Spatial Proteomics (CODEX) Yields Insights into Intratumoral Nanoparticle Reprogramming of Murine Melanoma

Introduction: Co-detection by indexing (CODEX) offers a multiplexed imaging strategy that provides insight into the spatial proteomics of the tumor microenvironment (TME), labelling 30+ markers in a single sample while preserving tissue architecture. Analysis of spatial proteomics in response to different treatments for melanoma can significantly aid in the development of novel immunotherapies. Here, we describe an approachable proteomics profiling pipeline to analyze a novel CODEX panel developed for murine tumor samples. This methodology encompasses cellular segmentation, quantitative analysis, phenotyping, and spatial characterization. We apply this pipeline to analyze B16F10 melanoma tumors treated with immune-reprogramming polymeric nanoparticles (NPs). Our analysis highlights spatially resolved modulation of the TME engendered by this reprogramming NP technology and lends insights into its mechanism of action.

Materials and

Methods: B16F10 murine flank tumors were injected with poly(beta-amino ester)-based NPs delivering control plasmids (luciferase) or plasmids encoding 4-1BBL and IL-12 on days 9, 11, 16, and 18 following tumor implantation. Tumors were harvested and fixed on day 20, and formalin-fixed paraffin-embedded (FFPE) samples were sectioned onto slides. A DNA barcode custom-conjugated antibody panel was validated, and CODEX images were generated on a PhenoCycler-Fusion 2.0 device (Akoya Biosciences). QPTIFF files were then imported into the digital pathology suite HALO v3.6.4134.362 (Indica Labs). For each sample, tissue-level annotations were drawn around the whole sample, and intratumoral annotations were drawn based on the expression patterns of SOX10+ (melanoma) cells (Figure 1). Cell segmentation was performed using the HighPlex FL v4.2.14 algorithm in HALO by adjusting nuclear detection parameters. For marker positivity analysis, signal intensity thresholds were set for each marker in the antibody panel. Summary analysis data for positive cell counts was exported from HALO and analyzed in GraphPad Prism. The control condition (luciferase NPs) was represented by n = 4 biological replicates and the treatment condition (4-1BBL/IL-12 NPs) was represented by n = 8 biological replicates. Unpaired t-tests were used to determine statistical significance between control and treatment conditions. Object analysis data was imported into R, and cells were phenotyped using Seurat v4. Uniform manifold approximation and projection (UMAP) was used for dimensionality reduction, and clusters were defined according to marker expression profiles. Phenotype classifications were added to the cellular metadata, and spatial analysis was conducted in R using SPIAT v1.4.2.

Results, Conclusions, and Discussions: Within the intratumoral compartment, 4-1BBL/IL-12 NPs significantly decreased SOX10+ tumor burden (p < 0.0001; Figure 2A) and significantly increased the infiltration of CD45+ immune cells (p = 0.0007; Figure 2B), CD3+ T cells (p = 0.0074; Figure 2C), CD4+ T cells (p = 0.0062; Figure 2D), CD8+ T cells (p = 0.0118; Figure 2E), and F4/80+ macrophages (p = 0.0342; Figure 2F). 4-1BBL/IL-12 NPs caused an increase in proinflammatory F4/80+ CD86+ M1 macrophages (p = 0.0042; Figure 2G) and no significant change in pro-tumoral F4/80+ CD206+ M2 macrophages (p = 0.675; Figure 2H) when compared to luciferase control. Phenotype classifications and expression profiles for a representative 4-1BBL/IL-12 treatment condition are shown in Figures 3A and 3B respectively. Colocalization analysis around the tumor cell population showed increased interactions between tumor cells and immune cells following injection with 4-1BBL/IL-12 NPs (Figure 3C). Within a 100 μm radius around the tumor cell population, pro-inflammatory immune cells made up 38.64% of the compositional breakdown after injection with 4-1BBL/IL-12 NPs (5.58% CD4 T Cells, 13.10% CD8 T Cells, 14.42% M1 Macrophages, 4.05% Dendritic Cells, 1.49% NK Cells) versus only 2.27% of the compositional breakdown after injection with luciferase control NPs (0.88% CD4 T Cells, 1.39% Dendritic Cells). Taken together, these results indicate an increased ability for proinflammatory immune cells to infiltrate and interact with tumor cells in the murine melanoma microenvironment after reprogramming with 4-1BBL/IL-12 NPs. Additionally, this work presents an approachable pipeline for analyzing spatial proteomics from murine CODEX data that can be applied to various pathological samples for aiding in the screening of novel therapies.

Acknowledgements (Optional):