The molecular ontogeny of follicular lymphoma

Background: Advanced-stage follicular lymphoma (FL) is considered incurable, with patients typically suffering from a chronic relapsing clinical course. Relapsed FL is thought to originate from common progenitor cells (CPCs) through divergent evolution. CPCs typically contain the hallmark BCL2/IGH translocation, in addition to other recurrent mutations. Targeting CPC-defining, early acquired (i.e., truncal) mutations offers the chance to eradicate/cure the disease. Aim: To define and functionally characterize early acquired mutations in FL. Results: We hypothesized that CPC-defining mutations can be acquired before the FL-hallmark BCL2/IGH translocation in hematopoietic stem and progenitor cells (HSPCs). To directly define the molecular ontogeny of FL, we used highly sensitive mutation detection of purified BCL2/IGH-negative (BCL2/IGH-) HSPCs from leukapheresis samples of 3 patients in clinical remission. However, using complementary approaches, we did not identify CPC-defining mutations in BCL2/IGH- progenitor cells. Vice versa, we identified B cells harboring the FL-specific BCL2/IGH translocation but without CPC-defining mutations in one informative case. We next used indirect approaches to identify truncal mutations for functional characterization. Using cancer cell fraction corrected (CCF)-variant allele frequency (VAF) from FL patients, we identified STAT6 (based on a previous study) and IKZF3 (based on our own data) as truncal events by high clonality. Previous work from our group (Boesl et al.) provided the first insights into the biology of STAT6 mutations, however, the mechanisms remained incompletely understood. In my thesis, I significantly contributed to this work by showing that STAT6 mutations drive a self-reinforcing microcircuit. Specifically, I showed that in the presence of interleukin-4 (IL-4) (i) STAT6D419G mutations are gain-of-function (by serial replating pre-B cell colony-forming-units (CFU) assays), that (ii) STAT6D419G (but not STAT6WT) binds to the PARP14 promoter and activates gene expression of PARP14 (which is a component of the STAT6 enhanceosome), and that (iii) increased nuclear accumulation of phosphorylated STAT6D419G drives increased expression of known STAT6 target genes, including FCER2, CCL17, CCL22, as well as PARP14 itself. Finally, using our own available DNA sequencing data, we identified IKZF3 mutations as truncal events by high clonality. Of note, mutations cluster in distinct hotspots, mostly L162R and S215R. To model early and later mutation acquisition, I utilized Cre-Lox recombination restricted IKZF3 expression in lineage negative (lin-) HSPCs from Emu-BCL2/Vav-Cre mice (IKZF3 expression before B-lineage commitment) vs. Emu-BCL2/Mb1-Cre mice (IKZF3 expression in B-lineage committed cells). My results indicate that IKZF3S215R mutations are primarily gain-of-function (i.e. provide a serial replating phenotype) when expressed in HSPCs. In contrast, IKZF3L162R mutations provide a serial replating phenotype when expressed in pre-B cells. Phenotypic analysis of IKZF3L162R mouse pre-B CFUs identified them to be in a highly proliferative differentiation state (B220+ CD24++ CD43+), characterized by ongoing pre-BCR signaling and increased SYK expression. In primary human FL-like B cells, we confirmed this phenotype, with IKZF3L162R cells showing increased SYK expression and an increased proliferative rate. RNA sequencing analysis of IKZF3L162R FL-like B cells vs. IKZF3WT identified B cell proliferation and activation to be enriched. Discussion and conclusion: The absence of mutations in HSPCs suggests that in human FL, the BCL2/IGH translocation can precede the acquisition of CPC-defining mutations. We also identified BCL2/IGH+ cells in a patient in ongoing clinical remission but without CPC-defining mutations. The absence of CPC-defining mutations suggests that these cells are not contributing to relapse. In the second part of my thesis, I studied early acquired STAT6 mutations. My results significantly contributed to further characterizing the STAT6 mutation phenotype. We propose that the STAT6D419G gain-of-function phenotype is strictly dependent on IL-4. STAT6D419G but not STAT6WT aberrantly increases PARP14 levels in lymphoma cells. Increased PARP14 levels then contribute to increased assembly/stabilization of the STAT6 enhanceosome complex, resulting in increased STAT6-dependent gene expression and thereby amplifying an IL-4 driven self-reinforcing microcircuit. Therefore, PARP14 represents an attractive therapeutic target in STAT6MUT FL. Lastly, I identified IKZF3 with mutation-specific phenotypes when expressed in HSPCs vs. B cells. My results indicate that IKZF3L162R induces proliferative expansion of pre-B cells and FL-like mature B cells, characterized by increased BCR signaling and SYK expression. Further functional experiments are required to better define the precise mechanism