Adenovirus-Based Phospholamban Antisense Expression as a Novel Approach to Improve Cardiac Contractile Dysfunction
Comparison of a Constitutive Viral Versus an Endothelin-1–Responsive Cardiac Promoter
Background—A decrease in sarcoplasmic reticulum Ca2+ pump (SERCA2) activity is believed to play a role in the impairment of diastolic function of the failing heart. Because the expression ratio of phospholamban (PL) to SERCA2 may be a target to improve contractile dysfunction, a PL antisense RNA strategy was developed under the control of either a constitutive cytomegalovirus (CMV) or an inducible atrial natriuretic factor (ANF) promoter. The latter is upregulated in hypertrophied and failing heart, allowing “induction-by-disease” gene therapy.
Methods and Results—Part of the PL cDNA was cloned in antisense and sense directions into adenovectors under the control of either a CMV (Ad5CMVPLas and Ad5CMVPLs, respectively) or ANF (Ad5ANFPLas and Ad5ANFPLs, respectively) promoter. Infection of cultured rat neonatal cardiomyocytes with Ad5CMVPLas reduced PL mRNA to 30±7% of baseline and PL protein to 24±3% within 48 and 72 hours, respectively. The effects were vector dose dependent. Ad5CMVPLas increased the Ca2+ sensitivity of SERCA2 and reduced the time to 50% recovery of the Ca2+ transient. A decrease of PL protein was also achieved by infection with Ad5ANFPLas, and the presence of the hypertrophic stimulus, endothelin-1, led to enhanced downregulation of PL. The adenovectors expressing PL sense RNA had no effect on any of the tested parameters.
Conclusions—Vector-mediated PL antisense RNA expression may become a feasible approach to modulate myocyte Ca2+ homeostasis in the failing heart. The inducible ANF promoter for the first time offers the perspective for induction-by-disease gene therapy, ie, selective expression of therapeutic genes in hypertrophied and failing cardiomyocytes.
The heart adapts to increased workload by responding to mechanical and endocrine, paracrine, and autocrine factors that induce hypertrophy of ventricular myocytes. Although this response is initially compensatory by increasing cardiac output, sustained excessive workload ultimately leads to heart failure.1 The β-adrenoceptor and sarcoplasmic reticulum (SR) Ca2+ pump (SERCA2) genes are downregulated and the atrial natriuretic factor (ANF) gene is strongly upregulated in hypertrophied and failing human myocardium.2 3 4 5 The decrease in SR Ca2+ pump activity is believed to be involved in the diastolic dysfunction of the failing heart.1 4 5 The SR Ca2+ pump in the normal heart is inhibited by the regulatory protein phospholamban (PL); this effect is relieved by β-adrenergic stimulation.6 Studies on (heterozygous) PL-knockout and PL-overexpressing mice have indicated that the expression ratio of PL to SERCA2 is a potential target for improvement of diastolic function of the failing heart.7 Cardiac overexpression of SERCA2 and PL by efficient adenovirus-mediated gene transfer indeed led to shortening and prolongation of the relaxation phase, respectively.8 9 These studies used viral promoters, resulting in ubiquitous and constitutive transgene expression. Cardiac troponin T and myosin heavy- and light-chain promoters have also been applied, but only the latter displayed reporter activity exclusively in ventricles.10 11
In this study, we used an antisense RNA strategy directed against de novo synthesis of PL under control of cytomegalovirus (CMV) or ANF promoter. The latter is strongly upregulated in hypertrophied and failing myocardium and therefore is used to evaluate the concept of “induction-by-disease” gene therapy.3 5 Many studies on signaling involved in hypertrophy and changes in gene expression have used the model of cultured neonatal rat cardiomyocytes stimulated by stretch or agonists such as endothelin-1 (ET-1).12 13 14 Therefore, this model was used to test the effectiveness of the newly developed adenovectors on PL expression, SR Ca2+ uptake activity, and Ca2+ transients.
Culture dishes (Costar catalogue No. 3506 and NUNC, catalogue No. 176740) from Becton Dickinson, culture media DMEM and M199 from Gibco-BRL, and fetal bovine and horse sera and all molecular biology enzymes from Boehringer were used. Cloning plasmids were from Promega and Microbix, ET-1 from Peninsula Laboratories, [32P]dCTP and 45Ca from Amersham, and 125I-labeled goat antimouse IgG from ICN.
Isolation of PL cDNA
PL cDNA was obtained by screening of a rat cDNA library (Lambda ZAP, Stratagene) with a probe derived from genomic DNA containing promoter and exon 1 of PL (unpublished). One cDNA (clone 321, 586 bp) appeared to be identical to previously published sequences, with minor differences in the 3′ trailer,15 and was used for construction of the adenovirus shuttle plasmids.
Development of Recombinant Adenoviral Vectors
Recombinant adenoviral vectors expressing a partial PL antisense mRNA (Ad5CMVPLas) or sense mRNA (Ad5CMVPLs), respectively, under the control of a CMV promoter were developed.16 17 A 230-bp EcoRI-TaqI fragment of PL cDNA containing the translation start codon (Figure 1⇓) was inserted into the polylinker of pCMVI in antisense or sense orientation. The resulting CMV promoter expression plasmids were cotransfected with circularized adenoviral genome pJM17 into 293 cells. Viral plaques appeared between 10 and 20 days and were screened for recombinant virus.16 17 Recombinant vectors expressing a partial PL antisense RNA (Ad5ANFPLas) or sense (Ad5ANFPLs) RNA, respectively, under control of an ANF promoter were developed as follows: The ANF promoter (680-bp EcoRI-BamHI), a kind gift from K.R. Chien (University of California, San Diego), was first cloned in pCAT3 (ANF-CAT) to prove its ET-1 inducibility in rat cardiomyocytes transfected by gene gun biolistics by use of a chloramphenicol acetyltransferase (CAT) ELISA assay (Boehringer).18 19 20 The partial PL cDNA was inserted into the ANF-CAT plasmid in antisense or sense orientation, thereby eliminating the CAT reporter and introducing a chimeric intron between ANF promoter and PL cDNA. Next, EcoRI-BamHI fragments (1521 bp) containing the ANF promoter, followed by the partial PL cDNA and the SV40 poly-A signal, were cloned into the plasmid pΔE1sp1A yielding transfer plasmids with structures similar to those used for Ad5CMVPLs and Ad5CMVPLas, but with ANF instead of CMV promoter. After cotransfection with pJM17 into 293 cells, viral plaques were screened for recombinant virus. Purified positive plaques were grown in large quantities on 293 cell cultures and purified by standard CsCl ultracentrifugation and desalting.16 17 Absence of replication-competent adenovirus was confirmed by polymerase chain reaction (PCR) for the wild-type adenoviral E1 region and by high-titer infections of nonpermissive EAhy.926 and CHO cells (multiplicity of infection [MOI] 10). Virus titers determined by plaque assay and vector doses are expressed as MOI, defined as plaque-forming units per cell.
Preparation and Adenoviral Infection of Myocytes
The investigation conforms to the Guide for the Care and Use of Laboratory Animals (NIH publication No. 85-23, revised 1996). Rat neonatal ventricular myocytes were isolated,21 preplated, and cultured in 1.8-cm2 (Western blotting), 10-cm2 (Ca2+ transient measurements), or 20-cm2 (Northern blotting and Ca2+ uptake measurements) dishes (7.5×104 cells/cm2 for the Ad5ANFPLas and Ad5ANFPLs infections and 1.5×105 cells/cm2 for the Ad5CMVPLas and Ad5CMVPLs infections, except for the Ca2+ transient measurements: 0.3×105 cells/cm2), up to 24 hours in DMEM/M199 (4:1) supplemented with 5% FCS and 5% horse serum and up to 64 hours with only 5% horse serum. Subsequently, infections were started in 200 μL (1.8-cm2 dishes) or in 600 μL (20-cm2 dishes) serum-free medium for 3 hours and continued for 72 hours after addition of 600 or 2600 μL serum-free medium, respectively. When appropriate, after 3 hours, ET-1 (10−8 mol/L) was added to induce hypertrophy.13 18 19 21 Analyses in addition to the plaque assays were carried out to confirm that the relative functional activities of the viruses closely matched: (1) CMV vector doses were tested at the RNA level by hybridizing Northern blots (Figure 2⇓) with SV40 probe specifically recognizing vector-derived RNA, and similar transcriptional activities (relative to GAPDH) were found; and (2) ANF vectors, because of their low expression levels, were tested instead by semiquantitative PCR of infected cells, followed by gel electrophoresis and quantification of the PCR products (not shown).
Total cardiomyocyte RNA was isolated by the guanidinium isothiocyanate method, separated on 1% denaturing formaldehyde-agarose gels, and blotted onto Hybond (Amersham). Probes were labeled by random priming with [α-32P]dCTP: PL; the TaqI-EcoRI fragment of the PL cDNA (excluding the part used for the constructs, Figure 1⇑); GAPDH, developed by RT-PCR on rat heart RNA; and SV40, the SV40 poly A fragment of pCMVI plasmid. Hybridization and washing were carried out as described.13 Hybridization signals were quantified by Molecular Imager (BioRad) and corrected for GAPDH. Two bands were detected with the PL probe and coquantified.
Cardiomyocytes were homogenized into 10 mmol/L Tris-HCl, 0.3 mol/L sucrose, 1 mmol/L dithiothreitol, and 3 mmol/L NaN3 by freeze-thawing. Protein (10 μg) in SDS loading dye (not boiled) was separated with an SDS-PAGE 7.5% to 15% gradient gel and blotted onto PVDF (Amersham). Blots were incubated with monoclonal anti-PL (1:2500) (clone 20/2 [Affinity Bioreagents]) or anti-SERCA2 (1:1000) (clone 2A7-A1, BIOMOL Research Laboratory) as primary antibody followed by 125I-labeled goat antimouse IgG (1:1000) as secondary antibody. After washing, blots were quantified in the molecular imager. The major band (pentameric form of PL, PLV) was quantified.
Ca2+ Affinity of the SR Ca2+ Pump
Oxalate-dependent and thapsigargin (1 μmol/L)–sensitive Ca2+ uptake activity at different free Ca2+ concentrations was measured in infected and noninfected cardiomyocyte homogenates.22 Briefly, cardiomyocytes were frozen in liquid N2 and homogenized in 100 μL 0.3 mol/L sucrose, 20 mmol/L imidazole (pH 7.0), and 10 mmol/L Na2S2O5 (to stabilize SR Ca2+ pump23 ) with a microdismembrator (Braun). Protein (20 μg) was preincubated at 37°C for 2 minutes in 410 μL 2.5 mmol/L ATP, 50 mmol/L KCl, 10 mmol/L imidazole (pH 7.0), 2.5 mmol/L MgCl2, 5 mmol/L potassium oxalate, and 10 μmol/L ruthenium red. The reactions were started by addition of 40 μL 45Ca2+ (0.2 Ci/mmol)–EGTA (1 mmol/L) buffers, of which the pCa’s were 7, 6.5, 6, and 5. The free Ca2+ concentrations were calculated by the computer program WinMAXC version 1.75 (www.stanford.edu/∼cpatton) with the following dissociation constants: Ca2+-EGTA, 3.88×10−7 mol/L; Ca2+-ATP, 8.16×10−5 mol/L; Mg2+-EGTA, 3.10×10−2 mol/L; and Mg2+-ATP, 6.86×10−5 mol/L. Aliquots (100 μL) were filtered after 1, 2, and 4 minutes through 0.45-μm Millipore filters, and 45Ca was measured by liquid scintillation counting.
Ca2+ Transients of Cardiomyocytes
Infected and noninfected beating cardiomyocytes were cultured on 5-cm2 glass coverslips in 10-cm2 dishes, washed with HEPES-buffered salt solution (HBSS) containing (in mmol/L) NaCl 125, KCl 5, MgSO4 1, KH2PO4 1, CaCl2 2.5, NaHCO3 10, HEPES 20, and probenecid 2.5, pH 7.4, and loaded in HBSS containing 2 μmol/L fura 2-AM (Molecular Probes, Eugene) and 5 mmol/L glucose for 45 minutes at 37°C. Subsequently, the cells were washed, and the coverslip was fitted in a Teflon culture dish,24 25 incubated in 1 mL 5 mmol/L glucose containing HBSS in a thermostatted chamber, and mounted on the stage of a fluorescence microscope (Leitz Diavert). Fluorescence was recorded at 360- and 380-nm excitation wavelengths with a 490-nm high-pass filter in the emission path with a sensitive video camera (Hamamatsu C2400-08). The mean gray level of video line segments enclosed by the region of interest was sampled at a frequency of 50 Hz by computer. First, the dark current of the camera was recorded and stored, for which fluorescence intensities were corrected. Second, fluorescence at 360 nm was recorded, by which fluorescence at 380 nm was divided and stored as F380/F360 ratios. These ratios produce signals nonlinearly related to [Ca2+]i. The time to 50% recovery of the Ca2+ transient, but not the time to peak of the Ca2+ transient, was dependent on beating frequency. Therefore, myocytes were always stimulated electrically at 1 Hz by 2 platinum electrodes.
One-way ANOVA was performed, followed by multiple comparison by Student-Newman-Keuls test, with the statistical computer program SigmaStat. Significance was set at P<0.05 for ≥4 independent experiments.
Cellular PL mRNA and Protein
Cardiomyocytes were infected with 2 different doses of the vectors Ad5CMVPLas and Ad5CMVPLs (MOI 5 and 10) to study their effects on PL expression. A marked and vector-dose–dependent reduction in PL mRNA is apparent after 24 hours of infection with Ad5CMVPLas (88±14% and 68±19% left at MOI 5 and 10, respectively) and becomes more pronounced after 48 hours (53±1% and 30±7% left, respectively) (Figure 2⇑). Subsequently, we examined the effect of the CMV vectors on PL protein. A significant decrease in PL protein content was first visible 48 hours after infection with Ad5CMVPLas (90±11% [P=NS], 63±3%, and 55±6% left at MOI 2.5, 5, and 10, respectively), becoming more pronounced after 72 hours (69±9%, 42±10%, and 24±3% left, respectively) (Figure 3⇓). Like control cells, incubation with Ad5CMVPLs virus produced no effects, indicating specificity of the antisense effect.
Cellular SR Ca2+ Pump Activity and SERCA2 Protein
We measured oxalate- and ATP-dependent Ca2+ uptake in homogenates of control and infected cardiomyocytes at different nonsaturating free [Ca2+]i (pCa 5 to 7). No significant effect of Ad5CMVPLas (MOI 10) on Ca2+ uptake activity was seen 48 hours after infection (Figure 4A⇓), but at 72 hours, the relative Ca2+ uptake activity at pCa 6.5 was significantly increased compared with control and Ad5CMVPLs-infected cells (Figure 4B⇓). No effect of Ad5CMVPLas was seen at pCa 6, when Ca2+ uptake activity is closer to maximum reached at pCa 5. After 72 hours, Ad5CMVPLas reduced maximal Ca2+ pump activity, which cannot be explained by cell loss, because total protein content remained unchanged (legend to Figure 4⇓). A partial reduction of SERCA2 protein after 72 hours of Ad5CMVPLas infection (57.9±6.4% left, P<0.05 versus uninfected or Ad5CMVPLs-infected cells) was observed, however, which may explain the decrease of maximal Ca2+ uptake activity (Figure 5⇓).
Ca2+ Transients After Infection With Ad5CMVPLas
To determine whether depletion of PL protein by Ad5CMVPLas also alters the duration of the Ca2+ transient of intact cardiomyocytes, the specific fluorescence of fura 2–loaded cells was recorded at fixed beating frequency (1 Hz). Time to 50% recovery of Ca2+ transients was already lower after 48 hours in Ad5CMVPLas (MOI 5)–infected cells (245±98 ms; n=42, P<0.05 versus uninfected and Ad5CMVPLs-infected cells) compared with uninfected (593±88 ms; n=4) and Ad5CMVPLs-infected cells (564±38 ms; n=7), indicating that by decrease of PL content, the removal rate of Ca2+ from myofilaments during relaxation of the cardiomyocytes is increased (see representative average Ca2+ transient recordings in Figure 6⇓).
Cellular PL Protein After Infection With Ad5ANFPLas
The ANF promoter fragment used was first tested for ET-1 responsiveness and displayed low activity in nonstimulated gene-gun–transfected cardiomyocytes.18 19 On ET-1 (10−8 mol/L) stimulation, the promoter activity increased 2.5-fold.18 Figure 6⇑ depicts PL protein expression obtained 48 and 72 hours after infection with Ad5ANFPLas (MOI 10) with or without ET-1 stimulation. Earlier, we showed that in unstimulated cells, ANF mRNA is present in low but significant amounts.13 19 This explains the decrease in PL protein that was measured in Ad5ANFPLas-infected cells also in the absence of ET-1 (Figure 7⇓, at 72 hours, 86±6% left versus noninfected cells). However, Ad5ANFPLas in the presence of ET-1 resulted in a significant further reduction of PL protein (72±7% left). Ad5ANFPLs had no effect, indicating specificity of the antisense effect.
Infection of cardiomyocytes with Ad5CMVPLas resulted in a dramatic decrease in endogenous PL mRNA expression after 48 hours, presumably by increased degradation of the message. The vector generating partial sense mRNA and perhaps truncated PL protein clearly had no effect on endogenous PL mRNA or protein, indicating the specificity of the antisense effect. Downregulation of PL protein lags ≈24 hours behind that of PL mRNA, indicating that no compensatory posttranscriptional mechanisms are present.
Downregulation of SERCA2 in hypertrophied and/or failing myocardium is believed to play a role in the diastolic dysfunction. Several studies reported downregulation of PL in failing heart as well, but generally not as much as SERCA2.4 5 Because PL in anomalous phosphorylation states may also occur in severe heart failure,26 antisense suppression of the endogenous PL gene may have an advantage over the enhancement of SERCA2 gene expression. Gene enhancement will not influence preexisting malfunctioning endogenous gene products, whereas antisense-based gene suppression can do so. The reduced expression of PL caused by Ad5CMVPLas results in increased Ca2+ affinity of the SR Ca2+ pump and in decreased time to 50% recovery of the Ca2+ transients of intact cells. This is consistent with a very recent report by He et al,27 who showed increased shortening as well as relengthening velocity in adult rabbit cardiomyocytes infected with an adenovector expressing a dominant negative mutant PL protein. However, in our present study on monolayer cultures of cardiomyocytes, we were unable to measure relative shortening of the cells during stimulation, because rhythmic displacement of cell boundaries is hard to relate to initial length of the myocytes concerned. Moreover, extrapolation of results on contractile motion of unloaded myocytes to the in vivo situation is doubtful.
Ad5CMVPLas infection (72 hours) of cardiomyocytes not only increased the Ca2+ affinity of SR Ca2+ pump but also partially reduced its maximal activity. The latter effect is probably due to decreased SERCA2 expression (Figure 5⇑). At present, it is unclear whether this adenovector effect indicates counterregulatory systems in cardiac Ca2+ homeostasis.
Conversely, by suppression of endogenous PL protein, the β-adrenergic response becomes blunted. In failing myocardium, however, β-adrenoceptors are reduced and the associated regulatory components functionally altered, causing the attenuated β-adrenergic response.2 Interestingly, myocytes isolated from hearts of PL-knockout mice retained the relaxant effect evoked by catecholamine.28 Troponin I phosphorylation may increase the “off rate” for Ca2+ exchange, with troponin C compensating for the effect of PL ablation on the β-adrenergic response.
Our development of inducible ANF promotor vectors for induction-by-disease gene therapy takes advantage of the fact that the cardiac hypertrophic phenotype in vivo and in vitro is characterized by transcriptional upregulation of distinct genes.3 5 12 13 19 20 Unlike the in vivo situation, a low expression of ANF is present in nonstimulated cardiomyocytes,12 13 14 19 which explains the small reduction by treatment with Ad5ANFPLas even in the absence of ET-1. In the presence of ET-1, however, a stronger reduction of PL protein was observed, consistent with the known ET-1 responsiveness of the ANF promoter.12 13 14 These results are a first demonstration in cardiomyocytes of the feasibility of using a heart-specific promoter that becomes activated when the cells become hypertrophic. Another advantage is the expected return of this promoter to minimal activity when the myocytes improve in contractile function because of treatment. Knowlton et al29 tested the same rat ANF promoter in transgenic mice with hypertrophy induced by pressure overload. In this model, the rat ANF promoter did confer cell-specific expression but not inducibility on pressure overload. Possibly, transgenic mice may compensate for the effects of the transgene during development, or species difference of rat promoter versus transgenic mice may be responsible for the lack of inducibility.29 Hajjar et al30 first described a method yielding high adenovector transfer rates in vivo specifically to the rat ventricles. We could confirm those findings using an adenovector expressing green fluorescent protein, thus opening a way to test the PL antisense adenovectors in vivo.31 In conclusion, the present study suggests that vector-mediated PL antisense RNA expression may become a feasible approach to improve contractile dysfunction of the failing heart. The inducible ANF promoter offers the perspective for induction-by-disease gene therapy by selective expression of therapeutic genes in the hypertrophied and failing myocytes.
This work was supported by grant 95.109 from the Netherlands Heart Foundation and by the Deutsche Forschungsgemeinschaft through a Heisenberg fellowship to Dr Poller, grants Po. 378/2-1,2.
- Received October 4, 1999.
- Revision received December 7, 1999.
- Accepted December 22, 1999.
- Copyright © 2000 by American Heart Association
Francis GS, Cohn JN. Heart failure: mechanisms of cardiac and vascular dysfunction and the rationale for pharmacologic intervention. FASEB J. 1990;4:3068–3075.
Takahishi T, Allen PD, Izumo S. Expression of C-type natriuretic peptide gene in failing and developing human ventricles: correlation with expression of the Ca2+ ATPase gene. Circ Res. 1992;71:9–17.
Arai M, Matsui H, Periasamy M. Sarcoplasmic reticulum gene expression in cardiac hypertrophy and heart failure. Circ Res. 1994;74:555–564.
Hasenfuss G, Just H. Myocardial phenotype changes in heart failure: cellular and subcellular changes and their functional significance. Br Heart J. 1994;72:510–517.
Simmerman HKB, Jones LR. Phospholamban: protein structure, mechanism of action and role in cardiac function. Physiol Rev. 1998;78:921–947.
Koss KL, Grupp IL, Kranias EG. The relative phospholamban and SERCA2 ratio: a critical determinant of myocardial contractility. Basic Res Cardiol. 1997;92(suppl 1):17–24.
Meyer M, Dillmann WH. Sarcoplasmic Ca2+ ATPase overexpressing by adenovirus mediated gene transfer and in transgenic mice. Cardiovasc Res. 1998;37:360–366.
Hajjar RJ, Schmidt U, Kang JX, et al. Adenoviral gene transfer of phospholamban in isolated rat cardiomyocytes: rescue effects by concomitant gene transfer of sarcoplasmic reticulum Ca2+-ATPase. Circ Res. 1997;81:145–153.
Franz WM, Rothmann T, Frey N, et al. Analysis of tissue-specific gene delivery by recombinant adenoviruses containing cardiac-specific promoters. Cardiovasc Res. 1997;35:560–566.
Inesi G, Lewis D, Sumbilla C, et al. Cell-specific promoter in adenovirus vector for transgenic expression of SERCA1 ATPase in cardiac myocytes. Am J Physiol. 1998;274:C645–C653.
Shubeita HE, McDonough PM, Harris AN, et al. Endothelin induction of inositol phospholipid hydrolysis, sarcomere assembly, and cardiac gene expression in ventricular myocytes: a paracrine mechanism for cell hypertrophy. J Biol Chem. 1990;265:20555–20562.
Van Heugten HAA, Van Setten MC, Eizema K, et al. Sarcoplasmic reticulum Ca2+ ATPase promoter activity during endothelin-1 induced hypertrophy of cultured rat cardiomyocytes. Cardiovasc Res. 1998;37:503–514.
Shanahan CM, Weissberg PL, Metcalfe JC. Isolation of gene markers of differentiated and proliferating vascular smooth muscle cells. Circ Res. 1993;73:193–204.
Gambaryan S, Wagner C, Smolenski A, et al. Cyclic AMP-stimulated renin release from rat isolated perfused kidney, microdissected glomeruli, and isolated juxtaglomerular cells is inhibited by endogenous or overexpressed cGMP-dependent protein kinases. Proc Natl Acad Sci U S A. 1998;95:9003–9008.
Eizema K, Van Heugten HAA, Bezstarosti K, et al. SERCA2 and ANF promoter-activity studies in hypertrophic cardiomyocytes using liposome-, gene gun- and adenovirus-mediated gene transfer. In: Takeda N, Nagano M, Dhalla NS, eds. The Hypertrophied Heart. Boston, Mass: Kluwer Academic Publishing; In press.
Van Heugten HAA, Bezstarosti K, Dekkers DHW, et al. Homologous desensitization of the endothelin-1 receptor evoked phosphoinositide response in cultured neonatal rat cardiomyocytes. J Mol Cell Cardiol. 1994;25:41–52.
Lamers JMJ, Duncker DJ, Bezstarosti K, et al. Increased activity of the sarcoplasmic reticular calcium pump in porcine stunned myocardium. Cardiovasc Res. 1993;27:520–524.
He H, Meyer M, Martin JL, et al. Effects of mutant and antisense RNA of phospholamban on SR Ca2+-ATPase activity and cardiac myocyte contractility. Circulation. 1999;100:974–980.
Wolska BM, Stojanovic MO, Luo W, et al. Effect of ablation of phospholamban on dynamics of cardiac myocyte contraction and intracellular Ca2+. Am J Physiol. 1996;271:C391–C397.
Knowlton KR, Rockman HA, Itani M, et al. Divergent pathways mediate the induction of ANF transgenes in neonatal and hypertrophic ventricular myocardium. J Clin Invest. 1995;96:1311–1318.
Hajjar RJ, Schmidt U, Matsui T, et al. Modulation of ventricular function through gene transfer in vivo. Proc Natl Acad Sci U S A. 1998;95:5251–5256.