Host sugars hide TB vaccine target Ag85, study finds

Nucleic acid vaccines, including DNA and mRNA platforms, work by telling the body’s own cells to make microbial antigens in place. That strategy can be powerful for many targets, but it also brings a potential pitfall: proteins made by human cells can receive chemical decorations that the original microbe never adds. Tuberculosis (TB), caused by Mycobacterium tuberculosis, is still a leading killer worldwide, and several nucleic acid vaccine efforts aimed at the antigen 85 (Ag85) complex failed to protect people in clinical trials. Concerned by this mismatch between microbial and host-made antigens, F.Y. Avci and colleagues asked whether modifications added by mammalian cells — specifically N-glycosylation, a common way cells attach sugar chains to proteins — might change how the immune system sees Ag85. Working from this hypothesis, the team set out to define, at structural, biochemical and immunological levels, whether and how host-imposed N-glycosylation remodels a bacterial immunogen when it’s expressed in mammalian cells used for vaccine antigen production or in situ expression from nucleic acid vaccines.

To test their idea, the researchers expressed the Ag85B protein in human Expi293 cells and examined the molecule in detail. They found that Ag85B is microheterogeneously N-glycosylated at four canonical sequons — N52, N224, N234, and N280 — and that the attached glycans are predominantly complex, highly fucosylated, and frequently sialylated. Molecular dynamics simulations showed these glycans explore substantial conformational space and effectively reduce solvent and antibody-accessible surface area, meaning parts of the protein that antibodies or T cells would normally contact become occluded. In lab binding tests, mammalian-expressed Ag85B showed markedly reduced binding to an Ag85-complex monoclonal antibody when measured by competitive ELISA and by biolayer interferometry. The team also found that the sialylated glycans on Ag85B enable binding to the inhibitory receptor Siglec-9, and that this Siglec-9 interaction can be eliminated by sialidase treatment, linking the sugar decorations to both shielding and an active immune-modulatory interaction.

Taken together, these findings define a clear structural and biochemical mechanism by which host glycosylation can remodel a bacterial vaccine antigen and interfere with immune recognition. By showing that N-glycans can sterically block established B-cell and T-cell epitope regions and can engage Siglec-9, the work explains a plausible pathway through which mammalian expression of a bacterial immunogen like Ag85B could weaken vaccine-induced responses. The study supports a design principle the authors call glycosylation-aware immunogen engineering: when designing nucleic acid vaccines against non-viral pathogens, researchers should anticipate and address host-imposed post-translational modifications. For TB vaccine development and other non-viral targets, this may mean altering antigen sequences, choosing delivery approaches that avoid problematic glycosylation sequons, or otherwise engineering immunogens so that they present the same molecular faces that the immune system sees during natural infection rather than host-modified versions that escape recognition.