The Direct Anti-Tumor Activity of Enzastaurin (LY317615.HCl) and Its Primary Metabolite (LY326020.HCl) Is Evident in Preclinical DLBCL Models Regardless of Molecular Subtype

Abstract 1336

Enzastaurin (Enza) is an ATP-competitive inhibitor of PKCβ and related AGC family kinases and blocks signaling through the PI3K/AKT/TOR pathway. Accordingly, in preclinical tumor models, clinically achievable Enza doses suppress phosphorylation of AKT, GSK3β, ribosomal protein S6, mTOR, p70S6 Kinase and 4EBP1.

Enza has shown significant activity in Diffuse Large B Cell Lymphoma (DLBCL) patients under clinical study. 4/55 DLBCL patients treated with Enza were progression- free after prolonged, continuous oral enza therapy, with 3/4 confirmed as complete responders (Robertson et al., JCO, 2007). Furthermore, preliminary analyses of intermediate or high-risk DLBCL patients treated with R-CHOP vs. Enza + R-CHOP suggest that Enza improves both progression-free survival and complete response rates (Hainsworth et al., JCO, 29: 2011, suppl; abstract 8016). Enza is currently being investigated in a phase 3 registration trial for DLBCL patients at high risk of relapse following R-CHOP therapy.

Gene expression profiling studies have revealed molecular alterations involved in DLBCL. For example, overexpression of PKCβ is particularly evident in recurrent DLBCL and has been associated with reduced patient survival. Similarly, BCL-6, a transcriptional repressor likely involved in the corruption of appropriate B Cell differentiation, is overexpressed in ∼ 70% of DLBCL and the target of translocation in nearly 40% of DLBCL. From these studies, distinct molecular subtypes of DLBCL have emerged. The Activated B Cell (ABC) subtype, which carries the poorest prognosis, is defined by constitutive activation of NFκB whereas the Germinal Center (GC) subtype, which carries a better prognosis, is not. 10% of ABC DLBCLs harbor CARD11 mutations leading to constitutive NFκB activation, while another 20% have mutations in CD79A/B, enabling chronic B Cell receptor signaling. In both ABC and GC subtypes, chronic activation of the B Cell receptor complex signals through PKCβ and the PI3K/AKT pathway.

Given its activity in DLBCL patients and its molecular mechanism of action, we chose to investigate the molecular basis for the direct anti-cancer activity of Enza, specifically in DLBCL, and to discern whether sensitivity to Enza is associated with distinct DLBCL subtypes. In multiple DLBCL lines, we show that Enza treatment at clinically achievable concentrations significantly reduces BCL-6 expression. We also show that Enza robustly induces apoptosis in both GC and ABC models, blocking signaling throughout the PI3K/AKT pathway. Moreover, we show in the same DLBCL models that the primary metabolite of enza (LY326020.HCl), which comprises ∼50% of circulating drug levels in patients, more potently inhibits these same signaling pathways and more robustly induces apoptosis. Importantly, we have also identified cell lines that are resistant to both enza and its metabolite. Profiling differences between these resistant and sensitive DLBCL lines, we now show that the pro-apoptotic activity of Enza and LY326020 requires control of the eIF4F translation initiation complex- the complex responsible for enabling translation of critical, malignancy-related mRNAs (e.g. VEGF, BCL-2, c-myc). In insensitive DLBCL cells, 4EBP1 expression levels are low to non-existent, even prior to treatment. In sensitive DLBCL cells, Enza and LY326020 reduce 4EBP1^ser65 phosphorylation in a dose and time-dependent manner, increasing 4EBP1: eIF4E binding thereby precluding eIF4E:eIF4G binding and preventing eIF4F complex assembly. Furthermore, selection of DLBCL cells for resistance yields cells with reduced 4EBP1 expression, increased eIF4G expression, or both, highlighting the critical importance of this complex for Enza activity. Collectively, these data show that: 1) Enza and LY326020 have direct anti-cancer activity in DLBCL cells regardless of ABC or GC subtype; 2) Enza and LY326020 directly impact numerous molecular alterations driving DLBCL (BCL-6 expression, PKCβ and PI3K-AKT pathway signaling); 3) the ability to control eIF4F complex assembly is critical for enza and LY326020 –induced apoptosis; and 4) LY326020 is more potent than Enza at blocking signaling and inducing apoptosis. Moreover, these data suggest that baseline expression levels of 4EBP1 and/or levels of phosphorylated eIF4E (an indicator of active eIF4F complex) may be particularly informative in delineating patients most sensitive to treatment.