Evi1 recruits polycomb complexes to repress PTEN. (A) Reporter assays that used Jurkat cells and PTEN4_wt promoter in the presence of Evi1, G9a, or dominant-negative (DN) G9a (n = 4). Error bars indicate SD; *P < .05. (B) Reporter assays that used Jurkat cells and PTEN4_wt promoter in the presence of EVi1, SUV39H1, or DN-SUV39H1 (n = 4). Error bars indicate SD; *P < .05. (C) Reporter assays that used Jurkat cells and PTEN4_wt promoter in the presence of Evi1, EZH2, or EZH2-H689 (n = 6). Error bars indicate SD; *P < .05, **P < .01, and ***P < .001. (D-E) shRNAs (shEZH2-A, -B, -C, -D or control) were retrovirally delivered to Evi1-transduced BM cells, and the efficacy of these constructs and the effects on the PTEN/AKT/mTOR signaling were examined by quantitative real-time PCR (D; n = 3) and immunoblotting (E; experiments were performed twice and the representative figures are presented). BM cells (1-2 × 10⁶) were used for Western blotting. Error bars indicate SD; *P < .05. (F) ChIP assays for PTEN promoter region as shown in Figure 2F using primers 3 (Figure 1D) (n = 3). Error bars indicate SD; *P < .05, **P < .01, ***P < .001, and ****P < .0001. (G) ChIP assays for PTEN promoter region that used human AML blasts (n = 5). Primers used in these assays amplify sequences, including a putative Evi1 binding site (5′-AGAAGATAA-3′ fragment at position 1875/1884 downstream of the initiation codon ATG [+1] of human PTEN). P value was calculated by comparing variables of patients with low-intermediate Evi1 expression (n = 3) and patients with high Evi1 expression (n = 2).