NF-κB Signalling Pathway: Canonical and Non-Canonical Routes

NF-κB is not one pathway, it is two clocks running at different speeds. The canonical arm fires within minutes of TNF-α, IL-1β or LPS reaching the surface, peaks, and is switched off by feedback it encodes itself. The non-canonical arm takes hours, needs new protein synthesis, and is driven not by a kinase cascade but by the disappearance of a ubiquitin ligase. Most inflammatory phenotypes worth modelling ask which clock is running, and for how long.

Many receptors, one hub: the inputs

Three input families dominate the canonical arm. TNF-α trimers engage TNFR1 and TNFR2; ligated TNFR1 recruits TRADD, which brings in RIPK1, TRAF2 and the ubiquitin ligases cIAP1 and cIAP2. IL-1β on IL-1R and LPS on the CD14–MD-2–TLR4 complex both use MyD88, which nucleates a helical Myddosome with IRAK4 and IRAK1 and hands off to TRAF6. TLR4 also signals from the endosome through TRIF.

Antigen receptors feed the same hub differently: CD3/TCR with CD28 co-stimulation activates PKCθ and the CBM complex. Layered on top is the TNF receptor superfamily — CD40L on CD40, RANKL on RANK, BAFF on BAFF-R, LTβ on LTβR, plus OX40 and 4-1BB. The practical consequence: nuclear p65 tells you the hub fired but never which receptor fired it, so attribution needs blocking upstream, not assaying downstream.

The canonical module: ubiquitin chains, TAK1 and IKK

What happens between receptor and kinase is a ubiquitin problem. cIAP1 and cIAP2 decorate RIPK1 with K63- and K11-linked chains; the LUBAC subunit HOIP adds linear M1 chains. These are not degradation signals but docking platforms. NEMO binds M1 chains with high specificity and delivers IKKα and IKKβ; in parallel TAB2 reads K63 chains and delivers TAK1. Co-localising kinase with substrate on one chain scaffold is what converts ligand binding into activation.

TAK1 phosphorylates the IKKβ activation loop, and IKKβ phosphorylates IκBα on Ser32 and Ser36, creating a βTrCP degron. The proteasome removes IκBα and unmasks the nuclear localisation signal on the p50/p65 (RelA) dimer — no transcription or translation required. The antigen-receptor route reaches the same IKK complex through the CBM: PKCθ phosphorylates CARMA1, seeding BCL10 filaments that recruit the paracaspase MALT1, whose substrates include A20 and CYLD. Lymphocytes cut the wires that would switch the pathway off.

The brakes: A20, CYLD and the IκBα loop

The IκBα gene is among the fastest NF-κB targets, so the pathway synthesises its own inhibitor as its first act. New IκBα enters the nucleus, strips p65 from DNA and exports it. The result is a damped oscillation rather than a clean on/off step, and the period matters: genes with slow, stable promoters integrate across cycles, while genes needing one sharp pulse respond only to the first wave.

Upstream, deubiquitinases do the real termination. A20 (TNFAIP3) strips K63 chains from RIPK1 and TRAF6, then adds K48 chains to mark them for destruction; CYLD trims K63 and M1 chains from TRAF2, TRAF6 and NEMO. Their loss is not subtle: A20 polymorphisms recur across rheumatoid arthritis, lupus and psoriasis, and biallelic inactivation is a recurrent lymphoma lesion. Ubiquitin status also sets cell fate — deubiquitinated RIPK1 leaves the membrane complex and nucleates death signalling.

The non-canonical arm: NIK, p100 and RelB

The second arm is regulated by proteolysis rather than by a cascade. In a resting cell NIK (MAP3K14) is translated normally but destroyed within minutes by TRAF3, TRAF2 and cIAP1/cIAP2, so it is effectively undetectable. Engaging LTβR, BAFF-R, CD40 or RANK recruits TRAF3 to the receptor and triggers its degradation; NIK then accumulates. Activation is removal of a brake, not a switch being thrown, which is why the arm is slow.

Accumulated NIK phosphorylates IKKα, and an IKKα homodimer — NEMO-independent, unlike the canonical complex — phosphorylates p100 (NF-κB2). The proteasome then processes p100 only partially, degrading its ankyrin-repeat half to generate p52, which partners RelB. That dimer drives a distinct programme: CCL19, CCL21 and BAFF for lymphoid organogenesis, and the osteoclast programme downstream of RANKL. Before processing, p100 acts as an IκB-like inhibitor, so one protein is both substrate and brake.

The output: inflammation, adhesion, survival — and disease

Nuclear p50/p65 turns on a coherent inflammatory module: TNF-α, IL-1β, IL-6 and IL-12p40; CXCL1 and CCL2 for neutrophil and monocyte recruitment; COX-2 and iNOS as effector enzymes; and ICAM-1 with VCAM-1 on endothelium. Because TNF-α and IL-1β are themselves inputs, this is an explicit feed-forward loop — the reason a local stimulus becomes systemic, and the reason neutralising one cytokine can collapse a whole circuit.

Running alongside is a survival module: Bcl-2, Bcl-xL, XIAP and cIAP1/2. This resolves an old paradox — TNF-α engages a death-domain receptor yet rarely kills a healthy cell, because the canonical arm outruns the death machinery. Block it and the same ligand turns cytotoxic. Persistent activation is itself a disease signature: canonical in rheumatoid arthritis and colitis, non-canonical in NIK-stabilised myeloma and RANKL-driven bone erosion.

Key targets and matching reagents

Studying NF-κB signalling in vivo

Because the pathway is entered almost entirely through cytokine and TNF-superfamily receptors, it is unusually tractable with functional-grade antibodies. Experiments fall into three groups.

1. Neutralising the cytokine inputs

Anti-TNF, anti-IL-1β and anti-IL-1R (CD121a) close the canonical arm at its two commonest entry points, while anti-RANKL removes the dominant non-canonical input in bone and lymphoid tissue. Pairing a ligand block with the matching receptor block — anti-TNFR1 and anti-TNFR2 — is the cleanest way to assign a phenotype to one receptor, since the two recruit overlapping but non-identical complexes. The In Vivo range is supplied low endotoxin throughout.

2. Engaging or blocking TNFRSF co-stimulation

Anti-CD40, anti-OX40 (OX-86) and anti-4-1BB are agonist-capable reagents that drive NF-κB deliberately rather than blocking it — useful when the question is whether forcing the pathway on rescues a phenotype. On the antigen-receptor side, anti-CD3 and anti-CD28 supply the PKCθ–CBM input. Agonist activity depends on Fc-mediated crosslinking, so isotype and endotoxin level change the outcome more here than almost anywhere else.

3. Quantifying the transcriptional output

Turn a blocking experiment into a quantitative one with ELISA readouts for the target genes: IL-6, TNF-α, IL-1β and IL-12p40 in serum, CXCL1 and CCL2 for recruitment, ICAM-1 and VCAM-1 for endothelial activation, COX-2 and iNOS in tissue lysate. For the non-canonical arm measure CCL19, CCL21 and BAFF, and confirm mechanism with p100/p52 and RelB beside IκBα and p65.

Featured products for this pathway

Frequently asked questions

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