NOTCH1 is a cell surface receptor, regulation of which depends on the integrity and subsequent cleavage of its inhibitory domain. Subtle mechanical forces transmitted after ligand-binding [Wang et al., 2013] or removal of Ca2+-ions [Rand et al., 2000] make the site accessible for cleavage, resulting in release of the transcription factor NICD1. Mutations in NOTCH1 that prolong NICD1 activity have been found in chronic lymphocytic leukemia (CLL) with enrichment in up to 30% of Richter transformation (RT). Clinical trials have revealed that NOTCH1 mutant CLL derives no benefit from the addition of type I anti-CD20 monoclonal antibodies (mAb) such as rituximab. This lack of benefit has not been observed with the type II anti-CD20 mAb obinutuzumab.

To understand how anti-CD20 mAbs influence NOTCH1, we studied intracellular signaling events induced by rituximab or obinutuzumab and assessed their implications on NOTCH1 signaling.

As a model for this, mass spectrometry-based phosphoproteomic analysis was applied to rituximab or obinutuzumab treated SU-DHL4 B-lymphoma cells (1h and 24h time-points) and >8500 phosphorylation sites measured relative to untreated controls. Activation of kinase pathways was inferred by kinase-substrate enrichment analysis (KSEA). NOTCH1 receptor activity was assessed by western blot for NICD1 and qPCR for HES1 expression (NICD1 target gene). We used rituximab F(ab')2 fragments, trastuzumab as an isotype control and SB2H2 to cross-link the IgG B-cell receptor (BCR). Where applicable, cells were pre-incubated with inhibitors for 48h before mAb exposure.

Rituximab or obinutuzumab treatment resulted in strong phosphorylation of all proximal BCR signaling proteins at 1h and 24h, strongly activating MAPK and PLCγ2-induced signaling and remodeling processes in the actin cytoskeleton. Cell cycle promoting proteins were inhibited and in addition, obinutuzumab promoted pro-apoptotic signaling. AKT signaling was inhibited due to SHIP activation following FcγRIIB binding.

We hypothesized that BCR-induced Ca2+-flux and/or mechanical forces inflicted by cytoskeletal rearrangements disrupt the integrity of the inhibitory domain of NOTCH1 and promote ligand-independent activation. A strong, immediate increase in NOTCH1 activity was observed after rituximab, rituximab F(ab')2 and SB2H2 treatment. This was preventable by SYK inhibition, which also abrogated Ca2+-flux. Neither Ca2+-flux nor a relevant increase in NICD1 activity were induced by obinutuzumab, clearly supporting our first hypothesis. To test, if cytoskeletal rearrangements also contributed to NOTCH1 activation, we treated cells with a Rac inhibitor and found significantly reduced basal HES1 expression.

PI-4,5-P2, which is the substrate for the Pi3-kinase (Pi3K), suppresses Ca2+-flux and cytoskeletal rearrangements. In vitro, the Pi3K inhibitors idelalisib, duvelisib and umbralisib reduced maximum HES1 expression after rituximab exposure as well as basal HES1 expression levels significantly more than BTK inhibition. Thus, Pi3K inhibitors likely act as suppressors of NOTCH1 signaling. Abnormally strong NOTCH1 signaling is one driver of RT and our observations can serve as explanation why idelalisib treated patients might develop RT at lower rates than patients treated with ibrutinib or venetoclax.

In summary, we demonstrate clear and direct links between BCR and NOTCH1 signaling and show that anti-CD20 mAbs induce cell cycle arrest and apoptosis after strong BCR signaling. With c-MYC as an established target gene, NICD1 release upon Ca2+-flux and cytoskeletal movements in B-cells generates survival and proliferation signals. BCR-triggered effects on cell fate depend on signal strength and vary between activation/survival and apoptosis, the latter occurring when signal strength exceeds a threshold. CLL cells have low CD20 expression and the threshold for apoptosis induction after anti-CD20 mAb treatment may thus be difficult to reach. NOTCH1 mutations leading to abnormally strong effects of NICD1 would then lead to a prominence of survival over apoptotic signals, thereby promoting the observed resistance to rituximab.

We believe that Ca2+-flux inducing type I anti-CD20 mAbs should be avoided in the treatment of NOTCH1 mutant CLL. Furthermore, we hypothesize that patients at increased risk for RT (mutant NOTCH1, c-MYC gain, highly autoreactive BCR) may benefit from Pi3K inhibition.

Disclosures

Cragg:Boehringer Ingleheim: Consultancy; Bioinvent: Consultancy, Patents & Royalties: Patent licenced to Bioinvent around CD32b blockade. Gribben:TG Therapeutics: Honoraria; NIH: Research Funding; Celgene: Consultancy, Honoraria, Research Funding; Novartis: Honoraria; Acerta Pharma: Honoraria, Research Funding; Janssen: Honoraria, Research Funding; Wellcome Trust: Research Funding; Kite: Honoraria; Roche: Honoraria; Unum: Equity Ownership; Cancer Research UK: Research Funding; Medical Research Council: Research Funding; Abbvie: Honoraria; Pharmacyclics: Honoraria.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution