ABT‐199 selectively inhibits BCL2 but not BCL2L1 and efficiently induces apoptosis of chronic lymphocytic leukaemic cells but not platelets
The anti-apoptotic BCL2 proteins comprising BCL2, BCL2L1 (also termed BCL-XL), BCL2L2 (BCLw) and MCL1 prevent apoptosis by inhibiting the release of mitochondrial cytochrome c. Their main function is to inhibit the activation of BAX and BAK1 (BAK), which once activated mediate the release of cytochrome c from mitochondria and initiate apoptosis (Cory & Adams, 2002). Like many other lymphoid malignancies, chronic lymphocytic leukaemia (CLL) cells are characterized by a high BCL2 expression. Thus, inactivation of BCL2, using small molecule inhibitors, may be a promising strategy for the development of targeted therapies. Consequently, several small molecule BCL2-inhibitors have been developed, amongst them ABT-737 and ABT-263, which are specific antagonists of BCL2 and the related anti-apoptotic proteins BCL2L1 and BCL2L2 (Oltersdorf et al, 2005; Park et al, 2008). While ABT-737 has only been used for preclinical studies, ABT-263 has entered clinical trials for leukaemia and lymphoma. Results from these trials indicate some single agent activity of ABT-263 in CLL but, compared with the remarkable susceptibility of isolated primary CLL cells in vitro, the results from these clinical trials have been rather disappointing (Roberts et al, 2012). The major dose-limiting toxicity of ABT-263 has been thrombocytopenia, which has been explained by platelet survival being dependent on BCL2L1 levels (Mason et al, 2007). Therefore, in order to achieve higher drug doses without unwanted toxicities a selective BCL2-antagonist that does not inhibit BCL2L1 may be favourable for the treatment of CLL. The first specific inhibitor (ABT-199), which selectively inactivates BCL2 but not BCL2L1, has recently been published (Souers et al, 2013). However, many supposed specific inhibitors of BCL2 proteins are promiscuous compounds and induce cell death independently of BCL2 proteins. In order to assess the specificity of ABT-199 (Fig 1A), we measured apoptosis in BAK1-deficient or reconstituted Jurkat cells. Given that Jurkat cells also lack expression of BAX, any specific inhibitor of BCL2 proteins should not induce apoptosis in BAK1-deficient cells. While ABT-199 caused a concentration-dependent induction of apoptosis in Jurkat cells reconstituted with BAK1, apoptosis was significantly inhibited in the absence of BAK1 and BAX, indicating that ABT-199 requires either BAX or BAK1 for apoptosis induction and specifically targets BCL2 proteins (Fig 1B). To investigate which anti-apoptotic BCL2 proteins were targeted by ABT-199, we used FL5.12 cells overexpressing either BCL2 or BCL2L1. Unless protected by overexpression of anti-apoptotic proteins, FL5.12 cells die upon withdrawal of interleukin-3 (Gottschalk et al, 1996). Therefore, they represent an ideal system to investigate whether BCL2 or BCL2L1 are inhibited by novel compounds targeting the BCL2 family. Upon interleukin-3 withdrawal, ABT-263 efficiently induces apoptosis in both BCL2- and BCL2L1-dependent FL5.12 cells, confirming that this compound targets both these anti-apoptotic proteins (Fig 1C–D). However, although ABT-199 induced apoptosis in BCL2-dependent FL5.12 cells more efficiently than ABT-263, ABT-199-induced apoptosis was significantly reduced in FL5.12 cells overexpressing BCL2L1. These data confirm that in cellular systems ABT-199 is a potent inhibitor of BCL2 but not BCL2L1. Some tumour cells, including CLL cells, are highly dependent on BCL2 expression for survival (Del Gaizo Moore et al, 2007). CLL therefore represents an ideal tumour to investigate the clinical potential of BCL2 inhibitors. After only 4 h of exposure, ABT-737, ABT-263 and ABT-199 all potently induced apoptosis in primary CLL cells with median effective concentrations (EC50s) of 12·4, 74·1 and 7·6 nM, respectively (Fig 2A). ABT-199 induced activation of caspases-9 (CASP9) and -3 (CASP3), as well as cleavage of the caspase substrate, poly(ADP)ribose-polymerase, consistent with activation of the intrinsic apoptotic pathway (Fig 2B). ABT-199 also induced striking morphological changes including chromatin condensation, mitochondrial swelling and rupture of the outer mitochondrial membrane (Fig 2C). These data are consistent with ABT-199 being a potent inducer of apoptosis in CLL cells by a novel paradigm involving permeabilization of the outer mitochondrial membrane, as described for both ABT-737 and ABT-263 (Vogler et al, 2008, 2010). The clinical potential of ABT-263 has been limited by dose-limiting thrombocytopenia due to the dependence of platelets on BCL2L1 for survival (Mason et al, 2007; Roberts et al, 2012). To investigate whether platelets are susceptible to ABT-199, isolated platelets from healthy volunteers were exposed to ABT-199. Notably, ABT-199 was much less toxic to platelets than either ABT-737 or ABT-263, with EC50s of 2, 0·2 or 0·009 μM, respectively (Fig 2D). To extend our studies from isolated cells to more physiologically relevant culture models, we cultured whole blood from CLL patients with ABT-199. Direct comparison of cell death in the platelets versus the CLL cells in samples from the same patient showed that ABT-199 selectively affects CLL cells without affecting platelets, in contrast to ABT-263 (Fig 2E). Higher concentrations of ABT-199 were required to kill CLL cells in whole blood than in isolated CLL cells cultured in RPMI (Fig 2A, E), probably due to binding to albumin, as observed for both ABT-263 and ABT-737 (Vogler et al, 2010). These higher concentrations are similar to those obtained following dosing of patients with ABT-263 (Roberts et al, 2012) and it is anticipated that they should be readily achievable following dosing with ABT-199. In summary, ABT-199 selectively inhibits BCL2 and not BCL2L1 and potently induces apoptosis in CLL cells but requires ~260-fold higher concentrations to induce platelet apoptosis. Our study indicates that the increased selectivity of ABT-199 should overcome the dose-limiting thrombocytopenia observed with ABT-263 and be highly efficacious in killing circulating CLL cells, in agreement with an initial study in three patients with CLL (Souers et al, 2013). Caution will be required in the clinical use of ABT-199 due to its high potency in killing CLL cells, which could result in marked tumour lysis and subsequent toxicity. However, the susceptibility of CLL cells in protected environments, such as lymph nodes, is not known. Given that BCL2L1 may be upregulated in certain microenvironments in the lymph nodes, the efficiency of ABT-199 may be limited and residual resistant cells may survive due to high expression of BCL2L1. To realise their full clinical potential, specific BCL2-inhibitors may require combination with agents, such as Btk or phosphatidylinositide 3 kinase δ inhibitors to mobilize these protected cells into the circulation, where they may be killed by ABT-199. We thank Drs. A. Letai, L. Boise and K. Schulze-Osthoff for cells and T. Smith and J. McWilliam for preparation of EM samples. MV. designed the study, performed experiments and wrote the manuscript, DD performed the electron microscopy experiments, MJSD and GMC designed the study and analysed the data.