A Jagged road to lymphoma aggressiveness

Abstract

In this issue of Cancer Cell, Cao and colleagues identifyanFGF4/Jagged1-driven crosstalk between tumor cells and their vascular niche that activates Notch signaling, sustaining the aggressiveness of certain mouse and human B cell lymphomas. These findings identify new therapeutic opportunities to target pathogenic angiocrine functions in cancer.

The Notch signaling pathway plays important roles in development, tissue homeostasis and cancer. In mammals, Notch signaling is mediated by four Notch receptors (Notch1-4) interacting with ligands of the Delta-like (Dll1, Dll3, and Dll4) or Jagged (Jagged1 andJagged2) families. During physiological signaling, ligand binding leads to intramembrane receptor proteolysis and release of intracellular Notch, which translocates to the nucleus and initiates target gene transcription. In cancer, activating NOTCH1 mutations were initially identified in a majority of T cell acute lymphoblastic leukemias (Fig. 1A)(Weng et al., 2004). A first class of point mutations in the extracellular heterodimerization domain disrupts stability of a negative regulatory region, leading to constitutive Notch activation even in the absence of ligand. A second type of mutations truncates the C-terminal PEST domain, leading to decreased proteasomal degradation and increased half-life of activated Notch. Although first associated with transformation of the T cell lineage, activating NOTCH mutations were subsequently identified in several subtypes of B cell malignancies, including Chronic Lymphocytic Leukemia, Mantle Cell Lymphoma, Splenic Marginal Zone Lymphoma and Diffuse Large B Cell Lymphoma (Fig. 1A) (Kiel et al., 2012; Kridel et al., 2012; Lee et al., 2009; Martinez-Trillos et al., 2013; Rossi et al., 2013). In almost all these cases, point mutations were identified only within the region encoding the PEST degron domain of either NOTCH1 or NOTCH2, suggesting sensitization to ligand-mediated receptor activation rather than true constitutive signaling. Moreover, systematic immunohistochemical scoring demonstrates Notch1activation at a frequency that markedly exceeds the reported prevalence of NOTCH1 mutations in specific lymphoid malignancies, suggesting that other mechanisms must exist to activate Notch signaling in these diseases(Kluk et al., 2013).

Figure 1. Emerging roles of Notch in lymphoma pathogenesis.

a. Structure of Notch1 and Notch2 receptors with sites of activating mutations previously reported in lymphoid malignancies. T-ALL - T cell acute lymphoblastic leukemia/lymphoma; CLL - chronic lymphocytic leukemia; MCL- mantle cell lymphoma; SMZL - splenic marginal zone lymphoma; ICN - intracellular Notch; HD - heterodimerization domain; TMD - transmembrane domain.

b. Proposed model for crosstalk between lymphoma cells (LC) and endothelial cells (EC), with Jagged1/Notch2-driven effects on lymphoma aggressiveness(Cao et al., 2014). FGFR1 - fibroblast growth factor receptor 1; FGF4 - fibroblast growth factor 4; ICN - intracellular Notch.

Results presented by Cao at al. in this issue of Cancer Cell identify the capacity of endothelial cells within the tumor microenvironment to induce ligand-mediated Notch activation in adjacent lymphoma cells, leading to enhanced tumor growth and aggressive in vivo behavior (Fig. 1B)(Cao et al., 2014). Using complementary in vitro and in vivo models of mouse and human Myc-driven lymphoma interacting with a vascular niche, the authors describe reciprocal interactions involving FGF4-dependent induction of the expression of the Notch ligand Jagged1 in endothelial cells, leading in turn to Notch2-mediated signaling in tumor cells with lymphoma initiating characteristics. As a first step, Cao et al. built on past work from their laboratory using E4ORF1-transduced endothelial cells, which can be maintained in culture without serum or recombinant angiogenic factors and allow for detailed analysis of their angiocrine functions(Butler et al., 2010). Co-culture of these endothelial cells with Eµ-Myc-driven mouse B lymphoma cells selected for cells with increased in vitro growth, chemoresistance, in vivo repopulation potential and invasiveness. Using a combination of genetic and pharmacological methods, the authors demonstrated that this phenomenon requiredJagged1 expression by the endothelial cells and Notch2 but notNotch1 expression in lymphoma cells. FGF4 release by the lymphoma cells induced Jagged1 expression, suggesting that specific lymphomas capable of FGF4 production might be uniquely sensitive to this mechanism (although alternative pathways might exist to induce Jagged1). The effects of Notch2 appeared entirely dependent on the Notch target gene Hey1, while mechanisms operating downstream of this transcriptional repressor remain to be investigated. To gain insight about the potential human relevance of these findings, Cao et al. analyzed a panel of primary human Burkitt’s lymphomas, demonstrating the existence of Hey1-positive tumor cells in proximity of Jagged1-expressing endothelial cells in these tumors. Moreover, knockdown of Notch2 in these primary human tumor cells followed by transfer into immunodeficient mouse recipients recapitulated observations made with Eµ- Mycmouse tumors, suggesting the existence of shared pathogenic mechanisms at least in these Myc-driven lymphoid malignancies.

It remains to be determined how broadly applicable the specific observations presented in this paper will be. Of note, tumor-vasculature interactions involving Notch signaling were reported recently in other contexts, such as glioblastoma multi for me and colorectal cancer. An interesting implication of these findings is that Notch activation might be at play in malignant tissues even in the absence of activating NOTCH mutations, provided the tumor microenvironment constitutes a good source of Notch ligands. In lymphoid malignancies, PEST domain mutations are predicted to matter only upon exposure of the tumor cells to Notch ligands, as these genetic events lead to stabilization of cleaved active Notch only after ligand-receptor interaction. Thus, PEST domain mutations may function as a sensitizer for the exposure of malignant cells to Notch ligands in their immediate environment. In other words, the type of ligand-dependent mechanisms identified by Cao and colleagues could synergize with PEST domain Notch mutants to potently activate the pathway. In terms of the overall importance of Notch as an oncogenic pathway in B cell neoplasms, another word of caution is that Notch signaling can have versatile functions. Tumor suppressive effects of the pathway have been reported, for example in myeloid neoplasms, squamous cell carcinomas and certain B cell malignancies(Zweidler-McKay et al., 2005). It is possible that Notch activation could have different effects in lymphomas originating from cells arrested at specific stages of differentiation, such as before, during, or after the germinal center reaction. To fully investigate the spectrum of Notch effects in lymphoma, future work should ideally focus on in vivo models and on the combined use of loss-of-function and gain-of-function experimental approaches.

The vascular niche represents an attractive potential source of Notch ligands in the tumor microenvironment. In lymphoma, past observations revealed increased Notch activity in a high proportion of aggressive and highly vascularized angioimmunoblastic T cell lymphomas (Kluk et al., 2013). However, other cellular sources of Notch ligands must be considered as well. Using standardized immunohistochemistry to specifically detect the cleaved and active form of Notch1 in tumor tissues, Aster and colleagues reported evidence of high Notch activity within secondary lymphoid organs, but with rapid loss of the signal in tumor areas that extend beyond the lymph node capsule(Kluk et al., 2013). Although anecdotal, these findings suggest that cellular elements specific to the lymph node microenvironment might represent an important source of Notch ligands in vivo. Moving forward, additional work may reveal specific mechanisms for Notch ligand induction in this context and cooperativity with genetic Notch activation, perhaps leading to the maintenance of lymphoma-initiating cells in vivo. As a practical implication, future investigations are predicted to underestimate or to altogether miss the importance of Notch signaling in lymphoid malignancies if tumor cells are not adequately exposed to Notch ligands in culture (e.g. in the absence of cocultured cells expressing relevant Notch ligands) or in vivo (e.g. in conventional subcutaneous xenografts).

The translational impact of the findings reported by Cao et al. is significant, as they implya larger array of potential tumor targets for the use of therapeutic Notch inhibition than predicted only by NOTCH mutational analysis. In any case, emerging data implicating a central role for Notch legends derived from the tumor vasculature or other sources suggest that our current understanding of Notch in lymphomagenesis has only exposed the tip of the iceberg. It is time to dive deeper.