Growth Factor Independent-1 (Gfi1) As a New Target for Human Leukemia Therapy

Abstract 560

More than 50% of patients diagnosed with B or T-cell leukemia and lymphoma will fail current treatment protocols. This highlights the urgent need for new and improved therapies. Since the transcription factor Growth factor independent-1 (Gfi1) plays an important role in lymphoid differentiation, we explored whether it might be a suitable target for therapy. Using mouse models in which T-cell leukemia can be induced by transgenic expression of mutated forms of Notch1, by injection of the carcinogen N-Ethyl–nitrosourea (ENU) or infection with a Murine Moloney Leukemia Virus, we found that Gfi1 knockout mice had a significantly lower incidence and a longer latency period of T-ALL. To verify whether targeting Gfi1 would be a novel approach to treat B- or T-cell lymphoma, we used Mx1Cre Gfi1^fl/fl mice. In these mice, injection of polyinosinic-polycytidylic acid (pIpC) activates the Mx1 promoter driven Cre expression, which ultimately leads to the deletion of the floxed Gfi1 alleles. As controls, we used Gfi1^fl/fl mice, which lack the Cre recombinase and thus still express Gfi1 after pIpC injection. To elicit a T- or B-cell lymphoma, we used ENU injection combined with expression of a mutated Notch1 transgene for T-ALL, or transgenic over-expression of c-Myc for B-cell leukemia (Eμ-Myc). Using in-vivo ultrasound supported imaging, we observed complete regression of tumour masses when Gfi1 was eliminated in Mx1Cre Gfi1^fl/fl mice, curing the mice of the either the T or B-cell malignancies. Strikingly, this effect was observed in the absence of any other treatment regimen.

To explore the mechanisms underlying this phenomenon, we explanted tumor samples from mice, in which Gfi1 expression was either maintained or deleted and performed gene expression arrays. A comparative analysis of the array data demonstrated that loss of Gfi1 affects pathways of key importance for leukemia such as metabolism, cell cycle progression, basal transcription and apoptosis but also the response to DNA damage. It has been shown previously for non hematological malignancies that oncogenic transformation in conjuction with dysregulated cell cycle induces DNA strand breaks (DSBs), which leads to an increased p53 dependent apoptotic response in tumours compared to non-transformed cells. This forces the tumor cells to counteract this effect – for instance by selection for the loss of p53. Consistent with this concept, we noted that leukemic cells from our tumor models displayed a greater amount of DSBs and also higher rates of spontaneous apoptosis than normal cells. Interestingly, the number of apoptotic cells was further increased in those tumors where Gfi1 had been deleted. We hypothesized that Gfi1 protects leukemia cells against DSB induced apoptosis. To test this hypothesis, we irradiated in Gfi1^+/+ and Gfi1^−/− thymocytes, which induces DSB in thymocytes. In line with our hypothesis we found that Gfi1^−/− thymocytes showed increased rates of apoptosis compared to irradiated Gfi1^+/+ thymocytes and that loss of Gfi1 led to an increased induction of pro-apoptotic genes such as Bax, Noxa and Puma after irradiation. Since Bax, Noxa and Puma are all p53 target genes, we investigated a possible link between Gfi1 and p53 and found that (I) Gfi1 binds to p53, (II) that Gfi1 inhibits the transcription of p53 target genes and (III) that Gfi1 occupies p53 target gene at the same sites as p53. In summary, Gfi1 antagonizes p53 function and loss of Gfi1 sensitizes cells to p53-mediated apoptosis.

Next, we used different human T-ALL cell lines and treated these cells either with sh-RNA lentivirus or morpholinos to abrogate GFI1 expression. In all cases, down-regulation of GFI1 expression led to increased apoptosis and impeded growth of human leukemia cells. Finally, we transplanted leukemic cells of T-ALL patients into NOD-Scid, IL2R^null (NSG) mice, waited for the leukemic cells to engraft, and then injected GFI1 specific- or control morpholinos. While mice treated with control morpholino died of leukemia, the animals treated with GFI1-specific morpholinos survived showing a significant reduction of human leukemic cells in the blood, bone marrow and spleen, even with samples from patients who did not respond to first line therapy. Since morpholinos have received approval for use in humans, our data suggest that targeting GFI1 in human T-ALL patients may be a promising therapeutic target and a feasible way to complement current T-ALL treatment regimens.