Genetic and Pharmacological Interruption of the SIX1/EYA2 Axis Impairs Leukemia Proliferation

Background: The CALM-AF10 translocation is found in 5-10% of T-cell acute lymphoblastic leukemias (T-ALL), and a subset of acute myeloid leukemias (AML). CALM-AF10 leukemias are characterized by elevated expression of proleukemic HOXA genes. Since HOXA genes are difficult to target, we sought to identify non- HOXA effector genes through RNA sequencing and Microarray analyses, and showed that SIX1 expression is increased in CALM-AF10 leukemia cells. Like HOXA genes, SIX1 is a homeobox gene involved in cell proliferation, embryogenesis, and apoptosis. SIX1 interacts with the cofactor Eyes Absent 2 (EYA2), a protein phosphatase, to transcriptionally activate developmental genes, the TGFβ pathway, and the epithelial-mesenchymal transition (EMT). SIX1 expression is elevated during development but decreases with maturation. Intriguingly, overexpression of SIX1 has been observed in numerous malignancies including breast cancer, ovarian cancer, gliomas and esophageal cancer, and SIX1 is involved in accelerating the epithelial mesenchymal transition (EMT) and metastasis. Knockdown of either SIX1 or EYA2 has been shown to reverse these effects, and two novel inhibitors - one of the SIX1/EYA2 complex (Compound 8430, “8430”) and one of the EYA2 phosphatase (Compound 9987, “9987”) - have replicated the effects of SIX1 or EYA2 knockdown.

Methods: SIX1 gene and protein expression were assessed in leukemia cell lines via RT-qPCR and Western Blot. SIX1 expression vectors were transduced into fetal liver hematopoietic stem cells (HSC) and immortalization was assessed using colony assays. CALM-AF10 leukemia cells lines were derived from irradiated mice transplanted with HSC retrovirally transfected with CALM-AF10 plasmids created in our lab. The established leukemia cell lines Jurkat, MOLT16, OCI-M2, TF-1, and SHI-1 were obtained from ATCC or DSMZ. shRNAs targeting SIX1 were lentivirally transduced into CALM-AF10 and SHI1leukemia cell lines. Cell-Titer-Glo assays and liquid culture were used to assess the effects of shRNAs, 8430, or 9987 on cell proliferation and cell growth. SIX1 gene expression in leukemias was determined from The Cancer Dependency Map from The Broad Institute using R.

Results: We first validated RNAseq and Microarray studies, showing that CALM-AF10 induces SIX1 expression in murine fibroblasts and murine HSC, and localizes to the SIX1 locus. Furthermore, SIX1 overexpression was sufficient to immortalize murine fetal liver HSC (FL-mHSCs), implying a role in leukemic transformation. Multiple murine derived CALM-AF10 cell lines displayed increased SIX1 expression compared to fresh murine bone marrow and isolated HSCs. shRNA knockdown of SIX1 decreased viability of murine CALM-AF10 cells in culture. In addition, both the SIX1/EYA2 complex inhibitor (8430) and the EYA2 inhibitor (9987) decreased CALM-AF10 leukemia cell proliferation in vitro.

To expand the relevance of SIX1 to other leukemias, we then queried The Cancer Dependency Map from the Broad Institute and identified other leukemia cell lines with increased SIX1 expression, including Jurkat and Molt16 (T-ALL), and SHI-1, OCIM2, and TF-1 (AML). SIX1 expression in these cells was validated by immunoblot and shRNA knockdown of SIX1 slowed the proliferation of SHI-1 cells. Finally, the 8430 and 9987 inhibitors reduced the proliferation of Jurkat, OCI-M2, and SHI-1 cells at doses similar to those used in CALM-AF10 leukemias.

Conclusions: We have identified involvement of the SIX1/EYA2 axis in several different leukemias, including T-ALL and AML. Furthermore, we have shown that inhibitors of the SIX1/EYA2 interaction and of EYA2 reduce the proliferation of SIX1 expressing leukemias, suggesting a potential new therapeutic target in these leukemias.