J Mol Cell Cardiol 30, 2683–2694 (1998)Article No. mc980830

Sub-Antihypertensive Doses of RamiprilNormalize Sarcoplasmic ReticulumCalcium ATPase Expression and Functionfollowing Cardiac Hypertrophy in RatsSamuel Y. Boateng1, Anne-Marie L. Seymour3, Nabeela S. Bhutta1,Michael J. Dunn1, Magdi H. Yacoub1 and Kenneth R. Boheler1,2

1Imperial College School of Medicine, National Heart and Lung Institute, Department of CardiothoracicSurgery, Dovehouse Street, London, SW3 6LY, UK, 2NIH/NIA/GRC/LCS, 5600 Nathan Shock Drive,Baltimore, MD 21224-6825, USA, 3Department of Biological Sciences, University of Hull,Hull HU6 7RX, UK

(Received 4 May 1998, accepted in revised form 16 September 1998)

S. Y. B, A.-M. L. S, N. S. B, M. J. D, M. H. Y K. R. B. Sub-Antihypertensive Doses of Ramipril Normalize Sarcoplasmic Reticulum Calcium ATPase Expression andFunction following Cardiac Hypertrophy in Rats. Journal of Molecular and Cellular Cardiology (1998) 30,2683–2694. We examined the hypothesis that the angiotensin converting enzyme inhibitor ramipril at sub-antihypertensive concentrations could improve sarcoplasmic reticulum (SR) CaATPase expression and functionin compensated hypertrophied rat hearts. Five weeks after abdominal aortic constriction, rats received a dailydose (50 lg/kg/day) of ramipril or vehicle for 4 weeks. Cardiac angiotensin-converting enzyme (ACE) activityincreased with cardiac hypertrophy (CH) but returned to normal following ramipril treatment. SR CaATPaseprotein levels and activity decreased with CH (P<0.05) and were normalized following ramipril treatment(P<0.05 for protein and activity). No change in phospholamban (PLB) protein levels could be demonstratedbetween any of the groups. In contrast, ramipril treatment specifically increased control SR CaATPase andPLB mRNA levels by >60% (P<0.01) and >30%, respectively. In the hypertrophied group, SR CaATPaseincreased by 35% (P<0.05 n=6) after ramipril treatment. Calsequestrin mRNA levels were unaffected byramipril administration. In conclusion, ramipril normalizes SR CaATPase protein expression and function inpressure-overloaded and compensated CH. The effects of ramipril are however multifaceted, affecting RNAand protein expression differentially. 1998 Academic Press

K W: SR CaATPase; Phospholamban; Cardiac hypertrophy; Regression; RNA; Protein.

reticulum (SR); therefore, any change in SRIntroductionfunction might have profound effects on theamplitude and duration of contraction. SRIn hypertrophy and heart failure, cardiac tissue

displays impaired contraction and relaxation (Mor- Ca ATPase (SERCA2a, the predominant isoformin cardiac myocytes) (Lompre et al., 1994; Araigan et al., 1990) and in human ventricles,

changes in contraction occur independently of et al., 1994) mRNA and/or protein expressionhave been shown to be decreased in animalmyosin heavy chain or actin gene transitions

(Hasenfuss et al., 1994). The main reservoir for models (de la Bastie et al., 1990; Feldman et al.,1993) following cardiac hypertrophy (CH). Thecalcium in the cardiac myocyte is the sarcoplasmic

Please address all correspondence to: Kenneth R. Boheler, NIH/NIA/GRC/LCS, 5600 Nathan Shock Drive, Baltimore, MD 21224-6825, USA.

0022–2828/98/122683+12 $30.00/0 1998 Academic Press

S. Y. Boateng et al.2684

SERCA2b isoform in heart (non-myocytes) has to distinguish between the neurohumeral and pres-sure effects on cardiac hypertrophy.been studied to a much lesser degree, and it is

unclear if changes in its abundance also occur. Inpatients with dilated cardiomyopathy, SR calciumuptake in heart homogenates was found to be Materials and Methodsreduced in some (Limas et al., 1987), but notall studies (Movsesian and Schwinger, 1998). The Model of cardiac hypertrophy and drug treatmentdiminished function of the SR and particularlythe decrease in CaATPase expression and activity, Four experimental groups were studied: sham op-provide a molecular basis for the prolonged cytosolic erated (controls)+vehicles, sham operated+Ca transients and slowed diastolic calcium decline ramipril, hypertrophy+vehicle and hyper-seen in isolated beating myocytes from hyper- trophy+ramipril. Male Sprague–Dawley ratstrophied and failing myocardium. As abnormalities (Charles River Inc.) weighing 225–250 g were an-in calcium handling can modify the contractile and esthetized with ketamine (60 mg/kg body weight)relaxation behaviour of the myocardium and may and medetomidine (0.25 mg/kg), and the des-promote the development of arrhythmias (Than- cending aorta between the left and right renaldroyen et al., 1991), any normalization of CaATPase branches was constricted with ethicon suture (Boat-expression and/or function might be beneficial to eng et al., 1997). Five weeks later, rats received athe hypertrophied myocardium. daily oral dose (50 lg/kg/day) of ramipril or vehicle

Autocrine, paracrine and circulating biologically (polyethylene glycol) for 4 weeks. Rats were killedactive mediators in addition to pressure overload by cervical dislocation and hearts stored at−70°Ccan initiate events that can result in the de- prior to protein and RNA extractions. Calcium up-velopment of CH, and by association might be take or tissue ACE activity was measued on freshlyresponsible for the altered abundance and activity isolated, left ventricular tissue.of the SR CaATPase (Baker et al., 1992; Dzau,1992). Of particular interest has been the role ofangiotension II (AgII) and angiotensin-converting Urinary cyclic guanosine monophosphate (cGMP)enzyme (ACE) inhibition on the development and analysisregression of hypertrophy (Schunkert et al., 1990;Miyata and Haneda, 1994). Rats administered AgII, Urinary cyclic guanosine monophosphate (cGMP)even at sub-hypertensive doses, develop CH, and was used as a non-invasive marker of cardiac hyper-AgII content has been shown to correlate with left trophy and failure (Michel et al., 1990). Twenty-ventricular weight (Mizuno et al., 1992). Addition four hour urine samples were collected at 5 weeksof the angiotensin type I (AT1) receptor blocker, and at 9 weeks after surgery and stored at−20°C.losartan, inhibits this response, suggesting that car- Samples were diluted 1/200, and cGMP con-diac growth mediated by AgII is independent of centration measured using a [3H]cGMP radio-blood pressure changes. ACE inhibitor treatment of immunoassay kit (Amersham). At 5 weeks post-rats with abdominal aortic constriction prevents surgery, only those banded animals which showedthe development of CH (Linz et al., 1992) and a significant increase in cGMP were used in theblockade of angiotensin-mediated responses induces hypertrophy+ramipril group.regression of CH (Rockman et al., 1994). ACE in-hibition can also prevent the decrease in SRCa ATPase mRNA with cardiac hypertrophy (Zi- Carotid artery blood pressure measurementserhut et al., 1996; Bruckschlegel et al., 1995; Angeret al., 1995), but it has yet to be demonstrated Rats were anesthetized as described and injectedthat SR CaATPase expression and function can be with heparin (200 IU). The carotid artery was can-normalized in the fully compensated and hyper- nulated and blood pressure measurements recorded.trophied myocardium. To test this hypothesis, we Measurements were made less than 6 h after thehave studied the effects of ramipril on the expression final dose of ramipril was administered.and function of the CaATPase and the expressionof its regulatory protein phospholamban (PLB) inhypertrophied rat myocardium. Specifically, this mRNA analysis by Northern and dot blotshas been done in a rat model of pressure-overloadinduced cardiac hypertrophy where sub-anti- Total RNA was isolated from rat left ventricles

(Chomczynski and Sacchi, 1987). SERCA2 (Lomprehypertensive doses of ramipril were administered

SR Calcium ATPase and Regression of Cardiac Hypertrophy 2685

et al., 1989), calsequestrin and atrial natriuretic on 6% polyacrylamide gels; ACE and Ca ATPaseproteins were separated on 7.5% gels; and PLB andfactor (ANF) (Petrou et al., 1995) mRNA levels

were determined by Northern blots (10 lg of total ANF on 12% gels. The membranes were blockedwith 3% non-fat milk (Marvel) in phosphate buf-RNA) and dot blots (containing 1.25, 2.5 and 5 lg

of total RNA). Phospholamban (PLB) mRNA levels fered saline and 0.05% polyethylene sorbitan mon-olaurate (Tween 20) and probed with appropriatewere quantitated only from dot blots since PLB

mRNAs are comprised of four transcripts of highly antibodies [PLB monoclonal Ab (provided by J.Wang); Ca ATPase (SERCA2a-specific) polyclonaldifferent molecular weights (Moorman et al., 1995).

Autoradiograms were quantitated only where sig- Ab (F. Wuytack); ACE monoclonal Ab (J.D. Ca-travus); ANF polyclonal Ab (Ian Milne) and myosinnals were not saturated (linear increases in signals

over the concentration range) and results were heavy chain (MHC) mouse monoclonal-cardiacmuscle specific (Novocastra Laboratories Ltd)] andalmost identical between Northern and dot blots.

DNA probes used in this study were all of rat origin exposed to ECL film (Amersham). The membraneswere probed with horse radish peroxidase con-and included a 2.1 kb EcoRI fragment insert from

pRH39 (Lompre et al., 1989) or full-length jugated secondary antibodies (rabbit anti-mouse orgoat-anti-rabbit, Amersham). The proteins werepRCSERCA6 (Koban et al., 1997); a 1.0kb EcoRI

fragment from pRCPLB1 (Moorman et al., 1995), visualized using enhanced chemiluminescence(ECL) (Amersham). Membranes were stripped withand a 1.1 kb EcoRI fragment from pANF kindly

provided by K. Knowlton. For calsequestrin, a 1% SDS, 15 m Tris buffer pH 6.8, 0.02% betamercaptoethanol at 50°C for 30 min and stained600 bp rat specific cDNA fragment was isolated

from pRCal-11 (provided by D. Wynne) after di- with 0.15% Amido black (BDH). Both the stainedmembranes and the ECL films were scanned atgestion with NcoI and SalI. Each insert was labelled

by random priming (Amersham) to a specific ac- 176 lm resolution using a Molecular Dynamicslaser densitomer and analysed by the One di-tivity of 5–8×108 cpm/lg. Hybridizations were as

described (Moorman et al., 1995) with final washing mensional software package (Dunn, 1987; Blose,1988). All data have been expressed as the den-conditions as follows: SERCA2, 0.1X SSPE, 0.1%

SDS, 65°C; PLB, ANF and Calsequestrin, 0.5X SSPE, sitometric ratio of ECL signal to total protein stainingand to MHC as previously described (Koban et al.,0.1% SDS, 55°C. Membranes were stripped of

labelled probe by addition of boiling 1% SDS and 1997).subsequently hybridized to an end-labelled oligo-nucleotide complementary to 18S. RNA signalswere normalized against 18S ribosomal RNA andcalsequestrin which is myocyte specific and un-changed in cardiac hypertrophy (Studer et al., Measurement of calcium uptake1994).

Oxalate stimulated calcium uptake was determinedon homogenates as described (Boateng et al., 1997).Left ventricles were homogenized in extractionProtein analysis by Western blottingmedium (30% glycerol, 20 m Hepes pH 7.4, 5 m

sodium azide, 1 m PMSF), and calcium uptakeTotal protein was isolated by homogenizing frozenleft ventricular tissue (>0.3–0.6 g) in 1.0% sodium was measured in 30 m MOPS, 100 m KCl, 5 m

ATP, 6 m MgCl2, 15 m K-oxalate, 0.2 m EGTA,dodecyl sulfate (SDS) containing 1 m phenylmethyl sulfonyl-fluoride (PMSF). Homogenates 5 m sodium azide, 20–50 lg homogenate and

0.150 m of 45CaCl2. Sodium azide was added towere centrifued at 14 000 rpm (4°C) and the super-natant kept for protein determination and Western inhibit mitochondrial uptake. After addition of cal-

cium, uptake was measured for a period of up toblotting. Protein concentrations were determinedusing the Bradford assay (Bradford, 1976). 15 min at 30°C, and the solution was filtered

through a 45 lm pore filter (Millipore). Filters wereProteins were separated using SDS-poly-acrylamide gel electrophoresis (PAGE) and trans- washed with 10 ml of a solution containing 100 m

KCl, 1 m EGTA and 10 m histidine. Data areferred by electroblotting onto nitrocellulosemembranes (Hybond-C, Amersham) for ACE, expressed as a function of total protein (nmol per

min per mg total protein). Control experiments wereCa ATPase, ANF and myosin heavy chain analysisas described previously (Latif et al., 1993). PLB performed with thapsigargin, in the presence or

absence of ATP and/or oxalate, the concentrationswas transferred onto polyvinyl difluoridine (PVDF)membranes. Myosin heavy chain was separated of which are indicated in the text.

S. Y. Boateng et al.2686

Tissue ACE activity Statistical analysis

Tissue ACE activity was determined on homo- All data are expressed as mean±... Data groupswere compared using a one way analysis of variancegenates using [3H] benzoyl-Phe-Ala-Pro (BPAP, 25

Ci/mmol) from Amersham (final activity 0.1 lCi/ followed by a Tukey-Kramer multiple comparisonstest. Significance between lung ACE activity forml) (Catravas and Watkins, 1985). Fifty microlitres

of heart homogenate were incubated with 940 ll vehicle v hypertrophy+ramipril and vehicle vcontrol+ramipril was tested using a Student’s t-of serum free Hanks medium for 30 min at 37°C.

reactions were started by the addition of 10 ll BPAP. test. Significance was taken at P<0.05.After 30 min of incubation at 37°C, 400 ll ofthe mixture was placed in Ecoscint (Packard) and400 ll in a mixture of 0.12 HCl and toluene Resultsscintillation fluid to measure the parent compoundand metabolite, respectively by liquid scintillation. Animal model

The ratio of heart weight to tibia length, sig-nificantly increased in both hypertrophy groupsafter 9 weeks of aortic constriction (Table 1). No

Serum ACE activity difference in body weight was seen between any ofthe groups. The percentage of water within the

Blood was collected at the time of sacrifice and lungs was not different, indicating the absence ofstored at−70°C. Serum ACE activity was measured heart failure. Five weeks post-surgery, urinaryby the Harefield Hospital Pathology Laboratory cGMP (concentrations) levels were unchanged inusing a Sigma diagnostics kit. The spec- control animals (176.48±13.27 pmols/ml, n=14)trophotometric method used the synthetic tripeptide but were increased in rats with aortic constrictionsubstrate N-[3-(2-furyl) acryloyl]--phenylalanyl- (326.01±38.30 pmols/ml, n=18). After treatmentglycylglycine. Results were expressed as units per with ramipril, elevated levels of urinary cGMP re-litre. turned towards normal (219.16±15.92 pmols/ml,

n=12) even in the absence of a significant re-gression of cardiac hypertrophy. Rats with cardiachypertrophy without ramipril treatment main-tained elevated levels of urinary cGMP

Isolated cardiomyocyte ACE activity (349.33±11.45 pmols/ml, n=5). No change inthese concentrations was seen between the controls

Cells were isolated as described (Naqvi and on ramipril and those on vehicle at any time.MacLeod, 1994). Briefly, hearts were perfused for 5 Carotid systolic blood pressure was significantlymin in a Krebs–Henseleit (KH) solution, composition elevated in all animals with an aortic constriction(m) NaCl 120, NaHCO3 25, KCl 4.7, MgSO4 0.97, (Table 1). No significant change in blood pressureKH2PO4 1.2, glucose 11, containing 1 m Ca2+ and could be demonstrated between the hypertrophypHed to 7.4. The perfusate was changed to low group on vehicle v those on ramipril. This dose ofcalcium (LC) solution [composition (m) NaCl 120, ramipril was therefore sub-antihypertensive. Car-KCl 5.4, MgSO4 5, pyruvate 5, glucose 20, taurine diac function was not measured in these animals.20, Hepes 10 and nitroacetic acid (NTA) 5], bubbledwith 100% oxygen for 5 min and a final freecalcium concentration of 12–14 l. The heart was

Expression studiesperfused with collagenase (0.3 mg/ml, Worth-ington) and hyaluronidase (0.6 mg/ml, Sigma). En-

RNA expressionzymes were dissolved in LC medium without NTAand a free calcium of 200 l. The ventricles were To determine the effects of sub-antihypertensive

doses of ramipril on SR-CaATPase and PLB mRNAschopped and further digested in the collagenasemixture. After filtering and several low speed cent- in hypertrophied rat hearts, total RNA was analysed

by Northern and dot blot assays and normalizedrifugations the myocytes pellet was homogenizedin a medium containing 30% glycerol, 20 m Hepes against 18S ribosomal RNA. From dot blot analyses,

a linear increase in signal for SERCA2 and PLB waspH 7.4, 5 m sodium azide, 1 m PMSF and storedat −70°C. seen between 1.25 and 5 lg of RNA. No significant

SR Calcium ATPase and Regression of Cardiac Hypertrophy 2687

Table 1 Animal data following aortic constriction and ramipril treatment

Treatment groups Heart weight (g) Heart weight/ % water in lung Body weight (g) Blood pressuretibial length tissue (mmHg)

(g/cm)

controls+vehicle 1.38±0.04 0.30±0.01 78.2±1.1 521±20 154.8±7.2controls+ramipril 1.40±0.04 0.31±0.01 77.2±1.2 551±19 nhypertrophy+vehicle 1.69±0.03∗∗ 0.37±0.01∗∗ 76.9±3.9 495±20 185±7.4∗hypertrophy+ramipril 1.57±0.04∗∗ 0.35±0.01∗∗ 77.0±1.1 533±10 190±5.3∗∗

The heart weight and ratio of heart weight to tibial length increased significantly following aortic constriction, but were notsignificantly reduced following ramipril treatment. Neither the body weight nor the percentage of water within the lungs alteredbetween any of the groups. Blood pressure was significantly elevated following aortic constriction but was not reduced following sub-antihypertensive dosing with ramipril.∗P<0.05 and ∗∗P<0.01 and represent differences compared with (controls+vehicle), n=9 to 12 animals in each group.

n—not measured.

difference in CaATPase transcripts could be dem-onstrated between controls on vehicle and hyper-trophied hearts [Fig. 1(a)]. Control andhypertrophied groups treated with ramipril showeda significant increase in CaATPase mRNA ex-pression [Fig. 2(a)]. Similarly, the abundance ofPLB mRNA increased in the ramipril treated groupsrelative to controls on vehicle, but no difference inPLB expression could be demonstrated betweencontrol+vehicle and hypertrophy+vehicle orcontrol+ramipril and hypertrophy+ramipril [Figs1(b) and 2(b)]. When standardized to 18S ribosomalRNA, no change in calsequestrin mRNA could bedemonstrated between the 4 groups (P>0.1 in allgroups) [Fig. 1(c)]. Calcium ATPase mRNA alsoremained significantly increased following ramipriltreatment of the hypertrophy group when stand-ardized to calsequestrin (P<0.05 n=6).

As an independent marker of cardiac hyper-trophy, tissue ANF mRNA was measured [Fig. 1(c)]. Relative to the controls on vehicle(0.214±0.09), ANF mRNA levels increased byalmost 10-fold following cardiac hypertrophy(2.04±0.5, P<0.001). The controls treated withramipril (0.08±0.03) gave similar results tocontrols on vehicle. Ramipril treatment of thehypertrophied group led to a small decrease in ANFmRNA (1.3±0.29) relative to the hyper-trophy+vehicle group (P>0.05), but its ex- Figure 1 (a) Northern blot analysis of SR CaATPasepression was not normalized to control levels and 18S RNA isolated from (A) control+vehicle, (B)

hypertrophy+vehicle, (C) control+ramipril, (D and E)(P<0.05).hypertrophy+ramipril, and from RNA isolated from(F) kidney and (G) liver. (b) PLB and 18S Dot Blotanalysis. (c) Northern blot of ANF, Calsequestrin, and

Protein abundance 18S ribosomal RNA. Lanes 1, controls+vehicle; Lane 2hypertrophy+vehicle; Lane 3, controls+ramipril; LaneProtein quantities were determined by Western blot4, hypertrophy+ramipril; Lane 5, rat kidney; Lane 6,

analysis and membranes stained with 0.15% Amido rat liver and; lane 7, atrium.black (BDH), [Figs 3(a),(b)]. The optical density ofthe amido black stained protein showed a linearresponse for total protein loading between 5 and

S. Y. Boateng et al.2688

Figure 2 (a) Relative abundance of CaATPase mRNAstandardized to 18S ribosomal RNA. CaATPase mRNAwas unchanged following cardiac hypertrophy, but in-creased significantly following ramipril treatment,hypertrophy+vehicle (n=6) v hypertrophy+ramipril(n=6) ∗P<0.05. SERCA also increased in the controlsfollowing ramipril treatment—control+vehicle (n=7) vcontrol+ramipril (n=5) ∗∗P<0.01. (b) PhospholambanmRNA standardized to 18S RNA from dot blot analysis.Phospholamban mRNA was unchanged following car-diac hypertrophy (n=6), but increased significantly fol-lowing ramipril treatment, control+vehicle orhypertrophy+ramipril (n=7) v hypertrophy+ramipril(n=6) ∗∗P<0.01. Phospholamban also increased in thesham operated group following ramipril treatment(∗∗P<0.01). Control+vehicle v control+ramipril (n=5).

40 lg. Similarly, the ECL signals for the antibodiesto PLB, CaATPase, ACE and ANF had a linearresponse between 10 and 40 lg [Fig. 3(c) and notshown]. When the parameters were plotted againsteach other, a linear response was found between Figure 3 (a) An Amido black stained membrane show-

ing proteins of 30 to 200 kDaltons. The stained membrane15 and 40 lg of total protein loading [Fig. 3(d)].was used to standardize against protein loading. (b)Similar experiments were performed using PVDFRelationship between Amido black staining and proteinmembranes and 0.2% Coomassie blue stain and theloading. (c) Relationship between the enhanced

results were identical (not shown). All subsequent chemiluminesence (ECL) signal for CaATPase and proteinexperiments were performed using 25 lg of total loading. (d) Relationship between optical density of totalprotein and expressed as a function of total protein protein staining and signal for CaATPase. All results are

from triplicate experiments.loading (Amido black staining) or to MHC as aninternal marker for cardiomyocytes.

SR Calcium ATPase and Regression of Cardiac Hypertrophy 2689

Table 2 CaATPase to phospholamban ratios

Treatment CaATPase/phospholambanproteins (arbitrary units)

controls+vehicle 1.05±0.21controls+ramipril 0.62±0.26hypertrophy+vehicle 0.29±0.04∗hypertrophy+ramipril 0.79±0.09∗∗

Following cardiac hypertrophy, the ratio of CaATPase tophospholamban was decreased by 70%. Controls+vehicle (n=7) v hypertrophy+vehicle (n=5) ∗P<0.05. Following ramipriltreatment, this ratio increased by more than 2.5-fold. Hypertrophy+vehicle v hypertrophy+ramipril (n=5)∗∗P<0.01. Ramipril did not significantly alter the ratio in thecontrol group receiving ramipril (n=5). Results are mean±.

fibrosis) signals were standardized to MHC. Whenexpressed in this manner, similar resultswere obtained (control+vehicle v hyper-trophy+vehicle, P<0.01 and hypertrophy+vehicle v hypertrophy+ramipril, P<0.05). Nochange was observed between any of the groupsFigure 4 (a) Western immunoblot of CaATPase. (b)when phospholamban was expressed as a ratio toQuantitation of CaATPase protein following Western blot-

ting standardized to total protein loading. In moderate MHC. Table 2 shows that the ratio of CaATPase tohypertrophy CaATPase protein was significantly reduced, phospholamban protein is significantly decreasedcontrols+vehicle (n=7) v hypertrophy+vehicle (n=5) following cardiac hypertrophy, (control+vehicle v∗P<0.01. Following ramipril treatment CaATPase protein

hypertrophy+vehicle P<0.05 n=5). Followingincreased significantly, hypertrophy+vehicle vramipril treatment this ratio increasedhypertrophy+ramipril (n=5) ∗P<0.05. Controls on

ramipril showed no change in CaATPase protein levels (hypertrophy+vehicle v hypertrophy+ramipril(n=5). P<0.01 n=5).

Functional studies

Calcium uptake into vesicles

Oxalate stimulated calcium uptakes were performedFigure 5 Quantitation of phospholamban protein fol- a minimum of five times on each homogenate inlowing western blotting, standardized to total protein either the presence or absence of ATP. Calciumloading. No significant difference was observed between

was used to initiate each reaction. Under theseany of the groups. Controls+vehicle (n=7),conditions calcium uptake is linear over a periodcontrols+ramipril (n=5), hypertrophy+vehicle (n=5)

and hypertrophy+ramipril (n=5). of 6–8 min and plateaus between 10 and 15 min(not shown). No difference could be demonstratedfor the time to plateau between any of the ex-perimental groups (not shown). In the absence ofIn contrast to the mRNA expression, no difference

in CaATPase protein levels could be demonstrated ATP or calcium, uptake was not detectable and thereaction could be completely inhibited by addition ofbetween controls on vehicle and those treated with

ramipril [Figs 4(a),(b)]. CaATPase levels sig- 10−7 thapsigargin. No non-specific uptake abovebackground levels (no ATP) could be demonstrated.nificantly decreased in the hypertrophied rat hearts,

but were normalized following ramipril treatment. The calcium precipitations measured under theseexperimental conditions were therefore the resultNo change in PLB protein expression could be

demonstrated between any of the groups (Fig 5). of SR CaATPase activity.Figure 6 shows the rate of calcium uptake (nmolTo determine whether the effects on CaATPase and

phospholamban proteins were the result of changes of Ca per mg of protein per min) into homogenatesof SR vesicles. When compared to controls onin the non-myocyte protein content (for example,

S. Y. Boateng et al.2690

could be demonstrated between any of the groupsin the levels of tissue ACE protein when measuredby Western blotting (Table 3). A 25% increase intissue ACE activity was found (Table 3; P<0.05) inhearts from the hypertrophied group relative tocontrols on vehicle. After ramipril treatment, car-diac ACE activity returned to control levels even inthe presence of an increased afterload. Serum ACEactivity was not significantly altered in any of thegroups, suggesting that ramipril at this con-centration inhibited tissue ACE more effectivelythan serum ACE.

Lung has high ACE activity and the pulmonaryFigure 6 Oxalate stimulated calcium uptake into the capillary bed drains into the left ventricle, thereforeSR from crude heart homogenates. In cardiac hyper- any change in ACE activity in this organ mighttrophy calcium uptake was significantly reduced, influence the amount of AgII reaching the heart.controls+vehicle (n=8) v hypertrophy+vehicle (n=5)

Rat heart homogenates have an ACE activity∗P<0.05. Following ramipril treatment, calcium uptake(Vmax/Km) of 0.138±0.01 (n=5) compared withwas normalized in hypertrophied hearts

[controls+vehicle v hypertrophy+ramipril (n=6) 5.85±0.19 (n=5) in lung. This represents aP>0.2]. 40-fold difference between the two organs. Lung

ACE activity was measured in the experimentalgroups and was expressed as Vmax/Km: vehicle

vehicle, the rate of oxalate stimulated calcium up- treated, 5.85±0.19 (n=5); controls+ramipril,take was decreased by 30% in hypertrophied hearts, 4.38±0.20 (n=5); and hypertrophy+ramipril,but was normalized following ACE inhibitor treat- 4.12±0.11 (n=11). (Vehicle treated vment. These data are very similar to that seen control+ramipril, P<0.001; and vehicle treatedfor the protein and indicate that the decrease in v hypertrophy+ramipril, P<0.001.) These resultsobserved activity was due to a decrease in SR show that lung ACE activity was 30% lowerCaATPase protein. following ramipril treatment. To determine which

site within the myocardium might be importantfor ACE inhibition, the enzyme levels were alsomeasured in isolated myocytes obtained fromACE activitycontrol hearts. Compared with whole heart homo-genate, myocytes isolated from rat hearts of theTo partially determine where the actions of sub-same age have negligible ACE activity (not shown).antihypertensive doses of ramipril might be re-These isolated myocyte preparations were func-sponsible for altering SR CaATPase abundance, ACEtional, since oxalate stimulated calcium uptakeprotein and activity were measured in serum, and

in heart and lung tissues. No significant difference showed activities greater than those in whole

Table 3 Levels of ACE and ANF

Treatment Cardiac tissue ANF Cardiac tissue ACE Tissue ACE activity Serum ACE activity(arbitrary units) (arbitrary units) (Vmax/Km/mg) (units per litre)

controls+vehicle 0.021±0.014 0.113±0.034 0.138±0.01 169±14.3controls+ramipril n n 0.149±0.006 157±20.0hypertrophy+vehicle 0.432±0.274∗∗ 0.141±0.01 0.163±0.006† 183.4±8hypertrophy+ramipril 0.134±0.064∗ 0.135±0.064 0.129±0.005‡ 175±9.8

In moderate hypertrophy the cardiac tissue ANF protein levels (determined by Western blotting) were significantly increased.Controls+vehicle (n=5) v hypertrophy+vehicle (n=7) ∗∗P<0.01. Following ramipril treatment ANF protein levels were significantlyreduced in hypertrophied hearts, hypertrophy+vehicle v hypertrophy+ramipril (n=6) ∗P<0.05. Using Western blotting the 140kDacardiac ACE protein was found to be unchanged. No significant differences could be found between the various groups. Controls+vehicle(n=6), hypertrophy+vehicle (n=6) and hypertrophy+ramipril (n=5). Following hypertrophy the rate constant of the enzyme wassignificantly increased, control+vehicle (n=5) v hypertrophy+vehicle (n=7) †P<0.05. After ramipril treatment the enzyme activitywas reduced to the level of controls, hypertrophy+vehicle (n=5) v hypertrophy+ramipril (n=8) ‡P<0.01. Serum ACE activity wasmeasured and no significant difference was observed between any of the groups. Controls+vehicle (n=10), controls+ramipril (n=10), hypertrophy+vehicle (n=10) and hypertrophy+ramipril (n=21).n—not measured.

SR Calcium ATPase and Regression of Cardiac Hypertrophy 2691

heart. Decreases in non-myocyte cardiac or lung we provide clear evidence that ACE inhibitioncan normalize CaATPase protein and calciumACE activity following treatment with ramipril

therefore appear to be the primary mechanism uptake to normal. The mechanism apparently isindependent of changes in blood pressure andwhereby SR CaATPase expression and activity are

modified in the hypertrophied myocardium. may involve a change in the mRNA abundance,as SR CaATPase transcripts can be increasedmerely through ACE inhibition.

What is the mechanism responsible for the dis-Discussioncrepancy between mRNA and protein levels inthis model? Ramipril might increase SR CaATPaseRamipril treatment of rats subjected to abdominal

aortic constriction normalizes SR CaATPase protein transcription or mRNA stability and/or decreasethe translation of its mRNA or the half life of SRabundance independently of changes in blood pres-

sure. The normalization in CaATPase protein con- CaATPase protein. In a severe model of pressure-overload induced cardiac hypertrophy, we havetent occurred without a significant decrease in

cardiac mass or relative to MHC protein expression. shown that SERCA2 gene transcription is activelydecreased (Ribadeau-Dumas et al., 1998): however,Additionally, treatment with ramipril does not re-

duce isolated cardiomyocyte cell size following car- during the perinatal period we find an increase inthe mRNA and protein abundance in the absencediac hypertrophy and, following a single twitch,

myocyte relaxation is about 40% slower compared of transcriptional activation (manuscript in pre-paration). One possible mechanism accounting forwith controls, but is normalized following ramipril

treatment (Boateng, unpublished data). These data altered mRNA stability might involve the 3′ splicingvariant of SERCA2a transcripts recently found insuggest that changes in SR CaATPase abundance

and function result in improved myocyte relaxation heart (Koban et al., 1997). Definitive proof is, how-ever, lacking. Another possibility is the potentialthat is not secondary to cellular regression. SR

CaATPase mRNA levels were also unchanged in for increased amounts of SERCA2b. It is clear thatACE inhibitors can modify the expression of manythis model of cardiac hypertrophy, but ramipril

increased the amount of these transcripts in both gene products. As we have not systematically meas-ured the quantities of SERCA2a and 2b from thesecontrol and hypertrophied hearts without altering

the expression of calsequestrin mRNA. Ramipril samples, the increase in mRNA seen in the groupstreated with ramipril might be due to a differentialtreatment therefore can selectively alter CaATPase

mRNA expression independently of its effects on effect on SERCA2 gene expression in either themyocytes (2a) or the non-myocytes (2b). Re-protein expression. Regulation of SR CaATPase ex-

pression and function thus involves at least two gardless, it is clear from these results that SERCA2mRNA abundance should not be used as a markermechanisms, both of which can be modulated by

ramipril. of CaATPase protein or activity. Finally, others havereported that mRNA levels of actin and MHCs doMeasurements of total SR CaATPase mRNA

should not be used as an indicator of SR CaATPase not necessarily produce corresponding differences intheir proteins in the adult human heart (Coumans etprotein expression or function. De la Bastie et al.,

(1990) previously showed that in moderate cardiac al., 1997). In interpreting these results the authorsfelt that differences in turnover of these sarcomerichypertrophy, CaATPase mRNA levels did not

change but CaATPase function decreased. A re- proteins might account for the observed dis-crepancy; however, any such potential mechanismduction in mRNA was only seen in severe

hypertrophy which may have been indicative of is not understood. Whether ACE inhibition playsa role on any of these mechanisms is entirelydecompensation and failure (Feldman et al., 1993).

Changes in mRNA expression have been taken unknown.Changes in the relative abundance of CaATPaseas an indication of altered CaATPase function.

Zierhut et al. (1996) showed in a 2 kidney 1 to PLB might also account for some of the observedfunctional changes. For example, a higher ratio ofclip model of rat cardiac hypertrophy that the

‘down regulation’ of SERCA mRNA could be CaATPase to phospholamban leads to significantlyshorter times to half relaxation in isolated musclereversed by antihypertensive doses of the ACE

inhibitor benazepril or by addition of the AT1 strips (Koss et al., 1995). The absence of phos-pholamban also results in enhanced myocardialreceptor antagonist Valsartan. Likewise, Bruck-

schlegel et al. (1995) showed that Losartan could performance and significantly decreases the timeto peak pressure and time to half relaxation (Luoblunt the decrease in SR CaATPase mRNA in a

model of severe left ventricular hypertrophy. Here et al., 1994). Based on these observations, our

S. Y. Boateng et al.2692

protein data predict that in cardiac hypertrophy a used in conjunction with ACE inhibition. Thesestudies imply that the effects of bradykinin duringdecreased ratio would lead to slower relaxation,

but following ramipril treatment the relaxation ACE inhibition may be of importance.In conclusion, ramipril has beneficial effects onwould be returned towards normal. This is exactly

what is seen. Phosphorylation of PLB may also SR CaATPase protein abundance and function inthe presence of high blood pressure. These findingsaccount for some of the changes in CaATPase

function seen in vivo. clearly explain some of the beneficial effects of ACEinhibitors on calcium handling in compensatedcardiac hypertrophy and may partially explainmechanistically why ACE inhibitors are associatedMechanism of diastolic improvement following ACE

inhibition with an improved prognosis in humans and animalswith cardiac hypertrophy and failure (Weinberg et

Upregulation of ACE expression and activity has al., 1994; Hirsch et al., 1992). We hypothesize thatramipril affects SERCA2 and PLB mRNA contentbeen implicated in the cardiac dysfunction as-

sociated with cardiac hypertrophy and failure either through transcriptional activation of theSERCA2 gene or through changes in the mRNA(Schunkert et al., 1990; Hirsch et al., 1991). ACE

activity is elevated following hypertrophy; however, half-life of either SERCA2a or 2b, whereas thedecrease in SR CaATPase protein in cardiac hyper-we cannot attribute this increase in activity to an

increase in enzyme expression. The protein levels trophy may be due to translational or post-trans-lational mechanisms. The precise signallingwere not significantly altered. A post-translational

modification such as phosphorylation of the ACE mechanisms through which ramipril exerts its ac-tions on SERCA mRNA and protein abundanceenzyme might explain the differential effect of

ramipril on hypertrophied v control hearts. Indeed however remain to be elucidated.Schunkert et al. (1993) observed that followingintravenous infusion of AgII in rats, ACE mRNAdecreased to a greater extent than the activity of

Acknowledgementsthe enzyme. The rat myocardium was also foundto have an endogenous inhibitor of ACE (Ikemoto et

This study was supported by the British Heartal., 1989) the effects of which are unclear. AlthoughFoundation (KRB/AMS/MD: PG/94013 and PG/inhibition of cardiac ACE activity might be involved95093). The authors would like to thank Davidin the increase in CaATPase protein followingShepherd for technical assistance, Fison Pharma-ramipril treatment in hypertrophy, this does notceuticals for the gift of the ACE inhibitor (ramipril)appear to explain all the effects seen in these ex-and Ruby Naqvi for isolation of myocytes.perimental groups.

The site of ACE inhibition within the myocardiummay not be the myocytes themselves, but endo-thelial cells or fibroblasts, as isolated myocytes have Referenceslittle or no ACE activity. Lung ACE activity wasreduced in all groups receiving ramipril. Since the A M, L F, C D, D P, S

E, L AM, L Y, 1995. Sarcoplasmiclung has high ACE activity, it might regulate thereticulum Ca2+ pumps in heart and diaphragm ofamount of Ag II and bradykinin reaching the leftcardiomyopathic hamster: effects of perindopril. Am Jventricle. Alternatively, ACE inhibitors may exertPhysiol 268: H1947–H1953.

their effects by altering the level of catecholamines A M, M H, P M, 1994. Sarcoplasmicand prostaglandins (Bristow and Abraham, 1995; reticulum gene expression in cardiac hypertrophy and

heart failure. Circ Res 74: 555–564.Schror, 1992). The use of ACE inhibitors in bothB K, B G, D D, 1992. Cardiac action ofexperimental and clinical settings results in the

angiotensin II: role of an intracardiac renin-angio-accumulation of kinins in addition to the effect oftensin system. Annu Rev Physiol 54: 227–241.

reducing Ag II synthesis (Gohlke et al., 1994; Linz B LH, 1988. Cellular protein databases: a new ap-and Scholkens, 1992a,b). If bradykinin receptor proach to the study of genomic organisation and func-

tion of the cell. Leukaemia 2: 559.antagonists are used in conjunction with ACE in-B SY, S AM, D MJ, Y MH, Bhibitors, some of the beneficial effects are lost

KR, 1997. Inhibition of endogenous cardiac phos-(Gohlke et al., 1994; Linz and Scholkens, 1992b).phatase activity and meaurement of sarcoplasmic

Holtz et al. (1992) also indicated that the up- reticulum calcium uptake: A possible role of phos-regulation of CaATPase mRNA was partially pholamban phosphorylation in the hypertrophied myo-

cardium. Biochem Biophys Res Comm 239: 701–705.blocked if a bradykinin receptor antagonist was

SR Calcium ATPase and Regression of Cardiac Hypertrophy 2693

B MM, 1976. A rapid and sensitive method for Ca(++)-ATPase in cardiac hypertrophy by an-giotensin converting enzyme? Basic Res Cardiol 87the quantification of microgram quantities of proteinSuppl 2: 191–204.utilising the principle of protein-dye binding. Anal Bio-

I F, S GB, T M, Y K, 1989.chem 72: 248–254.Endogenous inhibitor of angiotensin converting en-B MR, A WT, 1995. Anti-adrenergic effectszyme in the rat heart. Biochem Biophys Res Communof angiotensin converting enzyme inhibitors. Eur Heart159: 1093–1099.J 16: 37–41.

K MU, B SY, B KR, 1997. The role ofB G, H SR, J K, G D,the sarcoplasmic reticulum Calcium ATPase and itsM F, K EP, R GAJ, S H,regulatory protein phospholamban in the developing1995. Blockade of the renin-angiotensin system inmyocardium. In: Ostadal B, Nagano M, Takeda N,cardiac pressure-overload hypertrophy in rats. Hyper-Dhalla NS (eds) The Developing Heart. Lippincott-Ravenstension 25: 250–259.Publishers, New York, pp. 143–164.C JD, W CA, 1985. Plasmalemmal meta-

K KL, P S, J WK, G IL, K EG,bolic activities in cultured calf pulmonary arterial endo-1995. Differential phospholamban gene expression inthelial cells. Res Commun Chem Pathol Pharmacol 50:murine cardiac compartments. Molecular and physio-163–179.logical analyses. Circ Res 77: 342–353.C P, S N, 1987. Single-step method

L N, B CS, D MJ, R ML, B P, Yof RNA isolation by acid guanidinium thiocyanate-MH, 1993. Frequency and specificity of antiheart anti-phenol-chloroform extraction. Anal Biochem 162: 156–bodies in patients with dilated cardiomyopathy detected159.using SDS-PAGE and western blotting. J Am Coll CardiolC JVF, Y T, S RK, K A, B K,22: 1378–1384.G P, H E, R CG, 1997.

L CJ, O MT, G IF, L TB, BVariations in the relative mRNA levels of actins andDG, S A, 1987. Calcium uptake by cardiac sar-myosin heavy chains do not produce correspondingcoplasmic reticulum in human dilated cardio-differences in their proteins in the adult human heart.myopathy. Cardiovasc Res 21: 601–605.J Mol Cell Cardiol 29: 895–905.

L W, S J, W G, A G, S BA, B D, L D, R L, M J,1992. Ramipril prevents left ventricular hypertrophyM F, W C, B B, Swith myocardial fibrosis without blood pressure re-K, L A, 1990. Function of the sarcoplasmicduction: on one year study in rats. Br J Pharmacol 107:reticulum and expression of its Ca2+-ATPase gene in970–975.pressure overload-induced cardiac hypertrophy in the

L W, S BA, 1992a. Role of bradykinin inrat. Circ Res 66: 554–564.the cardiac effects of angiotensin-converting enzymeD MJ, 1987. Two-dimensional polyacrylamide gelsinhibitors. J Cardiovasc Pharmacol 20 Suppl 9: S83–90.for the construction of protein databases. In: Bryce L W, S BA, 1992b. A specific B2-bradykininCFA (ed) Microcomputers in Biochemistry. IRL Press, receptor antagonist HOE 140 abolishes the anti-

Oxford, pp. 215. hypertrophic effect of ramipril. Br J Pharmacol 105:D VJ, 1992. Autocrine and paracrine mechanisms in 771–772.

the pathophysiology of heart failure. Am J Cardiol 70: L A, B D, B KR, S K,4C–11C. 1989. Characterization and expression of the rat heart

F AM, W EO, R PE, L BH, 1993. sarcoplasmic reticulum Ca2+-ATPase mRNA. FEBS LettSelective changes in cardiac gene expression during 249: 35–41.compensated hypertrophy and the transition to cardiac L A, A M, L D, 1994. Sarco-decompensation in rats with chronic aortic banding. (endo)plasmic reticulum calcium pumps in the cardio-Circ Res 73: 184–192. vascular system: function and gene expression (Re-

G P, L W, S BA, K I, B view). J Mol Cell Cardiol 26: 1109–1121.S, S A, U T, 1994. Angiotensin-converting L W, G IL, H J, P S, G G, Denzyme inhibition improves cardiac function. Role of JJ, D T, K EG, 1994. Targeted ablationbradykinin. Hypertension 23: 411–418. of the phospholamban gene is associated with markedly

H G, R H, S R, M M, P B, enhanced myocardial contractility and loss of b-agonistH J, H C, P H, J H, D stimulation. Circ Res 75: 401–409.H, 1994. Relation between myocardial function and M JB, M JJ, G FX, U R, Dexpression of sarcoplasmic reticulum Ca2+-ATPase in JC, P M, C P, 1990. Urinary cyclic guan-failing and nonfailing human myocardium. Circ Res osine monophosphate as an indicator of experimental75: 434–442. congestive heart failure in rats. Cardiovasc Res 24:

H AT, T CE, S H, P M, D VJ, 946–952.1991. Tissue-specific activation of cardiac angiotensin M S, H T, 1994. Hypertrophic growth ofconverting enzyme in experimental heart failure. Circ cultured neonatal rat heart cells mediated by type IRes 69: 475–482. angiotensin II receptor. Am J Physiol 266: H2443–

H AT, T CE, S AD, S H, I- H2451. JR, D VJ, 1992. Differential effects of cap- M K, T M, H S, N S, S H,topril and enalapril on tissue renin-angiotensin systems W H, O M, F S, 1992. Effects ofin experimental heart failure. Circulation 86: 1566– losartan, a nonpeptide angiotensin II receptor ant-1574. agonist, on cardic hypertrophy and the tissue angio-

H J, S R, R H, J H, D H, 1992. tensin II content in spontaneously hypertensive rats.Life Sciences 51: 367–374.Modulation of myocardial sarcoplasmic reticulum

S. Y. Boateng et al.2694

M AF, V JL, K MU, S K, verting enzyme inhibitors. J Cardiovasc Pharmacol 20Suppl 9: S68–73.L WH, B KR, 1995. Patterns of expression

of sarcoplasmic reticulum Ca(2+)-ATPase and phos- S H, D VJ, T SS, H AT, A CS,L BH, 1990. Increased rat cardiac angiotensinpholamban mRNAs during rat heart development. Circ

Res 76: 616–625. converting enzyme activity and mRNA expression inpressure overload left ventricular hypertrophy. J ClinM JP, E RE, A PD, G W, G

JK, 1990. Abnormal intracellular calcium handling: a Invest 86: 1913–1920.S H, I JR, H AT, P Y, Rmajor cause of systolic and diastolic dysfunction in

ventricular myocardium from patients with heart fail- WJ, J H, D VJ, 1993. Feedback regulation ofangiotensin converting enzyme activity and mRNAure. Circulation 81 Suppl III: 21–32.

M MA, S RH, 1998. Calcium se- levels by angiotensin II. Circ Res 73: 312–318.S R, R H, R I, E T, Bquestration by the sarcoplasmic reticulum in heart

failure. Cardiovasc Res 37: 352–359. M, H G, H J, D H, 1994. Geneexpression of the cardiac NA+-CA2+ exchanger in end-N RU, ML KT, 1994. Effect of hypertrophy on

mechanisms of relaxation in isolated cardiac myocytes stage human heart failure. Circ Res 75: 443–453.T FT, M AC, H HK, Z B, Pfrom guinea pig. Am J Physiol 267 (36): H1851–

H1861. L, W JT, B LM, 1991. Intracellular calciumtransients and arrhythmias in isolated heart cells. CircP M, W DG, B KR, Y MH, 1995.

Clenbuterol induces hypertrophy of the latissimus dorsi Res 69: 810–819.W EO, S FJ, G D, K Y, Dmuscle and heart in the rat with molecular and pheno-

typic changes. Circulation 92: 483–489. PS, L SE, S H, B CR, LBH, 1994. Angiotensin-converting enzyme inhibitionR-D A, W C, B KR, T

K H, F MY, S K, 1997. The sarco- prolongs survival and modifies the transition to heartfailure in rats with pressure overload hypertrophy due(endo)plasmic reticulum Ca2+-ATPase gene is regulated

at the transcriptional level during compensated left to ascending aortic stenosis. Circulation 90: 1410–1422.ventricular hypertrophy in the rat. C.R. Acad. Sci. Paris

320: 963–969. Z W, S R, L D, K S, AP, W S, C F, B H-P, G M,R HA, W SP, M L, R, J. 1994. ANG

II receptor blockade prevents ventricular hypertrophy D H, 1996. Left ventricular wall stress andsarcoplasmic reticulum Ca2+-ATPase gene expressionand ANF gene expression with pressure overload in

mice. Am J Physiol 266: H2468–H2475. in renal hypertensive rats:dose-dependent effects ofACE inhibition and AT1-receptor blockade. CardiovascS K, 1992. Role of prostaglandins in the cardio-

vascular effects of bradykinin and angiotensin-con- Res 31: 758–768.