At presentation, 50% of T cell acute lymphoblastic leukemia (T ALL) cases harbor activating mutations of the NOTCH1 transmembrane receptor, virtually all of which have also sustained deletions of the CDKN2A (INK4A-ARF) locus. Although the concordance of these genetic anomalies has been well documented, its basis remains unclear. The CDKN2A gene cluster encodes two structurally and functionally distinct tumor suppressors, p16INK4A and p14ARF (p19Arf in the mouse), the activation of which modulates the activities of the retinoblastoma protein and p53, respectively, to induce cell cycle arrest and/or apoptosis in response to oncogenic stress. A novel approach for modeling T ALL in the mouse now reveals that the ability of Arf to suppress Notch1-induced tumors is conferred at a specific stage of T cell development. Bone marrow cells or thymocytes transduced with a vector encoding the constitutively active intracellular fragment of NOTCH1 (ICN1) together with green fluorescent protein (GFP) were cultured ex vivo under conditions that support T lymphocyte differentiation and proliferation. These cultures were quickly dominated by early T cell precursors that expressed Thy-1 but not B cell or myeloid markers, and which produced a rapidly fatal CD4+/CD8+ (”double-positive”) T ALL when transferred into healthy, non-irradiated syngeneic mice. A phenotypically identical disease resulted upon infusion of the transduced cells into athymic nude mice, demonstrating that the deregulated ICN1 signal is sufficient to drive T cell development to the CD4+/CD8+ stage without any requirement for input from the thymic microenvironment. In recipients of bone marrow-derived ICN1+, Arf+/+ progenitors, T ALLs arose at high frequency, retained the Arf locus, but universally failed to express p19Arf. In turn, T ALLs initiated with bone marrow-derived ICN1+, Arf-/- donor cells exhibited only a modest acceleration of disease progression. In marked contrast, retention of the intact Arf locus in thymocyte-derived ICN1+ donor cells significantly prolonged disease latency and decreased tumor penetrance, indicating that Arf tumor suppression can be activated in thymic progenitors, but not in their less mature bone marrow precursors. Polycomb complexes epigenetically silence the Ink4a-Arf locus in primitive hematopoietic progenitors and early T cell precursors, preventing p19Arf expression and licensing a robust proliferative response to ICN1 signals. However, the locus is subsequently remodeled before T cells reach the double-positive stage, and its activation enables the Arf-p53 axis to cull T cells exposed to aberrant ICN1-induced oncogenic signals. In order to directly demonstrate that ICN1 overexpression in thymocyte-derived progenitors can induce the Arf locus, thymocytes from a homozygous Arf-GFP “knock-in” mouse were transduced with a vector encoding ICN1 and cherry-fluorescent protein (CFP), cultured short-term, and infused into healthy recipient mice. Notably, the Arf-GFP allele, while functionally null, expresses GFP in lieu of p19Arf when the cellular Arf promoter is induced. Healthy syngeneic recipients of ICN1/CFP, ArfGfp/Gfp donor cells developed CFP-marked T ALLs that co-expressed GFP, providing direct evidence that Arf is induced during thymus-derived leukemogenesis in vivo. Notably, Arf-induction not only provides p53-dependent elimination of incipient tumor cells but also generates a selective pressure for the emergence and survival of rare clones that have sustained Arf deletions. We reason that deletion of INK4A-ARF observed in human NOTCH1-induced CD4+/CD8+ T ALL strongly argues for the obligate induction of the locus at an earlier stage of T cell maturation. Our mouse model recapitulates this requirement, and should provide a useful platform to further elucidate disease mechanisms and potential therapies in T ALLs arising in nonirradiated mice that retain hematopoietic and immune functions.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.