AbstractAbstract 150The recent demonstration that memory B-cells can re-enter germinal centers (GCs) and participate to new rounds of GC reactions has opened the possibility that multi-hit B-cell lymphomagenesis could be a much more dynamic process than initially anticipated, gradually progressing throughout the successive passages of memory B-cells in GCs during a lifetime of successive immunological challenges. Here, we provide evidence for this scenario in follicular lymphoma (FL), a GC derived B-cell malignancy initiated in the bone marrow by the hallmark t(14;18) BCL2/IGH translocation. To address this issue, we engineered an original sporadic BCL2tracer mouse model mimicking the rare occurrence of t(14;18) translocation in humans through V(D)J recombination errors (1 in a million B-cells) allowing to track the resultant BCL2-expressing clones; and underwent a molecular/immunofluorescent tracking of t(14;18)+ clones vs. normal memory B-cells in paired lymphoid tissue samples (spleen, lymph nodes, bone marrow) from healthy individuals.We first show that contrary to the current dogma, ectopic BCL2 expression is not sufficient to provoke the FL’s characteristic differentiation arrest of activated B-cells as GC B-cells, thereby suggesting that differentiated BCL2+ memory B-cells must return to the GC to acquire additional oncogenic hits and “fix” in situ growth. Strikingly, we further find that in a small fraction of “healthy” humans, such differentiation arrest already operated, and that a clonally expanded population of t(14;18)+ cells with FL-like features have widely disseminated in blood and in multiple lymphoid organs (spleen, lymph nodes, bone marrow), with unprecedentedly reported frequencies (from 1/million to 1/500 cells in some individuals), shaping the systemic disease presentation observed in FL patients. Using molecular/immunofluorescent backtracking of such clones in various paired and remote lymphoid organs, we further demonstrate that t(14;18)+ clones systematically display an extensive history of AID (activation-induced cytidine deaminase)-mediated events compatible with iterative rounds of GC co-opting, in sharp contrast to single memory B-cell clones from the same individuals. We thus show that BCL2-expressing memory B-cells require multiple GC transits to acquire the distinctive FL-like maturation arrest as centrocyte/centroblasts and to progress to advanced FL precursor stages. Altogether, our data argue for a model of lymphomagenesis, in which progression from precursor stages to FL occurs asymptomatically over an extended period of time by subverting the dynamic and plastic attributes of memory B-cells. This understanding of the pre-clinical phases driving FL development in asymptomatic patients should help rationalize prospective approaches designed to identify biomarkers of risk, and innovative therapeutic targets present in early, potentially more curable phases of the disease. Disclosures:No relevant conflicts of interest to declare.