Abstract

Cell death pathways are desired targets of small molecule inhibitors since their deregulation plays an important role in chemotherapy resistance. Obatoclax binds to the BH3 pocket of anti-apoptotic BCL-2 family proteins, inhibiting their interactions with pro-apoptotic BCL-2 family members. Before we found potent obatoclax activity in MLL/AF4+ cell lines and 6 MLL rearranged (MLL+) leukemias. In apoptosis assays of the cell lines obatoclax increases TUNEL staining but minimally activates caspase 3 (Rege ASH 2005; Zhang AACR 2007). In this study, we tested the cytotoxicity of obatoclax in additional MLL+ leukemias and investigated its mechanism of action with a focus on autophagy (ATG), since ATG proteins such as beclin 1 also interact with anti-apoptotic BCL-2 family members.

Methods:MLL partner genes were determined by molecular/cytogenetic methods. 17 primary MLL+ leukemias including the original 6 (12 ALL/11 infants, 1 child; 3 AML/1 each infant, child, adolescent; 2 bilineal/2 infants) were tested in MTT assays after 72 h obatoclax exposures. MTT assays on cytotoxic drug-obatoclax combinations were performed in a primary MLL/AF4+ ALL and interactions were studied by response surface modeling. MCL-1/BAK complex inhibition was tested in this ALL by co-IP/immunoblot analysis. Cell death and ATG were studied in obatoclax treated RS4:11 and/or SEM-K2 cells by PI flow, LC3 and p62 Western blot analysis and EM, using doxorubicin as a control for apoptosis. Gene expression changes after vehicle treatment vs. obatoclax treatment at the IC50 and IC90 for 6 h were studied using Affymetrix HG_U133 Plus2.0 arrays. Differentially expressed genes overall and those specifically associated with ATG were queried by ANOVA (p<0.01, ≥50% change in mean expression considered as significant). Q-RT PCR analysis of basal expression levels of select ATG genes (BECN1, WIPI1, MAP1LC3B) was performed in 10 of the 17 primary cases and correlations with the IC50 values were determined using Pearson correlation coefficients and their levels of significance. Changes in the expression patterns of these genes after vehicle treatment vs. treatment with obatoclax at the IC50 for 6 h and 48 h were compared in 6 cases by Q-RT PCR and cluster analysis.

Results:MLL partner genes were AF4, ENL and other in 6, 5 and 1 ALL, respectively; AF9 in 2 and AF6 in 1 AML; and AF4 and ENL in 1 each bilineal leukemia. The single agent IC50’s of obatoclax suggested greater sensitivity in ALL (13–834 nM; median 104 nM) than AML (243–488nM; median 341 nM), and were 79 nM and 508 nM in the bilineal leukemias. In addition to synergy with ARAC, ADR, VP16 and DEX (Zhang AACR 2007) there was synergy with LASP and VCR in the primary MLL/AF4+ ALL. In the same ALL obatoclax decreased MCL-1/BAK dimers, suggesting interaction with the MCL-1 target, and obatoclax increased high molecular weight MCL-1/BAK complexes and decreased MCL-1, the latter of which would also decrease MCL-1/BAK dimers. Obatoclax treatment of RS4:11 and SEM-K2 cells increased PI staining and LC-3I to LC-3II conversion; EM analysis of SEM-K2 cells revealed phagophores, autophagosomes and autophagolysosomes indicative of ATG induction. Lack of p62 accumulation showed that ATG was not blocked. EM findings of apoptosis occurred in SEM-K2 cells after doxorubicin exposure. ATG was not the most affected pathway in microarray analyses of obatoclax treated SEM-K2 and RS4:11 cells, but specific analysis of ATG-related genes showed WIPI1 and MAP1LC3B upregulation. Moreover, the basal expression of BECN1 was positively correlated with obatoclax activity in the 10 primary cases (r =0.659; p=0.038), which is consistent with importance of this pathway in the drug response. ATG gene expression analysis in obatoclax treated primary MLL+ cases identified 2 patient clusters; in one cluster obatoclax decreased BECN1 and increased WIPI1 expression; in the other ATG gene expression changes were more variable.

Conclusions: Obatoclax induces cell death in MLL+ leukemias via the ATG pathway even though they are apoptosis competent. This is distinct from the apoptosis activation in other cancer cell types and indicates that the targets of obatoclax are disease-specific. The activity in a broad spectrum of MLL+ leukemias indicates that obatoclax is a promising molecularly targeted agent for this population.

Disclosures: No relevant conflicts of interest to declare.

Author notes

Corresponding author