Granzyme B refill keeps exhausted-looking CD8+ T cells armed for serial killing
Human CD8+ T lymphocytes do not simply empty their lytic granules and run out of ammunition. In this preprint, investigators describe a selective molecular reset in which granzyme B is preferentially renewed after acute degranulation, while perforin and several other granule-associated proteins recover poorly. The result is a cytotoxic program that remains active during prolonged target engagement, even when mature perforin is limited.
The central observation is strikingly asymmetric. After the initial discharge of canonical lytic granules, mature perforin falls, yet granzyme B is replenished through de novo protein synthesis and redistributed into vesicular compartments. Those newly stocked vesicles overlap less with mature perforin than the original granules did, but they still deliver enzymatically active cargo to successive targets. In practical terms, the effector cell does not rebuild the full original arsenal; it selectively restores the component most immediately useful for continued killing.
A selective lytic reset rather than a wholesale rebuild
This pattern matters because cytotoxic lymphocytes are often framed as if their late-phase killing depends mainly on death receptor pathways or cytokine-driven effects once the primary granule burst has been spent. Here, the cell’s own intracellular machinery remains central. By preserving and resupplying granzyme B, the CTL retains the protease most closely tied to apoptotic execution, even as the broader granule architecture changes after release.
The biology is plausible. Perforin creates access for granzymes, but granzyme B is the catalytic driver that pushes the target toward apoptosis by cleaving intracellular substrates and amplifying death signaling. If a CTL can keep granzyme B available in a second wave, it can continue to strike new targets without waiting for a complete granule regeneration cycle. The preprint positions this selective renewal as a molecular switch for sustained cytotoxicity rather than a passive leftover of earlier activation.
A second layer of the story concerns compartmentalization. Replenished granzyme B was found in vesicular compartments with reduced overlap with mature perforin, suggesting that the cell reorganizes its lytic storage rather than simply refilling the original granules in identical form. That reorganization could explain why later killing events are productive but kinetically heterogeneous: the effector cell still works, but not with the same timing or stoichiometry seen at the outset.
Sustained killing despite limited mature perforin
One of the most practical observations comes from the functional readout. Under conditions where mature perforin was difficult to detect, CTLs still delivered enzymatically active granzyme B, triggered altered calcium responses in target cells, and supported productive killing. The death process did not disappear; it slowed and diversified in timing. Some targets were eliminated more readily than others, which fits with a model in which the cytotoxic machinery has been remodeled but not dismantled.
This is the kind of result that changes how late-phase immune killing is interpreted at the bench. If a CTL can continue to supply granzyme B after the first degranulation wave, then persistence in tissue is not just a matter of cytokines or extrinsic death pathways. It is also a matter of intracellular replenishment, trafficking, and selective preservation of one molecular weapon over another.
The authors extend that logic to a harder context: engagement with resistant melanoma cells. In that setting, the GzmB-preserved effector state was maintained and further accentuated. Resistant tumor targets may force prolonged synapse formation and repeated release attempts, creating exactly the kind of environment in which selective granzyme B renewal becomes decisive. A CTL facing a stubborn tumor cell may therefore rely less on a full granule reset and more on rapid top-up of the most lethal enzymatic component.
What this means for sustained immune surveillance
The broader implication is that sustained cytotoxicity is not a uniform afterburn of the first killing event. It is an adaptive remodeling process. The CTL appears to preserve the killing axis that can be rebuilt fastest and used most efficiently, while allowing other parts of the lytic package to lag behind. In a prolonged immune response, that asymmetry may preserve serial killing capacity when antigen exposure continues and fresh targets keep arriving.
At the bedside, the relevance is indirect but real. Tumor control, viral clearance, and tissue injury all depend on how efficiently effector T cells recycle their cytotoxic machinery. A pathway that selectively restores granzyme B could help explain why some T cells remain dangerous after repeated engagements, even when their granules no longer resemble their naïve state. It also raises the possibility that immunotherapies which expand or sustain CTLs may ultimately depend on this intracellular renewal process, not only on T-cell numbers or activation markers.
The mechanistic language here should stay grounded in the data. This is a preprint, so the findings remain provisional, and the article does not provide clinical outcome data in patients. Even so, the cellular logic is coherent: repeated target contact drains mature perforin, de novo synthesis replenishes granzyme B, vesicular redistribution preserves delivery, and target cells still die, albeit with altered kinetics.
A preclinical mechanism that deserves validation
The unanswered question is not whether CTLs can kill after degranulation; they clearly can. The harder question is how often selective granzyme B renewal governs that process in vivo, across different tissues, inflammatory states, and tumor microenvironments. The preprint points to one molecular lever, but it does not establish how broadly that lever operates in human disease or whether it can be therapeutically tuned without intensifying collateral tissue damage.
For now, the work reframes sustained CTL function as a selective maintenance program. Granzyme B is not merely spent and replaced; it is preferentially protected, replenished, and redeployed while other lytic components recover more slowly. In the architecture of prolonged cytotoxic immunity, that selective refill may be the difference between a transient hit and continued serial killing.