Single Nuclear RNA Sequencing of Pituitary Neuroendocrine Tumors Challenges Their Current Classification Schema

Introduction: While the 2022 WHO classification of pituitary neuroendocrine tumors (PitNETs) relies on staining for lineage-specific transcription factors (TFs; PIT1, TPIT and SF1), this classification assumes PitNET oncogenesis reflects normal pituitary development. Classification refinement could improve understanding and treatment of PitNETs.

Methods: We conducted the largest single-nucleus RNA sequencing analysis of PitNETs to date on 419,874 cells from two nontumor pituitaries and 11 PitNETs (somatotrophic, hormonally active and silent corticotrophic (SCAs), lactotrophic, silent gonadotroph (SGA), and null cell (NCA)).

Results: TF analysis identified previously undescribed secretory cells in normal and PitNET samples arising from stem populations with mixed identities (GH1/PRL/POMC/DIO2 positive). Tumor clustering showed that despite their TPIT lineage, SCAs grouped closer to SGAs than other TPIT-positive functional corticotroph adenomas. Trajectory analysis also revealed a diversity of cellular origins that underscored the limitations of a lineage-based classification - while most subtypes derived from their normal neuroendocrine counterparts and had significant copy number variations (CNVs), somatotroph and null cell PitNETs had minimal CNVs and arose from stem cells. We identified key transcription factors specific to the development of each PitNET subtype - somatotrophic (HOXB4, NKX2, NPAS4), secretory corticotrophic (TFCP2L1, BNC2, CUX2), NCA (ONECUT, ZNF804A, MKX), lactotroph (ETV5), SCAs (DMRTA2, IRF6, and EBF1) and SGAs (TCF7). Developmental maturity also varied among subtypes, where secretory corticotrophic PitNETs appeared earlier in development than SCAs and expressed genes associated with cell morphogenesis and development.

Conclusion : We discovered cell populations that carry multiple TF signatures and demonstrate how greater complexity and fluidity is needed for PitNET classification. We highlight the importance of underexplored metrics such as CNV instability, cellular origin, and developmental maturity among PitNETs. These findings represent important initial steps towards the development of a more robust classification schema rooted in cellular development and trajectory that would better account for PitNET biology.