Virtual staining enabled combined morphological and spatial transcriptomic analysis of individual malignant B cells and local tumor microenvironments

Abstract B-cell lymphomas, commonly classified by histological presentation and B-cell lineage, are in fact heterogeneous diseases with respect to gene expression and tumor microenvironment. Certain lymphomas, such as Hodgkin and T-cell rich B-cell lymphoma, may contain fewer than 1% malignant cells in a largely inflammatory background. These rare malignant cells have unique interactions with local tumor microenvironments (TME), in some cases driving distinct gene expression patterns for individual B-cells. This presents a challenging fundamental biology landscape, which can create uncertainty in directing patient care or developing new therapies. The lack of tools allowing subtyping of malignant B-cells while simultaneously preserving the context of local TMEs and providing deep spatial transcriptomic sequencing further complicates analysis. Here we have developed an integrated, single-slide workflow of virtual H&E staining using the inherent tissue autofluorescence (PictorLabs, Los Angeles, CA) of formalin-fixed, paraffin embedded lymphoma tissue sections on the GeoMx® Digital Spatial Profiler (Nanostring, Seattle, WA) combined with spatial expression analysis of 1,800 genes using the Cancer Transcriptome Atlas (CTA) panel. The H&E virtual staining allows for the assessment of B-cell lymphoma subtypes, identification and phenotyping of malignant B-cells and the immediately proximal (within 30µm) TME regions by a certified pathologist. These regions of interest (ROI) are spatially profiled for transcriptional expression determined by NGS sequencing of the RNA target specific DNA oligo tags. We benchmarked our system by analyzing four lymphoma cases: Nodular sclerosis classical Hodgkin, Nodular lymphocyte-predominant Hodgkin, T-cell/histiocyte rich large B-cell, and Diffuse large B-cell lymphomas. Using virtual H&E staining on the GeoMx Digital Spatial Profiler platform, a pathologist correctly identified the diagnosis and subtype based on the morphology. Regions of inflammatory B- and T-cells could be distinguished from lymphoma cells on virtual staining, allowing subset segregation for digital transcriptional expression profiling with the GeoMx CTA panel. Additionally, background TME more distant to B-cell lymphoma as well as regions of high T-cell infiltration were selected for profiling. Analysis of the ROIs revealed distinct transcriptional profiles of malignant B-cells (including cell-of-origin signature), the signaling contributions of local TMEs (versus non-adjacent regions) and confirmed that the expression patterns are further influenced by the presence of T-cell infiltration. Citation Format: Michael E. Kallen, Alyssa Rosenbloom, Zihang Fang, Raymond Kozikowski, Yair Rivenson, Kevin de Haan, Serge Alexanian, Mark Conner, Daniel Newhouse, Felicia New, Yan Liang, Kyla Teplitz, Joachim Schmid, Jaemyeong Jung. Virtual staining enabled combined morphological and spatial transcriptomic analysis of individual malignant B cells and local tumor microenvironments [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 6619.