Synergistic induction of cell death in haematological malignancies by combined phosphoinositide-3-kinase and BET bromodomain inhibition.

Abstract

There has been a marked increase in the development of targeted therapies for the treatment of malignant disease. Responses may be compromised by the presence of pre-existing or treatment-induced resistance mechanisms. The phosphoinositide-3-kinase (PI3K) pathway has been a major focus of drug development due to its function as a key regulator of cell growth and survival. PI3Ks may be dysregulated by mechanisms including mutations of the negative regulator phosphatase and tensin homologue (PTEN), of regulatory and catalytic subunits and of upstream regulators, such as tyrosine kinases and the RAS family (Fruman & Rommel, 2014). Key targets of PI3K include the serine/threonine kinase AKT and the mechanistic target of rapamycin (MTOR) pathway. Drugs targeting the PI3K/AKT/MTOR axis include pan-PI3K inhibitors, isoform-selective PI3K inhibitors, rapamycin and analogues, active-site MTOR inhibitors, dual-PI3K-MTOR and AKT inhibitors.(Fruman & Rommel, 2014) Early clinical trials with these agents have shown varying activity with incomplete understanding for the lack of response seen in some settings.( Klempner et al, 2013; Fruman & Rommel, 2014) A large body of work shows that PI3K and v-myc avian myelocytomatosis viral oncogene homolog (MYC) form interlinked but overlapping signalling pathways.(Dang, 2012) Both pathways are often dysregulated in acute leukaemias, lymphomas and myeloma. We and others have shown that MYC upregulation can impair the response to PI3K inhibitors and may be a frequent mechanism underlying resistance to this class of drugs.(Klempner et al, 2013; Shepherd et al, 2013) Until recently, it has been difficult to target MYC using pharmacological agents, but this has changed with the identification and development of bromodomain and extra terminal domain (BET) inhibitors, such as JQ1 and I-BET762.(Dawson et al, 2012) BETs, such as bromodomain containing 4 (BRD4), function as chromatin readers to alter transcription – inhibition leads to the loss of MYC expression with activity in several haematological tumours (Delmore et al, 2011; Zuber et al, 2011). Inhibition of BRD4 and PI3K is individually associated with consistent reductions in cancer cell proliferation, but with variable effects on cell death.(Delmore et al, 2011; Zuber et al, 2011; Fruman & Rommel, 2014) We postulated that the combination of PI3K and BRD4 inhibition may show enhanced cytotoxicity. As induction of cell death in ex vivo assays, rather than of cytostasis, is a key predictor of clinical activity, (Faber et al, 2012) we focussed on this parameter. We screened a panel of acute myeloid leukaemia (AML) cell lines with the BRD4 inhibitor JQ1, the dual PI3K/MTOR inhibitor BEZ235 or a combination of these to investigate effects on cell survival. Figure 1A shows enhanced killing with the combination of BEZ235 and JQ1 compared with either agent alone. The combination was effective in cells with differing cytogenetic and molecular abnormalities (KMT2A (MLL) translocated – ML2, MOLM13, MV4;11; FLT3 mutated – MOLM13, MV4;11, PL21; KIT mutated – HMC1.2; CBL mutated – GDM1). Formal combination effect analysis showed synergy in the majority of cell lines (Fig 1B). Next, we extended the analysis to cells of different haematopoietic cell origin including T-cell acute lymphoblastic leukaemia (T-ALL), myeloma and Burkitt lymphoma. We found similar effects to those seen in AML, with dual PI3K/MTOR and BRD4 inhibition leading to synergistically increased levels of cell death. (Fig 1B, C) We confirmed our results using an alternative BRD4 inhibitor, I-BET151, and an alternative PI3K/MTOR inhibitor, PI103 (data not shown). As data from cell lines may not be representative of results in primary tumours, we carried out dose ranging experiments in six separate primary AML samples. Figure 1D shows that the combination of JQ1 and BEZ235 is more potent than the either agent alone. This is statistically significant at higher concentrations – e.g., survival is 47 ± 10%, 66 ± 8% and 24 ± 7% of control for JQ1 (2 μmol/l), BEZ235 (1 μmol/l) and the combination respectively (P = 0·01 by one-way analysis of variance). As BEZ235 inhibits both PI3K and MTOR, we wanted to assess which of these components may be responsible for the synergistic effects seen. We compared the effect of BEZ235 with that of a PI3K inhibitor with minimal direct anti-MTOR activity (ZSTK474), and with an active-site MTOR inhibitor with no anti-PI3K activity (WYE354). Figure 1E shows that, although each inhibitor has activity in conjunction with JQ1, neither agent can fully replicate the effect of BEZ235, indicating that both anti-PI3K and anti-MTOR activities are required. To assess the role of MYC, we ectopically expressed it in SUP-T1 cells and repeated the cytotoxicity experiment. (Fig 1F) MYC only partly rescued the combined effects of BEZ235/JQ1, indicating that other pathways regulated by BRD4 are also involved in this process. Recent results indicate that BRD4 can mediate the expression of a number of oncogenes, regulated by so-called super enhancers (Loven et al, 2013). To analyse the potential pathways involved in mediating cytotoxicity, we carried out Western blotting for relevant downstream targets. In the majority of the cell lines, incubation with JQ1 eliminated the expression of MYC. As expected, BEZ235 inhibited phosphorylation of PI3K and MTOR targets, AKT, eukaryotic translation initiation factor 4E binding protein 1 (EIF4EBP1) and ribosomal protein S6 (RPS6). (Fig 2) The effects of single or combined inhibition on key apoptosis regulators showed variable effects. Some cell lines, such as RPMI8402 and MM1S, upregulated BCL2-like 11 (apoptosis facilitator) (BCL2L11, also known as BIM), whereas it was unaffected in others (HPBALL, KMS28BM). (Fig 2) Effects on BCL2-like 1 (BCL2L1, also known as BCL-X) were also variable. However, the majority of cell lines showed significant effects on the anti-apoptotic protein myeloid cell leukemia 1 (MCL1). In AML and T-ALL cells, full length MCL1 was downregulated, especially in response to combined JQ1 plus BEZ235. (Fig 2A, B and D) In myeloma cells, consistent upregulation of the short isoform of MCL1, which has pro-apoptotic activity antagonistic to full-length MCL1, was seen. In conclusion, we have shown that the combination of PI3K/MTOR and BRD4 inhibition can synergize across a range of haematological malignancies to induce high levels of cell death. Both classes of agents are in clinical trials and these results provide a rational basis for testing this combination. AK was the principal investigator and takes primary responsibility for the paper. ST, KM and CS performed laboratory work. AK wrote the letter, which was reviewed and revised by all the authors. This work was undertaken at UCLH/UCL who received a proportion of funding from the NIHR Biomedical Research Centres funding scheme of the UK Department of Health. ST was supported by a UCL Biomedicine Grand Challenge Studentship and CS by the Kay Kendall Leukaemia Fund. Research in the AK laboratory is supported by Leukaemia and Lymphoma Research. All authors report no conflict of interest.