Abstract
Rationale: About 75% of newly formed B-cells are autoreactive and express potentially harmful autoantibodies (Wardemann 2003). Hence, a powerful mechanism, termed central tolerance, is in place to eliminate millions of newly formed autoreactive B-cells every day.
Results: B-ALL, mantle cell lymphoma (MCL) and unmutated chronic lymphocytic leukemia (U-CLL) originate from early, pre-germinal center (pre-GC) stages of B-cell development that are subject to negative B-cell selection and central tolerance mechanisms. While designed to eliminate autoreactive clones during early B-cell development, we recently discovered that B-ALL, MCL and U-CLL fully retained sensitivity to central tolerance mechanisms, which are triggered by persistent PI3K-hyperactivation.
PI3K-signaling code to distinguish between normal and pathological signaling. Studying short transient pulses and chronic activation of PI3K-signaling, we discovered that pre-GC B-cells have evolved a "PI3K-signaling code" to distinguish between normal B-cell activation by antigen and pathological signaling: thereby, antigen encounter induces a short transient pulse of PI3K-activation which promotes survival and proliferation. Conversely, persistent activation of PI3K-activation reflects pathological signaling, either from an autoreactive B-cell receptor (BCR) or a transforming oncogene.
Pre-GC B-cell malignancies are exempt from oncogenic PI3K-lesions. PI3K-lesions in cancer result in permanent hyperactivation as in autoreactive B-cells. The PI3K pathway is targeted by oncogenic lesions in ~25% of human cancer. The phosphatases PTEN, SHIP1 and PP2A function as negative regulators of PI3K signaling and are frequently mutated in a broad range of cancers and also occur in some GC- and post-GC lymphomas (e.g. Burkitt's, DLBCL). However, our analysis in six clinical cohorts revealed that pre-GC B-cell malignancies, including B-ALL, MCL and U-CLL critically depend on PTEN, SHIP1 and PP2A function and do not tolerate persistent hyperactivation of PI3K-signaling for more than three hours. Loss-of-function mutations of these phosphatases and activating PI3K lesions were not detected in large clinical cohorts of patients with B-ALL, MCL and CLL. Likewise, phosphorylation of AKT-S473, reflecting PI3K signaling strength, is elevated throughout multiple cancer types including post-GC DLBCL, but not in B-ALL and MCL. This is in line with previous work demonstrating that inherited mutations that cause PI3K-activation predispose to various cancers but cause profound defects in human B-lymphopoiesis (Fruman 2014).
Pharmacological targeting of PI3K-dependent central tolerance mechanisms. We tested the hypothesis that PI3K-hyperactivation represents a unique vulnerability in pre-GC B-cell tumors including B-ALL, MCL and U-CLL. Sensitivity to PI3K-hyperactivation of reflects their pre-GC origin and central tolerance mechanisms during early B-cell development that are designed to eliminate autoreactive B-cells based on hyperactive PI3K-signaling. For this reason, we tested pharmacological PI3K-hyperactivation as a novel strategy to selectively target pre-GC B-cell malignancies. To this end, we tested 144 compounds for their ability to engage PI3K-dependent central tolerance mechanism in B-ALL, MCL and CLL. Small molecule inhibitors of SHIP1 (3AC, K118), PTEN (SF-1670), PP2A (LB-100) and a direct PI3K-agonist (SC79) achieved strong phosphorylation of known PI3K-substrates (AKT, S6K) in vitro and prolonged overall survival in NSG mice transplanted with refractory B-ALL and MCL PDX in vivo.
Conclusions and future directions: Current treatment regimens (kinase-inhibitor paradigm) use agents that apply selective pressure in one direction (e.g., PI3K-inhibitors; BCR-ABL1, SYK- or BTK-inhibitors). Here, we are pursuing a new concept (central tolerance paradigm) based on sequential treatment regimens that alternate between kinase-inhibitors (e.g., dasatinib, ibrutinib, idelalisib) and PI3K-hyperactivation (3AC, K118, LB100). By sequentially applying selective pressures in opposite directions, our approach will subvert clonal evolution and selection for drug-resistant mutants.
No relevant conflicts of interest to declare.
