Immune landscape characterization of neurofibromas with atypical features in Neurofibromatosis1 reveals PD-1 and the Tim-3/Galectin-9 pathway as potential therapeutic targets.

Abstract

Neurofibromatosis type 1 (NF1) is a genetic disorder that predisposes individuals to the development of plexiform neurofibromas (pNF), benign tumors of the nerve sheath that can progress to malignant peripheral nerve sheath tumors (MPNST). Some pNFs exhibit atypical characteristics, particularly atypical neurofibromatous neoplasms of uncertain biological potential (ANNUBP), which meet recently defined criteria. These are pNFs with concerning histological and/or molecular features and uncertain behavior, leading to their consideration as potential precursors of MPNSTs. In this study, we characterized the immune landscape of a series of atypical neurofibromas (exhibiting increased cellularity or atypical cells) and ANNUBPs (ANF/ANNUBP) compared to cutaneous neurofibromas (cNF), pNFs, and MPNSTs, all from patients with NF1, using immunohistochemistry and single-cell RNA sequencing (scRNA-seq). Our findings revealed that ANF/ANNUBPs are heavily infiltrated with CD3+ T-lymphocytes and CD163+ macrophages. Transcriptomic profiling using the NanoString technology demonstrated that ANF/ANNUBPs exhibit a distinct immune signature compared to both pNFs and MPNSTs. This includes overexpression of two inhibitory immune checkpoints (ICs): HAVCR2 (encoding the IC receptor Tim-3) and its primary ligand Galectin-9 (Gal9). Subsequent analyses of scRNA datasets corroborated these findings with Prss56^Cre/+, Nf1^fl/fl, R26^tdTom/+ mouse model-derived tumors. Using multiplex immunofluorescence, we showed that CD3+ T-cells express the inhibitory IC PD-1, while CD163+ macrophages express both Tim-3 and Gal9. Furthermore, tumor cells within ANF/ANNUBPs were positive for activated caspase-3, suggesting a possible role of the immune microenvironment in controlling tumor growth or inducing cell death. Through deep-targeted DNA sequencing, we detected loss of cell cycle regulators (especially CDKN2A) in ANF/ANNUBPs but no somatic alteration of polycomb repressor complex components EED and SUZ12, the latter typically associated with MPNSTs, even at low frequencies. Additionally, loss of H3K27me3 was associated with even lower densities of CD3+ T-cells and CD163+ macrophages in MPNSTs, and the loss of p16 seemed to act synergistically. Taken together, these findings suggest that an immune response involving both T-lymphocytes and macrophages may play a role in controlling or delaying the malignant transformation of ANF/ANNUBPs into MPNSTs and point to IC molecules as novel immunopreventive strategies in NF1 patients with ANF/ANNUBPs.