MAGNETIC RESONANCE IN MEDICINE

17,69-8 1 ( 1991 )

31PNMR Spectroscopy in Chronic Adriamycin Cardiotoxicity VERA

BITTNER,* RUSSELLc. REEVES,STANLEY B. DIGERNESS,~ JAMES B. CAULFIELD,$ AND GERALD M. POHOST

Division of Cardiovascular Disease, t Division of Cardiac Surgery, .#Department of Pathology, University of Alabama at Birmingham, Birmingham, Alabama 35294 Received August 19, 1989; revised December 5 , 1989 Abnormal cardiac energy metabolism has been postulated as a mechanism for adriamycin induced cardiotoxicity. This study was designed to determine high energy phosphate stores at rest and with hemodynamic stress in perfused rat hearts after animals had been chronically exposed to adriamycin ( 2 mg/ kg weekly for 14 weeks). Morphologic and hemodynamic changes were mild in this model. Phosphorus-3 1 NMR determined intracellular pH and levels of inorganic phosphate (Pi) and ATP were comparable in treated and control hearts. Phosphocreatine (PCr) levels were markedly decreased in treated hearts (0.89 k 0.07 units/g versus 1.7 f 0.13 units/g, p < 0.001). The PCr/Pi ratio decreased in both groups during hemodynamic stress. It recovered earlier in controls and there was a marked overshoot after cessation of rapid pacing in this group which was not present in adriamycin treated hearts. These results suggest that metabolic regulation in response to hemodynamic stress is impaired after chronic adriamycin exposure. PCr depletion and delayed metabolic recovery after hemodynamic stress appear to be potentially useful markers for the effect of adriamycin on the heart. 0 1991 Academic Press, Inc. INTRODUCTION

Doxorubicin hydrochloride ( adriamycin ) is a bacterium derived anthracycline antibiotic that is widely and successfully used as an antineoplastic agent. Its use is limited by dose related cardiotoxicity with an incidence of 1% in patients treated with cumulative doses below 500 mg/m2 and an incidence of over 30% in patients treated with greater than 600 mg/m2 (1, 2). Noninvasive methods to date have provided neither a reliable nor a sensitive tool for early identification of cardiotoxicity (3-8). Presently, the diagnostic procedure of choice is endomyocardial biopsy applied serially. The biopsy procedure is costly, can be uncomfortable, has a slight risk, and is not universally available ( 8-1 0). Drug administration is, therefore, frequently empirically limited to less than 550 mg/m2 in clinical practice (2). Since individual patient susceptibility vanes widely, this precaution does not prevent irreversible myocardial damage in all patients and results in premature withdrawal of potentially beneficial therapy in others who could tolerate substantially higher doses of the drug. Nuclear magnetic resonance (NMR) techniques have shown promise for the detection of adriamycin cardiotoxicity. Prolongation of the proton relaxation parameter, T1, has been demonstrated in toxic rat hearts (11). Abnormalities in high energy

* Supported in part by a Southern Medical Association Research Project Grant. 69

0740-3 194/91 $3.00 Copyright Q 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.

70

BITTNER ET AL.

phosphate levels have been demonstrated in acute adriamycin cardiotoxicity, while results are inconclusive in more chronic models (12-15). This study in rats was designed to expand on previous studies of myocardial high energy phosphate levels after chronic exposure to adriamycin by examining these levels not only at rest but also during hemodynamic stress. METHODS

Study Population and Induction of Adriamycin Toxicity The study utilized 18 4-week-old male Fisher rats (Charles River Laboratories) housed 5 to 6 animals per cage with unlimited access to food and water. Ten animals received 2 mg/ kg of adriamycin (Adria Laboratories) subcutaneously at rotating injection sites weekly for 14 weeks. Eight animals served as controls and received equivalent volumes of subcutaneous saline during the study period. All animals were weighed weekly during this induction period and again immediately prior to sacrifice. There were no deaths during the induction period. Cannulation of the aorta in preparation for the Langendorff perfusion was not successful in 4 animals (2 in each group). Thus, 8 adriamycin treated rats and 6 control rats form the basis of this report. This study was reviewed and approved by the Animal Use Review Committee of the University of Alabama at Birmingham. Animal use conforms to the guiding principles of the American Physiological Society.

Isolated Perfused Rat Heart Preparation Animals were heparinized and anesthetized with sodium pentobarbital. The beating heart was rapidly excised and immediately placed into iced Krebs Henseleit buffer. The aorta was cannulated and the heart perfused with modified phosphate-free Krebs Henseleit buffer (121 mMNaC1, 23 mMNaHC03, 5.9 mMKC1, 1.75 m M CaC12, 1.2 m M MgCI2, 0.5 m M H4EDTA, 11 m M glucose bubbled with 95% O2 and 5% C02) at 37°C and 100 mm Hg perfusion pressure. A fluid filled latex balloon was placed into the left ventricle through a left atriotomy and sutured to the mitral annulus. Left ventricular pressures were recorded on a Hewlett-Packard strip chart recorder. Pacing electrodes were inserted into the right atrium and through a pulmonary arteriotomy into the right ventricular outflow tract. A capillary tube filled with methylene diphosphonate solution (MDP) was affixed close to the heart as a phosphorus-31 ( 31P) reference standard. The heart was submerged in its effluent inside an NMR glass tube and placed into the vertical 3’PNMR probe. The effluent level was held constant by an adjustable roller pump. At the time of sacrifice open biopsies of lung tissue and liver were also obtained and fixed in formalin for histological evaluation.

Hemodynamic Measurements Heart rate, coronary blood flow, and left ventricular pressure were recorded after an initial equilibration period and during each stage of the experimental protocol. The 65-min protocol consisted of 13 consecutive 5-min stages (Fig. 1 ). In the “Preload Experiment,” hearts were paced at 300 beats/min. Balloon volume was increased over a 20-min period in 0.05 ml increments from “0 volume” to a maximum of 0.15

71

CHRONIC ADRIAMYCIN CARDIOTOXICITY

r

1

BalloonVolume (cc) Pacing Rate (beats/min)

0 I

i

1

1

,-, 1

0.05; 0.1 I0.151 0 I I

I I

i

I !

1

0

1

0 I

1

0

1

l

;

!

l

1

1

0

,

i

I

0

0

-- .

i

!

1

0

0

I

j

i

I

1

1

1

300 300 300 300 j 300 300 330 j 360 390 420 300 300 300

j

I

]

I

1

1 1

I ! 1

1

1

!

ml. This was followed by a 5-min recovery period at 0 volume and a pacing rate of 300 beats/min. In the “Pacing Experiment,” balloon volume was maintained at 0 volume while the pacing rate was increased in 30 beat increments from 300 beats/ min to a maximum of 420 beats/min. After the pacing experiment, the pacing rate was reduced to 300 beats/min and balloon volume maintained at 0 volume for three consecutive 5-min recovery periods (Recovery 1-3). After completion of the protocol, hearts were immediately removed from the perfusion apparatus. Aorta, remaining pulmonary arteries, and atria were removed and the hearts blotted dry prior to determining their wet weights. The hearts were then “breadloaved” from base to apex. The apical slices were weighed, dried at 100°C under vacuum for 72 h, and weighed again. The apical slice wet to dry weight ratio was used as the sample to represent the whole heart. The remaining heart slices were fixed in formalin and glutaraldehyde for light and electron microscopy.

Phosphorus-31 NMR Spectroscop.v Phosphorus-31 NMR spectra were obtained on a Bruker 360 AM vertical widebore spectrometer (Bruker Corp., Billerica, MA) utilizing a conventional 20-mm Bruker Helmholtz select frequency phosphorus coil which has a homogeneous Bl field across the entire probe. Hearts were carefully positioned within the sensitive volume. After the spectrometer was tuned and shimmed, data were acquired for 5 min during each stage of the protocol. Acquisition was not gated to heart rate. Hearts were pulsed using 33-ys (90”) RF pulses with a repetition time of 2 s. One hundred fifty scans were averaged over each 5-min period to optimize signal to noise ratio over the @adenosine triphosphate ( ATP) signal. The resultant spectra were plotted. The areas underneath the inorganic phosphate, phosphocreatine, and ATP peaks were determined by triangulation and normalized to the MDP reference standard and dry heart weight. Values are expressed in units per gram dry heart weight. The ATP/PCr ratio and the PCr/Pi ratio were computed for each stage of the experiment. The chemical shift of the Pi peak relative to PCr expressed in “ppm” was used to estimate intracellular pH utilizing a modification of the Henderson-Hasselbalch equation ( 1 6 ) :pH = 6.77 log((ppm - 3.29)/(5.68 - ppm)). To estimate peak saturation, fully relaxed spectra were obtained utilizing 33-ys

+

72

BITTNER ET AL.

pulses with a repetition time of 25 s before and after the experimental protocol in subsets of adriamycin treated and control animals. Saturation of the PCr peak remained constant during the experiment (39.9 4.4% prior to the preload experiment versus 38. I k 3.7%after the third recovery period; n = 7 , p = NS) and there was no significant differencebetween the two groups of animals (adriamycin group ( n = 5); 36.4 f 3.8%; control group ( n = 4); 4 1.7 4.1%; p = NS). Similarly, there was no significant difference between the two groups of animals in the degree of saturation of the ATP peak (adriamycin group ( n = 5); 34.5 9.9%; control group ( n = 4); 38.5 ? 5.9%; p = NS). The fully relaxed spectra had a lower signal to noise ratio since the number of scans had to be limited due to time considerations. The Pi peak was not consistently visible in the fully relaxed spectra and thus a saturation factor could not be calculated. Since the degree of ATP and PCr peak saturation was comparable in both groups, comparison of values obtained by peak integration should remain valid even without correcting for saturation. Therefore, “raw” data, uncorrected for saturation, are presented in this paper.

*

*

*

Statistical Methods Mean values and standard errors were computed for each variable. Values are expressed as mean standard error. The significance of differences between measurements at different experimental stages within each study group was assessed using paired Student’s t-tests and repeated measures ANOVA. Group mean values were compared by split plot ANOVA, Duncan’s multiple range test, and two tailed unpaired Student’s t-tests. A p value of less than 0.05 was considered significant.

*

RESULTS

Animals in both control and treatment groups were sacrificed a mean of 24 k 5 days after the final injection of adriamycin or saline (range 17 to 34 days). Heart Weight and Histology At sacrifice, overt heart failure was not observed in adriamycin treated animals nor was there any evidence of passive congestion on histologic examination of liver and lung samples. Body weight at sacrifice was significantly higher in the control group (305 f 4 g versus 212 -t 4 g, p < 0.001 ). Even though absolute heart wet weight was slightly higher in the control group (0.93 k 0.02 g versus 0.86 f 0.03 g, p = NS), heart wet weight as expressed as a percentage of body weight at sacrificewas significantly higher in the adriamycin group (0.4 0.0 14%versus 0.3 0.006%,p < 0.000 1 ). This difference is due to a higher wet to dry heart weight ratio in the adriamycin group than in the control group (6.84 -t 0.32 versus 5.68 k 0.26, p < 0.02). Histological findings were similar to those previously published ( 17-1 9 ) . Perfusion artifacts such as interstitial edema were also apparent in histologic sections from both groups of hearts. Toluidine blue stained light microscopic sections of adriamycin treated hearts (Fig. 2) showed vacuolization of the cytoplasm, myocyte atrophy, and an increase in interstitial matrix. A patchy increase of collagen deposition around myocytes and

*

*

CHRONIC ADRIAMYCIN CARDIOTOXICITY

73

FIG.2. Toluidine blue stained light microscopic section through an adriamycin treated heart (magnification of myocytes (V), myocyte atrophy, and increase in interstitial matrix ( M ) are seen. Myocyte necrosis is not apparent. BOOX). Vacuolization

groups of myofibers was seen by scanning electron microscopy (Fig. 3 ) . Mitochondria appeared normal by transmission electron microscopy.

Hemodynamic Findings Baseline heart rate, coronary blood flow, and developed left ventricular pressure were comparable in the two groups. Coronary blood flow did not change significantly during the protocol. Rate pressure products (heart rate times systolic pressure in mmHg/min) during the preload and pacing experiments and during recovery are shown in Fig. 4. Control hearts tended to have higher rate pressure products. The

74

BITTNER ET AL.

FIG.3. Scanning electron micrographs (magnification 19OOX) of a control heart ( a) and an adriamycin treated heart (b). A fine weave of collagen is seen overlying myocytes in the control heart. Collagen deposition is markedly increased in the adriamycin treated heart.

increase in rate pressure product with increasing pacing rates was more pronounced in control hearts ( p < 0.05 1.

Phosphorus-31 NMR Measurements Representative 31PNMR spectra are shown in Fig. 5. A marked decrease in the PCr peak was generally evident in the adriamycin treated hearts. High energy phosphate

a

1.04

-

0

0

005

01 LOAD

015

0

of

I

I

300

330

I

360

PACING-

I

I

t

I

I

390 420 Rec 1 R e c Z R e c 3 LRECOVERY 1

FIG.4. Mean rate pressure products (mmHg/min X 10,000) are plotted for both groups of rats for each stage of the protocol. Open circles represent adriamycin treated hearts and closed circles, control hearts. Error bars indicate standard errors. Results of the preload experiment are shown in A and those of pacing and recovery in B. Control hearts tended to have higher rate pressure products. The increase in rate pressure product with increasing pacing rates was more pronounced in control hearts ( p < 0.05).

75

CHRONIC ADRIAMYCIN CARDIOTOXICITY

20.0

15.0

20.0

15.0

10.0

10.0

5.0

5.0

0.0 PPM

-5.0

0.0 PPM

-10.0

-5.0

-15.0

-10.0

-20.0

-15.0

FIG. 5. Representative 3'P NMR spectra of an adriamycin treated heart ( A ) and a control heart (B). Spectral peaks are labeled Ref.; inorganic phosphate, Pi; phosphocreatine, PCr ; and the three adenosine triphosphate peaks, ATP. Note the significant decrease in PCr in the adriamycin treated heart.

levels in units per gram dry heart weight, metabolite ratios, and pH are shown in Table 1. In adriamycin treated hearts, PCr levels were significantly decreased to approximately 50% compared to those of control hearts. ATP levels in adriamycin treated hearts were slightly lower than those in control hearts but the difference did not achieve statistical significance. Thus, largely due to the decrease in PCr levels, the ATP/PCr TABLE I High Energy Phosphate Levels and Ratios Parameter

ADR ( n = 8)

CTRL (n = 6)

p Value

PCr W/g) ATP ( U / d Pi W/g) ATP/PCr PCr/Pi PH

0.89 & 0.07 1.45 f 0.07 0.52 f 0.04 1.78 f 0.15 2.18 f 0.28 7.1 ? 0.01

1.7 f 0 . 1 3 1.64 -t 0.15