High throughput proteomic profiling of tumor immunophenotypes by integrating microPOTS sample preparation, laser dissection, and implanted microtumor array

Tumours contain multiple cell types with distinct spatial organization, which forms heterogeneous microenvironments and immunophenotypes. It is increasingly recognized that the tumor immunophenotypes play key roles in tumor progression and therapeutic outcome, butit is still unclear what the main determining factors for these phenotypes are. A recently-established Skin Tumour Array by MicroPoration (STAMP) is an approach to generate hundreds of tumors in vivo to study immunophenotypes. Within a single mouse ear slice, there are typically 10-15 tumors originating from the same cell line that exhibit distinct levels of immune infiltration.

Recent advancements in mass spectrometry-based proteomics have revolutionized the analysis of complex biological samples. Spatial proteomics has emerged as a promising tool for investigating cellular heterogeneity, tumor microenvironments, and subcellular protein localization, thus enhancing our understanding of disease mechanisms. Herein, we applied spatial proteomics to study diverse immunophenotypes by profiling of individual microtumours in STAMP. For high-throughput profiling, we employed laser capture microdissection to isolate individual microtumours from mouse skin tissue, followed by microPOTS sample preparation and LC-MS analysis.

RNAscope analysis was first performed using T-cell and dendritic cell (DC) markers to characterize the inflammation of immune cells within the tumour. Three main phenotypes are established; EE (T-cell and DC excluded), II (T-cell and DC inflamed) and IE (T-cell inflamed and DC excluded). Tumor slices approximately 1 mm in diameter and 10 µm in thickness were collected from each phenotype using laser capture microdissection on a Zeiss LCM microscope and directly collected into a microPOTS chip. Sample preparation was performed in a volume of ~2 µL to improve sample recovery. The entire sample was injected for LC-MS/MS analysis using a 2-hr LC-gradient and data-dependent acquisition. FragPipe with IonQuant was used for database search and label-free quantification.

An average of 6900 proteins were identified per sample with >90% of proteins robustly quantified in all samples, providing deep proteome coverage from ~ 1-mm-diamter tumour slice. We observed distinct enrichment of proteins and pathways among these phenotypes, including the identification of T-cell and DC surface markers such as CD3 and CD4, which were enriched in the II tumours. Additionally II tumours were enriched for immunological-related processes, as expected. While phenotype-specific patterns were evident, an underlying biological variation driven by mouse background was also noted. However, this also highlights the sensitivity of our approach. Together, this study demonstrated spatial proteomics coupled with STAMP as a high-throughput and high-efficiency platform for large-scale tumor phenotyping and mechanistic study.