Abstract:
Tumor development and progression are shaped not only by the intrinsic properties of malignant cells but also by their complex and spatially organized interactions with the surrounding microenvironment. In this thesis, I established and advanced a high-dimensional imaging platform based on the MACSima imaging cyclic staining (MICS) technology for deep immune profiling of tumor tissue. By designing and validating a panel comprising over 121 markers and optimizing imaging and analysis workflows, we successfully applied the platform to hepatocellular carcinoma (HCC), uncovering spatially structured immunosuppressive niches.
Building on this foundation, the platform was adapted to study formalin-fixed paraffin-embedded (FFPE) samples of NUT carcinoma (NC), a rare and aggressive cancer driven by chromosomal rearrangements involving the NUTM1 gene. Due to the rarity of NC, tissue samples are scarce and often limited to FFPE material. A cohort comprising 16 tumor specimens from 14 patients was assembled, and a highly detailed spatial single-cell proteomics analysis of NC was performed. In addition, complementary technologies including immunohistochemistry (IHC), whole-genome sequencing (WGS) and RNA-sequencing (RNA-Seq), and flow cytometry were employed, enabling one of the most comprehensive molecular characterizations of this rare malignancy to date across the DNA, RNA, and protein levels. The data revealed significant heterogeneity within the tumor immune microenvironment (TIME), including “hot” immune-infiltrated and “cold” immune-excluded phenotypes, as well as immunoregulatory interactions between myeloid-derived suppressor cells (MDSCs) and CD8+ T cells. Furthermore, multi-omic profiling identified the immune checkpoint molecule CD276 as a consistently overexpressed and actionable therapeutic target in NUT carcinoma (NC), which was validated functionally, both in vitro and in xenograft models.
Overall, this work demonstrates the feasibility and value of spatial single-cell profiling in prevalent and rare cancers and provides novel biological insights and therapeutic implications for NC. This thesis presents a flexible and scalable framework for deep spatial tissue analysis, advancing the field of spatially resolved cancer research. By enabling comprehensive characterization of tumor ecosystems, it aims to support translational efforts and guide the development of future therapeutic strategies.