MSACL 2025 Abstract

Multimodal Molecular Imaging of Human Colorectal Cancer Biopsies Combining MALDI IMS, CODEX, and Spatial Transcriptomics

Martin Dufresne (1,2), Seug Woo Kang (2,3,4,5), Alan J Simmons (2,3,4,5), Lukasz G. Migas (6), Thai Pham (1,2), Harsimran Kaurt (2,3,4,5), Jamie L. Allen (1,2), Audra M. Judd (1,2), Anna J. Smith (1,2), Melissa A. Farrow (1,2), Raf Van De Plas (6), Ken Lau (2,3,4,5,7), and Jeffrey M. Spraggins (1,2,8,9,10)
(1) Mass Spectrometry Research Center, Vanderbilt University, Nashville, TN (2) Department of Cell and Developmental Biology, Vanderbilt University, Nashville, TN (3) Epithelial Biology Center, Vanderbilt University Medical Center, Nashville, TN (4) Vanderbilt-Ingran Cancer Center, Vanderbilt University Medical Center, Nashville, TN (5) Center for Computational Systems Biology, Vanderbilt University School of Medicine, Nashville, TN (6) Delft Center for Systems and Control, Delft University of Technology, Delft, Netherlands (7) Vanderbilt-Ingran Cancer Center, Vanderbilt University Medical Center, Nashville, TN (8) Department of Chemistry, Vanderbilt University, Nashville, TN (9) Department of Biochemistry, Vanderbilt University, Nashville, TN (10) Department of Pathology, Microbiology, and Immunology, Vanderbilt University, Nashville, TN

Martin Dufresne, Ph.D. (Presenter) Vanderbilt University

Relevant Financial Disclosures (within past 24 months, reported on Jul 18, 2025) No relevant financial relationship(s) to disclose. Conflict mitigation NOT REQUIRED.

INTRODUCTION
Matrix-assisted laser desorption/ionization (MALDI) is a key high spatial resolution (≤ 10 µm) imaging mass spectrometry (IMS) technology, owing to its broad molecular coverage and ability to target selected molecular classes through tunable sample preparation. Combined with complimentary spatial biology platforms (e.g., CODEX, Visium, and Xenium), we can link IMS-derived molecular profiles with specific cell types and neighborhoods to discover how systems are altered in normal aging and disease. This approach is well-suited for studying cancer where characterizing tumor heterogeneity and disease progression is key to understanding cancer across stages, from pre-cancerous to advanced malignancies. Here, we present our integrated multimodal molecular imaging approach for the study of human colorectal cancer as part of the Human Tumor Atlas Network (HTAN).

METHODS
Preliminary data was acquired on 9 serial sections of a single colorectal cancer patient biopsy and two control biopsies for multi-omics analysis. All serial sections were initially scanned for autofluorescence (AF) prior to further analysis allowing for high accuracy registration to be performed using a single non-destructive modality prior to all experiments. Serial sections were then subjected to MALDI IMS, histological staining (H&E), spatial transcriptomic (Visium and Xenium), or immunofluorescence (CODEX). MALDI IMS was performed using a timsTOF flex (Bruker Daltonics) at 10 µm spatial resolution using minimal laser power after sublimation of VANDY37 matrix using the SubliMATE system (HTX Imaging). Co-registration and analysis were performed using in-house software and SCILS lab (Bruker Daltonics).

PRELIMINARY DATA
HTAN aims to build molecular atlases using integrated multi-omics to define tumor microenvironments and characterize tumor evolution at the cellular level in 2- and 3-dimensions. We have started the work on early-stage and late-onset human colorectal cancer using a multi-omics approach to achieve this goal. Because most molecular imaging modalities, including those proposed here, are inherently 2-dimensional, our strategy for generating 3-dimensional data is to co-register multiple serial sections using common, low-cost modalities that can be collected in tandem with molecular assays, e.g. autofluorescence and stained microscopy. This allows us to quickly assess the quality of each of the biopsies and histologically score the sample, allowing the selection of the best candidates for full 3-D multimodal analysis. Serial tissue sections are then guided through different spatially resolved omics and co-registered back to their respective AF pre-acquisition scan, which is also used for 3-dimensional reconstruction.

Our current workflow involves MALDI IMS in both polarities followed by CODEX of the same section, MALDI IMS of glycans, and LCM for targeted bulk metabolomics and proteomics analysis. Further sections are reserved for spatial transcriptomics using Visium and Xenium, along with traditional histological staining for the first and last sections of the 3-D stack. Initial MALDI IMS data has shown a strong correlation between certain classes of lipids such as carnitines, ether phosphatidylcholine (ether PC), and some sphingomyelin (SM) to the proliferation region of the tumor. Other signals, such as long chain PCs, ether PCs, and SMs, have intriguing distribution that do not overlap with the tumor proliferation region or healthy tissue. Further integration of these data with CODEX and spatial transcriptomics is underway, enabling mechanistic information to be layered onto molecular distributions and cellular organization.

CONCLUSION
This workflow will now be applied to a larger cohort of human colorectal cancer samples in 3D using single tissue multimodal imaging approach using lipid, glycan, and glycogen MALDI IMS which will be combined with serial section CODEX imaging and spatial transcriptomic.