NK Cell Cytotoxicity Pathway: Receptors, Signalling and Assays

A natural killer cell decides by subtraction. It carries no rearranged antigen receptor and no memory of what it has seen. It presses a panel of germline-encoded receptors against a target and integrates the result: NKG2D, NKp46, DNAM-1 and CD16 pushing one way, inhibitory KIR and NKG2A pushing the other. If the sum crosses threshold, the cell polarises its granules and delivers perforin and granzyme B within minutes. Everything interesting about NK biology sits in how that sum is computed.

Missing self: how inhibitory receptors set the threshold

An NK cell is armed by default and held in check by ligation. Inhibitory KIR molecules in humans, and the structurally unrelated Ly49 family in mice, survey classical MHC class I on every cell they touch. NKG2A pairs with CD94 to read HLA-E, which is loaded almost exclusively with peptides cleaved from the leader sequences of HLA-A, -B and -C. HLA-E density is therefore a running report on whether the whole class I assembly line is working.

Hence the missing-self rule: a cell that stops displaying class I becomes visible to NK cells precisely because it has become invisible to cytotoxic T cells. Cytomegalovirus counters this with UL40, a synthetic HLA-E leader peptide that preserves NKG2A engagement while the rest of class I is destroyed. The less obvious half is education — an NK cell that never engages self class I becomes hyporesponsive rather than autoreactive, so the inhibitory repertoire sets sensitivity as well as restraint.

The activating arm: NKG2D, NKp46, DNAM-1 and CD16

NKG2D (KLRK1) is the archetypal induced-self receptor. Its ligands — MICA, MICB and the ULBPs in humans, RAE-1, H60 and MULT-1 in mice — are near-absent from healthy tissue and induced by the DNA damage response and proliferative stress. NKG2D reports cellular distress rather than foreignness, which is why it matters in tumour surveillance and why tumours shed soluble MICA to blunt it. NKp46 (NCR1) is the most conserved cytotoxicity receptor and the best pan-NK marker across species.

DNAM-1 (CD226) engages the nectin ligands CD155 and CD112, and its output is set by competition: TIGIT and CD96 bind the same ligands with higher affinity, so signalling depends on receptor ratio rather than DNAM-1 abundance. CD16 (FcγRIIIa) is the one receptor here whose specificity is not innate. Binding IgG already deposited on a target makes the NK cell the effector arm of antibody-dependent cellular cytotoxicity — the strongest single input in the panel, and able to override inhibition unaided.

Proximal signalling: DAP12, SYK and PI3K

None of those receptors signals through its own cytoplasmic tail; they borrow one. DAP12 (TYROBP) is a disulphide-linked homodimer carrying a single ITAM per chain, held to its partner receptor by a charged transmembrane interaction. Src-family kinases phosphorylate both tyrosines, creating the paired docking site that the tandem SH2 domains of SYK require. SYK then phosphorylates LAT and SLP-76, activating PLCγ2, Vav1 and PI3K; PIP₃ recruits AKT and drives granule polarisation and cytokine transcription. Human NKG2D takes a different route, signalling through DAP10 rather than an ITAM.

The brake is SHP-1. Phosphorylated ITIMs on KIR and NKG2A recruit it into the contact, and the instructive detail is its substrate: the dominant functional target in NK cells is Vav1, not the receptor-proximal kinases. Dephosphorylating Vav1 aborts Rac1-dependent actin remodelling, so an inhibited NK cell never builds a mature synapse rather than building one and declining to fire. Inhibition is fast, local and fully reversible — the cell disengages and moves on.

The lytic synapse: perforin and granzyme B

Crossing threshold triggers a stereotyped sequence: LFA-1 engages ICAM-1 to form an adhesion ring, actin clears from the centre of the contact, and the microtubule-organising centre migrates to the membrane dragging the secretory lysosomes with it. Only then do granules dock, prime and fuse. Each step is separable, and each carries a human immunodeficiency: MUNC13-4, syntaxin-11 and MUNC18-2 mutations all cause familial haemophagocytic lymphohistiocytosis by blocking priming or fusion while leaving granule content entirely intact.

Perforin is the delivery mechanism, held inactive at granule pH and oligomerising on the target membrane in a calcium-dependent manner once released into the near-neutral cleft. Granzyme B enters through those pores, cleaves BID to tBID and processes caspase-3 directly, converging on apoptosis from two directions at once. PRF1 mutations cause familial HLH type 2 — the cleanest demonstration that a single protein is rate-limiting for the whole cytotoxic programme.

The cytokine arm and where the pathway is targeted

Killing is only half the output. Activated NK cells are the dominant early source of IFN-γ in most infection models, releasing it hours before a T-cell response exists. IFN-γ polarises macrophages, supports Th1 differentiation and drives transcription of MHC class I — closing an elegant loop, because raising class I on surrounding cells makes them less visible to NK cells and more visible to CD8 T cells. The innate response actively hands its target over to adaptive immunity.

TNF-α is the second output, contributing to target apoptosis through TNFR1 and amplifying endothelial activation, and it is the arm most likely to drive pathology. That is why neutralising both cytokines is the standard test of whether a phenotype is cytotoxicity-driven or cytokine-driven. Every node on this map is now a drug target: anti-NKG2A and anti-KIR checkpoint agents, Fc-engineered antibodies and bispecific NK engagers working through CD16, and CAR-NK constructs grafted onto the DAP12–SYK core described above.

Key targets and matching reagents

Studying NK cytotoxicity in vivo

NK biology suits a mixed approach: functional-grade antibodies to perturb the system, quantitative immunoassays to read receptor balance and lytic content.

1. Neutralising the cytokine output

Anti-mouse IFN-γ (XMG1.2) and anti-mouse TNF-α separate the cytotoxic arm from the cytokine arm: if a phenotype survives neutralisation of both, it is contact killing. Both come low-endotoxin and azide-free from the In Vivo range, which matters more here than usual — trace LPS is a potent indirect NK activator through myeloid IL-12, IL-15 and IL-18, so contaminated material drives the very axis you are blocking.

2. Reading the activating–inhibitory balance

Because NK output is a sum, the informative measurement is a ratio. Quantify the activating side with NKG2D, NKp46, DNAM-1 and CD16, the inhibitory side with KIR3DL1 and NKG2A. Shifts in this balance distinguish an exhausted NK compartment from a numerically depleted one, and rising NKG2A against falling DNAM-1 is the classic tumour-escape signature.

3. Quantifying signalling and lytic machinery

DAP12/TYROBP, SYK, PI3K and SHP-1 cover the activating core and its principal brake, while perforin and granzyme B report loaded effector content. Run the granule assays on lysate and supernatant from the same experiment: content shows how the cells were primed, supernatant shows what they released, and the gap between them is usually where the biology is.

Featured products for this pathway

Frequently asked questions

For research use only. Not for use in diagnostic or therapeutic procedures.