BH3-mimetic drugs that antagonize Bcl-2 family prosurvival proteins are effective against some cancers, particularly those with abnormally high expression of the prosurvival protein target.1 Despite the recent success of BH3-mimetics targeting Bcl-2, Bcl-xL and Bcl-w, mechanism-based cell killing in vivo by targeting other important prosurvival proteins such as Mcl-1 or Bfl-1 has yet to be demonstrated. In the absence of small molecules targeting these prosurvival proteins, BH3-domain peptides, the prototypes for this class of drug, are useful because their binding specificity profile can be manipulated.2, 3 Such peptides have been applied in in vitro studies to validate the targeting of particular prosurvival proteins in certain tumor types.4, 5 However, in vivo applications require technically challenging chemical modifications of peptides, and while mouse xenograft models of virally-infected BH3 domain-expressing tumor cells can be used, this does not allow evaluation of effects of the ligand on normal tissues.
Here we have generated transgenic mice in which peptide-based BH3 ligands can be inducibly expressed to evaluate their effects in vivo and provide proof-of-principle for similar-acting drugs. To achieve this we adapted a previously described strategy allowing FLP-recombinase-mediated insertion of an expression cassette at an frt ‘landing pad’ at the type I collagen (Col1a1) locus in mouse embryonic stem cells.6 The cassette comprises sequences encoding a BH3-domain protein under control of a tetracycline (tet)-regulated element (TRE) promoter. To develop a system that is broadly applicable to BH3 domains with different specificities, we employed the BimS BH3-only protein as a scaffold in which BH3 domains with different specificities could replace the native BH3 sequence (Figure 1a). BimS is an intrinsically unstructured protein that tolerates extensive mutation of its BH3 sequence, and BimSBH3 chimeras display the prosurvival protein specificity profile of the replacement BH3 domain.7 In this study we focussed on the BH3 domain of Bad because it targets the prosurvival proteins Bcl-2, Bcl-xL and Bcl-w.7 Hence BimSBad expression should mimic the BH3-mimetics ABT-737 and ABT-263 that have been extensively studied in vivo.1
We generated TRE-BimSBad transgenic mice and crossed them to mice expressing the rtTA (tet-on) transactivator under the control of the cytomegalovirus (CMV) promoter, which provides high-level expression in many tissues including blood.8 Bitransgenic mice were then treated with doxycycline (Dox) to induce BimSBad expression (Figure 1a). Western blot analysis verified Dox-inducible transgene expression in white blood cells of BimSBad; CMV-rtTA bitransgenic mice (Figure 1b). As ABT-737/263 induces thrombocytopenia in mice and humans due to antagonism of Bcl-xL, we measured platelet levels as a biomarker for functional expression of the ligand.1 Notably, blood cell analysis of these mice revealed a significant reduction (~65% decrease) in platelet counts (Figure 1c and d), a degree of thrombocytopenia comparable with that seen in patients administered with ABT-263.9 Importantly, platelet counts rebounded to normal levels following removal of Dox for 7 days (Figure 1d), illustrating the reversibility of the system. We also generated an additional transgenic mouse strain allowing inducible expression of a BimSBad construct possessing a BH3 sequence mutation (BimSBadmut; Figure 1a) that decreases its affinity for Bcl-xL by >40-fold (KD 8.5 nM versus <0.2 nM as measured by surface plasmon resonance). Induction of BimSBadmut expression (Figure 1b) did not alter blood cell or platelet counts (Figure 1d), indicating that the thrombocytopenia observed in BimSBad mice is due to Bcl-xL inhibition. These data provide the first evidence that BH3-only proteins can be inducibly expressed in a mouse model, effectively mimicking at least one functional consequence (reduced platelet counts) seen in mice and humans treated with BH3-mimetic drugs of similar specificity. We envisage multiple applications for similarly engineered mice. For example, tet-regulated expression of different BimS variants in mice could reveal toxicities associated with neutralization of their prosurvival protein targets (individually or in combination), and tissue-specific effects could be addressed by crossing onto mice where rtTA expression is driven by different promoters. Moreover, mice crossed to different tumor-prone models could provide in vivo evidence for the tumor-killing efficacy associated with different BH3 specificities not yet available through small molecule drugs, and extended to studies on combination therapies with existing drugs.
Roy MJ et al Br J Pharmacol 2014; 171: 1973–1987.
Lee EF et al J Cell Biol 2008; 180: 341–355.
Foight GW et al ACS Chem Biol 2014; 9: 1962–1968.
Glaser SP et al Genes Dev 2012; 26: 120–125.
Del Gaizo Moore V, Letai A . Cancer Lett 2013; 332: 202–205.
Beard C et al Genesis 2006; 44: 23–28.
Chen L et al Mol Cell 2005; 17: 393–403.
Takiguchi M et al PLoS One 2013; 8: e54009.
Roberts AW et al J Clin Oncol 2012; 30: 488–496.
This work was supported by grants and fellowships from the NHMRC of Australia (Project Grant 1041936 to WDF and 575573 to RAD, Program grant 1016701 to DCSH, Career Development Fellowship 1024620 to EFL, Research Fellowship to DCSH), Cancer Council of Victoria Grant-In-Aid 1057949 to WDF and EFL, Leukemia and Lymphoma Society (Specialized Centers of Research Grant), and a Sylvia and Charles Viertel Senior Research Fellowship to RAD. Infrastructure support from NHMRC IRIISS grant #361646 and the Victorian State Government OIS grant is gratefully acknowledged.
The authors declare no conflict of interest.
About this article