogeneity of advanced PTC. The functions of APOE in PTC cell proliferation and invasion were confirmed through in vivo and in vitro assays. Pseudotime analysis and CellChat were performed to explore the the molecular mechanisms of the APOE in PTC progression.

We identified a subpopulation of tumor cells with lower expression levels of APOE, associated with advanced stages of PTC and cervical metastasis. APOE overexpression significantly reduced tumor cell proliferation and invasion, both in vitro and in vivo, by activating the ABCA1-LXR axis. APOE- tumor cells may promote tumor growth by interacting with dendritic cells and CD4+ T cells via CD99- rather than CD6-regulated signaling. We established a machine learning-based scRNA-seq data, 13-gene signature predictive of lymph node metastasis. Conclusions: We identified a distinct APOE- tumor cell population associated with cervical metastasis and poor prognosis. Our results and models have potential clinical, prognostic, and therapeutic implications for advanced PTC. Key points: A subpopulation of tumor cells with lower expression levels of APOE was strongly associated with more advanced stages and metastasis of PTC. APOE-negative (APOE-) cellsoverall exhibited weaker interactions with immune cells. A machine-learning bioinformatics model based on scRNA-seq data of in-situ thyroid cancer tissue was established to predict lymph node metastasis.