Spatially resolved transcriptomics of benign and malignant peripheral nerve sheath tumors.

Background: Peripheral nerve sheath tumors (PNSTs) encompass entities with different cellular differentiation and degrees of malignancy. Spatial heterogeneity complicates the diagnosis and grading of PNSTs in some cases. In malignant PNST (MPNST) for example, single-cell sequencing data has shown dissimilar differentiation states of tumor cells. Here, we aimed to determine the spatial and biological heterogeneity of PNSTs.

Methods: We performed spatial transcriptomics on formalin-fixed paraffin-embedded diseased peripheral nerve tissue. We used spatial clustering and weighted correlation network analysis to construct niche-similarity networks and gene expression modules. We determined differential expression in primary pathologies, analyzed pathways to investigate the biological significance of identified meta-signatures, integrated the transcriptional data with histological features and existing single-cell data, and validated expression data by immunohistochemistry.

Results: We identified distinct transcriptional signatures differentiating PNSTs. Immune cell infiltration, APOD, and perineurial fibroblast marker expression highlighted the neurofibroma component of hybrid PNSTs (HPNSTs). While APOD was evenly expressed in neurofibromatous tumor tissue in both, HPNST and pure neurofibromas, perineurial fibroblast markers were evenly expressed in HPNST, but restricted to the periphery in plexiform neurofibromas. Furthermore, we provide a spatial cellular differentiation map for MPNST, locating Schwann cell precursor and neural crest-like cells as well as those with mesenchymal transition.

Conclusions: This pilot study shows that applying spatial transcriptomics to PNSTs provides important insight into their biology. It helps establish new markers and provides spatial information about the cellular composition and distribution of cellular differentiation states. By integrating morphological and high-dimensional molecular data it can improve PNSTs classification in the future.