BCL-2 is dispensable for thrombopoiesis and platelet survival

Navitoclax (ABT-263), an inhibitor of the pro-survival BCL-2 family proteins BCL-2, BCL-XL and BCL-W, has shown clinical efficacy in certain BCL-2-dependent haematological cancers, but causes dose-limiting thrombocytopaenia. The latter effect is caused by Navitoclax directly inducing the apoptotic death of platelets, which are dependent on BCL-XL for survival. Recently, ABT-199, a selective BCL-2 antagonist, was developed. It has shown promising anti-leukaemia activity in patients whilst sparing platelets, suggesting that the megakaryocyte lineage does not require BCL-2. In order to elucidate the role of BCL-2 in megakaryocyte and platelet survival, we generated mice with a lineage-specific deletion of Bcl2, alone or in combination with loss of Mcl1 or Bclx. Platelet production and platelet survival were analysed. Additionally, we made use of BH3 mimetics that selectively inhibit BCL-2 or BCL-XL. We show that the deletion of BCL-2, on its own or in concert with MCL-1, does not affect platelet production or platelet lifespan. Thrombocytopaenia in Bclx-deficient mice was not affected by additional genetic loss or pharmacological inhibition of BCL-2. Thus, BCL-2 is dispensable for thrombopoiesis and platelet survival in mice.

Platelets are anucleate blood cells that play essential roles in haemostasis, wound healing and a range of other processes, including inflammation and immunity. 1 They are produced by megakaryocytes, large polyploid cells that develop primarily in the bone marrow, spleen and foetal liver. 2 Recent work has demonstrated that the survival of megakaryocytes and platelets is governed by the BCL-2 family proteins. 3 Both cell types possess a classical BAK/BAX-mediated intrinsic apoptosis pathway that must be restrained in order for them to develop and survive.
In platelets, BCL-X L is the critical pro-survival BCL-2 family member required to keep BAK and BAX in check. The first evidence of this came from Wagner et al., 4 who reported severe thrombocytopaenia in mice after MMTV-Cre-mediated deletion of Bclx in the haematopoietic system, skin and various secretory tissues. It has since been shown that megakaryocyte-restricted deletion of Bclx in mice reduces platelet lifespan from~5 days to~5 h, with a concomitant decrease in platelet counts to~2% of wild-type levels. 5,6 Pharmacological inhibition of BCL-X L with the BH3 mimetics ABT-737 7 or Navitoclax (ABT-263) 8 (which both also inhibit BCL-2 and BCL-W) triggers BAK/BAX-mediated platelet apoptosis. [9][10][11] As a result, these drugs cause dosedependent thrombocytopaenia in mice, dogs and humans. 9,[11][12][13][14] Indeed, thrombocytopaenia is the doselimiting toxicity for Navitoclax. [12][13][14] This fact provided additional impetus for the development of agents that specifically target BCL-2, beginning with ABT-199, 15 a BCL-2-selective antagonist currently in clinical trials for the treatment of a range of haematological malignancies including chronic lymphocytic leukaemia, non-Hodgkin's lymphoma, follicular lymphoma, mantle cell lymphoma, multiple myeloma and acute myeloid leukaemia. ABT-199 has already shown very promising antitumour activity, with little to no impact on platelet counts. 15,16 These data suggest that BCL-2 is dispensable for the development and survival of platelets.
In megakaryocytes, BCL-X L is also critical for survival. Although not absolutely required for their growth and maturation, BCL-X L is essential for megakaryocytes to proceed safely through pro-platelet formation and platelet shedding. 5 In addition to BCL-X L , megakaryocytes also depend on the pro-survival activity of MCL-1. Conditional deletion of Mcl1 alone has no effect on this lineage. In contrast, combined megakaryocyte-specific loss of Bclx and Mcl1 results in the failure of megakaryopoiesis, systemic haemorrhage and embryonic lethality. 5,17,18 These defects are rescued by deletion of Bak and Bax. 18 Consistent with the genetic studies, administration of ABT-737 to Mcl1 Pf4Δ/Pf4Δ mice, which lack MCL-1 in megakaryocytes and platelets, induces acute, fulminant BAK/BAXdependent megakaryocyte apoptosis. Given that, in addition to BCL-X L , ABT-737 also targets BCL-2, 7 these data suggested that BCL-2 might also contribute to the development and survival of the megakaryocyte lineage. This is supported by recent studies demonstrating that neonatal human platelets contain increased levels of BCL-2 relative to adult counterparts, 19 and that platelet lifespan is extended in transgenic mice expressing BCL-2 under the control of the pan-haematopoietic Vav promoter. 20 In light of these observations, and intense ongoing activity surrounding the development of novel BH3 mimetics, 21 we set out to elucidate the role of BCL-2 in megakaryocytes and platelets. Mice with a megakaryocyte-specific deletion of Bcl2, either alone or in combination with deletion of Mcl1 or Bclx, were generated. The effect of these mutations, and of BCL-2 or BCL-X Lselective BH3 mimetics, on the megakaryocyte lineage was assessed.

Results
Platelet production and platelet lifespan are normal in the absence of BCL-2. Mice lacking BCL-2 in the megakaryocytic lineage were generated by crossing animals carrying a floxed allele of Bcl2 22 with Pf4-Cre transgenic animals. 23 Bcl2 Pf4Δ/Pf4Δ mice were born at the expected Mendelian ratios, and were outwardly healthy. Deletion of BCL-2 in bone marrow-derived megakaryocytes and washed platelets was confirmed by western blotting (Figure 1a). Peripheral blood platelet counts ( Figure 1b) and platelet survival ( Figure 1c) in adult Bcl2 Pf4Δ/Pf4Δ mice were comparable with control animals. Megakaryocyte numbers and ploidy in bone marrow of Bcl2 Pf4Δ/Pf4Δ mice were normal (Figures 1d and e). Additionally, we assessed platelet and megakaryocyte counts in young (1-5-week old) mice with a constitutive deletion of Bcl2. 24 Despite their various phenotypic abnormalities, including kidney polycystic disease, growth retardation and lymphopaenia, blood platelet counts ( Figure 1f) and bone marrow and spleen megakaryocyte numbers (Figure 1g) in Bcl2 −/− mice were comparable with those of wild-type controls. Together, these results indicated that BCL-2 is dispensable for steady state platelet production. To establish whether this is also the case under conditions of stress, we induced transient thrombocytopaenia by injecting anti-platelet serum. This typically leads to platelet depletion in wild-type mice within 24 h, followed by recovery and rebound thrombocytosis at~5 days post injection. Bcl2 Pf4Δ/Pf4Δ and Bcl2 fl/fl mice responded similarly to anti-platelet serum treatment (Figure 1h), indicating that even under conditions of stress thrombopoiesis, BCL-2 is dispensable for the development and survival of megakaryocytes and platelets.
Combined loss of BCL-2 and MCL-1 does not affect platelet production or platelet survival. We and others have previously shown that platelet production and platelet counts are normal in Mcl1 Pf4Δ/Pf4Δ mice, whereas combined deletion of Mcl1 and Bclx in megakaryocytes results in haemorrhage and embryonic lethality. 17,18 To examine any potential functional redundancy between BCL-2 and MCL-1 in the megakaryocyte lineage, we conditionally deleted both of the genes encoding these proteins. We began by  megakaryocyte lineage in vivo. Bcl2 Pf4Δ/Pf4Δ Mcl1 Pf4Δ/Pf4Δ double knockout and control littermates were treated with a single dose of the BCL-X L -selective antagonist A-1155463.7 (A-463, 5 mg/kg). In Bcl2 fl/fl Mcl1 fl/fl mice, A-463 induced rapid thrombocytopaenia, but had no effect on bone marrow megakaryocyte numbers at 2 and 24 h post injection (Figures 3a, b  and d). In contrast, in Bcl2 Pf4Δ/Pf4Δ Mcl1 Pf4Δ/Pf4Δ animals, acute thrombocytopaenia was observed (Figure 3a). Additionally, apoptotic megakaryocytes with pyknotic nuclei were apparent in the bone marrow and spleen 2 h post administration (Figures 3c and d). Megakaryocyte numbers were reduced, and consistent with this, platelet counts were lower in A-463-treated Bcl2 Pf4Δ/Pf4Δ Mcl1 Pf4Δ/Pf4Δ mice compared with floxed control animals 24 h post injection (Figures 3a and b). These results prompted us to explore functional redundancy between BCL-2 and BCL-X L in the control of cell survival in megakaryocytes and platelets.
Pharmacological inhibition of BCL-X L , but not BCL-2, induces megakaryocyte apoptosis. The results from our in vivo studies suggested that genetic or pharmacological antagonism of BCL-2 has no adverse effect on megakaryopoiesis. To test this directly in isolated primary cells, we treated foetal liver or bone marrow-derived mouse megakaryocytes with ABT-737, ABT-199 or A-463. It has been previously reported that in cultured primary mouse megakaryocytes, deletion of Bcl-x or treatment with ABT-737 triggers loss of cell viability, Caspase-3/7 activity and a failure of proplatelet formation. 5 In line with these data, we found that the BCL-X L -selective inhibitor A-463 induced dosedependent Caspase-3/7 activation in wild-type foetal liverderived megakaryocytes (Figure 5a). Consistent with its increased potency against BCL-X L, 27 A-463 treatment at a concentration of 2.5 μM induced more caspase activity in megakaryocytes than ABT-737 at 5 μM. In contrast, treatment with the BCL-2 antagonist ABT-199 (2.5 μM) had no effect. Combination treatment with A-463 and ABT-199 did not amplify Caspase-3/7 activity beyond that seen with A-463 treatment alone. Similarly, loss of Bcl2 did not render bone

Discussion
In this study, we analysed platelet production in mice with a megakaryocytic lineage-specific deletion of Bcl2 alone or in combination with deletion of Mcl1 or Bclx. In addition, selective BH3 mimetics inhibiting BCL-2 or BCL-X L were assessed. Our genetic and pharmacological studies demonstrate that BCL-2 is dispensable for platelet production both at steady state and under conditions of stress. Moreover, platelet survival in vivo was not affected by genetic loss of BCL-2 or its pharmacological inhibition. This aligns with initial reports on the effects of ABT-199 in patients, 15   the importance of BCL-X L and the ancillary role of MCL-1 in maintaining megakaryocyte viability and BCL-X L in sustaining platelet survival. 5 Our data indicate that BCL-2 is dispensable for platelet survival in adults, and in neonates as well. The latter findings are consistent with a recent study from Liu and colleagues, 19 which demonstrated that although neonatal platelets display elevated BCL-2 protein levels and extended survival compared with adult platelets, BCL-2 was not the primary molecule facilitating this effect. We have further demonstrated that conditional deletion of Bcl2, on its own or in concert with Mcl1, had no impact on neonatal platelet counts.
BCL-2 has recently been linked to myeloid progenitor cell survival. 28 As Cre-dependent recombination utilising the Pf4-Cre transgenic mouse is restricted to megakaryocytes, platelets 23 and a small fraction of late megakaryocyte progenitors, 29 we could not use this model to address the role of BCL-2 in the earliest stages of megakaryopoiesis. However, mice carrying the constitutive deletion of Bcl2 did afford this opportunity. Germline loss of BCL-2 leads to polycystic kidney disease, lymphopaenia, grey fur (due to premature death of melanocytes), growth retardation and early mortality. 24,30,31 Even with the complex co-morbidities of this model, our results show that young Bcl2 −/− and control mice exhibited similar megakaryocyte and platelet counts, indicating that loss of BCL-2 in progenitor cells does not significantly affect the megakaryocytic lineage at steady state in young animals. Although it is increasingly clear that ABT-199 used as a single agent within the therapeutic window will not kill platelets and megakaryocytes, potential effects on progenitor cells with subsequent thrombocytopaenia may become apparent when used in combination with certain chemotherapeutics. This is pertinent as clinical trials are currently evaluating the safety and efficacy of ABT-199 in combination with proteasome inhibitors in multiple myeloma, 32 hypomethylating agents in acute myeloid leukaemia, 33 and alkylating or antimitotic drugs in non-Hodgkin's lymphoma, chronic lymphocytic leukaemia and follicular lymphoma 34,35 known to cause thrombocytopaenia through bone marrow suppression or inhibition of proplatelet formation. 36,37 Despite dose-limiting thrombocytopaenia, BCL-X L antagonism has shown encouraging results in certain solid tumours. 12 38 Combination trials with Navitoclax and kinase inhibitors (MEK, RAF and BRAF) are underway for advanced or metastatic solid tumours, including small-cell lung, colon, pancreatic, rectal and liver cancer. Limited information is available on the effects of this group of kinase inhibitors on megakaryocytes and platelets, although MEK inhibition has been associated with thrombocytopaenia. 39 Hence, for combination treatments using BCL-X L antagonists and kinase inhibitors, it will be imperative to closely monitor platelet counts, as both platelet production and platelet survival may be affected. With the recent development of BCL-X L -specific BH3 mimetics, [25][26][27] there is a need to find means of sustaining platelet counts in order to allow safe dose escalation without increased risk of bleeding. 40 Although BCL-2 is not required for platelet survival, we recently showed that overexpression of BCL-2 in blood cells extends platelet lifespan in adult mice, 20 similarly to that observed in animals lacking the essential mediators of intrinsic apoptosis, BAK and BAX. 5,9 One could imagine that one approach to facilitating the safe administration of BCL-X L antagonists might be transfusion of platelets either overexpressing BCL-2, or lacking BAK/BAX, thus rendering them resistant to BCL-X L inhibition. Recent advances in this field of research, including development of human-induced pluripotent stem cell-derived megakaryocytes generating platelets, 41 HLA-universal platelets, 42 synthetic micro-vessels 43 and a novel microfluidic bioreactor design enabling ex vivo platelet production, 44 might allow such manipulation in the not-too-distant future.
Platelet preparation. Murine blood was obtained by cardiac puncture into 0.1 volume of Aster-Jandl anticoagulant (85 mM sodium citrate, 69 mM citric acid, and 20 mg/ml glucose, pH 4.6). 48 Platelet-rich plasma was obtained by centrifugation at 125 × g for 8 min, followed by centrifugation of the supernatant buffy coat at 125 × g for 8 min. Platelets were washed by two sequential centrifugations at 860 × g for 5 min in 140 mM NaCl, 5 mM KCl, 12 mM trisodium citrate, 10 mM glucose and 12.5 mM sucrose, pH 6.0 (buffer A). The platelet pellet was resuspended in 10 mM Hepes, 140 Mm NaCl, 3 mM KCl, 0.5 mM MgCl 2 , 10 mM glucose and 0.5 mM NaHCO 3 , pH 7.4 (buffer B).
Platelet turnover studies. Mice were injected i.v. with 0.15 μg/g body weight of X488 (Emfret), a rat-derived IgG against the platelet CD42c (GPIbβ) receptor, and platelet lifespan was measured as previously described. 48 Platelets were identified in Platelet-rich plasma as being CD41 + by flow cytometry and the proportion of X488 + platelets remaining at each time point was assessed.
SDS-PAGE and western blot analysis. Platelets were lysed in NP40 lysis buffer and megakaryocytes were lysed in RIPA buffer. Proteins were separated on 4-12% Bis-Tris gels (NuPAGE; Invitrogen) under reducing conditions, transferred onto Immobilon-P membranes (Micron Separation), and immunoblotted with various Abs (see above), followed by incubation with secondary HRP-conjugated Abs and ECL.
Platelet counts in newborn mice. Newborn (day P2) mice were given paw tattoos to assist identification. Newborn mice were weighed on P5, 7 and 10 prior to blood collection. Genotyping was performed by PCR using DNA obtained on P2. For blood collection, the anterior facial vein was punctured using a 30-gauge needle. Five microlitres of blood were collected using a micropipette and dispensed into an EDTA blood tube (Sarstedt) with PBS to a total volume of 200 μl. Automated cell counts were performed on blood collected from newborn mice using an Advia 2120 haematological analyser (Siemens). Reticulated platelets in newborn mice were enumerated using staining with thiazole orange. 48 In vivo administration of BH3 mimetics. Mice were treated via oral gavage with 100 mg/kg ABT-199 administered as a single dose. A stock solution of ABT-199 (10 mg/ml) was diluted in 60% phosal 50 PG (standardised phosphatidylcholine concentrate with at least 50% PC and propylene glycol; Phospholipid GmbH, Cologne, Germany), 30% polyethylene glycol 400 and 10% ethanol. Blood and sterna were collected at 6 and 24 h after treatment. Alternatively, mice were injected i.p. with 5 mg/kg A-1155463.7 administered as a single dose. A stock solution of A-1155463.7 (12.5 mg/ml) in DMSO was diluted by addition of 30% final volume of dosing solution Cremophor ELP:ethanol (ratio 2 : 1) (Cremophor ELP, Sigma-Aldrich) and 5% dextrose in H 2 O (to reach final volume). Blood, sterna and spleens were collected at 2 and 24 h after treatment. Sternum and spleen sections were stained with H&E.
Statistical analyses. Statistical significance between two treatment groups was analysed using an unpaired Student's t test with two-tailed P values. One-way ANOVA with the Bonferroni multiple comparison test was applied where appropriate (GraphPad Prism Software Version 6.0b, La Jolla, CA, USA). *Po0.05; **Po0.005; ***Po0.001 or as otherwise stated. Data are presented as mean ± S.D.