Sialic acids are a family of nine-carbon acidic sugars that terminate glycan chains on the outer surface of virtually all vertebrate cells. Two linkages dominate: the α2,3 linkage, in which sialic acid attaches to the third carbon of a galactose residue, and the α2,6 linkage, which attaches at the sixth carbon. On cancer cells, overabundant sialylation is well known to dampen antitumor immunity, engaging inhibitory Siglec receptors on immune cells and creating a protective glycocalyx. But whether the T cells doing the attacking carry functionally meaningful sialic acid patterns of their own has remained murky, largely because existing tools could not resolve which linkage was present on which individual cell, let alone connect that information to gene expression in the same cell.

To break through that limitation, the group led by Xing Chen and Qi Tang developed a strategy they call scCAST_3&6, short for single-cell correlative analysis of sialoglycans and transcriptomes. The method uses mutant glycosidases and chemoenzymatic labeling tools—refinements of the team’s earlier work on linkage-specific sialoglycan labeling—to tag α2,3- and α2,6-linked sialic acids on single cells with distinct detectable labels. Those labels are then read out alongside single-cell RNA sequencing, allowing the researchers to pair, for every individual tumor-infiltrating immune cell, a quantitative measurement of its two sialic acid linkages with a full transcriptional profile. Applying the method to immune cells isolated from murine melanoma and other tumor models, they found that α2,3- and α2,6-sialylation are regulated independently and in strikingly cell-type-specific patterns across the tumor immune landscape, from macrophages and dendritic cells to natural killer cells and T cell subsets.

The most consequential discovery emerged when the team focused on CD8+ T cells, the cytotoxic lymphocytes responsible for directly killing cancer cells. Within tumors, these cells exist along a spectrum from naive bystanders to exhausted but tumor-reactive fighters, and the field has struggled to distinguish the truly antigen-reactive cells from the merely present ones. The scCAST analysis revealed that exhausted CD8+ T cells split into subpopulations based on their α2,6-sialylation levels, and that cells with low α2,6-sialylation—a state the researchers denote Sia2,6^lo—were dramatically enriched for tumor-specific antigen reactivity. When the team stained cells with peptide–MHC tetramers presenting a known tumor antigen, the Sia2,6^lo PD-1-positive fraction contained far more tetramer-binding cells than their Sia2,6^hi counterparts, and flow cytometry confirmed that the low-sialylation cells expressed higher levels of the activation markers TIM-3, CD137, granzyme B and interleukin-2.

Avidity—the collective strength of the interaction between a T cell receptor and the peptide–MHC complexes presented on a target cell—is a key determinant of antitumor potency. The researchers showed that Sia2,6^lo status tracks with high-avidity tumor antigen reactivity: expanded T cell clones within the low-sialylation compartment, and these cells killed antigen-bearing tumor cells more efficiently in vitro across multiple tumor models, including B16 melanoma, E0771 breast cancer and MC38 colon cancer. Critically, α2,6-sialylation did not drop indiscriminately during T cell activation. Naive CD8+ T cells stimulated with antibody-based activation reagents, which bypass the T cell receptor in a linkage-agnostic way, maintained their sialylation, whereas cells encountering antigen at high avidity specifically suppressed the gene encoding β-galactoside α2,6-sialyltransferase 1, or St6gal1, the enzyme that installs α2,6-linked sialic acids.

That mechanistic insight—high-avidity T cell receptor signaling drives St6gal1 downregulation—explains why the sugar signature is such a faithful proxy for tumor reactivity. Only T cells that have engaged their cognate antigen through a strong, multivalent receptor interaction pay the metabolic and transcriptional cost of remodeling their α2,6-sialylation, whereas low-avidity encounters leave the glycocalyx untouched. The team corroborated this in engineered systems where T cells were activated with variant peptides of graded affinity, showing that only high-avidity ligands suppressed St6gal1 expression. Datasets from human colorectal and lung cancer patients independently revealed that exhausted intratumoral PD-1-positive CD8+ T cells express lower levels of ST6GAL1 than their PD-1-negative counterparts, suggesting the mouse findings reflect a conserved biology.

Here the story takes a darker turn. Removing α2,6-linked sialic acids from glycan chains exposes the underlying galactose residues, which are preferred ligands for galectin-1, a β-galactoside-binding lectin abundantly produced by many tumors. The researchers demonstrated that Sia2,6^lo CD8+ T cells bind significantly more galectin-1 within the tumor, and that galectin-1 binding promotes apoptosis of these cells—effectively converting the very sugar change that marks a T cell as tumor-specific into a homing beacon for its own destruction. In tumors lacking galectin-1, the low-sialylation T cells persisted, and genetic deletion of St6gal1 in T cells enhanced galectin-1 binding without altering exhaustion markers, pinning the death pathway specifically on the exposed galectin ligands. Galectin-1 has long been implicated in creating immune-privileged tumor niches, and this work supplies a precise molecular handshake by which the tumor exploits the activation state of its attackers.

The therapeutic implications cut in two directions. On one hand, the Sia2,6^lo surface phenotype offers a practical, linkage-specific marker for isolating the highest-avidity tumor-reactive T cells from patient tumors—an advance that could sharpen adoptive cell therapy and tumor-infiltrating lymphocyte products, where identifying the right cells to expand has been a persistent bottleneck. Existing markers such as PD-1, CD39 and CD103 capture tumor reactivity only imperfectly, and a glycan-based readout adds an orthogonal, functionally grounded dimension. On the other hand, the galectin-1 vulnerability suggests that blocking galectin-1, or glycoengineering T cells to retain α2,6-sialylation while preserving their reactivity, could protect the most valuable antitumor cells from apoptosis. Combined with the growing arsenal of glycan-targeting immunotherapies—from sialidase-fused T cell engagers to inhibitors of sialyltransferases—the study expands the design space for manipulating the immune glycocalyx in cancer.

Technically, the work represents a milestone in single-cell multiomics: the ability to measure a chemically defined, linkage-specific glycan feature and a transcriptome in the same single cell, at scale, in a clinically relevant tissue. The scCAST platform and its associated datasets have been deposited in public repositories, and the tools build on a broader chemoenzymatic toolkit for glycan analysis that the Chen laboratory has developed over the past decade. As glycobiology moves from bulk descriptions of glycosylation toward cell-resolved, mechanistic accounts of how sugars regulate immunity in disease, this study offers both a template and a target: a sugar linkage that tells the true killers from the bystanders, and a lectin-mediated trap that tumors use to eliminate them. Understanding—and eventually interrupting—that trap may prove to be one of the more elegant ways to tip the balance of cancer immunotherapy in the patient’s favor.

Subject of Research: Linkage-specific sialylation of tumor-infiltrating CD8+ T cells and its role in antitumor immunity

Article Title: α2,6-Sialoglycan-regulated antitumor immunity of tumor-infiltrating CD8+ T cells

Article References: Liang, S., Tang, Q., Guo, Y., Zhang, S., Song, Q., Dai, P., & Chen, X. (2026). α2,6-Sialoglycan-regulated antitumor immunity of tumor-infiltrating CD8+ T cells. Nature Chemical Biology. https://doi.org/10.1038/s41589-026-02299-7

Image Credits: AI Generated

DOI: 10.1038/s41589-026-02299-7

Keywords: sialoglycans, α2,6-sialylation, CD8+ T cells, tumor microenvironment, galectin-1, St6gal1, single-cell multiomics, T cell avidity, cancer immunotherapy, glycosylation, glycobiology, Sialoglycan-regulated