Numerous studies, published over many years, have established the key role that the IκB kinase (IKK) subunits, α and β, play in regulating the Nuclear Factor κB (NF-κB) pathway. This research generally concluded that their functions can be separated, with IKKβ being the critical regulator of the canonical NF-κB pathway, while IKKα functions as the key activating kinase for the non-canonical pathway. However, other roles for these kinases have been described and several reports concluded that this separation of their functions may not always be the case. This commentary discusses the recent report by Biochem J. 479, 305–325, who elegantly demonstrate that in KRAS driven colorectal cancer cell lines, IKKα is an important regulator of the canonical NF-κB pathway. As is so often the case with trying to understand the complexity of NF-κB signalling, cellular context is everything.

Since its discovery in 1986, the NF-κB transcription factor pathway has become one of the most highly studied areas of cell signalling [1]. This reflects the central role it plays as a regulator of immunity, inflammation and the response of the cell to an incredible variety of stimuli and stresses. Moreover, aberrantly regulated NF-κB is a driver of many diseases, including cancer and numerous immune or inflammatory disorders. Inevitably this has resulted in a field that is often characterised by its intimidating complexity. Although, it can be argued that this complexity is required if the five subunits that make up the NF-κB subunit family are to correctly execute the myriad functions they perform.

There is, therefore, a natural tendency to try and simplify this complexity, to provide some ‘ground rules’ on which the field can be based. Or in other words, to generate dogmas. Central to this has been the role of the catalytic IKK subunits, α and β [2]. The IKK complex, consisting of IKKα (CHUK), IKKβ and the regulatory subunit NEMO (IKKγ), is a central node of the NF-κB pathway. The signalling pathways induced by the stimuli that activate NF-κB converge on the IKK complex and it is here that one of the central dogmas of this pathway is encountered. This dogma asserts that stimuli that activate the canonical (or classical) NF-κB pathway, do so in a manner dependent on NEMO and IKKβ. NEMO facilitates the linear/K63 ubiquitin chain dependent phosphorylation and activation of IKKβ by kinases such as Transforming Growth Factor-β-Activated Kinase 1 (TAK1, MAP3K7) [3]. IKKβ then phosphorylates the Inhibitor of NF-κB (IκB) α, resulting in its K48 linked ubiquitination and proteasomal degradation. This frees the NF-κB complex, typically a p50/RelA(p65) heterodimer, which then localises to the nucleus where it regulates target gene expression. In contrast the non-canonical (or alternative) pathway is activated by IKKα in a NEMO independent manner. Here, NF-κB Inducing Kinase (NIK) becomes stabilised, resulting in the phosphorylation and activation of IKKα, which in turn phosphorylates the p100 NF-κB subunit precursor, leading to its K48 linked ubiquitination and processing to p52 [4]. In this context and due to the lack of a requirement for NEMO, IKKα is often described as being a homodimer. This central dogma of the field is based, to a large extent, on early studies from knockout and knockin mice and cells [5,6]. And it is undoubtedly correct, at least in the contexts in which these models were analysed. But is this always true? In their study, Prescott et al. [1] convincingly demonstrate that it is not.

Prescott and colleagues used CRISPR-Cas9 genome engineering to delete IKKα and β in the HCT116 and SW620 colorectal cancer cell lines [1]. They then analysed activation of the NF-κB pathway, mostly focussing on stimulation with the inflammatory cytokine tumour necrosis factor (TNF) α but with similar results being seen with interleukin (IL) 1. With the IKKα and β double knockout (DKO) HCT116 cells, results were as expected with a complete ablation of TNF induced NF-κB activation, combined with an increase in apoptosis. In contrast, the single knockout (SKO) results were complex and surprising (Table 1). Firstly, deletion of IKKβ had little effect, with IκBα phosphorylation and degradation, RelA phosphorylation and nuclear translocation together with NF-κB transcriptional activity all being comparable to that seen in control cells. In contrast, deletion of IKKα, while also having little effect on IκBα phosphorylation and degradation, severely impacted RelA S468 phosphorylation, nuclear translocation and transcriptional activity. Further investigation confirmed that in the SKO cells, both IKKα and β could compensate for the loss of the other kinase.

Table 1
Summary of the effects of IKK gene deletions on the NF-κB pathway described in Prescott et al.
Gene knockout
Effect in HCT116 cells Control IKKα IKKβ IKKα & IKKβ 
IκBα phosphorylation +++ +++ +++ 
IκBα degradation +++ +++ +++ 
NF-κB transcriptional activity +++ +++ 
p65 nuclear localisation +++ +++ 
c-Rel nuclear localisation +++ 
p65 S536 phosphorylation +++ +++ +++ 
p65 S468 phosphorylation +++ +++ 
Apoptosis +++ 
Gene knockout
Effect in HCT116 cells Control IKKα IKKβ IKKα & IKKβ 
IκBα phosphorylation +++ +++ +++ 
IκBα degradation +++ +++ +++ 
NF-κB transcriptional activity +++ +++ 
p65 nuclear localisation +++ +++ 
c-Rel nuclear localisation +++ 
p65 S536 phosphorylation +++ +++ +++ 
p65 S468 phosphorylation +++ +++ 
Apoptosis +++ 

Effects shown are for TNF stimulation in HCT116 cells but similar results are seen with IL-1 and in SW620 cells. Scoring of effects performed by Neil Perkins, using data presented in Prescott et al.

As Prescott et al. [1] discuss in detail in the Introduction to their article, the ability of IKKα and β to both contribute towards NF-κB activation and compensate for each other has been reported by others previously. What I find particularly striking about this report is that by analysing multiple steps of the NF-κB pathway, the authors were able to establish a key regulatory step after IκB degradation that is primarily controlled by IKKα (Table 1). In these colorectal cancer cells, even under conditions of normal IκB degradation, IKKα is required for RelA nuclear localisation and NF-κB transcriptional activity. Interestingly though, this defect does not result in the TNF induced apoptosis seen in the DKO cells, implying that the low level of NF-κB activity remaining in the IKKα SKO cells is sufficient to prevent this. Alternatively, other IKK functions, such as regulation of mRNA stability [7] may still be intact in these cells. It would certainly be of interest for future studies to perform an RNA Seq analysis in this system to better define the transcriptional consequences of IKK subunit deletion.

It is a mark of the significance of this study that it raises more questions than it answers. For example, why in these cells does IKKα play such an integral role in the classical NF-κB pathway while in other contexts it does not? The molecular basis for this is currently unclear but there are some potential explanations.

Surprisingly, even after years of study, there is still much we don't know about the composition of the IKK complex outside of its core subunits and whether these can affect its ability to phosphorylate substrates. For example, RAP1, a protein normally associated with telomere function, can associate with the IKK complex and is required for its ability to phosphorylate and activate RelA but not IκBα [8]. The ability of RAP1 and other such adaptors to associate with IKK subunits could be regulated by post-translational modifications (PTMs). Commercially available antibodies that detect phosphorylation of the IKKα and β activation loops are frequently used as biomarkers of activity and consequently are the most frequently reported IKK PTMs (see Prescott et al. for example). However, the phosphosite database [9] (https://www.phosphosite.org/) indicates that both IKKα and β can be subject to numerous other PTMs including phosphorylation, acetylation and ubiquitination (Figure 1). The function and regulation of the majority of these is not known. However, it is plausible that some of them may regulate the interaction with adaptor proteins that in turn target the IKK complex to specific substrates or affect its function in other ways.

IKKα and β are subject to multiple PTMs.

Figure 1.
IKKα and β are subject to multiple PTMs.

Data from PhosphoSitePlus (https://www.phosphosite.org/) showing reported sites of PTMs on IKKα (A and B) and β (C and D). Also shown is a comparison between combined low and high throughput reports (A and C) versus high throughput reports (B and C) that reflects how the availability of commercial antibodies to sites of phosphorylation in the IKK activation loops biases reports towards these PTMs. Figure labelling: Pkinase = protein kinase domain. IKK = NEMO binding domain. A listing of available IKK phospho antibodies, together with their citation numbers, can be found at https://www.citeab.com/antibodies/search?q=IKK+phospho.

Figure 1.
IKKα and β are subject to multiple PTMs.

Data from PhosphoSitePlus (https://www.phosphosite.org/) showing reported sites of PTMs on IKKα (A and B) and β (C and D). Also shown is a comparison between combined low and high throughput reports (A and C) versus high throughput reports (B and C) that reflects how the availability of commercial antibodies to sites of phosphorylation in the IKK activation loops biases reports towards these PTMs. Figure labelling: Pkinase = protein kinase domain. IKK = NEMO binding domain. A listing of available IKK phospho antibodies, together with their citation numbers, can be found at https://www.citeab.com/antibodies/search?q=IKK+phospho.

Close modal

It is a feasible hypothesis that such regulatory PTMs on IKK subunits can be differentially regulated by parallel signalling pathways active in cancer cells. It is therefore possible that the presence of oncogenic KRAS in HCT116 (KRASG13D) and SW620 (KRASG12V) cells will differentially affect IKK or NF-κB subunit function when compared with non-transformed cells. For example, they could affect the pattern of IKK subunit PTMs or induce/suppress IKK adaptor proteins. Indeed, regulation of NF-κB signalling by mutant, oncogenic RAS proteins has been known for many years. Interestingly, IKKα has been shown to be required for KRAS(G12D) driven lung adenocarcinoma [10], while microarray analysis of NF-κB subunit dependent genes regulated by oncogenic HRASV12 in MEFs revealed that these were generally different from those seen upon inflammatory cytokine stimulation [11]. These examples demonstrate that the activity of cell signalling pathways regulated by oncogenes such as RAS can in turn affect and alter IKK/NF-κB activity.

Taken together, this suggests a model in which cancer cells with oncogenic RAS regulate IKK subunit PTMs that facilitate the recruitment of adaptor proteins (Figure 2). These will then alter the targeting of the catalytic IKK subunits, resulting in a function for IKKα in the canonical NF-κB pathway not seen in other contexts. It would be of great interest to perform a phosphoproteomic analysis to investigate how widespread such RAS induced IKK dependent phosphorylation changes might be.

A model to potentially explain altered IKK subunit function in oncogenic mutant KRAS (mKRAS) transformed cancer cells.

Figure 2.
A model to potentially explain altered IKK subunit function in oncogenic mutant KRAS (mKRAS) transformed cancer cells.

(1) The catalytic IKK subunits are subject to numerous post-translational modifications (PTMs) that include phosphorylation (P), acetylation (Ac) and ubiquitination (Ub). (2) These have the potential to regulate the interaction with Adaptor proteins such as RAP1, which can affect the substrate targeting of the IKK complex. (3) Oncogenic mKRAS has been shown to induce numerous cell signalling pathways. These may act on the IKK complex through direct regulation of subunit PTMs and adaptor proteins. They may also regulate activation of the IKK complex by upstream kinases such as TAK1, a process that requires the formation of K63 and linear (M1) ubiquitin chains. (4) Together these processes can alter the targeting and specificity of the IKK complex, resulting in IKKα regulation of the canonical NF-κB signalling pathway not seen in other cellular contexts.

Figure 2.
A model to potentially explain altered IKK subunit function in oncogenic mutant KRAS (mKRAS) transformed cancer cells.

(1) The catalytic IKK subunits are subject to numerous post-translational modifications (PTMs) that include phosphorylation (P), acetylation (Ac) and ubiquitination (Ub). (2) These have the potential to regulate the interaction with Adaptor proteins such as RAP1, which can affect the substrate targeting of the IKK complex. (3) Oncogenic mKRAS has been shown to induce numerous cell signalling pathways. These may act on the IKK complex through direct regulation of subunit PTMs and adaptor proteins. They may also regulate activation of the IKK complex by upstream kinases such as TAK1, a process that requires the formation of K63 and linear (M1) ubiquitin chains. (4) Together these processes can alter the targeting and specificity of the IKK complex, resulting in IKKα regulation of the canonical NF-κB signalling pathway not seen in other cellular contexts.

Close modal

There are other potential ways that IKK activity could be differentially regulated. For example, through the TNF receptor associated ubiquitination events required for its activation [3]. Moreover, Prescott et al. did not determine if NEMO was required for the IKKα dependent effects they see. It therefore remains possible that regulation of the TNF induced NF-κB pathway in these colorectal cancer cells is mediated by both the ‘canonical’ IKK complex and NEMO independent IKKα ‘homodimers’.

So, does this mean that IKKα inhibitors might be a useful treatment for colorectal cancers with an oncogenic KRAS mutation? IKKβ inhibitors have been the subject of much research but the side effects associated with their inhibition of NF-κB in normal cells, particularly on the normal colonic epithelium appears to have ruled this therapeutic strategy out, at least for the time being [12]. However, there is an excellent case for the development of IKKα inhibitors. Inhibition of the non-canonical NF-κB pathway is likely to be beneficial for the treatment of many cancer types, while avoiding the known IKKβ inhibitor side effects [13]. The added ability to inhibit the canonical NF-κB pathway specifically in cancers with KRAS mutations would only increase their therapeutic potential.

Some caution is required though. Firstly, it is not clear yet whether the IKKα dependent effects reported by Prescott and colleagues are specific for cancer cells with oncogenic KRAS and more studies are required. Secondly, although the phenotypes seen in mice predict relatively mild side effects for IKKα inhibitors relative to IKKβ inhibitors, it cannot be assumed that this will be the same in humans. Finally, as Prescott et al. [1] show, the induction of TNF induced apoptosis in HCT116 cells required the IKK subunit DKO and did not happen with the IKKα SKO. Furthermore, analysis of IKKα and β gene deletion, using data available at the Cancer Dependency Map [14] (https://depmap.org/portal/), reveals relatively little effect in a wide range of colorectal cancer cell lines (Figure 3). In contrast, haematopoietic malignancies, where NF-κB is known to play an important role [15], show a high degree of dependency on both IKKα and β. These screens, however, were not done in using an NF-κB inducing stimulus and this could be expected to be provided to a colorectal tumour by the natural microenvironment of the gut. Moreover, the beneficial effects of inhibiting NF-κB are not confined to induction of apoptosis. Therefore, should the development of IKKα inhibitors [16] reach the point where they can be used in mice, exploring their efficacy in pre-clinical models of KRAS driven colorectal cancer is definitely warranted.

Comparison of the effect on cell viability of deleting IKKα (CHUK) and β (IKBKB) in multiple lymphoma, myeloma and colorectal cancer cell lines.

Figure 3.
Comparison of the effect on cell viability of deleting IKKα (CHUK) and β (IKBKB) in multiple lymphoma, myeloma and colorectal cancer cell lines.

In the absence of NF-κB activating signals, colorectal cancer cell lines show low dependency on IKKα and β compared with haematological malignancies. Analysis performed using data available at the Cancer Dependency Map (https://depmap.org/portal/). The positions of the HCT116 and SM620 cell lines used in the study by Prescott et al. are shown.

Figure 3.
Comparison of the effect on cell viability of deleting IKKα (CHUK) and β (IKBKB) in multiple lymphoma, myeloma and colorectal cancer cell lines.

In the absence of NF-κB activating signals, colorectal cancer cell lines show low dependency on IKKα and β compared with haematological malignancies. Analysis performed using data available at the Cancer Dependency Map (https://depmap.org/portal/). The positions of the HCT116 and SM620 cell lines used in the study by Prescott et al. are shown.

Close modal

In conclusion, the study by Prescott et al. [1] opens up many exciting research and translational opportunities. It is to be hoped that the answers to the questions this manuscript raises will be answered soon. Moreover, it shows once again that despite the maturity of the NF-κB field, there are still surprises to be found and that we should all be careful of taking the dogmas we have established over this time too seriously.

The author declares that there are no competing interests associated with this manuscript.

I was the external examiner for Jack Prescott's PhD viva, which featured much of the data that is the focus of this commentary article. I was incredibly impressed with Jack and my recollection of the viva is that it was one of the most interesting and stimulating PhD exams I have participated in. It struck me at the time that Jack was a scientist of tremendous potential. I was therefore very upset to hear that he had recently passed away. My thoughts go out to his family, friends and colleagues who will have been devastated by this terrible news.

     
  • DKO

    double knockout

  •  
  • IKK

    IκB kinase

  •  
  • IL

    interleukin

  •  
  • PTMs

    post-translational modifications

  •  
  • SKO

    single knockout

  •  
  • TNF

    tumour necrosis factor

1
Zhang
,
Q.
,
Lenardo
,
M.J.
and
Baltimore
,
D.
(
2017
)
30 years of NF-κB: a blossoming of relevance to human pathobiology
.
Cell
168
,
37
57
2
Hinz
,
M.
and
Scheidereit
,
C.
(
2014
)
The IκB kinase complex in NF-κB regulation and beyond
.
EMBO Rep.
15
,
46
61
3
Chen
,
Z.J.
(
2012
)
Ubiquitination in signaling to and activation of IKK
.
Immunol. Rev.
246
,
95
106
4
Morgan
,
D.
,
Garg
,
M.
,
Tergaonkar
,
V.
,
Tan
,
S.Y.
and
Sethi
,
G.
(
2020
)
Pharmacological significance of the non-canonical NF-κB pathway in tumorigenesis
.
Biochim. Biophys. Acta Rev. Cancer
1874
,
188449
5
Gerondakis
,
S.
,
Grumont
,
R.
,
Gugasyan
,
R.
,
Wong
,
L.
,
Isomura
,
I.
,
Ho
,
W.
et al (
2006
)
Unravelling the complexities of the NF-κB signalling pathway using mouse knockout and transgenic models
.
Oncogene
25
,
6781
6799
6
Pasparakis
,
M.
,
Luedde
,
T.
and
Schmidt-Supprian
,
M.
(
2006
)
Dissection of the NF-κB signalling cascade in transgenic and knockout mice
.
Cell Death Differ.
13
,
861
872
7
Mikuda
,
N.
,
Kolesnichenko
,
M.
,
Beaudette
,
P.
,
Popp
,
O.
,
Uyar
,
B.
,
Sun
,
W.
et al (
2018
)
The IκB kinase complex is a regulator of mRNA stability
.
EMBO J.
37
,
e98658
8
Teo
,
H.
,
Ghosh
,
S.
,
Luesch
,
H.
,
Ghosh
,
A.
,
Wong
,
E.T.
,
Malik
,
N.
et al (
2010
)
Telomere-independent Rap1 is an IKK adaptor and regulates NF-κB-dependent gene expression
.
Nat. Cell Biol.
12
,
758
767
9
Hornbeck
,
P.V.
,
Kornhauser
,
J.M.
,
Latham
,
V.
,
Murray
,
B.
,
Nandhikonda
,
V.
,
Nord
,
A.
et al (
2019
)
15 years of phosphoSitePlus(R): integrating post-translationally modified sites, disease variants and isoforms
.
Nucleic Acids Res.
47
,
D433
D441
10
Vreka
,
M.
,
Lilis
,
I.
,
Papageorgopoulou
,
M.
,
Giotopoulou
,
G.A.
,
Lianou
,
M.
,
Giopanou
,
I.
et al (
2018
)
Iκb kinase α is required for development and progression of KRAS-mutant lung adenocarcinoma
.
Cancer Res.
78
,
2939
2951
11
Hanson
,
J.L.
,
Hawke
,
N.A.
,
Kashatus
,
D.
and
Baldwin
,
A.S.
(
2004
)
The nuclear factor κB subunits RelA/p65 and c-Rel potentiate but are not required for Ras-induced cellular transformation
.
Cancer Res.
64
,
7248
7255
12
Prescott
,
J.A.
and
Cook
,
S.J.
(
2018
)
Targeting IKKβ in cancer: challenges and opportunities for the therapeutic utilisation of IKKβ inhibitors
.
Cells
7
,
115
13
Paul
,
A.
,
Edwards
,
J.
,
Pepper
,
C.
and
Mackay
,
S.
(
2018
)
Inhibitory-κB kinase (IKK) α and nuclear factor-κB (NFκB)-inducing kinase (NIK) as anti-cancer drug targets
.
Cells
7
,
176
14
Pacini
,
C.
,
Dempster
,
J.M.
,
Boyle
,
I.
,
Goncalves
,
E.
,
Najgebauer
,
H.
,
Karakoc
,
E.
et al (
2021
)
Integrated cross-study datasets of genetic dependencies in cancer
.
Nat. Commun.
12
,
1661
15
Nagel
,
D.
,
Vincendeau
,
M.
,
Eitelhuber
,
A.C.
and
Krappmann
,
D.
(
2014
)
Mechanisms and consequences of constitutive NF-κB activation in B-cell lymphoid malignancies
.
Oncogene
33
,
5655
5665
16
Anthony
,
N.G.
,
Baiget
,
J.
,
Berretta
,
G.
,
Boyd
,
M.
,
Breen
,
D.
,
Edwards
,
J.
et al (
2017
)
Inhibitory κB kinase α (IKKα) inhibitors that recapitulate their selectivity in cells against isoform-related biomarkers
.
J. Med. Chem.
60
,
7043
7066

Author notes

Commentary on: Prescott, J. A., Balmanno, K., Mitchell, J. P., Okkenhaug, H. and Cook, S. J. (2022) IKKα plays a major role in canonical NF-κB signalling in colorectal cells. Biochem J. 479, 305–325