To investigate NPM1 wild-type and mutant interactors, we performed a comprehensive proteomic analysis. Comparing the proteins enriched in NPM1 wild-type and NPM1-mutated (NPM1c) samples, we identified the specific interactors of the two protein forms. In the NPM1 wild-type specific network, we observed proteins mainly localized into the nucleus/nucleoli, involved in ribosome biogenesis and rRNA processing such as TTF1, RPL18, RPL15 and SURF6. Conversely, the specific interactors of NPM1c were proteins involved in gene expression regulation, such as histone modifiers (KDM2A, KDM3B), transcription factors (NCOR1, FOXM1) and chromatin remodeling factors (HELLS). In addition, for NPM1c we found also proteins involved in the formation of nuclear pore complex (NUP42, NUP98, NUP214, KPNA3) and in mRNA export (GANP, ENY2). We are now validating these data by co-immunoprecipitation, immunofluorescence, and immunohistochemistry using NPM1-mutated AML cell lines and AML samples from partients.
We have generated mice carrying mutations of Npm1 and Flt3-ITD. These mice developed leukemia and we established that reduced levels of the transcription factor GATA1 promoted the development of AML in this model. We found that decreased levels of GATA1 mRNA and protein were secondary to an hypermethylated status on its promoter in the bone marrow and re-expression of GATA1 at early stages of AML development improved the myeloid and erythroid alterations. To evaluate the effect of hypomethylating agents on GATA1, we treated our compound Npm1/Flt3-ITD mutated mouse with 5-Azacytidine. We observed that 5-Azacytidine re-activated GATA1 expression and demonstrated significant anti-leukemic effect. (Sportoletti et al. Leukemia. 2019 Jul;33(7):1827-1832).
As part of the Objective 4, we investigated the relationship between NPM1-mutated AML and other hematological neoplasms to understand whether they had a common origin. In particular, we demonstrated that two completely different hematological neoplasms (i.e. RHOA-mutated angioimmunoblastic T-cell lymphoma and NPM1-mutated AML) could raise from a common clonal hematopoiesis driven by TET2 and ASXL1 mutations (Tiacci E.et al. N Engl J Med 2018 379(10):981-984). This implies that NPM1-mutated AML patients with persistent clonal hematopoiesis after chemotherapy should be monitored for the development of a second malignancy
The most ground-breaking result has been the demonstration that growth of NPM1-mutated AML closely depends on NPM1 mutant and its cytoplasmic localization. Indeed the loss of NPM1c from the cytoplasm, either through nuclear relocalization or targeted degradation, resulted in immediate downregulation of homeobox (HOX) genes followed by differentiation (Brunetti L et al. Cancer Cell 2018 34(3):499-512.e9). In contrast, the typical downregulation of CD34 in NPM1-mutated AML is not dependent on the cytoplasmic localization of NPM1c (Pianigiani G et al. Leukemia. 2022 36(7):1931-1934). Finally, we showed that XPO1 inhibition relocalized NPM1c to the nucleus, promoting differentiation of AML cells and prolonging survival of Npm1-mutated leukemic mice (Brunetti L et al. Cancer Cell 2018 34(3):499-512.e9). These findings provide the rationale for using the nuclear export inhibitors in NPM1-mutated AML (Falini B et al., Blood. 2020 136(15):1707-1721;Ranieri R. et al, Leukemia. 2022 Oct;36(10):2351-2367). More recently, we found that the second-generation XOP1 inhibitor Eltanexor is more active than selinexor since causes irreversible HOX downregulation, terminal AML differentiation and prolonged survival of leukemic mice (Pianigiani G et al., Blood Adv. 2022. PMID: 36037515). This information is of critical importance for designing clinical trials with XPO1 inhibitors in NPM1-mutated AML.
The trial we conducted with actinomycin D in older patients with NPM1-mutated AML unfit for intensive chemotherapy led to about 40% CR (Gionfriddo et al. Leukemia 2021 35(9):2552-2562; Falini B. et al. Blood. 2021 137(5):589-599). Interestingly, actinomycin D induced nucleolar stress both in vitro and in patients and this effect was more pronounced in AML cells expressing the NPM1 mutant as compared to NPM1 wild-type cells, suggesting that NPM1-mutated AML may be more sensitive to nucleolar stress (Gionfriddo I et al. Leukemia 2021 Sep;35(9):2552-2562). Finally, we identified the mitochondrial/ROS/PML/TP53 senescence pathway as an effector of Actinomycin D-based therapies in NPM1-mutated AML (Wu H-C, Cancer Discovery. 2021 11(12):3198-3213; Ranieri R. et al, Leukemia. 2022 36(10):2351-2367).
One of the patients in actinomycin D trial carried an NPM1-containing fusion transcript (NPM1/RPP30). This prompted us to screen n=387 NPM1-mutated AML patients and to discover novel NPM1 exon-5 mutations and fusion proteins (NPM1/RPP30, NPM1/SETBP1, NPM1/CCDC28A) (Martelli MP, Blood. 2021 138(25):2696-2701) that all shared the property to accumulate in the cytoplasm of leukemic cells. In an extended cohort of patients, we also showed the diagnostic utility of an anti-IDH1 R132H mAb (Falini B., Leukemia. 2019 33(4):1043-1047.).
