Avoidance of Transient Cardiomyopathy in Cardiomyocyte-Targeted Tamoxifen-Induced MerCreMer Gene Deletion Models
Cardiac myocyte targeted MerCreMer transgenic mice expressing tamoxifen-inducible Cre driven by the α-myosin heavy chain promoter are increasingly used to control gene expression in the adult heart. Here, we show tamoxifen-mediated MerCreMer (MCM) nuclear translocation can induce severe transient dilated cardiomyopathy in mice with or without loxP transgenes. The cardiomyopathy is accompanied by marked reduction of energy/metabolism and calcium-handling gene expression (eg, PGC1-α, peroxisome proliferator-activated α, SERCA2A), all fully normalized with recovery. MCM-negative/flox-positive controls display no dysfunction with tamoxifen. Nuclear Cre translocation and equally effective gene knockdown without cardiomyopathy is achievable with raloxifene, suggesting toxicity is not simply from Cre. Careful attention to controls, reduced tamoxifen dosing and/or use of raloxifene is advised with this model.
- inducible transgenic
- Cre recombinase
- selective estrogen receptor modulator
- ventricular function
- mouse models
The Cre-loxP system is widely used for selective cell-targeted manipulation of gene expression1 and has been further enhanced by generating tamoxifen-responsive fusion proteins for conditional Cre induction.2,3 Targeted cells constitutively express Cre flanked by mutated estrogen receptor ligand-binding domains (MerCreMer [MCM]) insensitive to endogenous estrogen but sensitive to tamoxifen (TAM). MCM is cytoplasmic via binding to heat shock protein 90 complex, but this complex dissociates on TAM-Mer binding, whereupon the MCM targeting sequence sends the construct to the nucleus for Cre-mediated excision of loxP flanked sequences.1 Sohal et al linked MCM with an α-myosin heavy chain (Myh6) promoter to create cardiomyocyte-specific gene targeting.3 However, Cre recombinase displays dose-dependent cytotoxicity impairing growth and causing DNA fragmentation,4,5 and a recent review raised a caution that TAM-stimulated MCM in adult hearts may also adversely influence heart function.6 Here, we report on these cardiac effects and provide methods to avoid them.
Materials and Methods
Myh6-MerCreMer+/+ transgenic mice (no. 005650, Jackson Labs, Bar Harbor, Me) were used. Myh6-MCM+/−/no-flox were generated by mating to C57Bl/6 mice. Two strains with floxed alleles coding for either R2 (Tgfbr2fl/fl) or R1 (Alk5fl/fl) transforming growth factor (TGF)β receptors (both on C57Bl/6 backgrounds) were crossed with Myh6-MCM+/− mice to study gene knockdown. Cardiac function was assessed by serial echocardiography and invasive pressure–volume analysis. Gene expression was determined by real-time PCR, gene knockdown was determined by analysis of mRNA and TGFβ-stimulated Smad2 phosphorylation, and nuclear Cre was determined by immunohistochemistry and immunoblot. All animal protocols were approved by The Johns Hopkins University Animal Care and Use Committee and followed established NIH guidelines. Details are provided in the Online Data Supplement, available at http://circres.ahajournals.org.
In both MCM+/−/Tgfbr2fl/fl and MCM+/−/Alk5fl/fl mice, TAM administered at 20 mg/kg body weight (BW) IP for 5 days (proposed dose3) was insufficient for gene and functional knockdown (latter assessed by suppression of TGFβ-stimulated Smad2 phosphorylation; Figure 1A and 1B; Online Figure I, A and B). Increasing the dose to 80 mg/kg BW per day for 5 days (IP) resulted in 60% mortality by 6 days after TAM treatment because of severe cardiomyopathy (Online Figure II). Oral delivery of the same dose for 7 days was tolerated (no mortality) and effective for gene and functional TGFβ-receptor knockdown (Figure 1A and 1⇓B; Online Figure I, A and B). However, a marked though reversible dilated cardiomyopathy (Figure 1C and Online Figure III) was also observed in both floxed/MCM+/− models and MCM+/− mice without a floxed transgene. MCM-negative controls with or without floxed genes (eg, MCM−/−/Tgfbr2fl/fl) developed no myopathy at any TAM dose. Cardiac-depression peaked ≈3 days after terminating TAM (day 10 of protocol) with fractional shortening declining from 61±1% to 26.5±5% (P<0.01) and end-diastolic dimension increasing (3.2±0.1 to 4.1±1.4 mm; P<0.01) in MCM+/− mice (with or without floxed alleles; n=19). In vivo pressure–volume analysis in MCM+/−/no flox mice confirmed marked transient systolic and diastolic depression (Figure 1D and Online Table I), with full recovery observed by day 28 (3 weeks after stopping TAM). MCM−/− controls had no TAM-induced dysfunction. Myocardium displayed patchy interstitial mononuclear infiltration at day 10 (mild myocarditis) that resolved by day 28 and no myocyte hypertrophy (Online Figure IV).
TAM/MCM-induced cardiomyopathy was accompanied by marked changes in stress response, energy/metabolism, and calcium-handling genes (Figure 2). Natriuretic peptide expression (Nppa and Nppb) rose markedly in MCM+/− versus MCM−/− by day 10 and then returned to normal, although β-myosin heavy chain (Myh7), which typically rises with cardiac stress, was unchanged. Peroxisome proliferator-activated receptor (PPAR)α, PPARγ-coactivator (PGC)1α, and transcription factor A-mitochondrial (TFAM) genes, which are centrally involved with coordinating mitochondrial function, energetics, and metabolism7 and suspected to play a key role in dilated human cardiomyopathy,8 all declined substantially with cardiac depression and then fully recovered to normal levels. Lastly, both sarcoplasmic reticular ATPase and phospholamban expression declined transiently, correlating with cardiac function. Although these changes are observed with various cardiac failure conditions, the insult in this instance started in the nucleus and its striking reversibility unusual. Although Cre toxicity might be suspected, reversibility would be less anticipated from DNA fragmentation, particularly in differentiated tissue without a high rate of cell regeneration.
Raloxifene (RAL) is an alternative selective estrogen receptor modulator with similarities but also differences when comparing TAM-regulated transcription. This may be attributable to differential binding to estrogen-related receptors (eg, ERRγ)9 and/or recruitment of different coactivators.10 Because RAL interacts with Mer, albeit at lower binding affinity,11 we tested whether RAL could induce gene knock-down without cardiomyopathy. Because of poor solubility, DMSO was required for IP dosing, limiting the dose to ≤40 mg/kg BW per day, which was suboptimal for gene knockdown in MCM+/−/Alk5fl/fl. However, higher oral doses were feasible, tolerated, and effective. Myh6-MCM+/− mice fed 160 mg/kg BW per day RAL PO displayed effective gene knockdown but without cardiac dysfunction (Figure 3A and 3B; Online Figure V). Nuclear Cre targeting was similar with TAM or RAL treatments, as shown by histochemistry (Figure 3C) and nuclear fraction immunoblot (Online Figure VI). Stress, metabolic, and Ca2+-handling gene changes (eg, Figure 2) were not observed (data not shown). Importantly, Cre recombinase activity (Online Figure VII), gene knockdown efficacy (Figure 3B), and corresponding functional suppression of TGFβ-induced SMAD2 phosphorylation (Figure 3D) with RAL were similar as with 80 mg/kg BW per day TAM, although required longer exposure (21 days). Recombination (lox-P site excision) was observed earlier with RAL (7 days), although at lower levels. Both RAL and TAM resulted in a similar ≈10% decline in BW during the first week that subsequently recovered (n=6 to 7/group; P=0.95, 2-way ANOVA). Lower oral TAM dose (20 mg/kg BW per day, one-third previously reported12) for 21 days also induced effective gene knockdown without dysfunction (Online Figure VIII).
Our study did not precisely define the mechanism for TAM-MCM cardiac effects, and such analysis falls outside the scope and intent of this report. However, the data raise a novel hypothesis that the cardiotoxicity is not simply attributable to Cre. First, the striking reversibility is difficult to reconcile with mechanisms of cell damage attributed to Cre, namely targeting pseudo loxP sites to cause DNA fragmentation, cell growth arrest, and/or death.4,5 Second, the finding that both TAM and RAL induced similar nuclear Cre localization and gene suppression, yet with striking differences in cardiac phenotype, further questions a Cre-toxicity mechanism. An alternative relates to the specific nature of the ligand–MCM complex. In addition to recruiting different nuclear coactivators that can differentially target transcription,10 TAM but not RAL can inhibit ERRγ, which, along with ERRα, plays a central role in bioenergetic regulation.13 These differences could alter nuclear interactions that depend or are independent of Cre recombinase. Although TAM exposure at the same dose was not toxic, the MCM construct increases nuclear levels >4-fold, which could amplify interactions. Although similar energy/metabolic changes often accompany pathological cardiac stress,7,8 here, the triggering mechanism involved altered nuclear signaling; therefore, these hibernation-like reversible changes may indeed be primary. Further studies are needed to clarify this hypothesis. Lastly, it remains possible that differences in the time course and/or nuclear Cre exposure between TAM and RAL play some role, and the finding that lower prolonged dosing of TAM was also effective without myopathy might suggest this. However, this could also reflect less nuclear exposure to TAM-MCM.
Our results have implications for existing and ongoing research with the MerCreMer model. Studies lacking Myh6-MCM+/−-flox− controls should be viewed cautiously, particularly if a significant cardiac phenotype is found within the 1 to 2 weeks after starting TAM. In such models, gene deletion without cardiodepression (eg, using RAL or longer-term low-dose TAM) is required. Both TAM and RAL dosing may need to be individualized depending on the floxed gene (perhaps related to gene accessibility and/or expression rate). For TAM, care to include MCM+ controls and provide sufficient recovery time is strongly advised. Although RAL avoids the myopathy, the dose required was fairly high in the present floxed models (lower doses might work for other models). However, mice treated mice with ≈10× this dose for up to 3 months had no systemic limiting effects.14 Because RAL may have antihypertrophic effects when given chronically,15 MCM+ controls are also advised.
Sources of Funding
Supported by National Heart, Lung, and Blood Institute grants HL-59480, HL-77180, and HL-98297 (to D.A.K.); an American Heart Association Fellowship Award (to M.Z.); and an American Heart Association Scientist Development Grant (to E.T.).
This manuscript was sent to David Eisner, Consulting Editor, for review by expert referees, editorial decision, and final disposition.
Original received April 1, 2009; revision received May 26, 2009; accepted June 1, 2009.
Sohal DS, Nghiem M, Crackower MA, Witt SA, Kimball TR, Tymitz KM, Penninger JM, Molkentin JD. Temporally regulated and tissue-specific gene manipulations in the adult and embryonic heart using a tamoxifen-inducible Cre protein. Circ Res. 2001; 89: 20–25.
Huss JM, Kelly DP. Nuclear receptor signaling and cardiac energetics. Circ Res. 2004; 95: 568–578.
Greschik H, Flaig R, Renaud JP, Moras D. Structural basis for the deactivation of the estrogen-related receptor gamma by diethylstilbestrol or 4-hydroxytamoxifen and determinants of selectivity. J Biol Chem. 2004; 279: 33639–33646.
Shang Y, Brown M. Molecular determinants for the tissue specificity of SERMs. Science. 2002; 295: 2465–2468.
Paulmurugan R, Gambhir SS. An intramolecular folding sensor for imaging estrogen receptor-ligand interactions. Proc Natl Acad Sci U S A. 2006; 103: 15883–15888.
Ogita H, Node K, Liao Y, Ishikura F, Beppu S, Asanuma H, Sanada S, Takashima S, Minamino T, Hori M, Kitakaze M. Raloxifene prevents cardiac hypertrophy and dysfunction in pressure-overloaded mice. Hypertension. 2004; 43: 237–242.