Notch receptors are transmembrane proteins that participate in a signaling pathway in which ligand binding induces cleavage and nuclear translocation of the Notch intracellular domain (NICD). This leads to formation of a transcriptional activation complex that turns on Notch target genes and mediates downstream effects. Gain-of-function mutations leading to Notch hyperactivation are common in a subset of hematologic malignancies. However, the distribution of mutations within Notch receptors varies. In T-ALL, mutations involving the extracellular domain of NOTCH1 that lead to ligand independent receptor cleavage and activation are frequently observed. By contrast, mutations in NOTCH1/2 in B-cell lymphomas are largely restricted to the C-terminal PEST domain, which normally ensures rapid turnover of ICD. This implies that oncogenic Notch signaling in B-cell lymphomas is ligand- dependent, with PEST deletions serving to turn up signaling within particular ligand-rich microenvironments.

We previously observed that NOTCH1 activation in CLL was sharply increased in the lymph node microenvironment and was not confined to NOTCH1 mutated tumors, though NOTCH1 mutated CLLs did have higher levels of NICD1. By analogy, we hypothesized that NOTCH2 activation in splenic marginal zone lymphoma (SMZL) would be increased in splenic marginal zones, a microenvironment rich in the Notch ligand DLL1, and that NOTCH2 activation would be stronger in, but not restricted to, NOTCH2 mutated tumors. We tested these predictions by doing immunohistochemistry (IHC) using an antibody directed against NICD2 in archival formalin-fixed paraffin embedded cases of SMZL (39), chronic lymphocytic leukemia (CLL, 43), non-splenic marginal zone lymphoma (MZL, 57), and mantle cell lymphoma (MCL, 47). Two experienced pathologists scored the stained slides for intensity (0-3) and the percentage of positive cells. Positive staining was defined as at least focal nuclear staining in ³5% of cells. Positive and negative controls were included in all staining runs, which were performed on an automated Leica Bond immunostainer. Image analysis with a customized IHC nuclear V9 algorithm (Aperio technologies, Inc.) was used to quantify nuclear staining in different histologic compartments. Targeted next-generation sequencing (BWH OncoPanel 447-gene panel) was performed on a subset of SMZL cases (n=20).

The vast majority of SMZL cases (36/39; 92%) were positive for NICD2 staining. By contrast, only weak NICD2 staining was seen in small subsets of other B cell tumors - 2% of CLL (1/42), 19% of MCL (9/38) and 16% of non-splenic MZL (9/48) cases. Most SMZL cases showed nuclear staining of variable intensity (weak to strong) in a variable number of neoplastic cells (5-80%). The majority of NICD2-positive cases (26/36, 72%) showed weak NICD2 staining in a subset of neoplastic cells (< 40% of cells), while a minority (10/36, 28%) showed more diffuse NICD2 staining (³ 40% of cells). In addition, NICD2 staining was accentuated in marginal zones relative to the inner half of white pulp nodules and the red pulp (1.6x, p<0.05). Targeted next-generation sequencing performed on subsets of NICD2-high and NICD2-low cases (n=10 of each from n=19 patients) revealed gain-of-function NOTCH2 PEST domain mutations in 9 cases. All but one of the cases (8/9) with high NICD2 staining had a gain-of-function NOTCH2 PEST domain mutations. Overall, the pattern of NICD2 staining correlated with NOTCH2 mutation status in 85% (17/20) cases. There was no significant difference (p>0.05) in overall survival between NOTCH2-wild type and -mutated or NICD2-high and -low groups.

In conclusion, among low-grade B cell lymphomas, NOTCH2 activation is much more prevalent in SMZL than in CLL, MCL, and nonsplenic MZL and is accentuated in (but not confined to) tumors with NOTCH2 PEST domain mutations. These findings suggest that NOTCH2 signaling is broadly important in the pathogenesis of SMZL, where it likely acts following ligand-mediated activation in the microenvironment of splenic marginal zones to drive expression of currently unknown target genes that support SMZL growth and survival.


Benhadji: Eli Lilly and Company: Employment. Oakley: Eli Lilly and Company: Employment. Aster: Eli Lilly and Company: Membership on an entity's Board of Directors or advisory committees.

Author notes


Asterisk with author names denotes non-ASH members.