Antigen presentation deficiency, mesenchymal differentiation, and resistance to immunotherapy in the murine syngeneic CT2A tumor model
Frontiers in Immunology. 2023;14:1297932. DOI 10.3389/fimmu.2023.1297932. PMID 38213329. PMCID PMC10782385.
How to cite
AMA
Iorgulescu JB, Ruthen N, Ahn R, Panagioti E, Gokhale PC, Neagu M, et al. Antigen presentation deficiency, mesenchymal differentiation, and resistance to immunotherapy in the murine syngeneic CT2A tumor model. Front Immunol. 2023;14:1297932. doi:10.3389/fimmu.2023.1297932APA
Iorgulescu, J. B., Ruthen, N., Ahn, R., Panagioti, E., Gokhale, P. C., Neagu, M., et al. (2023). Antigen presentation deficiency, mesenchymal differentiation, and resistance to immunotherapy in the murine syngeneic CT2A tumor model. Frontiers in Immunology, 14, 1297932. https://doi.org/10.3389/fimmu.2023.1297932BibTeX
@article{iorgulescu2023antigen,
title = {Antigen presentation deficiency, mesenchymal differentiation, and resistance to immunotherapy in the murine syngeneic CT2A tumor model},
author = {Iorgulescu, J. B. and Ruthen, N. and Ahn, R. and Panagioti, E. and Gokhale, P. C. and Neagu, M. and others},
journal = {Frontiers in Immunology},
volume = {14},
pages = {1297932},
year = {2023},
doi = {10.3389/fimmu.2023.1297932},
pmid = {38213329}
}Two mouse models of glioma sit on opposite sides of the same treatment: GL261 tumors respond to immune-checkpoint therapy, while CT2A tumors resist it. Both are workhorses for testing brain-cancer immunotherapies, so knowing why one responds and one does not shapes how their results should be read.
Here the authors profiled both models across exome and RNA sequencing, global and phospho-proteomics, flow cytometry, and tissue imaging. CT2A’s resistance tracked with a broken antigen-presentation machinery, including mutations that lower MHC class I, together with a mesenchymal differentiation program. Interferon-gamma partially restored MHC class I, suggesting the resistance is layered rather than fixed.
Key findings
- CT2A resisted the checkpoint therapy that benefited GL261. GL261 tumors responded to single-agent anti-PD-1 and to anti-PD-1 plus anti-CTLA-4 (adjusted p ≤ 0.01 versus IgG), while CT2A gained a survival benefit only from the anti-PD-1 plus anti-CTLA-4 combination (adjusted p < 0.001) and none from single agents (adjusted p ≥ 0.12).
- Resistant CT2A tumors carried defects in antigen-presentation machinery. The model harbored clonal mutations in Tap1 (p.Y488C) and Psmb8 (p.A275P) alongside reduced MHC class I, and adding 50 ng/mL interferon-gamma partially restored MHC class I expression.
- The mesenchymal phenotype was visualized in the tumor tissue itself. Whole-slide fluorescence imaging localized the mesenchymal marker vimentin and the proliferation marker Ki67 in situ across GL261 and CT2A sections, complementing transcriptomes that diverged at 24.9–46.4% of detected genes (FDR-adjusted p < 0.05).
CyteFinder in the methods
“Formalin-fixed paraffin-embedded tissue sections of brains implanted with GL261-luc2 or CT2A-luc tumors were counterstained and immunolabeled using a 1:5000 dilution of Hoechst dye (10 mg/ml stock) in Odyssey Blocking Buffer to which anti-vimentin (AF594-conjugated, clone: D21H3; Cell Signaling #7675S) and anti-Ki67 (AF488-conjugated, clone: D3B5; Cell Signaling #11882S) primary antibodies were added at 1:25 dilutions. Tissue sections were incubated with the resulting counterstain/antibody solution for 1 hour in the dark at room temperature, rinsed in opaque Coplin jars containing fresh 1X PBS for 10 minutes in triplicate, and cover slipped in a 50% v/v glycerol solution diluted in 1X PBS immediately prior to imaging. Image tiles were acquired using a CyteFinder slide-scanning fluorescence microscope (RareCyte Inc.) at 20x magnification with 2x2 binning then stitched, registered, and flatfield-corrected using the MCMICRO image processing pipeline to generate whole-slide mosaic images.”
— Iorgulescu et al., Frontiers in Immunology (2023), Methods, “Immunofluorescent staining”
Why it matters for CyteFinder users
If you are considering CyteFinder for tissue work, look at the job it did here. This study’s central claim — that the resistant CT2A model is mesenchymal — needed to be seen in the tissue, not just inferred from sequencing. That is what the team used CyteFinder for: they imaged FFPE brain-tumor sections at 20x, stained for vimentin and Ki67, and let the whole-slide scan carry the readout. Two things make that useful for your own workflow. Because the instrument acquires image tiles across the entire section and stitches them into one mosaic (here through the MCMICRO pipeline), you read marker distribution across a whole tumor rather than a hand-picked field, so a mesenchymal marker’s spread through the tissue is something you can show, not just describe. And because the acquisition is fluorescence-based, you can place two protein markers and a nuclear counterstain in the same section and keep them spatially registered. Here that turned a transcriptional signature into a picture of the tissue it came from.
