Work

Decoding the Immunologic Networks of the AML Marrow Microenvironment Using Codex

Year: 2024

Type: article

Abstract: Spatial organization of immune cell hubs in tumor microenvironments are increasingly recognized across diverse histologies and tissues. Bone marrow provides a rich platform for investigating immunologic hubs in the context of relapsed acute myeloid leukemia (AML) hematopoietic stem cell transplant (HSCT), which may provide fresh insights into mechanisms relapse and maintenance of remission after donor lymphocyte infusion (DLI). Multiple strategies for spatial localization transcriptomic and protein-based analysis at single cell resolution are now available. However, a recognized challenge is that conventional processing of clinical bone marrow samples involves formalin fixation and paraffin embedding (FFPE) and decalcification, thereby degrading RNA. We first confirmed feasibility of CODEX, a protein-based approach for multiplexed spatial analysis, to AML marrow FFPE samples using a custom 44-marker panel targeting T-, B-, NK-, and myeloid cells. We next evaluated 14 marrow FFPE sections from 7 patients with post-HSCT relapsed AML treated with DLI (4 responder [R], 3 nonresponder [NR]) on whom prior analysis of single cell RNA (scRNA-seq) and TCR (scTCR-seq) had revealed post-DLI expansion of 3 immune cell populations in DLI-Rs: cytotoxic CD8+ T cells, cytolytic NK cells, and naïve B cells (Maurer et al, manuscript in preparation). We hypothesized that spatial characterization of these immune cell types in relapse and post-DLI marrow microenvironments would define leukemia-immune cell interactions critical to DLI response. CODEX images were obtained on an Andor Dragonfly 200 Spinning Disk Confocal Microscope with a Plan Apochromat 20x/0.75 NA objective (Nikon, MRD00205). Images were aligned by antspyx using DAPI nuclear localization. We leveraged DeepCell for segmentation of individual cells (Figure 1). Average pixel intensity of all markers was calculated for each cell to generate a cell X marker matrix. All windows were combined from all samples and each marker was normalized across all cells and samples. All cells were then projected onto a 2D UMAP visualization and clustered using Phenograph (Figure 2). Immune cell subtypes were identified based on combinations of surface markers. Co-localization of cells was calculated using spatial cross correlation index (SCI) and “niche types” defined as recurrent combinations of neighboring leukemic and immune cell types were inferred (Figure 3). In confirmation with our scRNA-seq data, CD8 CTLs (identified by CD8 and Granzyme B) were found to expand in R post-DLI (post/pre fold change: 3.90). We now show the first spatial analysis of AML marrow microenvironment with CODEX and find compelling evidence of post-DLI T cell expansion concordant with our scRNA-seq analysis. Further evaluation of these spatial interactions may lead to novel targets for therapeutically optimizing adoptive cellular immunotherapy.

Source: Transplantation and Cellular Therapy

DOI: 10.1016/j.jtct.2023.12.341