Rituximab Activates NOTCH1 Signaling in CLL Cells and Induces Changes in the Cytokine Repertoire Favoring a Tumor-Protective Microenvironment

Introduction

The addition of the anti-CD20 monoclonal antibody (mAb) rituximab to chemotherapy improves responses in almost all chronic lymphocytic leukemia (CLL) patients, with the exception of those having NOTCH1 mutations. The reasons for this are largely unknown.

NOTCH1 is a cell surface receptor releasing its intracellular domain (NICD) upon activation. The NICD is a transcription factor and NOTCH1 mutations in CLL lead to longer lasting transcription factor activity.

We studied the effects of rituximab on NOTCH1 receptor activity and aimed to identify mechanisms of resistance towards chemoimmunotherapy based on NOTCH1 signaling.

Material & Methods

Freshly isolated peripheral blood mononuclear cells from CLL patients attending St. Bartholomew's Hospital, London were enriched for CD19 positive cells by negative selection using EDTA free buffer. Cells were treated with mAbs for various time periods. Whole protein lysates were obtained after 15, 30 and 60 min of mAb treatment; RNA was isolated after 150 min. NICD was semi-quantitatively assessed by western blot and NICD target gene expression, such as expression of HES1, was quantified by TaqMan^® probe-based quantitative PCR. Global gene expression profiling was done using Affymetrix^® Human Gene ST 2.0 Arrays on 3 biological replicates of treated and non-treated samples each deriving from the same patient.

Results

CLL cells freshly isolated from peripheral blood showed low NOTCH1 signaling activity. Following in-vitro treatment of cells with rituximab for 15, 30 or 60 min we observed NOTCH1 receptor activation with NICD release in a time-dependent manner, regardless of NOTCH1 mutation status. We used appropriate controls to distinguish whether NOTCH1 activation was associated with CD20 engagement and/or Fcγ-receptor (FcγR) IIB binding. Trastuzumab was used as an isotype control and led to stronger NOTCH1 activation than F(ab')2 fragments of rituximab, suggesting that NICD release mainly occurs upon engagement of FcγRIIB.

In line with this, in-vitro treatment of cells with rituximab was followed by up-regulated expression of NICD target genes such as HES1 . Global gene expression analysis when maximum HES1 gene expression was observed revealed 139 up-regulated and 124 down-regulated probesets in an analysis comparing rituximab treated with non-treated samples (fold-change ≥1.5; p-value ≤0.05) with the majority being non-coding RNAs. Only 28 differentially regulated probesets were actually coding for known genes. Gene set enrichment analysis against the curated genesets of the MSigDB collection (C2) identified 9 enriched genesets (false discovery rate <25%), mainly related to cytokine secretion.

The two most up-regulated genes were CCL2 and CCL4 with fold-changes of 6.5 and 3.2, respectively (p<0.001). The gene locus of CCL2 was previously identified as a direct NICD target and CCL2 and CCL4 were both shown to be produced by CLL cells. Numerous studies have linked CCL2 and CCL4 with the recruitment of tumor-protective immune cells such as tumor associated macrophages and T-cell subsets thereby changing the composition of the tumor microenvironment.

While wild-type NICD induced only short-lived gene expression changes after transient NOTCH1 activation, mutant NICD elicited more sustained changes. Since NOTCH1 activation following rituximab binding is likely to be temporary, we hypothesise that expression changes of cytokines such as CCL2 and CCL4 last long enough only in NOTCH1 mutant CLL to induce profound changes in the tumor microenvironment.

Conclusion

In summary, our data shows for the first time that rituximab binding to CLL cells can activate NOTCH1 signaling and influence the cells' transcriptome. We believe that NOTCH1 induced changes to the cytokine repertoire contribute to the formation of a tumor-protective microenvironment and that the extent of this is dependent on the durability of NICD's transcription factor activity, which is prolonged in NOTCH1 mutant CLL.

In ongoing studies we are testing the effects of ofatumumab and obinutuzumab on NOTCH1 signaling and we are using sequential samples taken before and after anti-CD20 mAb infusion to define the extent to which rituximab can influence NOTCH1 signaling in-vivo. We also aim to characterise the cellular composition of the CLL microenvironment with regard to NOTCH1 mutation status and NOTCH1 activity.