EMP1, a Novel Poor Prognostic Factor In Pediatric Leukemia Regulates Prednisolone Resistance, Cellular Proliferation, Migration and Adhesion

20% of pediatric precursor B-acute lymphoblastic leukemia (BCP-ALL) and 25% of pediatric T-ALL patients relapse on or after current treatment strategies. Treatment failure has been associated with cellular resistance to prednisolone, the leading drug in multi-drug treatment of ALL. This study aimed to find new druggable targets that may modulate prednisolone resistance.

Leukemic cells from children with newly diagnosed ALL were isolated from bone marrow aspirates, only samples with ≥ 90% leukemic blasts were used. RNA of 178 leukemic patients were hybridized to Affymetrix U133 Plus 2.0 gene expression microarrays and limma analysis was performed. EMP1 knockdown was achieved by lentiviral shEMP1 spin-occulation in four BCP-ALL cell lines and two T-ALL cell lines. Knockdown was confirmed by qRT-PCR and Western blot. Cell proliferation, cell cycle, and cell survival was analyzed by an AnnexinV/PI-staining and by a BrdU/7-AAD-staining. Response to prednisolone was evaluated in an in vitro3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS) assay. Irradiated human mesenchymal stromal cells (hMSC-TERT) were co-cultured overnight with maroon-labeled leukemic cells, whereafter non-adherent and adherent leukemic cells were counted with the MACSQuant. Maroon-labeled leukemic cells were also inserted in the upper compartment of a 3 µM pore-size transwell system and were allowed to migrate overnight to hMSC-TERT. Migrated leukemic cells were evaluated by trypan blue and the MACSQuant. To assess MSC-mediated prednisolone resistance, leukemic cells alone or in combination with irradiated hMSC-TERT were cultured with or without prednisolone and cell survival was assessed after 72 hours. Cell signaling after shEMP1 knockdown was studied using Milliplex 8-plex Human Src Family kinase kit and the 9-plex Multi-Pathway kit with GAPDH and β-tubuline loading control beads, following the manufacturers’ protocol. Cumulative incidence of relapse was calculated as time to relapse with death as a competing event in R using the method of Fine and Gray. Multivariate analysis of event free survival were performed in IBM SPSS Statistics 20 using Cox regression.

Gene expression microarray data from 178 primary pediatric ALL patients showed a 3.6-fold increase in epithelial membrane protein 1 (EMP1) expression between in vitro prednisolone-resistant and -sensitive patients (p=0.002). This difference was confirmed with qRT-PCR (p<0.0001). EMP1 is a small hydrophobic four-transmembrane glycoprotein of which the biological function is largely unknown. We show that lentiviral silencing of EMP1 in six different BCP-ALL and T-ALL cell lines increased apoptosis and induced G0/G1 and G2/M arrest leading to 84.1 ± 4.5% reduction in leukemic survival compared to non-silencing control cells (shNSC , p=0.014). Moreover, EMP1 silencing sensitized three out of six cell lines to prednisolone up to 8.8-fold (p<0.001) compared to shNSC. Silencing of EMP1 also abrogated migration and adhesion to mesenchymal stromal cells (MSCs) by 78.3±9.0% and 29.3±4.1% respectively compared to shNSC (p<0.002). Notably, we also discovered that EMP1 contributes to MSC-mediated prednisolone resistance, as prednisolone resistance triggered during leukemic MSC coculture was partially reversed after EMP1 knockdown. Pathway analysis indicated that EMP1 signals through Src kinase family phosphorylation to activate JNK, STAT3, STAT5, CREB and NF-κB. Finally, BCP-ALL patients with high EMP1 mRNA expression showed a higher cumulative incidence of relapse and non-response (p=0.02) and a poorer 5-year event-free survival compared to EMP1-low patients (82%±2 vs. 91%±9 ; p<0.01). Most importantly, Cox regression multivariate analysis with white blood cell count, age, and EMP1 expression as co-variates identified EMP1 as an independent predictor for poor outcome in BCP-ALL (p<0.01, hazard ratio: 2.36 (1.31-4.25) ).

Our results indicate that EMP1 is a novel poor prognostic factor in pediatric leukemia that regulates prednisolone resistance, cell survival, migration and adhesion. With this study we provide preclinical evidence that EMP1 is an interesting new druggable target to treat pediatric leukemia and sensitize to prednisolone.

No relevant conflicts of interest to declare.