SHIP1 Is a Novel Tumor Suppressor in Late-Stage Myelodysplastic Syndromes and Is Silenced by Mir-210 and Mir-155.

Abstract 3824

Poster Board III-760

INTRODUCTION. Early stages of myelodysplastic syndromes (MDS) are characterized by rapid apoptosis, however, a loss of apoptosis occurs as the disease transforms into acute myeloid leukemia (AML). The molecular basis surrounding this transformation remains unknown. Activation of oncogenes and inactivation of tumor suppressor genes have not been well established in MDS. We hypothesized that aberrant activation of the PI 3'kinase (PI3'K)/Akt pathway underlies this transformation. PTEN is a well-established tumor suppressor gene (TSG) that acts as a lipid phosphatase to negatively regulate the PI 3'kinase (PI3'K)/Akt pathway. Another lipid phosphatase, SHIP1, also regulates the levels of phosphoinositides and is expressed specifically in hematopoietic cells. To date, SHIP1 has not been shown to be a TSG in human cancer. METHODS. We performed western blotting on CD3+/CD19+-depleted bone marrow mononuclear cells from patients with late stage MDS (blast count >20%). We performed immunohistochemistry on bone marrow sections from patients with various stages of MDS. We performed qPCR for SHIP, PTEN, and microRNAs on CD34+ cells from patients or normal controls. We performed luciferase reporter assays and evaluated the effects of transfection of pre-miR or antagomir to miR-210 on SHIP1 levels in cell lines. Lastly, we developed a deterministic model involving Monte Carlo simulations for determining the effects of miR expression on protein levels. RESULTS. There was constitutive activation of Akt, but not ERK1/2, in MDS enriched cells. As disease progressed, bone marrow specimens showed increasing amounts of phospho-Akt. Protein expression of PTEN was decreased, and there was an absence of SHIP1 (in contrast to de novo AML). This was confirmed by immunohistochemistry. Using a SHIP1 construct with a mutation in its phosphatase domain (SHIP1D673G), we demonstrated a reduction in phosphatase activity and an increase of phospho-Akt. Transient transfection of 293T cells, resulted in a greater proportion of apoptotic cells with wild-type SHIP1 as compared to the phosphatase-defective SHIP1 mutant. Transfection of primary AML cells with WT, but not the phosphatase-defective mutant, resulted in a decreased number of spontaneously growing colonies suggesting that SHIP1 can act as a tumor suppressor in human myeloid disease. Since transcripts for SHIP1 were modestly decreased in CD34+ cells from MDS patients, we postulated that the reduction in SHIP1 protein levels may be due instead to translational repression by a microRNA (miR). We found that the levels of miR-210 and miR-155, both predicted to bind to the 3' UTR of SHIP1, were significantly higher (2-3 fold) in CD34+ cells from late stage MDS patients as compared to control donors. miR-155 has been shown to translationally repress SHIP1 levels. To test whether miR-210 was able to repress translation of SHIP1, we studied cells expressing a luciferase reporter fused to the 3' UTR of SHIP1. We found a 30% reduction in luciferase activity in cells transfected with pre-miR-210, as compared with control miR, suggesting miR-210 acts to silence SHIP1 translation. Antagomir to miR-210 increased SHIP1 expression in THP-1 cells, and pre-miR for miR-210 decreased levels in U937 cells. Based on experimental and published data, we constructed a deterministic model for protein expression that predicts that miRs have a substantial impact on protein levels (v. transcription rates and protein degradation). The model also predicts that miRs have an additive effect on protein level, reducing the levels of SHIP-1 in MDS patients by up to 50%. CONCLUSIONS: There was absence of SHIP1 and decreased levels of PTEN in MDS enriched cells from late stage disease. Silencing of SHIP1 occurs in late stage MDS by elevated expression of mir-210 and mir-155. As predicted by the mouse knockout, SHIP1 behaves as a myeloid tumor suppressor in human cancer. Our studies suggest that loss of SHIP1 could serve as a clinical biomarker for disease progression and that targeting of microRNAs might serve as a novel target in MDS.