Endocrine Abstracts

Background: While gonadotroph tumors are lacking both medical treatment options and prognostic markers, the tumor microenvironment represents a promising tool. The aim of this study was to 1) identify new tumor microenvironment actors in gonadotroph tumors, 2) gain insight into the tumorigenesis mechanisms driven by these tumor microenvironment cells, and 3) identify new tumor microenvironment-related prognostic markers, as well as new therapeutic targets.

Methods: Single-cell transcriptomics of six gonadotroph tumors, spatial transcriptomics of three gonadotroph tumors, followed by immunohistological validation on 49 gonadotroph tumors.

Results: 24,471 cells from the six tumors passed the quality control. Clustering analysis revealed expected tumor and tumor microenvironment cell types, as well as mast cells as a new population in pituitary tumors. Regarding mast cells’ phenotype, 132/133 mast cells expressed tryptase, of which three co-expressed chymase; no mast cell expressed only chymase. Immunohistochemistry for tryptase, followed by analysis of whole slides, confirmed mast cells were present in all 49 gonadotroph tumors (0.01-4.33%). Next, we investigated the ligand-receptor networks. Using CellPhoneDB, bilateral crosstalk was identified between mast and endothelial cells. The top ranked secreted ligand from mast cells signaling towards endothelial cells was VEGFA, suggesting mast cells have a pro-angiogenic role in gonadotroph tumors. In line with this hypothesis, there was a very strong and positive correlation between the percentages of endothelial cells and of mast cells in the 6 tumors (rho=0.94, P=0.01). In addition, analysis of spatial transcriptomics data showed mast cells were usually found in the vicinity of endothelial cells, which we confirmed with double immunofluorescence staining (tryptase/CD34). Furthermore, mast cells expressed more frequently VEGFA than TNF, suggesting a pro-tumorigenic role for mast cells in gonadotroph tumors. Indeed, mast cells were associated with a bad prognosis in the clinical cohort, notably tumors that progressed/relapsed more rapidly after surgery had more abundant mast cells (P=0.002). Moreover, for a same patient, recurrent tumors had more mast cells than the initial tumor (P=0.007). Finally, Nichenet analysis identified KITLG (acting via KIT) as one of the probable up-regulators of VEGFA expression in mast cells. Interestingly, the top ranked interaction identified between endothelial and mast cells was also KITLG-KIT.

Conclusion: We identified mast cells as a new population in gonadotroph tumors. In these tumors, mast cells appear to have pro-angiogenic and pro-tumorigenic roles. Moreover, mast cells were associated with a bad prognosis in the clinical cohort, suggesting that KIT targeting might be beneficial for the treatment of gonadotroph tumors.