Intrinsic and Extrinsic Apoptotic Pathways

Apoptosis is executed by caspases, a family of cysteine-aspartate proteases synthesized as inactive zymogens (procaspases) and activated by proteolytic cleavage or by forced proximity within a multi-protein activation complex. Caspases are divided functionally into initiator caspases (caspase-8, -9, -10), which sit at the top of a pathway and are activated by dimerization within a signaling platform, and executioner (effector) caspases (caspase-3, -6, -7), which are activated by cleavage from an initiator caspase and carry out the bulk of cellular demolition — cleaving nuclear lamins, activating the DNase CAD/DFF40 by destroying its inhibitor ICAD/DFF45, and cleaving cytoskeletal proteins to produce the characteristic membrane blebbing and cell shrinkage of apoptosis. The intrinsic (mitochondrial) pathway is triggered by signals that arise inside the cell: DNA damage detected by p53 and the DNA Damage Response, oxidative stress, ER stress, cytokine or growth-factor withdrawal, and improper cell-cycle entry. These signals converge on the Bcl-2 family of proteins, which control mitochondrial outer membrane permeabilization (MOMP). When pro-apoptotic Bcl-2 family members prevail, the mitochondrial outer membrane is breached, releasing cytochrome c and other pro-apoptotic factors from the intermembrane space into the cytosol; cytochrome c then nucleates the apoptosome, activating caspase-9. The extrinsic (death-receptor) pathway is triggered from outside the cell: a ligand such as Fas ligand (FasL/CD95L), TNF-α, or TRAIL binds a death receptor on the plasma membrane, recruiting adaptor proteins to build the death-inducing signaling complex (DISC), which activates caspase-8. In many cell types (so-called "type I" cells), caspase-8 activation alone generates enough caspase-3 to kill the cell; in others ("type II" cells, including most hepatocytes), the signal must be amplified through the mitochondrial pathway via cleavage of the BH3-only protein Bid. This cross-talk means the two pathways are not always independent — they are wired together, with the extrinsic pathway able to co-opt the intrinsic one for amplification.

Step 1 — Sensing stress: the Bcl-2 rheostat (intrinsic pathway). The Bcl-2 family comprises three functional classes, all sharing Bcl-2 homology (BH) domains: (1) anti-apoptotic guardians (Bcl-2, Bcl-xL, Mcl-1, Bcl-w, A1), which have four BH domains (BH1–4) and reside in the mitochondrial outer membrane, restraining death; (2) pro-apoptotic effectors (Bax, Bak, and Bok), which also have multiple BH domains and, when activated, oligomerize to form pores in the outer mitochondrial membrane; and (3) BH3-only proteins (Bid, Bad, Bim, Puma, Noxa, Bik), which sense specific stress signals and either directly activate Bax/Bak or neutralize the anti-apoptotic guardians by occupying their hydrophobic BH3-binding groove. DNA damage stabilizes p53, which transcriptionally induces Puma and Noxa; growth factor withdrawal dephosphorylates and activates Bad; ER stress activates Bim. The balance among these competing binders determines whether Bax and Bak remain monomeric and inert or oligomerize into membrane-spanning pores.

Step 2 — Mitochondrial outer membrane permeabilization (MOMP). Once Bax and Bak are activated (by direct BH3-only "activators" such as Bid and Bim, with other BH3-only proteins acting as "sensitizers" that free them from anti-apoptotic restraint), they insert into the outer mitochondrial membrane and oligomerize into large pores. This releases intermembrane-space proteins into the cytosol:

• cytochrome c, normally an electron carrier of the respiratory chain,
• Smac/DIABLO, which neutralizes inhibitor-of-apoptosis proteins (IAPs),
• AIF and endonuclease G, which act in a caspase-independent manner on chromatin.

MOMP is widely considered the point of no return for the intrinsic pathway: once a threshold fraction of mitochondria are permeabilized, cell death proceeds essentially irreversibly.

Step 3 — Apoptosome assembly and caspase-9 activation. In the cytosol, cytochrome c binds Apaf-1 (apoptotic protease activating factor-1), inducing an ATP/dATP-dependent conformational change that exposes a nucleotide-binding domain and allows seven Apaf-1 molecules to oligomerize into a wheel-shaped heptamer called the apoptosome. The apoptosome recruits procaspase-9 via CARD (caspase recruitment domain) interactions, clustering multiple procaspase-9 molecules together. Because caspase-9 activation depends on proximity-induced dimerization rather than proteolytic cleavage per se, the apoptosome acts as an activation platform, analogous to an "activation-by-induced-proximity" template.

Step 4 — Death receptor ligation and DISC assembly (extrinsic pathway). Death receptors are members of the TNF-receptor superfamily bearing a cytoplasmic death domain (DD): Fas (CD95), TNFR1, and the TRAIL receptors DR4/DR5. Ligand binding (FasL is typically trimeric, clustering three receptors) causes receptor trimerization and recruitment of the adaptor FADD via DD–DD interactions. FADD in turn recruits procaspase-8 (and procaspase-10) via death-effector domain (DED) interactions, assembling the DISC. As with caspase-9, caspase-8 is activated by induced proximity — clustering multiple zymogens together allows low-level intrinsic protease activity to cleave neighboring procaspase-8 molecules, generating active caspase-8 heterotetramers.

Step 5 — Regulation at the DISC: FLIP. The cellular FLICE-inhibitory protein (c-FLIP), structurally similar to procaspase-8 but catalytically dead (or weakly active), competes for DISC recruitment. The ratio of c-FLIP to procaspase-8 determines whether DISC assembly produces a fully active caspase-8 dimer or an inactive heterodimer, providing a tunable checkpoint on death-receptor signaling — important in T-cell activation and tumor immune evasion.

Step 6 — Cross-talk: Bid links extrinsic to intrinsic. Active caspase-8 cleaves the BH3-only protein Bid to generate truncated Bid (tBid), which translocates to mitochondria and directly activates Bax/Bak, triggering MOMP. This defines the type II amplification loop: in hepatocytes and pancreatic beta cells, Fas-induced death requires this mitochondrial amplification because DISC-generated caspase-8 alone is insufficient (in type I cells such as lymphocytes, direct DISC-to-caspase-3 activation suffices and Bcl-2 overexpression does not block Fas-induced death).

Step 7 — Convergence: executioner caspase activation and demolition. Both caspase-9 (from the apoptosome) and caspase-8 (from the DISC) cleave and activate caspase-3 and caspase-7. These executioners:

• cleave ICAD/DFF45, releasing the nuclease CAD/DFF40 to fragment chromosomal DNA into nucleosomal ladders (multiples of ~180–200 bp),
• cleave nuclear lamins, causing nuclear envelope breakdown,
• cleave ROCK1, causing constitutive membrane blebbing,
• cleave cytoskeletal and adhesion proteins, producing cell shrinkage, and
• expose phosphatidylserine on the outer leaflet of the plasma membrane, a signal recognized by macrophages for silent, non-inflammatory phagocytic clearance.

Step 8 — Inhibitor-of-apoptosis proteins (IAPs) as a brake. XIAP directly binds and inhibits caspase-9, -3, and -7. Smac/DIABLO released during MOMP binds XIAP and displaces it from the caspases, relieving this brake — another point of intrinsic/mitochondrial control over the whole system, reinforcing that MOMP does double duty: it releases both the activator (cytochrome c) and the de-repressor (Smac) of the caspase cascade.

Genetic dissection in C. elegans: Horvitz and the cell-death genes. The molecular logic of apoptosis was first cracked using the nematode C. elegans, whose development is invariant: exactly 1090 somatic cells are generated and exactly 131 die by programmed cell death in every wild-type animal. Working from Sydney Brenner's genetic toolkit, H. Robert Horvitz and colleagues screened for mutants in which the normally doomed 131 cells survived. They identified ced-3 and ced-4 ("cell death abnormal") as genes required for essentially all programmed cell deaths — loss-of-function mutants in either gene produced viable adults carrying an extra 131 cells that should have died. A third gene, ced-9, acted in the opposite direction: loss-of-function ced-9 mutations caused ectopic, lethal cell death even in cells that should have survived, and this phenotype was suppressed by simultaneous loss of ced-3, placing ced-9 upstream of ced-3 as a negative regulator. This established the core genetic circuit — a repressor of death (ced-9) restraining a downstream killing machine (ced-4/ced-3) — that would later map directly onto the mammalian Bcl-2/Apaf-1/caspase-9 axis. Ced-3 was subsequently shown to encode a cysteine protease homologous to interleukin-1β converting enzyme (ICE, the founding member of the caspase family), and Ced-9 was shown to be structurally and functionally homologous to human Bcl-2. This work earned Horvitz, together with Sydney Brenner and John Sulston, the 2002 Nobel Prize in Physiology or Medicine.

Bcl-2 as an oncogene that blocks death, not promotes division. Bcl-2 was originally cloned from the chromosomal translocation t(14;18) recurrent in human follicular lymphoma, where it is juxtaposed to the immunoglobulin heavy-chain enhancer, causing overexpression. Initially assumed to be a conventional proliferative oncogene, Bcl-2 was shown by David Vaux, Suzanne Cory, and Jerry Adams (1988) to promote the survival of a growth-factor-dependent hematopoietic cell line when the growth factor was withdrawn, without increasing proliferation rate — demonstrating for the first time that a human oncogene could act by inhibiting cell death rather than by driving cell-cycle entry. This reframed cancer genetics: tumors could arise not only from too much division but from too little death.

Biochemical reconstitution of the apoptosome: Xiaodong Wang's cell-free system. In a landmark series of papers in the mid-1990s, Xiaodong Wang and colleagues fractionated cytosol from HeLa cells and asked what was necessary and sufficient to activate caspase-3 in a test tube. They found that cytochrome c, released from isolated mitochondria, together with dATP, was required to activate procaspase-3-cleaving activity in cytosolic extracts. Purification of the responsible factors identified Apaf-1 as the cytochrome c- and dATP-dependent activator of procaspase-9, and biochemical reconstitution with purified cytochrome c, dATP, Apaf-1, and procaspase-9 alone was sufficient to generate active caspase-9 and caspase-3 — directly proving the cytochrome c → Apaf-1 → caspase-9 → caspase-3 cascade in a defined, cell-free system, independent of any other cellular component. Cryo-EM structures later (Yigong Shi and colleagues, 2011–2015) confirmed the wheel-shaped heptameric apoptosome architecture predicted from this biochemistry.

Discovery of Fas/CD95 and the DISC. The Fas antigen was identified in the late 1980s (Shin Yonehara; Peter Krammer's group) as a cell-surface protein whose antibody-mediated cross-linking triggered rapid apoptosis, and its natural ligand FasL was cloned shortly after. Marcus Peter and Peter Krammer's laboratory subsequently immunoprecipitated the ligand-bound Fas receptor from activated cells and identified a rapidly assembling multi-protein complex — the DISC — containing FADD and procaspase-8, directly cleaved and activated within it. This established that death receptors trigger apoptosis through a proximity-induced activation platform analogous to, but molecularly distinct from, the mitochondrial apoptosome — the two pathways converging only downstream, at the level of executioner caspases.

Knockout mice confirm non-redundant roles. Apaf1-knockout and caspase-9-knockout mice display strikingly similar, severe brain overgrowth phenotypes (excess neural tissue due to failed developmental apoptosis) and perinatal lethality, confirming the intrinsic pathway's essential role in neurodevelopment. Fas-deficient (lpr) mice and FasL-deficient (gld) mice, in contrast, develop lymphoproliferative disease and autoimmunity, showing that the extrinsic pathway is essential specifically for eliminating self-reactive and activated lymphocytes — a genetically clean dissociation of the two pathways' distinct biological jobs.

Example 1 — Predicting a lymphoma cell's response to chemotherapy from its Bcl-2 status. A follicular lymphoma biopsy carries the classic t(14;18) translocation and shows Bcl-2 protein levels roughly 20-fold higher than a normal B lymphocyte, by immunoblot densitometry. The oncologist wants to know whether standard genotoxic chemotherapy (which acts by inducing DNA damage and p53-dependent transcription of Puma and Noxa) will be effective, and whether adding a BH3-mimetic drug would help. Reasoning: chemotherapy raises the level of pro-apoptotic BH3-only proteins (Puma, Noxa), but Bcl-2 overexpression provides a 20-fold excess buffering capacity — most of the newly induced Puma/Noxa will simply be sequestered by the excess Bcl-2, so few Bax/Bak molecules will be freed to trigger MOMP, and killing will be blunted. A BH3-mimetic drug (such as venetoclax, which occupies the BH3-binding groove of Bcl-2 directly) does not need transcriptional induction — it acts stoichiometrically to displace pre-bound BH3-only proteins and Bax/Bak immediately. Combining chemotherapy (raising pro-apoptotic BH3-only levels) with a BH3-mimetic (neutralizing the excess Bcl-2 buffer) should synergize, because the two interventions attack both sides of the same rheostat — this is precisely the rationale for combining venetoclax with standard chemo-immunotherapy in clinical protocols for chronic lymphocytic leukemia and some lymphomas.

Example 2 — Distinguishing type I from type II death-receptor signaling experimentally. A researcher engineers two cell lines to overexpress Bcl-2 at 15-fold normal levels, then treats each with agonistic anti-Fas antibody and measures the fraction of cells apoptotic at 6 hours.

Reasoning: in Jurkat cells, Bcl-2 overexpression barely changes the outcome (78% → 74%), indicating that DISC-generated caspase-8 directly activates enough caspase-3 without needing mitochondrial amplification — a type I response. In hepatocytes, the same Bcl-2 overexpression nearly abolishes Fas-induced death (82% → 11%), showing that death in this cell type absolutely requires the mitochondrial amplification loop (caspase-8 → Bid cleavage → tBid → Bax/Bak → MOMP → caspase-9/-3), which Bcl-2 blocks upstream of MOMP — a type II response. This exact experimental logic (using Bcl-2 transgenic mice and Bid-knockout mice) was used to establish that Fas-induced fulminant hepatitis in mice is entirely Bid-dependent, while Fas-induced T-cell death is largely Bid-independent — a result with direct clinical relevance to Fas-antibody-induced liver toxicity in early clinical trials.

Cancer therapy. BH3-mimetic drugs, most notably venetoclax, directly occupy the BH3-binding groove of Bcl-2, displacing sequestered pro-apoptotic proteins and tipping cancer cells (particularly chronic lymphocytic leukemia, which is often exquisitely Bcl-2-dependent) toward MOMP without requiring DNA damage — a rational, structure-based drug built directly on the mechanistic understanding of the Bcl-2 rheostat.

Immunotherapy and death-receptor engineering. TRAIL-receptor agonists have been explored clinically to selectively trigger apoptosis in tumor cells that overexpress DR4/DR5, since normal tissues are relatively resistant; understanding c-FLIP-mediated resistance at the DISC has informed combination strategies that lower FLIP levels to sensitize tumors.

Autoimmune disease. Mutations in FAS or FASLG cause autoimmune lymphoproliferative syndrome (ALPS) in humans, phenocopying the lpr/*gld* mouse phenotypes, directly linking extrinsic-pathway failure to lymphocyte accumulation and autoimmunity — a clinical validation of the mouse genetics.

Neurodegeneration. Excessive intrinsic-pathway activation, driven by oxidative stress and mitochondrial dysfunction, is implicated in neuronal loss in Parkinson's and Alzheimer's disease, motivating caspase and Bax/Bak inhibitor research as neuroprotective strategies.

Viral immune evasion. Many viruses (e.g., cytomegalovirus, adenovirus) encode Bcl-2 homologs or FLIP homologs (viral FLIPs) to block one or both pathways in infected cells, buying time for viral replication — a direct evolutionary signature of how important both pathways are to host defense, connecting to broader themes in Apoptosis and the Machinery of Programmed Cell Death and to viral strategies covered alongside The Cell Cycle and Cancer.

Developmental biology and tissue sculpting. Interdigital webbing is removed by developmentally programmed intrinsic-pathway apoptosis; failure of this process (e.g., in some Bcl-2/Bax pathway mutant mice) produces retained webbing between digits, directly demonstrating the pathway's role in normal morphogenesis.

Ask the tutor: How does the balance of BH3-only proteins actually get "read out" quantitatively by Bax and Bak — is it a strict threshold or a graded response?