|Home | About | Journals | Submit | Contact Us | Français|
To establish reversibility of cardiac phenotypes in hypertrophic cardiomyopathy (HCM) by generating bigenic mice in which expression of the mutant transgene could be turned on and off as needed.
Advances in molecular therapeutics could ultimately lead to therapies aimed at correcting the causal mutations. However, whether cardiac phenotypes, once established, are permanent, or could be reversed, if expression of the mutant protein is turned off, is unknown.
We generated ligand-inducible bigenic mice, turned on and off expression of cardiac troponin T-Q92 (cTnT-Q92), responsible for human HCM, and characterized molecular, histological, and functional phenotypes.
We established 6 lines and in dose-titration studies showed treatment with 1000 ug/kg of mifepristone consistently switched on cTnT-Q92 expression in the heart. 16 days of expression enhanced myocardial systolic function without changing myocardial cAMP levels. Levels of PTEN, a regulator of cardiac function; phospho-PKC-Zλ-Thr538 and phosphor-PKD-Ser744-748 were reduced, while mRNA levels of NPPA, NPPB (hypertrophic markers), and ATP2A2 and procollagen COL1A1, COL1A2, and COL3A1 were unchanged. 70 days of expression increased COL1A1 and COL1A3 mRNAs levels and collagen volume fraction and reduced levels of NPPA and NPPB. Switching off expression of the cTnT-Q92 reversed the molecular and histological phenotypes completely.
The initial phenotype induced by cTnT-Q92 is enhanced myocardial systolic function followed by changes in signaling kinases and interstitial fibrosis. Established phenotypes in HCM are reversible if expression of the mutant protein is turned off. These findings provoke pursuing specific therapies directed at correcting the underlying the genetic defect in HCM.
Human hypertrophic cardiomyopathy (HCM) is a genetic disease characterized by unexplained cardiac hypertrophy, myocyte disarray, interstitial fibrosis and increased left ventricular ejection fraction (LVEF) (1-3). Mutations in at least 10 different sarcomeric proteins cause HCM (2) and those in cardiac troponin T (cTnT) account for approximately 15 to 20% of HCM cases (2). HCM due to cTnT mutations is generally characterized by an increased LVEF, minimal or mild hypertrophy, severe disarray, fibrosis and a high incidence of sudden cardiac death (SCD) (4;5).
Elucidation of the molecular genetic basis of HCM has shifted the focus toward elucidation of the pathogenesis of HCM and developing therapies aimed at reversing or attenuating the evolving phenotypes and ultimately correcting the primary genetic defects. Currently, it is unknown whether turning off expression of the mutant protein could lead to reversal of the established phenotypes; or the phenotypes, once established, are permanent. Proving the reversibility of the HCM phenotypes is fundamental and pre-requisite to pursuing specific molecular therapies directed at correcting the underlying genetic mutation in HCM. Thus, to determine the reversibility of the evolving phenotypes in HCM and gain insight into the pathogenesis of cardiac phenotypes, we generated an inducible bigenic mouse model in which expression of the cTnT-Q92, known to cause HCM in humans (4), could be switched on and off as needed. The inducible bigenic model is based on ligand-dependent activation of an inactive tripartite regulatory protein containing the ligand-binding domain (LBD) of a truncated human progesterone receptor (PR) that upon activation with mifepristone binds to the DNA sequences in the target transgene and induces expression of the target gene (6).
The Animal Subjects Committee of Baylor College of Medicine approved the experiments. The regulator is a tripartite protein comprised of the LBD of a C-terminal truncated (deletion of 42-amino acid) human PR, DNA-binding domain of yeast GAL4 protein and the transcription activation domain of the p65 subunit of human NFκB transcription factor. It is inactive in its native state because the modified LBD lacks the ability to respond to progestins but responds to synthetic anti-progeston mifepristone (RU486) (7). The target transgene is comprised of 4 tandomly placed 17-bp (CGGAGTACTGTCCTCCG) GAL4 upstream activating sequences (UAS) and a TATA box positioned 5’ to a mutant cTnT-Q92 cDNA. Expression of the target transgene could be turned on and off by adding or withdrawing mifepristone. The α-myosin heavy chain (α-MyHC) promoter affords cardiac-restricted expression of the regulator protein and consequently, that of the cTnT-Q92 (Figure 1A).
The α-MyHC promoter (clone 26) was a kind gift from Dr. Jeffery Robbins (University of Cincinnati) and the PAP-CMV-GLp65-SV and p17x4-TATA-CAT plasmids were generous gifts from Dr. Sophia Y Tsai (Baylor College of Medicine). To clone the regulator transgene, the GLp65 fragment containing the coding sequences for the tripartite regulator protein was excised from PAP-CMV-GLp65-SV at the unique Asp718 and BamH1 restriction sites, blunt ended and ligated to the blunt-ended SalI and HindIII cloning sites of clone 26, downstream to the α-MyHC promoter (Figure 1A). To clone the target transgene, a previously described full-length (1.1 kb) human cTnT-Q92 cDNA (8) was released from the pSP73/cTnT-Q92 clone and inserted at XhoI and KpnI sites, replacing the CAT sequences in the p17x4-TATA-CAT clone. The correct orientations and the sequence of the transgenes were verified by direct sequencing. The regulator transgene (~7 kb) was released from the vector by digesting the clone at the encompassing Not1 sites and the target transgene by digesting the clone at HindIII and KpnI sites (~2 kb). Transgenes were purified from the gel and injected into fertilized mouse zygotes (FvB/N) and the pups were screened by PCR for the regulator and target transgenes and for the β casein, as an internal control.
Mifepristone (BIOMOL, PA) was initially dissolved in dimehtyl sulfoxide (DMSO) at a concentration of 40 mg/ml and the working solution was prepared in sesame oil at a concentration of 1 mg/ml. To induce target gene expression, dose-titration and time-course experiments were performed and mifepristone was injected intra-peritoneally (n= 5 to 12 adult mice per group) at 3 dosages of 300, 500 and 1000 μg/Kg/day for 2, 4, 8 or 16 days (short-term) or 70 days (long-term). To switch off expression of the transgene, mifepristone was discontinued for at least 16 days, which is longer than 5 half-life of cTnT in the heart (9) and for 70 days in long-term experiments. Parallel experiments were performed in non-transgenic (NTG) injected with mifepristone and in bigenic mice injected with a placebo (sesame oil). Expression of the target transgene was determined by reverse transcription-polymerase chain reaction (RT-PCR) and immunoblotting after switching on or off expression of the transgene at the above time points. Short-term (2, 4, 8, and 16 days) induction of expression of cTnT-Q92 transgene was achieved through daily intra-peritoneal injections. To induce long-term expression of cTnT-Q92 transgene, mifepristone controlled-released pellets (Innovative research of America, Sarasota, FL) were prepared to deliver 1000 μg/Kg per day and implanted subcutaneously.
RT-PCR and immunoblotting were performed using transgene-specific oligonucleotide primers and monoclonal anti-troponin T antibody 2D10 (Research Diagnostic, MD), respectively, as described (8). Relative expression levels of the transgene and endogenous cTnT mRNAs was determined by RT-PCR using primers (forward 5’TTCATGCCCAACTTGGTGCC and reverse 5’CTCTCTTCAGACAGGCGGTTC) to amplify a 260 bp of endogenous and transgene cTnT mRNAs followed by digestion with 10 units of BstU1 restriction enzyme, which digests the endogenous (mouse) but not the transgene (human) cTnT into 2 fragments of 184 and 76 bp.
Expression levels of molecular markers of cardiac hypertrophy, namely A-type natriuretic peptide (NPPA), B-type natriuretic peptide (NPPB), skeletal α-actin (ACTA1) and sarcoplasmic reticulum calcium ATPase 2a (SERCA2a or ATP2A2) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as well as expression levels of procollagen COL1(α1), COL1(α2) and COL3(α1) mRNAs (COL1A1, COL1A2 and COL3A1, respectively) were determined by real-time quantitative RT-PCR using specific Taqman probes and primers in an 7900HT SDS unit (Applied Biosystem, Inc).
Expression levels of selected phosphorylated and total signaling kinases, implicated in affecting cardiac structure and function, were detected using 20 μg of cardiac protein extracts and phospho-specific and pan-specific antibodies, as described (10). Specific information about antibodies is available upon request.
Echocardiography was performed prior to and following switching on and off expression of the transgene at each dose and time point. Mice were anesthetized with intra-peritoneal injection of anesthetics (avertin 0.2 ml /10 gram body weight, intraperitoneal injection of a 1.25% solution in the long-term experiments) and trans-thoracic echocardiography was performed using an Acuson Sequoia Cardiac System equipped with 15-MHz linear transducer (Acuson Co., Mountain View, CA) per conventional methods (11). Echocardiographic parameters were measured in at least 3 consecutive cardiac cycles by an investigator totally unaware of the genotypes or the treatment groups.
The extent of cardiac myocyte disarray and interstitial fibrosis were quantified as published (10;12). Percentage of myocardium showing disarray was determined in a semi-quantitative manner in 12 fields per section and 10 sections per mouse. Similarly, collagen volume fraction (CVF) was calculated using quantitative automated planimetry in 12 fields per section and 10 sections per mouse. Myocyte cross sectional area (CSA) was determined by quantitative indirect immunofluorescence in approximately 2500 myocyte per group, as published (10;12). All histological quantifications were performed by an examiner who had no knowledge of group assignment.
Since cAMP is a major mediator of myocardial contractile function, we measured total myocardial cAMP levels by enzyme immunoassay per instruction of the manufacturer (Amersham Biosciences, UK). In brief, 100 mg of ventricular myocardium (n=4 to 8 mouse per group) was homogenized in Hank’s balanced salt solution with 5 mM EDTA without calcium and magnesium and centrifuged at 1000 g at 4°C for 15 minutes. The homogenates were subjected to ion-exchange chromatography using columns containing anion exchange silica sorbents followed by analyte elution with acidified methanol. The eluted samples were added to an assay buffer containing rabbit anti-cAMP antibody in 0.05 M acetate, 0.02% (w/v) bovine serum albumin and mixed thoroughly. Then, an equal volume of cAMP-horseradish peroxidase conjugate in 0.05 M acetate buffer, pH 5.8 was added to each reaction and washed with 0.01 M phosphate buffer pH 7.5 containing 0.05% (v/v) tween 20. The mixtures were added to microplates coated with a donkey anti-rabbit IgG antibody, incubated at 4°C for 60 minutes. Following washing, the enzyme substrate was added to the wells and the reactions were terminated after 1 hour of incubation at room temperature with 1.0 M sulfuric acid and the optical density was determined immediately in a plate reader at 450 nm.
Differences between the baseline and follow up values in each group were compared by paired t-test. Differences in variables among NTG, placebo, and mifepristone groups were compared by ANOVA, followed by Bartlett’s test to determine the homogeneity of the variances. Variables with unequal SD were compared by the non-parametric Kurskall-Wallis test.
Three lines of regulator (α-MyHC/GLp65-GAL4-ΔPR) and 7 lines of target (4xGAL4UAS-TATA-cTnT-Q92) transgenic mice were generated. The target transgenic lines were tested for background expression of the cTnT-Q92, after confirmation of the absence of background leak, mated to the regulator transgenic mice and 6 lines of bigenic mice were established. Monogenic and bigenic mice were identified by PCR using transgene-specific oligonucleotide primers (Figure 1B).
Expression of the transgene mRNA and protein was detected consistently after daily administration of mifepristone at 1000 μg/Kg/day for 16 days. Shorter durations or lower doses of administration of mifepristone did not consistently induce expression of the transgene. The results of RT-PCR (in three lines, Figure IC) and immunoblotting (in one line, Figure 1C) confirmed expression of the cTnT-Q92 mRNA and protein, respectively. No significant cTnT-Q92 expression was detected in bigenic mice injected with a placebo or in the target monogenic mice. The relative expression levels of the induced transgene cTnT was low in all lines and comprised < 10 % of the total cTnT mRNA in the heart. Transgene protein and mRNA was not detected 16 days after discontinuation of mifepristone (Figure 1D).
In the long-term study, 2/9 bigenic mice died 3 and 9 days after mifepritstone injection, in contrast to 0/11 NTG mice injected with mifepristone or 0/9 bigenic mice injected with the placebo. Baseline echocardiographic indices were unremarkable in the two mice died after induced gene expression. Since these mice were found dead, no histological or molecular studies were performed.
Echocardiographic indices of myocardial function were compared at the baseline prior to injection of mifepristone; 16 and 70 days after induction of expression of the cTnT-Q92 in the heart; and 16 days and 70 days after discontinuation of mifepristone administration. The results were remarkable for a smaller left ventricular (LV) end systolic diameter (ESD), increased LV fractional shortening (FS), LVEF, velocity of circumferential fiber shortening (Vcf) and maximum aortic flow velocity in induced state at both time points (16 and 70 days) as compared to the baseline values. There was no significant difference in the LV wall thickness, end diastolic diameter (EDD), or mitral inflow early (E), late (A) and E/A. However, deceleration time (DT) of E velocity was prolonged in the bigenic mice injected with mifepristone. Table 1 shows the echocardiographic data at the baseline, 16 days after induced gene expression and 16 days after discontinuation of mifepristone injection (data for 70 days induced expression were largely similar, except the mean heart rate was 462 ± 71 bpm, and are not shown). Discontinuation of mifepristone led to reversal of the echocardiographic phenotypes. There were no significant changes in other echocardiographic indices. Injection of mifepristone, at 1000 μg/Kg/day, to NTG mice alone had no significant effects on LV systolic function or other echocardiographic indices (Table 1).
To determine whether increased cAMP levels was responsible for the enhanced myocardial systolic function, we measured total myocardial cAMP levels in the experimental groups. There were no significant differences in myocardial cAMP levels among the experimental groups (NTG: 15.9 ± 3.9; target transgenic: 17.5 ± 3.0; regulator transgenic: 17.0 ± 3.3; bigenic placebo: 16.3 ± 3.4; bigenic mifepristone on: 13.5 ± 4.8; bigenic mifepristone off: 13.1 ± 1.5 fmol/gram, p=0.162).
To determine whether induced expression of cTnT-Q92 also led to changes in expression of activated signaling molecules, we assayed for levels of selected signaling kinases, implicated in cardiac function, by immunoblotting using phospho-specific antibodies. The results, partially shown in Figure 2, were notable for a significant reduction in the expression level of PTEN and to a lesser extent, in levels of p-PKC-Zλ-Threonine 538 and p-PKD-Serine 744-748, when transgene was turned on for 16-day. There were no significant changes in levels of PI3K-p110-α, β, δ; and p85 or p-PKC-δ-Serine 643, p-PKA, p42/p44 MAPKs, GSK3β or calcineurin. Discontinuation of mifepristone led to normalization of the activated signaling molecules. There were no statistically significant changes in the expression levels of selected signaling kinases following shorter (2,4 and 8 days) or longer (70 days) duration of induction of expression of cTnT-Q92 in the heart. To determine, whether changes in the expression levels of signaling kinases was independent of administration of mifepristone, NTG mice were injected with 1000 μg/Kg/day of mifepristone (n=4) and immunoblotting was repeated. Mifepristone injection alone did not induce significant changes in levels of p-PKC-Zλ-Threonine 538 and p-PKD Serine 744-748 or PTEN in the heart.
Expression levels of NPPA, NPPB, ACTA1, ATP2A2, COL1A1, COL1A2 and COL3A1 were determined using real-time quantitative RT-PCR 16 and 70 days after induction of expression of cTnT-Q92 and after switching off expression of the transgene. After 16 days of induced expression of cTnT-Q92, myocardial levels of molecular markers of hypertrophy and fibrosis were largely unchanged with the exception of a modest decrease in ATP2A2 levels (data not shown). However, long-term (70 days) induced expression of cTnT-Q92 led to significant reductions in expression levels of NPPA and NPPB without a significant change in ATP2A2 levels (Figure 3A). Furthermore, expression levels of COL1A1 and COL1A3, encoding the predominant cardiac collagen, were increased significantly, while expression level of COL1A2 was largely unchanged (Figure 3A).
To detect whether induction of expression of cTnT-Q92 was associated with morphological or histological phenotypes characteristics of human HCM, we determined heart weight, heart weight/body weight ratio, LV weight/body weight ratio, myocyte CSA, number of myocytes per microscopic fields, % CVF and the extent of myocyte disarray. Short-term induced expression of the cTnT-Q92 did not lead to discernible changes in histological or morphological phenotypes, since no significant differences in morphometric histoligical indices were detected among NTG, NTG injected with mifepristone, target and regulator monogenic and bigenic mice in the placebo and mifepristone groups (Table 1). However, as shown in Figure 3B, induced expression of cTnT-Q92 for 70 days led to 2.1 fold increase in CVF, in accord with the observed increased expression of the procollagen mRNAs. Turning off expression of cTnT-Q92 led to normalization of CVF (NTG: 2.5 ± 0.66; switch on: 4.2 ± 1.0; switch off: 2.0 ± 0.24, p=0.007). With regard to indices of hypertrophy, neither heart weight/body weight ratio, nor echocardiographic indices, nor myocyte CSA was significant different among the experimental groups. Myocyte disarray comprised 6.3 ± 3.6 % of the myocardium in the bigenic mice induced with mifepristone for 70 days, while it was 2.8 ± 2.6 % in NTG mice (T=2.02, p=0.08) and 5.0 ± 4.7 % in bigenic mice following turning off expression of the transgene (F= 1.05, p=0.373 among the three groups).
We describe a switch on – switch off bigenic mouse model in which expression of the mutant sarcomeric protein cTnT-Q92, responsible for human HCM (4), was turned on and off by administering and withdrawing mifepristone, the activating ligand for the otherwise inactive regulatory protein. Short-term induced expression of the mutant cTnT-Q92 enhanced myocardial systolic function, a characteristics phenotype of HCM, independent of myocardial cAMP levels and in the absence of histological or molecular phenotypes of hypertrophy, disarray or fibrosis. Enhanced myocardial contractile performance was associated with reduced levels of selected active signaling kinases, implicated in regulating cardiac function. The long-term induced expression of cTnT-Q92 led to increased expression of procollagen genes and interstitial fibrosis, while levels of the molecular markers of hypertrophy were reduced. The induced phenotypes, namely, interstitial fibrosis, increased expression of procollagen genes and signaling kinases and enhanced cardiac systolic function were completely reversed upon switching off expression of cTnT-Q92. These results, in a genetic animal model of human HCM mutation, establishes the reversibility of HCM phenotypes, a pre-requisite in designing future therapies targeted at correcting the underlying genetic mutations. The results also identify the primary defect conferred by cTnT-Q92 as enhanced myocardial systolic function, independent of myocardial cAMP levels. Thus, the potential utility of the switch on-switch off bigenic mouse model are at least two-fold: First, to establish the reversibility of the cardiac phenotypes induced by HCM mutations; and second, to identify the initial defects conferred by the mutations followed by delineation of the sequence and pathogenesis of ensuing functional, molecular, histological and morphological phenotypes in HCM.
The finding of reversal and normalization of interstitial fibrosis following switching off expression of the mutant troponin T was confirmed by two complementary methods of real time RT-PCR analysis of expression of major procollagen genes in the heart as well as by quantitative morphometric analysis of collagen protein in myocardial sections stained with collagen-specific dye. Reversal and normalization of interstitial fibrosis in the heart has potential clinical implications, since interstitial fibrosis is a potential risk factor of SCD in humans with HCM (13). In contrast to interstitial fibrosis, however, cardiac hypertrophy was absent in the bigenic cTnT-Q92 mice, as was also the case in the conventional transgenic cTnT-Q92 mice (8;14). Notably, humans with HCM due to cTnT-Q92 usually exhibit mild or minimal hypertrophy (15). The absence of discernible hypertrophy in the bigenic mice or even the presence of smaller myocyte and heart size in the conventional cTnT-Q92 mice (11;14) could simply reflect the enhanced cardiac performance, which affords the heart the ability to handle the normal load at a much lesser cardiac mass. In accord with the above, expression levels of markers of cardiac mass, namely NPPA and NPPB, were reduced significantly following long-term induced expression of cTnT-Q92. Regarding myocyte disarray, it was increased modestly in the bigenic mice followed long-term induced expression. However, changes were of borderline statistical significance. Therefore, it remains to be determined whether switching off expression of the mutant protein could also lead to reversal of hypertrophy and disarray, as we have shown previously through pharmacological interventions in genetically engineered animal models (10;12).
The molecular basis of enhanced myocardial contractile performance following induction of expression of cTnT-Q92 was not discerned, suffice that it was independent of myocardial cAMP levels or injection of mifepristone. Previous studies have shown that myofibrils isolated from the hearts of the conventional cTnT-Q92 transgenic mice exhibit enhanced calcium sensitivity of tension generation, evidenced by a higher values of half-maximally activating –log of the molar free Ca+2 concentration (pCa50) (16). The latter in conjunction with the results of studies in skinned fibers isolated from rabbit and pig hearts showing enhanced Ca+2 sensitivity (17;18), is likely to explain the observed enhanced cardiac systolic function in our bigenic mice in the induced state. In addition to enhanced Ca+2 sensitivity; other mechanisms, including changes in signaling molecules, as observed following induced expression of cTnT-Q92, could also affect cardiac function. Regarding the potential impact of signaling molecules, our study was limited to analysis of a limited number of signaling molecules, implicated in regulating cardiac structure and function, and was not comprehensive. Nonetheless, a significant change in the expression of tumor suppressor PTEN, implicated in regulating cardiac function (19), was detected. PTEN, in addition to affecting cardiac function, also plays an essential role in proper cell adhesion, alignment and migration (20;21). Thus, given the essential role of PTEN in cellular attachment and alignment, reduced expression of PTEN could be involved in subsequent evolution of myocyte disarray in HCM. The molecular mechanisms that regulate expression and activation of PTEN, p-PKC-Zλ-Threonine 538 and p-PKD-Serine 744-748 in the heart remain unknown. We showed reduction in the expression levels of selected signaling molecules was independent of mifepristone administration. The role of changes in the activation state of selected signaling kinases, noted above, in the pathogenesis of cardiac phenotypes in HCM merits additional investigations.
The findings of the present study are restricted to the cTnT-Q92 mutation. We note that the initial impetus and the ensuing early and intermediary phenotypes, imparted by the mutant sarcomeric proteins are likely to be diverse and vary according to the topography of the causal mutations and genes. The bulk of experimental data implicate enhanced myofibrillar Ca+2 sensitivity as the initial defect imparted by mutations in the cTnT (16-18). However, with regard to mutations in the β-MyHC and other genes, reduced myofibrillar and myocardial contractile performance and/or reduced ATPase activity have emerged as the prime candidates as the initial defects (reviewed in (2)). The complexity of the pathogenesis of HCM phenotypes is further elucidated by the divergence of phenotypic expression of mutations in the same causal gene, which lead to morphologically contrasting phenotypes of hypertrophic, dilated and restrictive cardiomyopathies (22;23). Therefore, a single animal model is insufficient to elucidate the pathogenesis of HCM caused by a diverse array of mutations. It is also noteworthy that there are several differences between transgene (human) and endogenous (mouse) cTnT sequences that could confound the effect of the mutation. Our previous studies, however, have shown that these differences do not induce discernible histological or echocardiographic phenotypes (8). Hence, the observed phenotypes in the switch on model likely reflect the effect of the mutation.
In conclusions, we describe a switch on-switch off bigenic mouse model in which expression of cTnT-Q92, responsible for human HCM, could be turned on and off as needed. The results show that the established cardiac phenotypes in HCM are not permanent and could be reversed upon turning off expression of the mutant protein. The findings also indicate that the initial phenotype induced by the cTnT-Q92 is enhanced myocardial systolic function followed by changes in activation of signaling molecules implicated in cell organization and myocyte function followed by interstitial fibrosis. These findings provoke pursuing specific therapies directed at correcting the underlying the genetic defect in HCM.
Supported by grants from the National Heart, Lung, and Blood Institute, Specialized Centers of Research P50-HL54313, RO1 HL68884, and a TexGen grant from Greater Houston Community Foundation.