Acute myeloid leukemia (AML) and myelodysplastic syndromes (MDS) with monosomy 7 or deletion of chromosome segment 7q [-7/del(7q)] are poor-prognosis, chemoresistant myeloid neoplasms that are often associated with prior treatment with alkylating drugs such as cyclophosphamide, melphalan, or cisplatin. Haploinsufficiency of one or more tumor suppressors has been postulated as critical to disease pathogenesis, but until now, experimental evidence implicating specific candidate gene(s) has been lacking.
In a multi-institutional collaboration led by Dr. Scott Lowe from Memorial Sloan Kettering Cancer Center, the gene encoding mixed lineage leukemia 3 (MLL3), located on chromosome 7q36.1, is identified as a key haploinsufficient tumor suppressor that cooperates with additional genetic lesions to mediate leukemogenesis. MLL3 encodes a histone methyltransferase and the gene is frequently mutated in human cancers; however, the mechanistic relationship of such mutations to tumorigenesis and the role of MLL3 in human leukemias is not well understood.
Interrogation of genomic profiles from AML patients revealed a significant association between -7/del(7q) cases and mutations of the NF1/RAS pathway and inactivation of the TP53 gene. To assess the cooperativity of these genetic lesions, a transplantation mouse model using retroviral gene transfer of hematopoietic stem and progenitor cells (HSPC) was undertaken. Short-hairpin RNAs (shRNAs) against MLL3 and NF1 were used to suppress levels of these proteins in the background of mice with TP53-/- HSPCs that already exhibit decreased survival. In mice with this TP53 null background, additional suppression of NF1 alone only mildly reduced survival, and MLL3 suppression alone generated no survival difference. In contrast, suppression of both NF1 andMLL3 significantly reduced the median survival of mice by approximately 50 percent.
In contradistinction to the predominant phenotype of CD3+T-cell lymphomas observed in mice transplanted with only TP53-deficient HSPCs or with additional shRNA knockdown of either NF1 or MLL3, tandem shRNA knockdown of NF1and MLL3 resulted in a phenotypic switch to a myeloid lineage leukemia. These mice exhibited marked leukocytosis with increased blasts, massive organomegaly, and anemia. In recipient mice transplanted with HSPCs with a null TP53background and shRNA knockdown of NF1, disruption of MLL3 by CRISPR/Cas9 genome editing technology recapitulated the AML phenotype produced by knockdown of MLL3 by RNA interference. Evaluation of AML clones subjected to CRISPR-Cas9–directed mutagenesis revealed heterozygous mutations of MLL3 in almost all cases, consistent with the notion that haploinsufficiency of MLL3 is adequate to drive leukemogenesis.
Investigators next sought to determine whether suppression of MLL3 impairs myeloid differentiation, a characteristic of both AML and MDS. Indeed, compared with controls, TP53-/- mice with shRNA knockdown of MLL3 exhibited reduced numbers of committed myeloid progenitors and reduced bone marrow cellularity, and marrow cells displayed reduced colony formation capacity. In addition, mice exhibited decreased white blood cell and platelet counts, and multilineage dysplasia. The transcriptional profile associated with suppression of MLL3 consisted of down-regulated genes linked to hematopoietic differentiation. The gene-expression profile was also similar to previously published “early leukemia stem cell” signatures, and signatures of genes down-regulated in hematopoietic stem cells derived from MDS patients, including the -7/del(7q) subtype. Chromatin immuno-precipitation assays confirmed that suppression of MLL3 was associated with reduced methylation of lysine 4 on histone H3 (H3K4) and increased methylation of lysine 27 on histone 3 (H3K27), changes expected to induce transcriptional repression at multiple loci, including genes involved in hematopoietic differentiation.
Similar to the historically poor responsiveness of -7/del(7q) AML, mice harboring MLL3-suppressed leukemia exhibited resistance to the standard chemotherapy agents doxorubicin and cytarabine. When chemosensitive AML1-ETO–induced AML cells were modified by suppression of MLL3, therapeutic resistance emerged. However, AML cells and leukemic mice with suppressedMLL3 were sensitive to the BET inhibitor JQ1, as were human AML cell lines with -7/ del(7q). JQ1 treatment-reduced MYC gene transcription and protein expression consistent with prior data showing that inhibition of the BET family protein Brd4 reverses the aberrant self-renewal programs in stem cells through blockade of Myc-dependent pathways.
In Brief
These elegant experiments provide compelling evidence that MLL3 is a key tumor suppressor whose haploinsufficiency recapitulates the phenotype of -7/del (7q) AML and MDS. However, heterozygous loss of MLL3 alone is not sufficient to drive leukemogenesis in murine models. A differentiation block in the myeloid lineage manifests only when suppression of MLL3 is combined with loss of p53, whose deficiency usually drives development of lymphoid neoplasias. Thus, further investigation of the cooperative role of other haploinsufficient genes in the 7q region is needed. The observed preclinical efficacy of the BET inhibitor JQ1 suggests an opportunity for efficacious drug development that has thus far been lacking for patients with therapy-related AML or MDS.
