The role of polycomb repressive complex 2 in early T-cell precursor acute lymphoblastic leukemia

ABSTRACT

Genetic lesions affecting polycomb repressive complex 2 (PRC2) have been found in more than 40% of pediatric cases of early T-cell precursor acute lymphoblastic leukemia. The functional role of these PRC2 alterations has been obscure. Our recent data suggest that compromise of PRC2 blocks differentiation and accentuates growth and survival signaling.

The importance of epigenetic gene regulation in cancer and development is becoming increasingly evident. Major mechanisms of epigenetic gene regulation involve direct methylation of DNA or methylation of histones on arginine and lysine residues. PRC2 is an important developmental regulator complex. The canonical function of PRC2 is to catalyze the di- and tri-methylation of lysine residue 27 of histone 3. This epigenetic modification is associated with low transcriptional output and is thought to reinforce (rather than initiate) epigenetic gene silencing.¹ PRC2, and in particular its major methyltransferase EZH2, was identified as oncogenic in 2002. Subsequently, high-level expression of Ezh2 was found in a number of cancers, and forced expression of Ezh2 was shown to enhance self-renewal of hematopoietic cells. Activating mutations of Ezh2 were discovered in B-lineage lymphomas, and small-molecule inhibitors of Ezh2 were shown to be of potential benefit in germinal center lymphomas.² In MLL-fusion mediated AML, the structural PRC2 component EED was shown to be strictly required for leukemogenesis. Ezh2, although not absolutely required, contributed to the progression of disease.³ This initial body of work supported an oncogenic role for hyperactive and wild type PRC2.

The complexity of PRC2 became apparent when inactivating genetic lesions in PRC2 components were discovered in myeloid malignancies, including myelodysplastic and myeloproliferative disorders. Subsequently, several groups found inactivating genetic lesions in PRC2 components in T-lineage leukemia, including more than 40% of cases of pediatric early T-cell precursor ALL (ETP-ALL). ETP-ALL is a genetically diverse subtype of T-lineage acute lymphoblastic leukemia (T-ALL) that has been characterized in particular detail by Charles Mullighan and Jinghui Zhang.⁴ The clinical diagnosis relies on immunophenotyping. ETP-ALL was thought to be particularly difficult to treat, although recent outcomes data are somewhat more encouraging. A mouse model of ETP-ALL has been reported, but the role of PRC2 in ETP-ALL was not addressed in this model.⁵ Also, loss of Ezh2 was shown to lead to T-ALL in mice with several months latency, but leukemic cells with and without PRC2 compromise were not investigated side by side in this model.⁶

Since PRC2 has documented oncogenic and tumor suppressive roles we developed a mouse model of ETP-ALL to characterize otherwise isogenic leukemia cells in the presence and absence of a fully functional PRC2 complex.⁷ We transformed early murine hematopoietic stem cells/progenitor cells with oncogenic NRASQ61K in the context of a floxed locus for Ezh2 and constitutively inactivated Cdkn2a. Cdkn2a is an important silencing target of PRC2 and its inactivation is required for establishment of the murine model. Oncogenic Ras was chosen for transformation because direct or indirect Ras pathway activation frequently occurs in human ETP-ALL together with PRC2 alterations. NRASQ61K-transformed cells were then expanded on OP9DL1 feeder cell lines, which constitute an established model of T-lineage differentiation. After 14 d in culture, cells were transplanted into syngeneic sublethally irradiated recipient mice. All mice developed leukemia with co-expression of lymphoid (CD5, low levels of CD4/8) and myeloid (Kit, Gr1, Mac1) markers, thus recapitulating the immunophenotypic features of human ETP-ALL. Inactivation of Ezh2 and inactivation of the structural essential PRC2 component Eed both led to acceleration of leukemia development.

Transcriptional profiling of leukemic cells isolated from mice or preleukemic cells prior to injection revealed that Ezh2 inactivation accentuated transcriptional programs associated with hematopoietic stem cells and normal murine ETPs. One of the deregulated genes is HoxA9. Our studies confirmed the functional relevance of HoxA9 as forced co-expression of NRASQ61K and HOXA9 recapitulated the phenotype of Ezh2 inactivation. We also noted transcriptional enrichment of cytokines and their cognate receptors in the Ezh2-null cells. One example is the Il6ra receptor, which was found to be a direct PREC2 silencing target in our model and in normal mouse thymopoiesis. Il6ra was overexpressed in Ezh2-null leukemic cells and this correlated with hyperactive STAT3 phosphorylation in response to IL6. Finally, we found that both Ezh2-null cells and their Ezh2-floxed counterparts respond in vitro to Jak inhibition with the approved inhibitor ruxolitinib. This is relevant because ruxolitinib has been proposed as a therapy for ETP-ALL with mutations affecting IL7ra or Jak3.⁸ Our data suggest that ETP-ALL with PRC2 alterations will also potentially benefit from therapies targeting Jak-stat signaling.

Our data dovetail with other reports demonstrating co-occurrence of PRC2 alterations and hyperactive Ras-signaling in neurofibromatosis 1 associated malignant peripheral nerve sheath tumor (MPNST)⁹ and juvenile myelomonocytic leukemia (JMML).¹⁰ Accentuated Ras signaling as a result of PRC2 compromise has been reported by DeRaedt et al.⁹ and we found the same in our Ezh2-null ETP-ALL model. It will be important to further characterize the similarities and differences between the different systems. A more detailed understanding of the molecular wiring of Ras-driven malignancies will undoubtedly open new avenues toward more efficacious and less toxic therapies for these highly problematic cancers.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.