Protocol for spatial proteomic profiling of tonsil cancer microenvironments using multiplexed imaging-powered deep visual proteomics

Abstract

Here, we present a protocol for spatial proteomic profiling of the tumor microenvironment in tonsil cancer using multiplexed imaging-powered deep visual proteomics (mipDVP). We describe steps for automated 22-plex immunofluorescence staining and imaging on formalin-fixed paraffin-embedded (FFPE) tissue sections, automated single-cell laser microdissection, and single-cell-type mass spectrometry. This workflow enables the spatially resolved isolation of distinct cell populations for proteomic analysis. We optimized this protocol for studying tumor-immune interactions, where it facilitates the systematic identification of biomarkers and functional cellular networks. For complete details on the use and execution of this protocol, please refer to Zheng et al.¹.

Keywords: Biotechnology and bioengineering; Cancer; Microscopy.

Graphical abstract

Figure 1

Sample preparation workflow for MACSima imaging system (A) Sample alignment guide showing the sectioning area compatible with the two-well imaging frame. (B) Guide for tissue sectioning onto PEN membrane slides with sections aligned to imaging frame wells. (C) Assembly of the two-well imaging frame, designed to secure tissue sections while enabling reagent access during automated staining cycles. (D) Deepwell plate (96-well format) preloaded with fluorescently conjugated antibodies for automated dispensing. (E) System rack setup, showing the imaging frame and deepwell plate. (F) Sample regions of interest (ROI) and autofocus configuration. The screenshots originate from the interface of MACSima Imaging System (#130-121-164 ) with software v.0.13.0.

Figure 2

Image analysis workflow for spatial characterization (A) MACS iQ View software (v.1.2.2) interface highlighting key modules: image preprocessing tools, processed image loading, and interactive panels for segmentation and gating-based classification. Example regions demonstrate cell segmentation (yellow masks) and immune cell gating (cytotoxic T cells (CTLs), green masks). (B) Spatial analysis workflow categorizing tumor cells into three proximity zones relative to cytotoxic T cells: close (≤30 μm), intermediate (30–60 μm), and far (>60 μm; ≤90 μm). (C) Integration of MACS iQ View and BIAS software for mask alignment of downstream laser microdissection. Export masks of phenotypically defined cells (tumor cells with intermediate distance to CTLs) from iQ View and import them into BIAS for reference point set up. Scale bars: 30 μm.

Figure 3

Settings in DIA-NN for mDIA experiments Screenshot of the DIA-NN (v.1.8.1) interface configured for dimethyl-labeled proteomic data analysis.

Figure 4

Workflow for spatial analysis and cell isolation (A) Overview of tumor-immune spatial mapping in tonsil carcinoma. The composite multiplex immunofluorescence image (top left) shows a tonsil carcinoma tissue section stained for cytokeratin (tumor cells), CD3 and CD8 (CTLs), CD4 (helper T cells), smooth muscle actin (SMA; stroma), and DAPI (nuclei). Automated segmentation (top right) identifies individual cells, with classification masks (middle left) distinguishing tumor cells (cyan) and CTLs (orange). Categorize tumor cells into three proximity zones relative to CTLs: close (≤30 μm, pink), intermediate (30-60 μm, blue), and far (60–90 μm, green; middle right). A density gradient map (lower left) visualizes tumor cell distances to stromal interfaces (blue: proximal, red: distal). A 3D volumetric rendering (lower right) depicts CD8+ T cell clusters (magenta isosurfaces) localized near stromal-tumor boundaries. Scale bars: 25 μm. (B) Automated laser microdissection workflow. Spatial classification masks recognize cellular contours and precisely excise them. Collect microdissected single cells via gravity into a 384-well plate (right) for downstream proteomic analysis. Panel B is adapted and modified from Zheng et al. under the terms of the CC-BY license.

Figure 5

Proteomic profiling of spatially resolved tumor cell populations (A) Quantification of peptide precursors and protein groups across technical replicates (n=3) for tumor cells stratified by proximity to cytotoxic T lymphocytes (CTLs): far (60-90 μm), intermediate (30-60 μm), and close (≤30 μm). The data are presented as mean ± SEM. (B) Protein abundance rank plot summarizing cumulative identification of >4,900 protein groups across all tumor cell populations. (C) Represent technical reproducibility of proteomic data as coefficients of variation (CVs, %) for quantified protein groups. (D) Principal component analysis (PCA) of tumor cell proteomes, illustrating distinct clustering of far-proximity populations compared to intermediate and close groups. Panel D is adapted and modified from Zheng et al. under the terms of the CC-BY license.