Bcl-2 phosphorylation confers resistance on chronic lymphocytic leukaemia cells to the BH3 mimetics ABT-737, ABT-263 and ABT-199 by impeding direct binding

Abstract

Background and purpose: Although the ongoing clinical trials of ABT-263 and ABT-199 in chronic lymphocytic leukaemia (CLL) have indicated that BH3 mimetics hold considerable promise, understanding the mechanism of CLL resistance to BH3 mimetics remains a challenge.

Experimental approach: The LD50 values of ABT-737, ABT-263 and ABT-199 in a number of primary CLL cells from 40 patients, were determined. The levels of Bcl-2 family proteins, including phosphorylated Bcl-2 (pBcl-2) and their interactions were measured by immunoblotting and co-immunoprecipitation. In vitro binding assays were performed by isothermal titration calorimetry and ELISA. BH3 profiling in isolated mitochondria was analysed.

Key results: The ratio of (Mcl-1 + pBcl-2) to Bcl-2 expression provided the most significant predictive marker for the cytotoxic potential of ABT-737, ABT-263 and ABT-199 in the panel of CLL samples. Mechanistically, pBcl-2 inhibited the effects of the ABT compounds on the displacement of Bax and Bim from Bcl-2, thereby suppressing mitochondrial apoptosis. The ABT compounds exhibited 100-300-fold lower binding affinity to the glutamic acid, phosphomimetic, mutant of Bcl-2 (T69E, S70E and S87E; EEE-Bcl-2). BH3 peptides exhibited different rank orders of binding affinities to full-length WT-Bcl-2 and full-length EEE-Bcl-2.

Conclusions and implications: Our study suggested that a structural alteration in the BH3-binding groove was induced by phosphorylation of Bcl-2. Our data also provided a framework to overcome resistance of CLL cells to the ABT compounds by combining pBcl-2 kinase inhibitors with the ABT compounds.

Figure 1

The ratio of (pBcl‐2 + Mcl‐1) to Bcl‐2 provided the most significant linear correlation with sensitivity to ABT‐737, ABT‐263 and ABT‐199 in primary CLL cells. (A) The total protein extracts (100 μg) from 40 different primary CLL samples were analysed by immunoblotting for Mcl‐1, Bcl‐xL, TBcl‐2, pBcl‐2 (Ser⁷⁰) and β‐actin expression. (B) The sensitivity (LD₅₀) of CLL cells, assessed by annexin V staining after 48 h of treatment with ABT‐737, ABT‐263 or ABT‐199, was plotted against the pBcl‐2/Bcl‐2, Mcl‐1/Bcl‐2 and (pBcl‐2 + Mcl‐1)/Bcl‐2 ratios. Relative protein quantification was carried out with kodak carestream molecular imaging software and normalized to β‐actin. Spearman's correlation (r) and P values are shown. Data shown are representative of five independent experiments.

Figure 2

Bcl‐2 phosphorylation inhibited the activities of ABT‐737, ABT‐263 and ABT‐199 in disrupting Bcl‐2 interactions, thereby suppressing mitochondrial apoptosis. (A) Hs 505.T cells were treated with gradient concentrations of ABT‐199 (0.01, 0.02 and 0.05 μM) for 12 h. Bcl‐2, Bcl‐xL, Mcl‐1 and pBcl‐2 (Ser⁷⁰) immunoprecipitations were performed, and the immunoprecipitated fractions were analysed by immunoblotting for the indicated proteins. Cytochrome c release and PARP cleavage were analysed in parallel. (B) CLL‐AAT cells were treated with ABT‐199 alone or together with 20 μM PD98059 for 12 h, followed by assays as described in panel A. Relative protein quantification in co‐immunoprecipitation (IP) of TBcl‐2 and pBcl‐2 was carried out with kodak carestream molecular imaging software and normalized to β‐actin. (C) Hs 505.T cells stably expressing AAA‐Bcl‐2 or EEE‐Bcl‐2 proteins were treated with the indicated concentrations of ABT‐737, ABT‐263 and ABT‐199 for 48 h, and viability was assessed by annexin V staining. The indicated protein levels were examined by immunoblotting. (D) Transfected Hs 505.T cells were treated with ABT‐737, ABT‐263 and ABT‐199 for 12 h, after which haemagglutinin (HA) immunoprecipitations were performed, and immunoprecipitated fractions were analysed by immunoblotting for the indicated proteins. Relative protein quantification in co‐IP of AAA‐Bcl‐2 and EEE‐Bcl‐2 was carried out with kodak carestream molecular imaging software and normalized to β‐actin. Data shown are representative of five independent experiments.

Figure 3

ABT‐737, ABT‐263 and ABT‐199 had a 100–300‐fold lower binding affinity to full‐length EEE‐Bcl‐2 than to full‐length WT‐Bcl‐2 and lower competition for Bax BH3 binding to full‐length EEE‐Bcl‐2 than to full‐length WT‐Bcl‐2. (A) Isothermal titration calorimetry of the Bcl‐2 isoform 2, full‐length WT‐Bcl‐2 and full‐length EEE‐Bcl‐2, with BH3 peptides of Bad, Bmf, Bax, Bim, Noxa and the ABT compounds. The upper panels display the raw titration data. Each peak corresponds to one injection. The K _d value was calculated from the integration of the raw data. (B) The competition for Bax BH3 binding to Bcl‐2 proteins was determined by an ELISA assay. Bcl‐2 isoform 2, full‐length WT‐Bcl‐2 and full‐length EEE‐Bcl‐2 were pre‐incubated with BH3 peptides or compounds for 1 h and then transferred to a biotin‐Bid/streptavidin complex, followed by absorbance measurement at 450 nm. The results are expressed as the mean ± SEM of five independent experiments.

Figure 4

BH3 profiling in mitochondria revealed an apoptosis block conferred by pBcl‐2 that could be overcome by Bmf BH3 but not by Bad BH3 or the ABT compounds, alone or in combination with Noxa BH3. (A) Mitochondria that were isolated from the Hs 505.T cells stably expressing AAA‐Bcl‐2 or EEE‐Bcl‐2 proteins were collected as described and incubated with a panel of BH3 peptides (100 μM), ABT‐737 (10, 50 and 100 μM), ABT‐263 (10, 50 and 100 μM) or ABT‐199 (1, 5 and 10 μM). The release of cytochrome c was determined by a comparison of cytochrome c in the pellet and supernatant as quantitated by ELISA. The results are shown as the percentage minus solvent control (DMSO) values and the mean (± SD) of triplicates. (B) Mitochondria from patient no. 4, no. 7 and no. 15 were treated with BH3 peptides (100 μM), ABT‐737 (100 μM), ABT‐263 (100 μM) and ABT‐199 (10 μM) alone or in combination, followed by assays as described in panel A. The expression of Bcl‐2 family proteins and complexes of pBcl‐2, Bcl‐2 and Mcl‐1 in isolated mitochondria were assayed by immunoblotting and co‐IP assays. Data shown are representative of five independent experiments.