NOTCH ligands JAG1 and JAG2 as critical pro-survival factors in childhood medulloblastoma

Abstract

Medulloblastoma (MB), the most common pediatric malignant brain cancer, typically arises as pathological result of deregulated developmental pathways, including the NOTCH signaling cascade. Unlike the evidence supporting a role for NOTCH receptors in MB development, the pathological functions of NOTCH ligands remain largely unexplored. By examining the expression in large cohorts of MB primary tumors, and in established in vitro MB models, this research study demonstrates that MB cells bear abnormal levels of distinct NOTCH ligands. We explored the potential association between NOTCH ligands and the clinical outcome of MB patients, and investigated the rational of inhibiting NOTCH signaling by targeting specific ligands to ultimately provide therapeutic benefits in MB. The research revealed a significant over-expression of ligand JAG1 in the vast majority of MBs, and proved that JAG1 mediates pro-proliferative signals via activation of NOTCH2 receptor and induction of HES1 expression, thus representing an attractive therapeutic target. Furthermore, we could identify a clinically relevant association between ligand JAG2 and the oncogene MYC, specific for MYC-driven Group 3 MB cases. We describe for the first time a mechanistic link between the oncogene MYC and NOTCH pathway in MB, by identifying JAG2 as MYC target, and by showing that MB cells acquire induced expression of JAG2 through MYC-induced transcriptional activation. Finally, the positive correlation of MYC and JAG2 also with aggressive anaplastic tumors and highly metastatic MB stages suggested that high JAG2 expression may be useful as additional marker to identify aggressive MBs.

Figure 1

Expression of NOTCH ligands in MB primary tumors. (a) Relative mRNA expression of the indicated NOTCH ligands in a representative gene expression dataset of 62 human MB tumors [23]. Dot plots showing relative expression of JAG1(b), JAG2(c), DLL1(d), DLL3(e), and DLL4(f) in three independent gene expression datasets of human MB tumors: 62 samples [23]; 57 samples [22]; 76 samples [26].

Figure 2

JAG1 mediates pro-survival signaling through activation of canonical NOTCH2 signaling. (a) Relative JAG1 expression in 47 fresh frozen MB primary samples. Values represent the fold-change in JAG1 mRNA expression compared to that in normal cerebellum samples (defined as 1). (b) JAG1 mRNA (upper panel) and protein (lower panel) expression in the indicated MB cell lines. mRNA values represent the fold-change in JAG1 mRNA expression compared to that in normal cerebellum samples (defined as 1). (c) Protein expression of JAG1, NICD2, and HES1 in DAOY cells at 72 hours after treatment with JAG1 siRNA compared to control siRNA. (d)HES1 mRNA relative expression in DAOY cells upon JAG1 siRNA treatment at the indicated time-points. Values represent the percent decrease in HES1 mRNA relative to the control. (e) Correlation between JAG1 and HES1 mRNA expression in a dataset of 103 MB tumors [25]. r: Pearson’s value; p: p values. Proliferation (f), cell viability (g), caspase 3/7 activation (h), and histone-associated DNA fragments (i) of DAOY cells at 48 hours after JAG1 siRNA treatment or at the indicated time-points compared to control siRNA. Percent decrease in the number of colonies (j) and neuro-spheres (k) formed by DAOY cells 72 and 120 hours, respectively, following JAG1 siRNA treatment compared to control siRNA. (k, right panel) Representative image of DAOY-derived neuro-spheres upon treatment with JAG1 siRNA and control siRNA.

Figure 3

Expression of NOTCH ligands across MB molecular subgroups and MYC/JAG2 expression correlation in MB primary tumors, in histological MB subtypes, and in distinct metastatic MB stages. Dot plots showing relative expression (log2) of JAG1(a), JAG2(b), DLL1(c), and DLL3(d), across MB subgroups in a representative dataset of 103 human MB tumors [25]: SHH, n = 33; WNT, n = 8; Grp3 (Group 3), n = 27; Grp4 (Group 4), n = 35. (e) Correlation between MYC and JAG2 mRNA expression (log2) in a representative dataset of 103 human MB tumors [25]. r: Pearson’s value; p: p values. (f) Correlation study of JAG2 and MYC levels in Group 3 MB cases [25]. Box plots showing JAG2(g) and MYC(h) expression (log 2) according to the following MB histological variants [26]: classic (n = 51), desmoplastic (desmop.) (n = 6), and (LCA) large cells/anaplastic (n = 17); center line = median. Box plots showing JAG2(i) and MYC(j) expression (log 2) in MB tumors clustered according to metastatic stage [23]: M0, n = 42; M1, n = 7; ≥M2, n = 9; center line = median.

Figure 4

MYC-dependent regulation of NOTCH ligand JAG2 expression. (a) Relative mRNA expression (upper panel) of MYC (black bars, left Y axis) and JAG2 (orange bars, right Y axis) in the indicated cell lines. Values represent the fold-change in mRNA expression relative to cerebellum (defined as 1). Protein expression (lower panel) of MYC and JAG2 in the corresponding cell lines. (b) Relative mRNA expression (upper panel) of MYC and JAG2 in DAOY cells and stable clones of DAOY V11, and DAOY M2.1 cells. Values represent fold-change in mRNA expression relative to cerebellum (defined as 1). Protein expression (lower panel) of MYC and JAG2 in corresponding cells is shown. (c) MYC and JAG2 mRNA (upper panel) and protein (lower panel) expression in DAOY M2.1 cells transfected with MYC siRNA compared to control siRNA. mRNA values represent the percent decrease in JAG2 and MYC expression relative to control. Expression of β-actin was used as a control for western blot analysis. Statistical analysis: black stars indicate p values relative to MYC expression; orange stars indicate p values relative to JAG2 expression. (d) ChIP-on-chip data show occupancy of the JAG2 genomic sequence by MYC in four MB cell lines (from top to bottom: DAOY, UW-228, D458, and D341). (e) Time-dependent caspase 3/7 activation in DAOY M2.1 cells upon JAG2 siRNA treatment at the indicated time-points; values represent the percent increase in caspase 3/7 activity relative to control. (f) Proliferation status of DAOY cells, DAOY V11, DAOY M2.1, and DAOY M2.1 cells at 48 hours after JAG2 siRNA treatment compared to control siRNA. Caspase 3/7 activation at 48 hours after treatment with GSI in the indicated cell lines (g) and DAOY-derived clones (h). Low MYC cells: DAOY; high MYC cells: DAOY M2.1. Values represent the percent increase in caspase 3/7 activity relative to control siRNA.

Figure 5

MYC-dependent NOTCH signaling in MB cells. Simplified scheme illustrating the MYC-dependent NOTCH molecular network involving JAG1 and JAG2 in MB cells. (Left panel) Low MYC MB cell. (Right panel) High MYC MB cell.