The stress-inducible transcription factor, NF-κB, plays an important in role in B-cell malignancies, regulating expression of a plethora of genes including those that drive proliferation, apoptosis and survival. Non-canonical ('alternative') NF-κB signalling, resulting in processing of the p100 subunit to p52 and dimerization with RelB, is less well-studied than the canonical pathway. Previous studies on NF-κB signalling, particularly in CLL, showed that the RelA (canonical) subunit is highly expressed and associated with poor outcome. In multiple myeloma and lymphoma, mutations in genes that alter NF-κB activity (NOTCH1, BIRC3, MYD88, TRAF2/3) can promote cell survival, by increasing activation of the CD40 receptor or NIK (NF-κB-inducing kinase), resulting in constitutive non-canonical NF-κB signalling. More recently, a comprehensive study showed that recurrent mutations in TRAF3, BIRC3 and NFKB2, all of which regulate non-canonical NF-κB signalling, also occur in CLL (Puente XS et al, Nature 2015). Taken together, these data suggest that non-canonical NF-κB signalling is important in CLL and that this pathway represents a novel therapeutic target. We hypothesised that receptor-activated proliferation in CLL cells is reliant on non-canonical NF-κB-regulated gene transcription, and that targeting this pathway would decrease CLL cell survival. Since the IKKα kinase phosphorylates p100, which in turn leads to active p52/RelB dimers in the nucleus, it is a key driver of the non-canonical signalling pathway. We used cell line models and patient-derived CLL cells to evaluate a series of selective, first-in-class IKKα inhibitors.

Initially, we assessed 20 novel IKKα inhibitors in MEC-1 (CLL-like, BIRC2/3 and TP53 deleted) and Maver-1 (mantle cell lymphoma, TRAF3 deleted) cell lines, as these are representative of B-cell malignancies with constitutive activation of non-canonical NF-κB. Compounds S and U were found to be potent (cell-free IC50 values < 50nM) and selective (e.g. >50-fold over IKKβ). In growth inhibition studies, compound S inhibited cell proliferation with GI50 values of 4.3 μM (MEC-1) and 2.7 μM (Maver-1) whereas compound U gave GI50 values of 0.47 μM (MEC-1) and 0.5 μM (Maver-1). We then examined selected cases from our CLL cohort. Western blotting of nuclear extracts showed that expression of NF-κB subunits in patient-derived CLL cells was heterogeneous: RelA was constitutively expressed in most cases, and although expression of the non-canonical subunits, p52 and RelB, varied, high levels of p52 were associated with high RelB expression, indicating that non-canonical signalling is active in CLL. The panel of 20 IKKα inhibitors were also tested in viability and apoptosis assays in patient-derived CLL cells, revealing LC50 values ranging from 0.5 μM (Compound U) - 5μM (48 hr exposure). Decreasing cell survival correlated with increased caspase 3/7 activation. To model the tumour microenvironment, we used CD40L-expressing fibroblasts in co-culture with primary CLL cells to stimulate NF-κB signalling and CLL cell proliferation. CD40L stimulation led to increased levels of nuclear RelB and p52 levels after 2-8 hrs co-culture, concomitant with activation of IKKα, as demonstrated by rapid phosphorylation of p100 subunit. Compound S was evaluated in CD40L-stimulated CLL cells where it was shown to inhibit p100 processing and decrease levels of nuclear RelB and p52. Compound S significantly reduced CD40L-stimulated proliferation of CLL cells following 7 day treatment (1μM), in a similar fashion to the clinically-used PI3Kδ inhibitor, Idelalisib, and was also effective at reducing proliferation of TP53-mutated patient-derived CLL cells. Compound S had little or no effect on accumulation of nuclear RelA (p65), indicating selectivity of these inhibitors for IKKα over IKKβ and suggesting that targeting receptor-stimulated IKKα in B cells may avoid 'global' toxicity anticipated by use of IKKβ-targeting therapies.

Future studies will examine the gene expression profile following IKKα inhibition in CLL, and also interrogate the underlying genotypes among patients that may drive NF-κB signaling. These data provide proof of concept that targeting non-canonical NF-κB signaling is a valid therapeutic strategy in CLL.

Disclosures

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution