Distinct tissue niches contribute to prostate tissue-resident memory CD8+ T cell differentiation and heterogeneity
Immunity. 2026;59(5):1325-1343.e10. DOI 10.1016/j.immuni.2026.03.003.
How to cite
AMA
Takehara KK, Monell A, Luna V, Melisso B, Mempel KM, Cheung KP, et al. Distinct tissue niches contribute to prostate tissue-resident memory CD8+ T cell differentiation and heterogeneity. Immunity. 2026;59(5):1325-1343.e10. doi:10.1016/j.immuni.2026.03.003APA
Takehara, K. K., Monell, A., Luna, V., Melisso, B., Mempel, K. M., Cheung, K. P., et al. (2026). Distinct tissue niches contribute to prostate tissue-resident memory CD8+ T cell differentiation and heterogeneity. Immunity, 59(5), 1325-1343.e10. https://doi.org/10.1016/j.immuni.2026.03.003BibTeX
@article{takehara2026distinct,
title = {Distinct tissue niches contribute to prostate tissue-resident memory CD8+ T cell differentiation and heterogeneity},
author = {Takehara, Kennidy K. and Monell, Alexander and Luna, Vida and Melisso, Baxter and Mempel, Kianoosh M. and Cheung, Kitty P. and others},
journal = {Immunity},
volume = {59},
number = {5},
pages = {1325--1343.e10},
year = {2026},
doi = {10.1016/j.immuni.2026.03.003}
}The prostate is a barrier organ and a common site of cancer, yet almost nothing was known about the CD8+ T cells that guard it. Where these cells settle, how long they last, and what holds them in place had not been mapped.
Working in mice and confirming the pattern in human tissue, the team tracked prostate tissue-resident memory CD8+ T cells after infection. These cells were long-lived and protective, but they were not uniform. The glandular epithelium and the surrounding stroma formed separate niches with different local signals, and three cytokines, IL-15, IL-7, and TGFβ, emerged as the regulators that pushed T cells toward distinct states.
The result is a spatial map of where prostate memory T cells form and why their location shapes what they become.
Key findings
- Long-lived, protective tissue-resident memory CD8+ T cells formed in the prostate after infection. The CD69+CD103+ population peaked at 150 days post-infection and held steady in absolute number from 50 to 150 days.
- Distinct epithelial and stromal niches drove Trm heterogeneity through differential cytokine signaling. IL-15 and TGFβ were highest in the prostate epithelium, where CD8+ T cells were most persistent and cytotoxic, while a 480-gene spatial-transcriptomics panel traced these cytokine niches across days 7 and 30 of infection.
- Orion 17-plex whole-slide imaging localized CD8α+CD103+ Trm cells in human prostate biopsies. Across n = 4 benign human core needle biopsies, the 17-marker panel resolved the distance of CD8α+CD103+ versus CD8α+CD103− T cells to the epithelium at single-cell resolution.
Orion in the methods
“Images were acquired with a 20x 0.75 dry objective (0.325μm/pixel) using whole-tissue tiling. The Orion system (RareCyte, Seattle, Washington, USA) captured ∼10nm emission bands via angled dichroic mirrors. Processing used Rarecyte's algorithm for tile stitching, non-linear channel alignment, and spectral unmixing based on single-color controls”
— Takehara et al., Immunity (2026), Methods, “Multiplex immunofluorescence on human biopsies”
Why it matters for Orion users
If you are weighing Orion for a tissue study, look at what its spatial arm was asked to carry here. The central claim, that a memory T cell's fate depends on which niche it occupies, is a claim about position, and you cannot make it from a dissociated sample or a single field of view. It needs single-cell locations across whole tissue, in enough colors to separate several immune and structural cell types at once. Here the authors stained benign human core needle biopsies with a 17-marker panel plus a nuclear counterstain in a single round, imaged the whole slide, then measured how far each CD8+CD103+ T cell sat from the epithelium. Two things follow for your own workflow. Because the staining happens in one round, a scarce core biopsy is stained and scanned once rather than carried through repeated stain-and-strip cycles before you have your data. And because the run resolves 17 markers together at single-cell resolution, you can ask the spatial question this paper turned on, who sits next to whom and how far apart, that a lower-plex or dissociated readout cannot answer.
