NF-κB Explained: The Switch Behind Inflammation Research

Almost every anti-inflammatory claim made for a research peptide eventually arrives at the same four characters: NF-κB. It is offered as the reason a tripeptide calms colitis in mice, the reason a copper complex is called protective, and the reason an antimicrobial peptide is described as immunomodulatory. NF-κB is real, central and extraordinarily well studied — which is exactly why the shorthand is worth unpacking, because several of the claims hung on it do not survive a look at the primary papers.

This guide explains what NF-κB is (it is not one protein), how the switch is thrown, what it turns on, why there is still no approved drug that inhibits it directly, and — compound by compound — what the published research on the peptides most often linked to it actually measured.

What NF-κB is

NF-κB is a transcription factor: a protein that binds DNA and changes which genes are read. It is not a peptide, not a receptor and not an enzyme, which matters because it is routinely discussed in the same sentence as research peptides that are all three.

More precisely, NF-κB is a family. Mammals have five Rel-homology DNA-binding subunits, and they function as dimers rather than alone. Two of them, p50 and p52, are not made as such: they are proteolytic products of the larger precursors p105 (NFKB1) and p100 (NFKB2), and the gene records at UniProt list the precursor as the gene product with the short form contained inside it. Only RelA (better known as p65), RelB and c-Rel carry transactivation domains, so p50 and p52 need one of those three as a partner to activate transcription. On their own, p50:p50 and p52:p52 homodimers are repressive.

When a review, a product page or an abstract says “NF-κB activation”, it almost always means the canonical p50/RelA heterodimer. That is worth knowing, because “NF-κB went up” is a less specific statement than it sounds: twelve or more dimer combinations are possible, they do different things, and a phospho-p65 western blot reports on one of them.

A small nomenclature note, since both forms appear in the literature: p50/p65 and p65/p50 are the same heterodimer, and the order encodes nothing. UniProt writes “p65-p50”, one widely cited review writes “p65:p50” and another writes “p50/RelA”. Pick a convention and be consistent.

How it was found, and why the name misleads

In 1986 Ranjan Sen and David Baltimore published two papers in Cell. The first described multiple nuclear factors binding immunoglobulin enhancer sequences, and named one of them NF-κB: the nuclear factor binding the κ light-chain enhancer in B cells. The second, later the same year, reported something more important than the name: that NF-κB is induced by a post-translational mechanism. The protein is already there. Something switches it on.

That is the whole regulatory logic of the pathway, and it explains the kinetics. A cell facing a bacterial product does not have time to transcribe and translate a new transcription factor. It keeps a stock of them restrained in the cytoplasm and releases them on signal.

The name, meanwhile, has misled two generations of readers into thinking of NF-κB as an immune-cell or B-cell protein. It is not. RelA is described in the gene record as widely expressed, the pathway is induced by a reported 150-plus distinct stimuli, and one review of its target genes argued that it functions less as a mediator of immunity specifically than as a central regulator of stress responses generally.

How the switch is thrown

The restraint is IκB — inhibitor of κB — and chiefly IκBα, encoded by NFKBIA. Its gene record puts the mechanism plainly: it traps RelA and p50 dimers in the cytoplasm by masking their nuclear localisation signals, and when immune or pro-inflammatory signals stimulate the cell, IκBα is phosphorylated, ubiquitinated and degraded, which frees the dimer to enter the nucleus and activate transcription.

The phosphorylation sites are Ser-32 and Ser-36. Both sit in a destruction-box motif; once phosphorylated, IκBα is recognised by the SCF ubiquitin-ligase complexes, polyubiquitinated on Lys-21 and Lys-22, and destroyed by the proteasome. Two further serines near the C-terminus are casein-kinase-2 sites and are not part of the degradation signal — a detail worth keeping straight, because “IκBα phosphorylation” on its own does not specify which event is meant.

The kinase that puts those phosphates on is the IKK complex: two catalytic subunits, IKKα and IKKβ, plus the regulatory subunit NEMO. One naming trap here catches writers routinely — IKKα’s approved gene symbol is CHUK, not “IKKA”, and its previous symbol was TCF16.

Canonical and non-canonical

Two routes lead to NF-κB, and the cleanest published distinction between them is dependence on NEMO rather than which catalytic subunit is involved.

The canonical pathway is the one in the figure: triggered by TNF receptors, the IL-1 receptor, Toll-like receptors and antigen receptors, it culminates in IKKβ activation, it is NEMO-dependent, and its primary mechanism is the inducible degradation of IκBα.

The non-canonical pathway does not degrade IκBα at all. Induced by a subset of TNF-receptor-superfamily members — CD40L, BAFF, lymphotoxin-β and RANK are the usual examples — it is IKKα-dependent and NEMO-independent, and it works by processing the precursor p100 into p52, producing RelB:p52 complexes. It is slower and more persistent than the canonical response.

The kinase driving it, NIK (gene symbol MAP3K14), is regulated in an unusual way that is worth knowing because it inverts the normal intuition: NIK is constitutively active, and the pathway is controlled by destroying it. NIK is continuously ubiquitinated and degraded; signalling disrupts the complex that does the destroying, so active NIK accumulates. Activation here is a stay of execution, not a switching-on.

TNF-α: the best-known input

If NF-κB has a signature stimulus it is tumour necrosis factor. The gene symbol, worth getting right, is simply TNF — TNFA is the previous symbol, although “TNF-α” remains perfectly good as a protein name. The protein record describes it as mainly secreted by macrophages, notes its alternative name cachectin, records that it acts as a potent pyrogen and that it is implicated in cachexia, and lists two receptors: TNFR1 (TNFRSF1A, CD120a) and TNFR2 (TNFRSF1B, CD120b).

The genuinely interesting fact about TNF is that the same ligand at the same receptor can drive opposite cell fates. TNFR1 can assemble a death-inducing complex through TRADD, RIP kinases, FADD and caspase-8, starting a caspase cascade toward apoptosis — or it can route to NF-κB and survival. And the default, according to one widely cited review, leans the second way: in most cell types TNFR1 induces the activation of NF-κB rather than apoptosis. NF-κB activation is largely what tips the balance toward survival, which is why an account of NF-κB as purely a driver of inflammation is incomplete.

TNF is also, unambiguously, a validated drug target. Regulators have approved biologic medicines that bind TNF and block it from engaging its receptors: etanercept’s label describes a dimeric soluble form of the p75 TNF receptor that binds TNF molecules and renders them biologically inactive, and adalimumab’s describes an antibody binding TNF-alpha and blocking its interaction with the p55 and p75 cell-surface receptors. Those are prescription medicines, used under medical supervision, and nothing here describes or recommends their use — the point is simply that blocking TNF works well enough to clear regulatory review, which is a useful calibration when reading claims about compounds that have never been near one.

What it switches on — including its own brake

The curated target-gene catalogue maintained by the Gilmore laboratory at Boston University sorts NF-κB targets into fourteen functional categories, and the list itself is admirably sceptical — it flags genes that have a κB site in the promoter but have not clearly been shown to be controlled by NF-κB, and separately flags genes whose expression merely tracks NF-κB activity. The categories include cytokines and chemokines (IL-6, TNF-α, CCL5), immunoreceptors (CD40, TLR2, TLR9), cell-adhesion molecules (ICAM-1, VCAM-1, E-selectin), acute-phase proteins (CRP, serum amyloid A), stress-response genes (COX-2, iNOS, SOD2) and regulators of apoptosis (BCL2, FAS, XIAP, TRAF1).

The most elegant item on that list is NFKBIA, the gene for IκBα. NF-κB switches on its own inhibitor. That was established in part by the observation that fibroblasts lacking RelA failed to induce IκBα messenger RNA in response to TNF, even though basal levels were normal.

But the feedback loop is more than a tidy diagram, and a 2008 study in mouse embryonic fibroblasts showed why. Comparing wild-type cells, IκBα-null cells and IκBα-null cells rescued with a constitutively expressed IκBα transgene, the authors found that constitutive expression could not restore wild-type dynamics. Their conclusion was that in the case of IκBα, negative feedback is required for function: without inducible IκBα, the first phase of NF-κB activity ballooned from minutes to around three hours. It is not enough for the inhibitor to be present. It has to be switched on by the thing it inhibits. A second NF-κB-induced brake, A20 (TNFAIP3), deubiquitinates RIP1, TRAF6 and NEMO to destabilise the IKK complex.

Why there is no NF-κB inhibitor

Given a transcription factor sitting at the centre of inflammation, with over 150 known inducers and a catalogue of inflammatory target genes, the obvious move is to block it. It has not worked, and the reason is instructive.

A 2022 review of the field states it flatly: no specific NF-κB inhibitor has been clinically approved. The authors attribute this to the dose-limiting, on-target toxicities of systemic NF-κB blockade and to the difficulty of achieving cell-type specificity, given NF-κB’s ubiquitous and pleiotropic physiological functions. The sharpest illustration of why came much earlier: mice engineered without RelA die at around embryonic day 15 to 16 with massive liver apoptosis. NF-κB is not only an inflammation switch. It is a survival signal, and in development a non-negotiable one.

So what is approved acts upstream, or elsewhere. TNF biologics neutralise an input. BTK inhibitors and anti-RANKL antibodies hit upstream pathway components. And two drugs commonly described as NF-κB inhibitors are worth correcting:

What the peptide research actually measured

This is where a sceptical reading pays off, because the four compounds most often linked to NF-κB in research-peptide writing sit at four quite different evidence levels.

α-MSH: the strongest data in this cluster, and it belongs to the parent hormone

A 1998 study in The Journal of Immunology reported that α-melanocyte-stimulating hormone completely blocked TNF-induced NF-κB activation in a dose- and time-dependent way, and also suppressed activation triggered by LPS, okadaic acid and ceramide. It used four readouts — NF-κB DNA binding by electrophoretic mobility shift assay, IκBα degradation by western blot, p65 nuclear translocation by western blot, and an NF-κB-dependent reporter gene — and the effect appeared to run through cAMP, since adenylate cyclase and protein kinase A inhibitors reversed it. Notably, TNF-induced AP-1 activation was unaffected, so this was selective rather than a global shutdown. It is cell-culture work, it says nothing about clinical effect in people, and the selectivity is the interesting part.

KPV: real preclinical signal, and the two headline papers do different things than people claim

KPV is Lys-Pro-Val, the C-terminal tripeptide of α-MSH — residues 11 to 13 of the 13-residue hormone. Two 2008 papers carry nearly all the weight.

The first, in Gastroenterology, is titled “PepT1-mediated tripeptide KPV uptake reduces intestinal inflammation”, and its abstract reports that KPV acts via the PepT1 transporter in intestinal epithelial and immune cells, where it suppresses NF-κB and MAP kinase inflammatory signalling and reduces pro-inflammatory cytokine secretion, with oral KPV reducing colitis in two mouse models. We could not obtain the full text, so we make no claim about which NF-κB assay was used or whether NF-κB was measured in the mice as well as in cells.

The second, in Inflammatory Bowel Diseases, is the one most worth correcting. It tested KPV in DSS colitis, in T-cell-transfer colitis, and in mice carrying a non-functional melanocortin-1 receptor, and reported faster recovery, more weight regained, fewer histological infiltrates and lower colonic myeloperoxidase activity, with KPV rescuing all animals from death in the MC1R-deficient group. Its readouts were weight, histology and myeloperoxidase. It did not measure NF-κB at all, so it cannot support an NF-κB claim, though it is a genuinely useful mouse efficacy paper and the source of the proposal that KPV’s action is at least partly independent of MC1R signalling.

Two further points of honesty. The PepT1 route is a proposal, not a settled mechanism: a 2023 review still describes KPV uptake via PepT1 as hypothesised, and the transporter’s own gene record notes a preference for dipeptides within a two-to-four-residue range. And we are not aware of any published human clinical trial data for KPV — two dedicated reviews of melanocortins in inflammatory bowel disease report none. Our KPV research guide goes into the compound in detail, and short peptides explained covers why a tripeptide behaves so differently from its parent.

LL-37: dual, and “it depends” is the verified answer

LL-37’s relationship with NF-κB genuinely runs both ways, and both directions have direct measurements behind them. At low concentrations in human monocytic cells and peripheral blood mononuclear cells, a 2006 study found LL-37 reduced TNF-α release from LPS-stimulated cells, suppressed several LPS-upregulated genes including NFKB1, and — measured directly, not inferred — reduced nuclear translocation of the p50 and p65 subunits. At higher concentrations in human umbilical vein endothelial cells, a 2019 study found the opposite: LL-37 increased p65 serine-536 phosphorylation and drove ICAM-1 expression, with p65 translocating from cytoplasm to nucleus. Anyone presenting LL-37 as simply anti-inflammatory is citing half the literature. See our LL-37 research guide and antimicrobial peptides explained.

GHK-Cu: one unreplicated paper — which is still more than the Nrf2 claim has

Here precision matters, because the honest answer differs from the one we reached on a neighbouring claim. A 2016 paper in Oncotarget did measure NF-κB directly: in murine macrophages pretreated with GHK-Cu and then stimulated with LPS, it reported inhibition of p65 serine-536 phosphorylation by western blot and blocked p65 nuclear translocation by immunofluorescence, and in mice given GHK-Cu by intraperitoneal injection before intratracheal LPS, reduced lung phospho-p65. That is real functional data. It is also one study, from one laboratory, in one model, with no independent replication we could find, no IκBα or reporter data to pin the step, and no human data. The sweeping “GHK suppresses NF-κB” line in the widely cited review by the compound’s commercial proponents is uncited in its abstract and, where cited in the body, points back to that same single paper; a separate 2017 mouse study of GHK attributed its effects to TGF-β1/Smad signalling and never measured NF-κB at all.

Worth stating alongside it: that 2016 paper did not measure Nrf2. It is the study most often reached for to defend the claim that GHK-Cu activates Nrf2, and the claim is not in it. Our Nrf2 and oxidative stress guide sets out why that separate claim has no published functional support, and our GHK-Cu research guide covers the compound itself.

What’s established and what isn’t

Sourcing reagents for signalling work in Canada

If you are buying material for inflammatory-signalling experiments, the documentation questions are the usual ones, with one addition specific to this area: NF-κB readouts are sensitive to what else is in the vial.

• Ask about endotoxin, not just purity. LPS is one of the canonical NF-κB stimuli in the figure above, so endotoxin carried into a culture is a confound that looks exactly like a result. Our guide to endotoxin and research peptides covers what to ask and why.
• Match the batch number on the certificate to your vial. Our COA guide covers what else a useful certificate contains, and reading an HPLC chromatogram covers how to read the purity trace rather than just the headline number.
• Account for net peptide content. Synthetic peptides arrive as salts carrying counter-ions and water, so labelled mass is not peptide mass — see net peptide content.
• Store cold, dry and dark, and keep material lyophilized until needed: the storage guide and what lyophilized means.
• Read the claim before the price. Our guide to reading a peptide study works through the substitution this article keeps flagging: “compound X suppressed NF-κB signalling” and “compound X works because it suppresses NF-κB” are different claims needing different evidence.

KPVThe α-MSH(11–13) tripeptide at the centre of the colitis literatureView product LL-37The human cathelicidin whose NF-κB effects run both waysView product GHK-Cu + KPV BlendBoth compounds discussed above, co-formulated in one vialView product Coremend BlendTB-500, KPV and BPC-157 together as a research blendView product

Frequently asked questions