Skip to content
NKCELLThe NK guide

The biology

How Do NK Cells Kill Target Cells?

A step-by-step look at recognition, contact, degranulation and the granzyme B–perforin pathway—and why a lab signal that NK cells have released their granules is not the same as proof that a target cell died.

Natural killer (NK) cells kill target cells by making close contact, releasing cytotoxic granules and delivering granule proteins that push the target into apoptosis—a controlled form of cell death. The key distinction is that recognition is not killing, and degranulation is not lysis. An NK cell can recognize a cell that looks suspicious—for example, a virus-infected cell or a tumor cell—but the target does not die just because an NK cell has found it. The kill depends on a sequence: signal integration, contact, granule release and effective delivery into the target cell.

Recognition is a balance, not a single marker

The NCI Dictionary describes natural killer cells as immune cells with granules containing enzymes that can kill tumor cells or cells infected by a virus. Before that happens, the NK cell reads surface proteins on the candidate target. A major inhibitory input comes from MHC class I molecules—think of them as identification tags displayed on many healthy cells. A study of human peripheral blood NK cells identified a subpopulation that lacks inhibitory receptors for self-MHC and appeared developmentally immature. That study supports the importance of MHC class I inhibition and missing-self biology, but it also warns against a cartoon version: NK-cell subsets differ, and loss of MHC class I does not automatically trigger killing.

Missing-self is the idea that NK cells can be more responsive when a cell displays too little MHC class I. A useful explanatory model—not a claim that one study measured every possible outcome—is that the NK cell integrates inhibitory and activating signals. The study specifically supports inhibitory receptors for self-MHC as one part of that picture. It does not turn loss of MHC class I into a guaranteed kill signal.

Contact makes the mechanism local

Once the signals favor action, the NK cell comes into close contact with the target. That contact matters because the cytotoxic granules—small sacs inside the NK cell—are released toward the target rather than acting as a general poison. The next step is degranulation: the fusion of granule membranes with the NK cell's outer membrane and the release of granule contents. In plain terms, contact and release are separate events. An NK cell can touch a target without completing the kill, and a release signal can occur without the target yet being destroyed.

Degranulation is a release step, not proof of target-cell death

The NCI Dictionary says NK-cell granules contain enzymes that can kill tumor cells or virus-infected cells. In the lab, one sign that an NK cell has degranulated is CD107a on its surface: when granule membranes fuse with the outer membrane, CD107a becomes detectable. A methods paper on functional analysis of human NK cells by flow cytometry describes CD107a degranulation after stimulation with target cells such as K562, a laboratory target cell line, alongside cytokine readouts. Cytokines are signaling molecules. The same paper explicitly cautions that degranulation does not equal target-cell lysis. Lysis means the target cell is broken down or destroyed. So a CD107a result and a killing assay answer different questions: one shows release, the other asks whether the target actually died.

Perforin and granzyme B: delivery into the target

Among the granule proteins, granzymes are enzymes and perforin is a protein in the granzyme B–perforin pathway. An in vitro study of granzyme B entry and trafficking traced granzyme B entering target cells and moving through intracellular compartments during granzyme B–perforin-mediated apoptosis. In vitro means the work was done in laboratory systems, not in a person. The study's focus on entry and trafficking makes an important point: release granules and the target dies is too compressed. Granzyme B has to reach the target cell and move inside it. The paper describes an intracellular route, not an instant effect.

This is also where a simple story can overreach. The in vitro study supports a granzyme B–perforin route to apoptosis; it is not evidence that every NK-cell kill uses only this pathway. And even a successful delivery step is not the same as a clinical outcome.

What NK-cell function means in a lab result

Research on NK-cell function often separates recognition, degranulation, cytokine production and target-cell death. A flow-cytometry experiment may measure CD107a, K562 stimulation or cytokines; a killing experiment asks whether target cells actually die. The methods paper's caution is a good rule for reading: a signal of release is not the same as lysis, and a laboratory readout is not the same as a clinical outcome. The NCI definition describes what NK cells can do—kill certain tumor or virus-infected cells—not a measure of an individual's cancer risk or infection outcome. That distinction matters because immune function is context-dependent.

This is general educational information, not medical advice.

Sources and further reading

Keep reading

Follow the next question.

The biology