Experimental Tuberculosis Vaccine Targets Rare Variable Antigens to Protect Lungs Without Cutting Bacterial Burdens

A tuberculosis vaccine built from an unexpected set of bacterial proteins has produced one of the more intriguing results seen in preclinical TB research in recent years. In a study published in Frontiers in Tuberculosis, a team at the University of California, San Francisco reports that vaccinating mice with so-called rare variable antigens of Mycobacterium tuberculosis dramatically reduced lung damage without lowering the number of bacteria in the animals’ lungs. The finding challenges a core assumption that has shaped TB vaccine development for decades: that a successful vaccine must reduce bacterial burden. Instead, the work points toward a different goal—preventing the tissue-destroying pathology that makes tuberculosis deadly and transmissible.

Tuberculosis remains the leading cause of death from a single infectious agent worldwide, yet the only licensed vaccine, BCG, offers unreliable protection against pulmonary disease in adults, the form of illness that drives most transmission. The leading subunit candidate, M72/AS01E, showed roughly fifty percent efficacy against TB disease in a major trial, leaving an enormous gap. Most vaccine candidates in testing aim to induce CD4 T cell responses, which are known to be essential for controlling Mtb. But the UCSF researchers, led by Joel Ernst, started from an observation that inverts conventional thinking about antigen selection. Unlike most pathogens, Mtb’s CD4 T cell epitopes are strikingly conserved across strains, suggesting that T cell recognition of these dominant antigens may actually benefit the bacterium, perhaps by promoting the inflammatory lung destruction that facilitates transmission through coughing and cavitation.

From that reasoning, the team focused on a small group of Mtb antigens that do show sequence diversity within human T cell epitopes—proteins they had previously identified and dubbed rare variable antigens, or RVMAs. The diversity within these epitopes hints that T cell responses against them are more detrimental to the bacterium than responses to conserved antigens, placing the genes under diversifying selection. The researchers built a DNA vaccine, named pCas4, encoding a fusion of four such antigens: Rv0010c, RimJ, Rv2719c, and Rv0990c, using sequences from the Mtb Erdman strain cloned into a commercial pVAX1 vector. Mice received three intramuscular immunizations spaced three weeks apart before being challenged with an aerosolized dose of bacteria.

The vaccine proved immunogenic. Spleen cells from vaccinated C57BL/6 mice produced interferon-gamma in response to three of the four antigens, with RimJ inducing the strongest responses. During actual infection, the same three antigens were recognized by CD4 T cells, though the responses were weaker than those mounted against the classic immunodominant antigens ESAT-6 and Ag85B. After challenge with the matched Erdman strain, vaccinated animals showed significantly increased IFN-gamma and TNF co-producing CD4 T cell responses to Rv0010c and RimJ. Notably, the team found a naturally occurring amino acid substitution, Y105C, in RimJ between the Erdman and H37Rv strains that fell within a predicted T cell epitope and altered the strength of T cell recognition—direct experimental support for the idea that antigenic variation genuinely exists in Mtb, even between closely related lineage 4 strains.

At twelve weeks of chronic infection in standard C57BL/6 mice, pCas4 vaccination produced no reduction in bacterial burdens in lungs or spleens. What it did change was pathology: vaccinated animals had fewer and smaller lymphoid aggregates in their lungs, structures composed largely of B cells that arise in chronic infection in mice and are also seen in human and non-human primate granulomas. Flow cytometry confirmed reduced frequencies of both B cells and T cells in the lungs of vaccinated animals. The researchers interpreted this as evidence of infection tolerance—a state in which the host tolerates the pathogen with minimal tissue damage rather than eliminating it.

The most striking results came from a hypersusceptible mouse model. SP140-deficient mice, generated on the otherwise resistant C57BL/6 background, develop severe, necrotizing lung lesions resembling those seen in human tuberculosis, which normally makes them difficult to study past five weeks of infection. When the team vaccinated these animals with pCas4 and challenged them with Mtb Erdman, bacterial burdens were again unchanged at five weeks post-infection. But a sensitive necrosis assay, in which the cell-impermeable dye Sytox Green is injected intravenously to stain extracellular DNA in dead tissue, revealed significantly less necrosis in vaccinated lungs. Histology confirmed the difference: mock-vaccinated mice had large necrotic lesions filled with acellular debris, while vaccinated mice had large cellular lesions entirely devoid of necrosis, though lacking clear organization.

Immunofluorescence microscopy of lung sections revealed how the lesions differed. In mock-vaccinated SP140-deficient mice, lesions had the expected architecture of human-like granulomas: a necrotic core rich in neutrophil and macrophage markers, surrounded by a lymphocytic cuff, with CD4 T cells almost entirely excluded from the core. In vaccinated mice, intact neutrophils and pockets of macrophages coexisted with CD4 T cells distributed throughout the lesion, including the core. Quantification showed significantly reduced neutrophil staining and significantly increased CD4 T cell numbers within lesion cores. Critically, markers of tissue integrity told a story of preservation: E-cadherin staining marking alveolar epithelium was retained throughout lesions in vaccinated mice, and alpha-smooth muscle actin staining suggested the presence of myofibroblasts associated with fibrotic containment, whereas mock-vaccinated lesions showed loss of epithelium and no fibrosis. Bacterial staining within the lesion core did not differ between groups, though vaccinated animals showed increased bacterial staining in the lesion cuff, possibly reflecting a redistribution of infected cells in less organized lesions.

Earlier in infection, at twenty-one days, before necrosis would normally develop, vaccinated mice already showed a distinct immune landscape: increased frequencies of RORγt-expressing CD4 T cells, including both Th17 and Tbet-positive Th1/17 subsets, along with decreased frequencies of conventional Tbet-positive Th1 cells, neutrophils, and both subsets of monocyte-derived macrophages, MNC1 and MNC2. The monocyte-derived cells also expressed lower levels of CD11c, potentially indicating a less inflammatory state. These shifts preceded the pathological differences, suggesting they contribute to preventing necrosis. B cells turned out to be dispensable: depleting them with anti-CD20 antibody during vaccination did not abolish the reduction in necrosis, and the effect also persisted when mice were vaccinated with a plasmid encoding RimJ alone, which increased RORγt-positive CD4 T cells similarly to the four-antigen cassette. Adoptive transfer experiments confirmed that RimJ-specific CD4 T cells traffic to infected lungs and localize within lesions near bacteria.

The authors are careful about the limits of their work. They did not establish a definitive mechanism, could not prove the observed T cell and myeloid changes are required for the protection, and did not compare the RVMA approach directly to candidates like M72. Not all variable antigens worked—an earlier cassette containing Rv0012 and LrfD2 antigens failed to alter infection outcomes, apparently because those proteins are poorly antigenic in mice. Still, the implications are substantial. The findings suggest that preclinical vaccine screening based solely on bacterial burden may discard candidates that could be genuinely effective at preventing disease. They also raise a caution: because some CD4 T cell responses to Mtb may promote the pathology that drives transmission, vaccines could theoretically worsen cavitary disease. If future work confirms that responses to rare variable antigens enhance disease tolerance in humans, a prevention-of-disease vaccine built around RimJ and its relatives could offer something TB has lacked for over a century—a way to keep infected lungs intact even when the bacteria persist within them.

Subject of Research: A rare variable antigen DNA vaccine against Mycobacterium tuberculosis that reduces lung pathology without lowering bacterial burden

Article Title: A Mycobacterium tuberculosis rare variable antigen vaccine reduces lung pathology in hypersusceptible SP140−/− mice

Article References: A Mycobacterium tuberculosis rare variable antigen vaccine reduces lung pathology in hypersusceptible SP140−/− mice. (n.d.). https://doi.org/10.3389/ftubr.2026.1814646

Image Credits: AI Generated

DOI: 10.3389/ftubr.2026.1814646

Keywords: tuberculosis, Mycobacterium tuberculosis, vaccine, rare variable antigens, CD4 T cells, lung pathology, necrosis, SP140, disease tolerance, RimJ, granuloma, prevention of disease