Notch2 controls non-autonomous Wnt-signalling in chronic lymphocytic leukaemia

Abstract

The Wnt signalling pathway, one of the core de-regulated pathways in chronic lymphocytic leukaemia (CLL), is activated in only a subset of patients through somatic mutations. Here we describe alternative, microenvironment-dependent mechanisms of Wnt activation in malignant B cells. We show that tumour cells specifically induce Notch2 activity in mesenchymal stromal cells (MSCs) required for the transcription of the complement factor C1q. MSC-derived C1q in turn inhibits Gsk3-β mediated degradation of β-catenin in CLL cells. Additionally, stromal Notch2 activity regulates N-cadherin expression in CLL cells, which interacts with and further stabilises β-catenin. Together, these stroma Notch2-dependent mechanisms induce strong activation of canonical Wnt signalling in CLL cells. Pharmacological inhibition of the Wnt pathway impairs microenvironment-mediated survival of tumour cells. Similarly, inhibition of Notch signalling diminishes survival of stroma-protected CLL cells in vitro and disease engraftment in vivo. Notch2 activation in the microenvironment is a pre-requisite for the activation of canonical Wnt signalling in tumour cells.

Fig. 1

Activation of CLL cells by BSMCs. a Volcano plot showing the differentially up- and down-regulated genes in CLL cells after 48 h of co-culture on EL08-1D2 cells compared to cells cultured for 4 h in mono-culture (4 h was chosen to avoid gene expression changes related to cell death). RNA-sequencing was performed on samples from 6 individual patients. b Transcriptomic data were subjected to Gene Set Enrichment (GSEA) analyses to identify pathways in CLL activated by contact to stromal cells. Gene sets are listed in order of Normalised Enrichment Scores (top black X-axis). FDR q values for each gene set are indicated by the red dotted line (lower red X-axis). c Heat map showing the 50 most significantly up- and down-regulated genes in CLL cells in response to contact with stromal cells. d Cell extracts from CLL mono-culture or from cells cultured for 6 h on EL08-1D2 cells were analysed using a human intracellular phosphorylation antibody array. Representative results from four different patients and experiments are shown. e CLL cells were cultured in medium only (red circles) or on EL08-1D2 cells (black squares) for 5 days before analysing apoptotic cells by Annexin-V/PI staining. Transwells were used to disrupt direct cell–cell contacts (blue triangles). Error bars show mean ± SEM from 9 patients; 
 ****p < 0.0001

Fig. 2

CLL cells activate Notch2 in BMSCs. a Constitutive expression of the Notch ligands Delta-1, Jagged-1 and Jagged-2 was analysed in primary CLL cells by flow cytometry. Representative results from three different patients are shown. b Constitutive expression of Notch1-4 in primary EL08-1D2 cells was assessed by flow cytometry. Two separate experiments revealed identical results. c Notch1 activity in EL08-1D2 cells was assessed by immunoblotting for Notch1^ICD 48 h after co-culturing primary CLL cells from 4 different patients on stromal cells. d Notch2 activity in EL08-1D2 cells was assessed by immunoblotting for Notch2^ICD 48 h after co-culturing primary CLL cells from the same 4 patients as shown in (c). e Notch2^ICD levels were analysed by western blotting in human bone marrow stromal-derived cells and human bone marrow stromal-derived cells co-cultured with CLL primary cells. f Confocal microscopy analysis of Notch2 expression in EL08-1D2 following 48 h co-culture with primary CLL cells or mono-culture. Images were captured with identical exposure time and settings between mono-cultures and co-cultures. Representative images from three independent experiments are shown. g Notch2 expression on EL08-1D2 cells was assessed on mono-cultured cells or on cells which were co-cultured for 48 h in the absence or presence of 20uM DAPT. Scale bar = 100 μm. One representative experiment out of three is shown

Fig. 3

Notch2 gene regulation in BMSCs. a Schematic representation of the experimental model. b Flow cytometry and immunoblot analyses of Notch2 in Notch^fl/fl stromal cells following in vitro CRE recombination. c Principal component analyses of the transcriptomes from three Notch2-proficient (orange circles) and Notch2-deficient (light blue circles) BMSCs and from five Notch2 wild-type (red circles) or knockout (dark blue) BMSCs cultured with CLL cells for 5 days. d GSEA comparing the expression of genes associated with the presence or absence of Notch2 expression in BMSCs co-cultured with primary CLL cells. Gene sets are listed in order of Normalised Enrichment Scores (left black Y-axis). FDR q values for each gene set are indicated by the red dotted line (right red Y-axis). e Pie-chart of Notch2-induced (left) and Notch2-repressed genes (right). f Heat map showing the 200 most significantly up- and down-regulated genes in Notch2-deficient BMSCs in response to contact with CLL cells. Genes listed in red have known functions in regulating Wnt signalling

Fig. 4

Stromal Notch2 regulates β-catenin in CLL cells. a β-Catenin expression in primary CLL cells co-cultured on murine BMSCs for 5 days or mono-cultured. b β-Catenin expression in four different primary CLL cells co-cultured on human primary BMSCs. c Lane 1+2: β-catenin expression was evaluated in primary CLL cells after 24 and 48 h in direct co-culture with EL08-1D2. Lane 3: conditioned media (CM) from the 48 h co-cultures or fresh medium (lane 4) was used as culture medium for freshly thawed cells of the same patient. β-Catenin expression was assessed after 24 h. Representative results from three different patients are shown. d β-Catenin localisation was assessed in primary CLL cells co-cultured on EL08-1D2 cells for 24 h. Specific non-phospho antibodies were used to detect the active form of β-catenin. Representative results from two different patients are shown. e β-Catenin expression in primary CLL cells cultured for 5 days on Notch2^+/+ or Notch2^-/- mBMSCs. f Analysis of cytoplasmic and nuclear β-catenin levels in primary CLL cells co-cultured for 48 h on BMSCs in which Notch2 was deleted by CRISPR/Cas9. Representative results from three different patients are shown. g Phospho-GSK3-β expression in primary CLL cells co-cultured on EL08-1D2 cells for 24 h. Representative results from three different experiments are shown. h β-Catenin expression in CLL cells co-cultured on Notch2-deficient EL08-1D2 cells. After 24 h on stromal cells, co-cultures were exposed to increasing doses of lithium chloride (LiCl) for additional 3 h before CLL cells were harvested. One representative experiment out of three is shown. i Quantitative reverse-transcription polymerase chain reaction analysis of the C1q complex in Notch2-deficient EL08-1D2 (blue symbols) cells normalised to expression in control cells (transfected with a control guide RNA, dark symbols). Shown is the mean ± SEM of six independent experiments, using different primary CLL cells; *p < 0.05, **p < 0.01. j β-Catenin expression in CLL cells co-cultured on Notch2-deficient EL08-1D2 cells. After 24 h on stromal cells, co-cultures were exposed to increasing doses of C1q for additional 24 h before CLL cells were harvested. Representative results from three different experiments are shown

Fig. 5

Stabilisation of β-catenin is partially dependent on N-cadherin. a CLL cells were co-cultured on EL08-1D2 cells. After the time points as indicated, N-cadherin and β-catenin expression were analysed. Two additional experiments revealed similar results. b N-cadherin was immunoprecipitated from CLL lysates derived from mono-cultures or from co-cultures with EL08-1D2 cells. One out of three experiments is shown. c Quantitative reverse-transcription polymerase chain reaction analysis of β-catenin (red circles) and N-cadherin (ruby squares) mRNA expression in primary CLL cells 24 h after co-culture with EL08-1D2, normalised to expression in mono-cultured cells. Shown is the mean ± standard deviation of five independent primary CLL cells; ***p < 0.001. d Cas9-expressing EL08-1D2 cells were transfected with two different guide RNAs (sgRNAs) against N-cadherin. Constitutive expression of N-cadherin and Notch2 is shown. e CLL cells were co-cultured on N-cadherin-proficient or -deficient EL08-1D2 cells. Expression of β-catenin and N-cadherin in CLL cells is shown after 72 h. One representative experiments out of three is depicted. f Partial least square (PLS) analysis of all proteins detected and quantified by mass spectrometry. Three biological replicates were performed with one CLL patient sample. g Volcano plot showing the differentially up- and down-regulated plasma proteins in EL08-1D2 cells after 48 h of co-culture with CLL cells compared to EL08-1D2 cells in mono-culture. h Volcano plot showing the differentially up- and down-regulated plasma proteins in EL08-1D2 cells after 48 h of co-culture with CLL cells. Comparison between Notch2-proficient and Notch2-deficient stromal cells. i Analysis of N-cadherin and β-catenin levels in primary CLL cells co-cultured for 48 h on BMSCs in which Notch2 was deleted by Cas9. Representative results from three different patients are shown. j Quantitative reverse-transcription polymerase chain reaction analysis of N-cadherin mRNA expression in primary CLL cells 48 h after co-culture on Notch2-deficient EL08-1D2 cells (ruby triangles), normalised to expression in control cells (black circles). Shown is the mean ± standard deviation of five independent experiments with individual primary CLL cells; **p < 0.01

Fig. 6

Increased sensitivity to Wnt inhibitors in primary CLL cells in the presence of BMSCs. a CLL cells were cultured for 5 days on mouse-derived BMSCs, either proficient (black squares) or deficient (blue triangles) for Notch2 (combining results from Cre-mediated and CRISPR/Cas9-mediated deletion of Notch2). Apoptotic cells were analysed by Annexin-V/PI staining. Error bars show mean ± SEM from 30 patients; ***p = 0.0004. b Schematic representation of the targeted proteins used in this study. c Three different CLL cells were co-cultured on EL08-1D2 cells for 72 h in the absence or presence of DAPT. Notch signalling was assessed by immunoblotting EL08-1D2 cell lysates for the expression of HES-1. d CLL cells were co-cultured on EL08-1D2 cells for 72 h. Then, increasing doses of DAPT were added for 48 h. Apoptotic CLL cells were detected by flow cytometry and staining for Annexin-V/PI. Error bars show mean ± SEM from 11 patient samples. e Analysis of β-catenin levels in primary CLL cells co-cultured for 48 h on BMSCs in the presence of DAPT. A representative result from three different patients is shown. f Similarly to data shown in (d), CLL co-cultures were exposed to increasing doses of the Wnt inhibitors XAV939, Dvl-PDZ3 or ICG-001 before cell death was assessed. Error bars show mean ± SEM from 11 patient samples; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Fig. 7

Notch targeting in vivo. a Schematic representation of the experimental design: mice transplanted with 40 × 10⁶ splenocytes from diseased TCL1⁺-Tg mice were treated after 4 weeks with 4 consecutive doses of 30 mg/kg DAPT or vehicle control (i.p. injection; once daily). b Notch2 expression analysis in bone marrow stromal Sca1⁺ cells of transplanted mice, treated with DAPT (ruby circles) or vehicle control (blue circles); (n = 5 for each treatment group; two different primary tumours were transplanted). Relative quantification of the MFI of treated versus untreated mice is shown on the right panel, *p < 0.05. c RT-qPCR analysis of the C1q complex in BMSC-Sca1⁺ cells of mice treated with DAPT, normalised to the expression in mice treated with vehicle. Because of low RNA yield, Sca1⁺ cells were sorted from three individual mice and then combined for RNA extraction. Shown is the mean ± standard deviation of three technical repeats. d Quantification of bone marrow CD19⁺CD5⁺ cells of mice treated with DAPT (red symbols) or vehicle control (blue symbols) using flow cytometry (n = 5 per treatment group, two different primary tumours were transplanted); **p < 0.01. e Schematic representation of the experimental design. f Far-Red CellTracker fluorescence of bone marrow-derived CD5⁺CD19⁺ cells, 7 days following transplantation (n = 3 per treatment group). g Quantification of the MFI of treated and untreated mice shown in (f); *p < 0.05. h Schematic scheme of the CLL-PDX model. Freshly isolated PBMCs from four untreated CLL patients were injected intravenously into four mice respectively. At 5 days after transplantation, mice were treated once daily with DAPT at a dose of 30 mg/kg body weight or vehicle control (i.p. injection). Mice received a total of 10 doses over the course of 12 days. i Lymphoma burden in the spleen was assessed by flow cytometry (staining for CD5⁺, CD19⁺ cells) and histology. The thin white arrow indicates areas of extramedullary haematopoiesis; thick white arrow: mitotic figure. *Cells with morphology and CD20 staining consistent with neoplastic cells. Scale bars: lines 1 and 3 = 300 μm; line 2 = 100 μm, insert in line 2 = 30 μm. j Number of splenic CD19⁺CD5⁺ presents in mice treated with DAPT (yellow triangles) or vehicle control (ruby squares); **p < 0.01

Fig. 8

a β-Catenin was assessed in lymph node sections of CLL patients by microscopy. A specific non-phospho-antibody for β-catenin was used to detect the active form of the protein. One representative result from six different patients is shown. Scale bar = 100 μm. b β-Catenin and N-cadherin expression in primary B cells from a patient with leukaemic DLBCL (left panel) or from three patients diagnosed with Mantle cell lymphoma (MCL) after 5 days of co-culture on EL08-1D2 cells (right panel). High-grade NHL cell lines (Namalwa = Burkitt lymphoma; SU-DHL-4 and OCI-LY8 = DLBCL) were cultured on stromal cells for 48 h (middle panel). c Schematic presentation of the mutual activation of BMSCs and CLL cells. (Left) CLL cells induce Notch2 activation in BMSCs. (Middle) Stromal Notch2 in turn regulates the expression of complement C1q and other soluble factors, required for the inhibition of GSK3-β and stabilisation of β-catenin in malignant B cells. (Right) In addition, up-regulated N-cadherin in CLL cells interacts with β-catenin and further contributes to its stabilisation. This figure was partly produced using the Smart Servier Medical Art, available from https://smart.servier.com/image-set-download/ and licensed under a Creative Common Attribution 3.0 Generic License. http://smart.servier.com/