Cytotoxic T Cell Killing

The assassin that checks ID before it kills

Every nucleated cell in your body is forced to advertise what it is doing inside. Once a second, the cell's proteasome shreds a sample of its own proteins into short fragments, ships them to the surface on a molecule called MHC class I, and holds them up like a continuously updated billboard. A cytotoxic T cell — a CD8+ T lymphocyte, also called a killer T cell or CTL — spends its life walking that billboard. If the peptides are all self, it moves on. If it finds a fragment of a virus, an intracellular bacterium, or a mutated cancer protein, it stops, latches on, and kills the cell that displayed it.

What makes this remarkable is the precision. The killer T cell never sees the whole pathogen. It infers infection entirely from short peptide windows the infected cell is compelled to show. And when it strikes, it does not lyse the cell open and spill virus everywhere — it triggers apoptosis, the clean, contained, suicide program, so the cell quietly packages itself for disposal with its contents (and any virions inside) sealed in. This is targeted, self-restrained cellular execution, and it is the backbone of our defense against viruses and cancer.

How the killing works, step by step

The full kill is a tightly choreographed sequence:

1. Antigen presentation. Inside an infected cell, viral proteins are tagged with ubiquitin and degraded by the proteasome into peptides of roughly 8–10 amino acids. The TAP transporter pumps them into the endoplasmic reticulum, where they are loaded onto newly made MHC-I (a heavy chain plus β2-microglobulin). The peptide-MHC-I complex traffics to the cell surface.
2. Scanning and recognition. The CTL's T-cell receptor (TCR) — a unique, genetically rearranged heterodimer — samples these complexes. The CD8 co-receptor grips a conserved (non-variable) part of MHC-I, stabilizing weak interactions long enough for the TCR to read the peptide. Recognition is exquisitely sensitive: as few as ~1–10 specific peptide-MHC complexes out of ~100,000 MHC molecules can trigger a response (a single agonist complex can suffice for early signaling).
3. Immune synapse formation. A productive TCR engagement triggers LFA-1 on the T cell to clamp onto ICAM-1 on the target, forming a sealed, bull's-eye contact called the immune synapse (a central TCR cluster, the cSMAC, ringed by an adhesion zone, the pSMAC). The gap narrows to ~15 nm and a tiny secretory cleft is walled off from surrounding tissue.
4. Polarization. The CTL's centrosome (microtubule-organizing center) and its preformed lytic granules travel along microtubules to dock right at the synapse — a striking reorientation of the cell's entire machinery toward the contact point in minutes.
5. Degranulation. A rise in intracellular Ca²⁺ triggers the granules to fuse with the membrane and release their cargo into the sealed cleft: perforin and a suite of granzymes (chiefly granzyme B).
6. Delivery and execution. Perforin, in the presence of Ca²⁺, polymerizes into pores that permeabilize the target's endosomal membrane, letting granzyme B reach the cytosol. Granzyme B is a serine protease that cleaves the executioner caspase, caspase-3, and the BH3-only protein Bid (generating tBid, which permeabilizes mitochondria). Both routes converge on apoptosis: chromatin condenses, DNA fragments, the membrane blebs, and phosphatidylserine flips outward as an "eat me" signal.
7. Detachment and serial killing. The CTL protects itself (serpin PI-9/serpinB9 neutralizes stray granzyme B in its own cytosol; its membrane resists perforin and is quickly repaired after degranulation), then disengages undamaged and moves to the next target.

There is also a slower, contact-dependent backup: the CTL displays Fas ligand (FasL/CD95L), which binds Fas (CD95) death receptors on the target and recruits FADD and caspase-8 to launch apoptosis through the extrinsic pathway — important for killing when granule contents are exhausted and for regulating the immune response itself.

The molecular cast

• CD8+ T cell (CTL). The effector. Differentiates from a naïve CD8 T cell after a dendritic cell cross-presents antigen and provides costimulation (CD28-B7) and CD4 helper signals (the "licensing" of dendritic cells via CD40).
• TCR. The recognition unit. Each clone has one specificity, generated by V(D)J recombination, producing a repertoire of ~10⁷–10⁸ distinct TCRs in a human.
• MHC class I (HLA-A, -B, -C in humans). The display platform on all nucleated cells. Highly polymorphic — thousands of HLA alleles exist — which is why transplant matching is hard and why populations vary in resistance to specific viruses.
• Perforin (PRF1). Pore-former; ~20 monomers oligomerize into a ~13–20 nm ring. Loss-of-function mutations cause familial hemophagocytic lymphohistiocytosis (FHL2), a fatal hyperinflammatory disease.
• Granzyme B (GZMB). The lead protease; uniquely cleaves after aspartate residues, like caspases, letting it shortcut the death program.
• FasL / Fas (CD95L / CD95). The death-receptor backup pathway.
• Co-receptors and adhesion: CD8, LFA-1, ICAM-1. Stabilize recognition and build the synapse.

Cytotoxic T cell vs helper T cell vs NK cell

Property	Cytotoxic T cell (CD8+)	Helper T cell (CD4+)	NK cell
Immune branch	Adaptive	Adaptive	Innate
Antigen receptor	Rearranged TCR (one specificity)	Rearranged TCR (one specificity)	Germ-line activating/inhibitory receptors
Reads which MHC	MHC class I (all nucleated cells)	MHC class II (antigen-presenting cells)	Senses absence of MHC-I ("missing self")
Primary job	Directly kill the displaying cell	Secrete cytokines, orchestrate response	Kill stressed / MHC-I-low cells
Killing machinery	Perforin + granzymes, FasL	Usually none (some kill via FasL)	Perforin + granzymes (same toolkit)
Target peptide source	Intracellular / viral proteins	Engulfed extracellular proteins	Not peptide-specific
Memory	Yes — long-lived memory CD8 cells	Yes — long-lived memory CD4 cells	Limited ("adaptive NK" exists)
Escape it dodges	Sees cells that keep MHC-I up	Needs professional APCs	Covers cells that drop MHC-I

The key complementarity is in the last two rows: a virus or tumor that hides from CD8 CTLs by switching off MHC-I makes itself a prime target for NK cells, which kill precisely the cells that have lost their MHC-I billboard.

By the numbers

Quantity	Value	Note
Presented peptide length	8–10 amino acids	Fits the closed MHC-I groove
MHC-I molecules per cell	~10⁴–10⁵	On a typical nucleated cell
Peptide-MHC needed to trigger	~1–10 specific complexes	Extreme TCR sensitivity
Synapse gap	~15 nm	Sealed secretory cleft
Perforin pore diameter	~13–20 nm	~20-monomer ring
Time to first apoptotic signs	~5–10 min	After stable synapse
Time to full target death	~30–60 min	Caspase cascade completes
Serial kills per CTL	A few to dozens / day	Up to ~10 in a few hours when crowded
Naïve→effector expansion	~10⁴-fold over ~1 week	Clonal burst after activation
Human TCR repertoire	~10⁷–10⁸ specificities	From V(D)J recombination

Where it shows up — disease and medicine

• Antiviral defense. CD8 CTLs clear influenza, hepatitis B, EBV, and SARS-CoV-2-infected cells. The size and breadth of the CD8 response often predicts how well a person controls a viral infection.
• HIV. CTLs control HIV early, but the virus's Nef protein downregulates MHC-I to hide infected cells, and the virus mutates the very peptide epitopes CTLs recognize ("CTL escape mutants") — a major reason HIV is never cleared.
• Cancer immunotherapy. CTLs are the cells that checkpoint inhibitors (anti-PD-1/anti-CTLA-4) release from the brakes, and the cells that CAR-T therapy re-engineers with a synthetic receptor to recognize tumor antigens (e.g., CD19 for leukemia/lymphoma) independently of MHC. Tumors frequently delete MHC-I or β2-microglobulin to escape — a known mechanism of immunotherapy resistance.
• Transplant rejection. Recipient CTLs recognize donor MHC-I as foreign and attack the graft — why HLA matching and immunosuppression are essential.
• Type 1 diabetes and autoimmunity. Autoreactive CTLs kill insulin-producing β cells in the pancreatic islets; CTL-mediated tissue destruction also drives some viral hepatitis and graft-versus-host disease.
• Genetic failure. Mutations in perforin (PRF1) or in granule-trafficking genes (e.g., Munc13-4, syntaxin-11) cause familial hemophagocytic lymphohistiocytosis — without functional killing, T cells cannot turn off an infection and the immune response runs away into lethal hyperinflammation.

Common misconceptions

• "Killer T cells lyse the target like a popped balloon." No — that would be necrosis and would spill live virus. CTLs trigger apoptosis, a contained suicide program; granzyme B activates caspases so the cell neatly fragments itself and is cleared by phagocytes with its contents sealed.
• "The T cell sees the virus." It never sees an intact pathogen. It sees only short 8–10-residue peptides the infected cell is forced to present on MHC-I. The whole system is indirect inference from the cell's protein billboard.
• "Perforin does the killing." Perforin is the delivery truck, not the assassin. It makes pores so granzymes can enter; granzyme B is what cleaves caspases and launches apoptosis. Perforin alone is far less lethal.
• "The T cell dies after one kill." The opposite — it is reusable. It protects itself with serpin PI-9 and a perforin-resistant membrane, detaches intact, and serially kills many targets over hours to days.
• "Lowering MHC-I always helps a virus hide." Only from T cells. NK cells follow the "missing-self" rule and specifically attack cells that have dropped MHC-I, so MHC-I downregulation trades one assassin for another.
• "CD4 helper T cells also kill cells directly." Rarely. Helper T cells mostly secrete cytokines to coordinate the response; the dedicated killing of MHC-I-displaying cells is the CD8 cytotoxic T cell's job (NK cells share the toolkit on the innate side).