Despite decades of global vaccination campaigns, Mycobacterium tuberculosis (Mtb) continues to kill more people each year than nearly any other infectious agent, claiming roughly 1.25 million lives in 2023. While the Bacillus Calmette–Guérin (BCG) vaccine protects infants from severe tuberculosis (TB), the century-old treatment offers inconsistent protection for adults.
A new study published in Science Translational Medicine describes a major step toward developing the next generation of TB vaccines using immunopeptidomics—a tool for understanding how T cells process and recognize antigens. Researchers from the Massachusetts Institute of Technology (MIT) and the Ragon Institute of Mass General MIT and Harvard used immunopeptidomics to identify bacterial proteins that human immune cells display to mobilize an attack against Mtb.
Leddy and colleagues designed nucleoside-modified mRNAs that encode key bacterial peptides and are translationally fused to signals that direct them to the subcellular component that processes and presents them during infection, which were developed and validated in vitro and in silico. This research presents a framework for a bottom-up, objective approach to designing TB vaccines, combining quick iterative testing of mRNA vaccine designs with focused assessment of antigen presentation in human cells to maximize antigen presentation prior to animal research or human clinical trials.
Mapping the hidden signals of infection
Immunity to TB depends heavily on CD4+ T cells to recognize infected cells through molecules called MHC class II (MHC-II). However, we still lack essential information for selecting effective vaccine targets, specifically which of Mtb’s 4,000 proteins infected human cells process and present.
To fill this gap, researchers led by Leddy and colleagues developed a high-resolution immunopeptidomics pipeline, using mass spectrometry to identify Mtb-derived peptides displayed on MHC-II molecules of infected human phagocytes. These are the cells that normally engulf bacteria and trigger immune responses. Using human monocyte-derived dendritic cells (hMDCs) from six donors representing 44 distinct MHC-II variants, the team isolated and analyzed thousands of peptides, including 27 that originated from 13 Mtb proteins.
Several peptides came from well-known TB vaccine targets such as EsxB and EsxG, but others emerged from previously overlooked bacterial proteins. No single antigen was present in all donors, which means that effective TB vaccines may need to include more than one antigen to work with the genetic diversity of immune systems around the world.
When the researchers examined where these Mtb proteins come from inside the bacterium, they found a striking pattern: the immune system preferentially displayed peptides from secreted or membrane-associated proteins, particularly those secreted by Mtb’s type VII secretion systems (T7SS). These include the proline-glutamate (PE) and proline-proline—lutamate (PPE) proteins that are found on the surface of the Mtb outer membrane and interact directly with host cells. These secreted proteins are essentially what the immune system ‘sees’ during infection. That finding not only reinforces the focus on these protein families as vaccine targets but also helps explain why other abundant bacterial proteins remain immunologically invisible—they never reach the compartments where MHC-II molecules are loaded.
Conserved yet vulnerable
An important question for any vaccine target is whether the pathogen can mutate to escape immune recognition. To test this, the researchers analyzed more than 51,000 global Mtb genomes. They found that nearly all the identified MHC-II epitopes—the small peptide fragments recognized by T cells—were highly conserved, with mutations affecting fewer than 0.15% of strains. That level of conservation suggests to the authors two things: the bacterium cannot easily change these sequences without harming itself, and vaccines targeting them would likely work across most Mtb lineages. Interestingly, these same regions also showed signs of “purifying selection”—evolutionary pressure that keeps the sequence stable—further strengthening their candidacy for vaccine development.
To confirm that these MHC-II peptides were truly recognized by human immune systems, the team exposed blood cells from individuals previously infected with Mtb to four of the identified peptides. One peptide derived from the EsxB protein, a well-documented Mtb virulence factor necessary for phagosome escape, triggered strong immune responses in six of seven donors, suggesting it is naturally processed and recognized during infection—and that it could be a vaccine candidate. Another peptide from a bacterial membrane protein called TatA did not elicit a similar response, which the authors speculate may have occurred because it is too similar to human proteins, potentially dampening immune activation through self-tolerance. These results highlight the fine balance vaccine designers must strike between choosing potent targets and avoiding self-reactivity.
A new blueprint for TB vaccines
Armed with this peptide “map,” the researchers then used their mass spectrometry platform to prototype and optimize experimental mRNA vaccines. Similar to COVID-19 mRNA vaccines, these deliver genetic instructions encoding selected Mtb antigens into human immune cells. They discovered that directing these antigens specifically to lysosomes—the acidic compartments where MHC-II molecules are loaded—dramatically improved presentation of the same peptides observed during real infection.
For example, when the team delivered mRNA encoding the EsxB and EsxA proteins together, mimicking their natural partnership inside Mtb, human cells efficiently displayed the correct MHC-II peptides. A single mRNA encoding a fused EsxB–EsxA protein performed just as well, suggesting a compact and efficient design for future vaccine candidates. Leddy and colleagues stated, “This shows we can use the biology of infection itself to guide mRNA vaccine engineering.”
The study provides a foundational blueprint for rational TB vaccine design based on human antigen presentation rather than animal models or guesswork. By revealing which bacterial fragments are truly visible to human T cells, the research bridges a critical gap between immunology, genomics, and vaccine technology. Beyond TB, this immunopeptidomics-guided approach could be adapted to other intracellular pathogens, where knowing what the immune system actually “sees” is key to designing better vaccines. After a century of relying on BCG, the world may finally have a path toward a new generation of TB vaccines—ones precisely tuned to the human immune system.
