Overexpressed MELK correlates with EZH2 expression in NKTL. (A) Gene expression profiling data showing MELK expression in NKTL cell lines and patient samples of NKTL (extranodal) and EBV⁺ peripheral T-cell lymphoma (nodal) cases. (B) Expression of MELK and EZH2 in normal NK and a panel of NKTL cell lines. Densitometry analysis was used to quantify average changes in 3 individual experiments. (C) IP showing MELK-EZH2 interaction. (D-E) EZH2 protein level change with MELK inhibitor OTSSP167 treatment of 48 hours in NKYS (D, densitometry analysis was used to quantify average changes in 3 individual experiments) and (E) NK-S1. (F-G) EZH2 protein level change with MELK knock-down using siRNA in (F) NKYS and (G) NK-S1. Cells were harvested for immunoblots 48 hours after knock-down. (H) Linear correlation was obtained by comparing percentage of EZH2 positive staining (EZH2⁺) to average percentage of cytoplasmic and nuclear MELK positive staining (MELK⁺) in NKTL cells (CD3⁺) for each core from NKTL patient tissues. N = 83 cores from 52 NKTL patients were stained (R = 0.4549, P < .0001). (I) Representative images indicating expression of EZH2 and MELK in NKTL tissue microarray sample (left) with corresponding segmented image masks (right). CD3 marks tumor cells. A total of 80.8% of CD3⁺ cells in sample 1 are positive for nuclear expression of EZH2, whereas 38.27% CD3⁺ cells in sample 2 are positive for EZH2. Seventy-three percent of CD3⁺ cells in sample 1 are positive for MELK and only 36.42% CD3⁺ cells are positive for MELK in sample 2. All immunoblots were performed in at least 3 individual experiments; representative images are shown. Ig, immunoglobulin