Cytosolic bacterial antigens drove the strongest mucosal CD4 T cell response
Antigen location, not just antigen identity, determined how the intestinal immune system responded to Citrobacter rodentium. In a mouse model of barrier infection, the same CD4 T cell epitope generated sharply different outcomes depending on whether it remained on the bacterial surface, localized to the apical membrane of infected epithelial cells, or reached the epithelial cytosol. The most consequential signal came from cytosolic delivery. Antigens injected into colonocyte cytosol produced robust clonal CD4 T cell expansion, persistence in the infected colon, and development of tissue-resident memory T cells within the epithelium itself.
A single-sentence result captures the hierarchy: membrane exposure was not enough, but cytosolic access changed the immune fate.
This distinction matters in mucosal immunology because epithelial surfaces do not behave like lymphoid organs. Barrier tissues encounter pathogens first, and epithelial cells can express MHC class II under inflammatory conditions even though they are not professional antigen-presenting cells. In this study, those infected epithelial cells did more than display antigen; they shaped whether responding CD4 T cells remained local sentinels or drifted toward a nonresident memory program. The work places nonhematopoietic cells in an active role in sculpting memory fate at the site of infection.
Antigen compartmentalization separated weak responses from durable tissue residency
Investigators engineered C. rodentium strains to carry the same LCMV gp66-80 epitope fused to different virulence proteins, allowing antigen recognition to be compared while holding peptide affinity constant. The carrier proteins spanned distinct compartments: intimin on the bacterial surface, Tir on the host epithelial membrane after translocation, and NleA and EspZ in the host cell cytosol. Confocal imaging confirmed the expected distributions in colonocytes isolated on day 8 after infection. Intimin remained bacterial, Tir concentrated at the apical plasma membrane, and both NleA and EspZ entered the cytoplasm.
The immunologic readout was unambiguous. At the peak of the mucosal response on day 14, SMARTA CD4 T cell accumulation in the mid-distal colon was minimal when the epitope stayed on the bacterial surface or on the epithelial membrane. By contrast, the two cytosolic antigens drove robust local expansion. NleA-gp66 and EspZ-gp66 produced 28-fold and 47-fold more SMARTA T cells in the colon, respectively, than intimin-gp66. Those numbers speak to a steep biological threshold: antigen has to enter the host cell interior to generate a productive CD4 T cell program in this setting.
The effect was not simply a function of abundance. In vitro and in vivo measurements showed that Tir was among the most abundant proteins detected, while EspZ was present at lower levels, yet Tir elicited little response and EspZ still drove strong expansion when delivered to the cytosol. Abundance alone could not explain the pattern. Compartmental access was the deciding variable.
Direct epithelial presentation shaped memory fate
To isolate translocation from every other property of the carrier protein, the study compared full-length EspZ with a truncated EspZ variant lacking its secretion signal. Both versions carried the same epitope and FLAG tag, but only the full-length protein was delivered into the epithelial cytosol. The truncated form remained associated with bacteria. This design stripped the comparison down to one question: does access to the host cytosol alter the CD4 memory program?
It did.
When naive SMARTA T cells were transferred into infected mice, the cytosol-delivered form promoted stronger accumulation in the colon and more effective retention in the infected mucosa. The nontranslocated version produced a much weaker response and favored a nonresident fate. The biological logic is straightforward. Cytosolic antigen inside infected epithelial cells can be sampled by MHC class II machinery in a context that sustains antigen presentation at the barrier, allowing prolonged contact between T cells and infected tissue. That persistent engagement appears to favor epithelial-resident memory T cells rather than memory cells that recirculate elsewhere.
The study also showed that this was not a one-way presentation event. Single-cell transcriptomic analysis revealed sustained bidirectional communication between epithelial cells and T cells. The epithelial compartment and the responding CD4 population jointly expressed programs linked to barrier protection and tissue residency. In practical terms, the infected epithelium was not a passive platform. It participated in deciding whether immune memory would stay at the front line.
Barrier protection depended on a resident memory program rather than central memory drift
The most important consequence of cytosolic antigen delivery was not merely expansion, but positioning. CD4 T cells that entered a tissue-resident memory trajectory were retained where reinfection would occur first: the intestinal epithelium. Those cells are ideally placed to respond rapidly to a repeated barrier challenge, especially in a tissue where pathogen containment depends on immediate local cytokine production. The work connects this residency program with the previously observed IL-22-producing CD4 response in distal colonocytes, strengthening the argument that epithelial antigen presentation is integral to mucosal defense.
This is a mechanistic correction to a common assumption in adaptive immunity. Not all antigen presentation is equivalent, and not all infected cells play the same instructive role. In this model, nonprofessional antigen-presenting epithelial cells did not merely report infection to the immune system; they determined whether the immune response matured into a durable barrier defense. The infected colonocyte became a memory-shaping cell.
The implications for mucosal vaccine design are clear within the logic of the data. If barrier immunity depends on antigen delivery into epithelial cytosol and local MHC class II presentation, then subcellular routing may be as important as antigen choice. Vaccines that target epithelial antigen processing in the intestine may be better positioned to induce local CD4 tissue residency than formulations that keep antigen extracellular or confined to bacterial surfaces. The study does not answer how such routing should be engineered in humans, and it does not establish whether this mechanism generalizes beyond C. rodentium and the mouse colon. It does, however, identify a concrete biological constraint: mucosal CD4 memory is not simply generated by infection. It is programmed by where the antigen lands inside the infected barrier cell.