Journal of Pharmacological and Toxicological Methods 66 (2012) 43–51
Contents lists available at SciVerse ScienceDirect
Journal of Pharmacological and Toxicological Methods
j ourna l homepage: www.e lsev ie r .com/ locate / jpharmtox
Original article
Optimising conditions for studying the acute effects of drugs on indices of cardiaccontractility and on haemodynamics in anaesthetized guinea pigs
Laura Mooney a, Louise Marks b, Karen L. Philp b, Matthew Skinner b, Susan J. Coker a,⁎, Susan Currie a
a Strathclyde Institute of Pharmacy and Biomedical Sciences, University of Strathclyde, 161 Cathedral Street, Glasgow, G4 0RE, UKb Safety Assessment UK, AstraZeneca R&D, Alderley Park, Macclesfield, SK10 4TG, UK
Introduction: Detecting adverse effects of drugs on cardiac contractility is becoming a priority in pre-clinical safety pharmacology. The aim of this work was to optimise conditions and explore the potential ofusing the anaesthetized guinea pig as an in vivo model. Methods: Guinea pigs were anaesthetized withHypnorm/Hypnovel, isoflurane, pentobarbital or fentanyl/pentobarbital. The electrocardiogram (ECG),heart rate, arterial blood pressure and indices of cardiac contractility were recorded. In further experimentsin fentanyl/pentobarbital anaesthetized guinea pigs the influence of bilateral versus unilateral carotid arteryocclusion on haemodynamic responses was investigated and the effects of inotropic drugs on left ventricular(LV) dP/dtmax and the QA interval were determined. Results: Pentobarbital, given alone or after fentanyl, pro-vided suitable anaesthesia for these experiments. Bilateral carotid artery occlusion did not alter heart rate orarterial blood pressure responses to isoprenaline or angiotensin II. Isoprenaline and ouabain increased LVdP/dtmax and decreased the QA interval whereas verapamil had opposite effects and strong inverse correlationsbetween LVdP/dtmax and the QA interval were found. Discussion: Conditions can be optimised to allow thepentobarbital-anaesthetized guinea pig to be used for simultaneous measurement of the effects of drugson the ECG, haemodynamics and indices of cardiac contractility. The use of this small animal model in
early pre-clinical safety pharmacology should contribute to improvements in detecting unwanted actionson the heart during the drug development process.
In recent years there has been increasing awareness that drugs canhave serious adverse effects on cardiac contractile function (Force &Kerkelä, 2008). Traditionally, cardiac safety pharmacology studieshave focused on exploring whether drugs in development may haveproarrhythmic activity but there is now a need to also consider thepotential for cardiac contractile dysfunction in pre-clinical testing. Arelatively simple, inexpensive, small animal model is therefore re-quired. Rats are often used for haemodynamic measurements andcardiac contractility can be measured in this species. However, as re-polarization in the rat heart is mainly dependent on the transient out-ward potassium current (Ito) rather than the rapid delayed rectifierpotassium current (IKr) it is not suitable for screening forproarrhythmia. The anaesthetized guinea pig is a model which has
d pressure; ECG, electrocardio-he maximum rate of rise of leftum current; Ito, transient out-
4 7866 869219 (mobile).).
rights reserved.
been used to investigate the proarrhythmic potential of variousdrugs by assessing their ability to prolong the QT interval (Batey,Lightbown, Lambert, Edwards, & Coker, 1997; Hamlin, Kijtawornrat,Keene, & Hamlin, 2003; Hauser, Stade, Schmidt, & Hanauer, 2005;Testai et al., 2004; Yao et al., 2008) or induce torsade de pointes(Michael, Kane, & Coker, 2008). As far as we are aware only onestudy (Hauser et al., 2005) has examined the effects of drugs onboth QT prolongation and cardiac contractility (via measurement ofthe maximum rate of rise of left ventricular pressure (LVdP/dtmax))in anaesthetized guinea pigs.
In order to measure arterial blood pressure and left ventricular(LV) pressure simultaneously, Hauser et al. (2005) cannulated bothcommon carotid arteries thus eliminating any input from the carotidsinus baroreceptors. This could substantially impair reflexes and af-fect responses to drugs that alter arterial blood pressure. As drugsthat have inotropic activity may also affect arterial pressure, it is im-portant to determine whether or not bilateral carotid occlusion (bycannulating both carotid arteries) alters the effects of drugs onheart rate and arterial blood pressure in guinea pigs. This is particu-larly important when planning to use LVdP/dtmax as an index of cardi-ac contractility because this parameter can be altered by changes inheart rate, preload or afterload (Wallace, Skinner, & Mitchell, 1963).
44 L. Mooney et al. / Journal of Pharmacological and Toxicological Methods 66 (2012) 43–51
Some time ago the QA interval was proposed as an alternativeindex of cardiac contractility which was not affected by changes incardiac loading (Cambridge & Whiting, 1986) and more recentlythere has been a revival of interest in the potential of this parameter.Comparisons of the QA interval and LVdP/dtmax as indices of cardiaccontractility in the dog (Norton, Iacono, & Vezina, 2009) and the rat(Adeyemi et al., 2009) have been published but no studies havebeen performed in guinea pigs.
Although some studies comparing different anaesthetics in guineapigs have been published (Brown, Thorne, & Nuttall, 1989; Buchanan,Burge, & Ruble, 1998; Schwenke & Cragg, 2004) none of these hasmeasured cardiac contractility or ECG intervals. The previous ECGstudies in guinea pigs cited above employed a variety of anaestheticregimes: pentobarbital (Batey et al., 1997; Testai et al., 2004), keta-mine/xylazine (Hamlin et al., 2003), thiobutabarbital/urethane (Yaoet al., 2008) and urethane (Hauser et al., 2005). It is therefore notclear which anaesthetic may be best for these types of experiment.
To test the hypothesis that the anaesthetized guinea pig is a suitablemodel for the simultaneous assessment of the effects of drugs on theECG, haemodynamics and cardiac contractility, the aims of the presentwork were: to choose an anaesthetic regime suitable for the natureand duration of the experiments; to determine if bilateral carotid can-nulation compromises drug effects on heart rate and blood pressure;and to compare the usefulness of the QA interval and LVdP/dtmax asindices of cardiac contractility.
2. Methods
2.1. Animals
All animal experiments were performed in accordance with theUK Animals (Scientific Procedures) Act 1986, approved by institution-al ethical review committees and conducted under the authority ofProject Licences held at the University of Strathclyde or atAstraZeneca. Fifty-five male Dunkin Hartley guinea pigs (450–720 g)were purchased from Harlan (Bicester, UK) and allowed a minimumof 1 week acclimatisation before use. They were housed in smallgroups, on aspen chip bedding and sizzle nest (Datesand), or hay, inrooms held at temperatures between 16 and 23 °C, with 40 to 70% rel-ative humidity and a 12 hour light/dark cycle. Food (Special Diet Ser-vices FD1 guinea pig diet or Teklad global higher fibre guinea pig diet2041, plus fresh fruit and vegetables) and water were available adlibitum.
2.2. Anaesthesia
Preliminary experiments focused on determining suitable anaes-thetic regimes. The anaesthetics tested were isoflurane (n=3), acombination of Hypnorm® and Hypnovel® (n=4), and sodium
Table 1Doses and routes of administration of agents used for the induction and maintenance of an
a Isoflurane was given by inhalation in 100% oxygen delivered at 1 L min−1 for inductionb A solution containing 1 part Hypnorm (a solution containing fentanyl citrate 0.315 mg m
parts water made up freshly each day.c Sodium pentobarbital was dissolved at 60 mg mL−1 in distilled water daily for inducti
water) for maintenance, just before the start of each experiment.d Fentanyl (Sublimaze®) was supplied as a solution of 50 μg mL−1 and was given 5 min
pentobarbital either given alone (n=4) or after premedication withfentanyl (n=4). Experiments with Hypnorm/Hypnovel and pento-barbital alone were carried out at AstraZeneca whereas those withisoflurane and fentanyl/pentobarbital were performed concurrentlyat the University of Strathclyde. Subsequently a different isofluraneprotocol was assessed at AstraZeneca (n=4). Animals anaesthetizedwith isoflurane were breathing 100% oxygen, the first 3 breathingspontaneously, the latter 4 artificially ventilated. All others were arti-ficially ventilated with room air. Details of the doses and routes of ad-ministration for the induction and maintenance of anaesthesia withall agents are given in Table 1. All subsequent experiments at the Uni-versity of Strathclyde were performed in guinea pigs anaesthetizedwith the fentanyl/pentobarbital combination. Body temperature wasmeasured via a rectal thermometer and maintained at ~37 °C with aheating lamp.
2.3. Surgical preparation
Immediately after induction of an adequate level of anaesthesia, asdetermined by the absence of corneal and/or pedal withdrawal re-flexes, the trachea was cannulated and artificial ventilation withroom air was commenced. At the University of Strathclyde a Biosci-ence pump (Harvard Apparatus, Edenbridge, Kent, UK) was used,set at a rate of 60 breaths min−1 and a stroke volume of7–8 mL kg−1. Oxygen saturation and expired CO2 were measuredcontinuously using a Medair Lifesense™ Vet pulse oximeter/capnograph (Kruuse UK Ltd., Sherburn in Elmet, UK). If necessary,during the preparation and stabilisation phases, the stroke volumeof the pump was adjusted to keep the expired CO2 value between35 and 45 mm Hg and oxygen saturation above 95%. At AstraZenecaa TOPO™ dual mode ventilator (Kent Scientific, West Malling, UK)was used with the rate set at 20–40 breaths per minute, and thepeak inspiration pressure set at 19.5 cm H2O. If necessary theseranges were adjusted to achieve arterial blood gas values within aset range: PO2>70 and b110 mm Hg, PCO2>25 and b45 mm Hg(Gem Premier Blood Gas analyser, Instrumentation Laboratory, War-rington, UK). Oxygen saturation was also assessed continuouslyusing a pulse oximeter (Medair PulseSense™ Vet, Kruuse UK Ltd.,Sherburn in Elmet, UK), where a desirable reading was >90%.
Needle electrodes were inserted subcutaneously to record limblead ECGs. Both jugular veins were isolated and cannulated, one forinfusion of maintenance doses of anaesthetic and the other for testdrugs. A fluid filled cannula was placed in the right carotid arteryfor measurement of arterial blood pressure and/or blood sampling.All cannulae were filled with normal saline (0.9% w/v NaCl) con-taining 10 units mL−1 heparin. The left carotid artery was isolatedand a 2F or 3F Millar Mikro-tip® catheter pressure transducer (LintonInstrumentation, Diss, UK) was advanced through the artery so thatits tip lay in the lumen of the left ventricle. A stabilisation period of
and at 0.5 L min−1 for maintenance of anaesthesia.L−1 and fluanisone 10 mg mL−1); 1 part Hypnovel (midazolam HCl 5 mg mL−1) and 2
on of anaesthesia and diluted to 6 or 12 mg mL−1 in normal saline (0.9% w/v NaCl in
before sodium pentobarbital.
45L. Mooney et al. / Journal of Pharmacological and Toxicological Methods 66 (2012) 43–51
at least 20 min was allowed following completion of the surgicalpreparation.
2.4. Data acquisition and analysis
At the University of Strathclyde the ECG electrodes were connectedto Gould 6615-65 ECG or 6615-58 bioelectric amplifiers and Leads I, II,and III of the electrocardiogram (ECG) were recorded simultaneously.The fluid filled cannula for arterial blood pressure measurement wasconnected via a Bell and Howell Type 4-442 transducer to a Gould6615-30 DC bridge amplifier whereas the Millar catheter was con-nected via a Millar TC-510 control box to another bridge amplifier. Allsignals from the amplifiers were recorded at 1000 Hz using PonemahP3 Plus software (Linton Instrumentation, Diss, UK). At AstraZeneca car-diovascular data were recorded using a PharmLab on-line 5.0 system(developed in-house). The lead II ECG and the pressure signals wererecorded at rates of 500 and 250 Hz respectively.
Heart rate was derived from the arterial blood pressure signaland recorded along with systolic, diastolic and mean arterial pres-sure. LV systolic and end-diastolic pressures were recorded alongwith LVdP/dtmax as an index of cardiac contractility. The QA interval,which is the interval between the onset of the Q wave of the ECG andthe beginning of the rise in the following arterial blood pressurepulse, was also recorded. At time points of interest, data were aver-aged over a 5 s interval. All ECG intervals were measured manuallyand in some experiments QA intervals were alsomeasuredmanually.Where intervals were measured manually each value was the aver-age of 5 consecutive beats.
2.5. Experimental protocols
The first group of experimentswas designed to compare responses todrugs with either one or both carotid arteries occluded by cannulation.Before the LV catheter was inserted but after a 20 minute stabilisationperiod, either isoprenaline (0.1, 0.3 and 1.0 nmol kg−1 min−1) or an-giotensin II (0.03, 0.1 and 0.3 nmol kg−1 min−1) was administered asa continuous intravenous (i.v.) infusion for 10 min per dose (n=4per group). A recovery period of 10 min was allowed and then theLV catheter was inserted followed by a further 20 minute stabilisationperiod. After this the infusions of either isoprenaline or angiotensin IIwere repeated.
To investigate the reproducibility of responses to vasoactive drugs,anaesthetized guinea pigs were given the same i.v. bolus dose of ei-ther phenylephrine (75 nmol kg−1, n=4) or sodium nitroprusside(336 nmol kg−1, n=4) three times at 15 minute intervals. In sepa-rate animals, to assess the sensitivity of baroreceptor reflex re-sponses, changes in arterial blood pressure were evoked inanaesthetized guinea pigs (n=4) by giving four increasing i.v. bolusdoses of phenylephrine (7.5–75 nmol kg−1) followed by sodiumnitroprusside (33.6–336 nmol kg−1) at 5 minute intervals. For eachdose of phenylephrine and sodium nitroprusside peak values formean arterial blood pressure and corresponding heart rate valueswere selected. The changes from baseline values were plotted and alinear relationship between heart rate and mean arterial blood pres-sure was determined for each animal. The slope of the mean linear re-lationship represents the baroreflex sensitivity. These experimentswere performed in animals where only one carotid artery wasoccluded.
A further group of experiments focused on comparing the effects ofdrugs on cardiac contractility. After completing the full surgical prepa-ration and allowing a 20 minute stabilisation period, guinea pigswere given either isoprenaline (0.1, 0.3 and 1.0 nmol kg−1 min−1),ouabain (1, 3 and, 10 nmol kg−1 min−1), verapamil (14, 42 and140 nmol kg−1 min−1) or vehicle (equal volumes of normal saline;20, 60, 200 μl kg−1 min−1), n=4per group, as continuous i.v. infusionswith each dose being infused for 15 min. The doses of ouabain and
verapamil were selected from within the ranges used by Hauser et al.(2005). For isoprenaline, lower doses were selected than those usedby Hauser et al. (2005) because their lowest dose had the maximal ef-fect and our previous experience with isoprenaline in pentobarbital-anaesthetized rats suggested that lower doses would be sufficient.
2.6. Statistical analysis
Data are presented as mean±s.e. mean. Paired t tests were usedto compare values at two time points within groups. The effects of in-creasing doses of individual drugs with time were assessed using one-way ANOVA followed by Dunnett's test. For comparison of the effectsof a drug under conditions of unilateral versus bilateral carotid occlu-sion the areas under the dose/time-response curves were calculatedand compared using paired t-tests. One-way ANOVA followed byTukey's tests, or Kruskal–Wallis tests (for data that were not distrib-uted normally), were used for comparisons among groups. Correla-tions were calculated using SigmaPlot software.
2.7. Drugs
Isoflurane and fentanyl were obtained from veterinary whole-salers through the Biological Procedures Unit at the University ofStrathclyde. Hypnorm and Hypnovel were obtained from Vet Pharma(Chennai, India) and Roche (Welwyn Garden City, UK), respectively.Angiotensin II acetate, isoprenaline HCl, ouabain, phenylephrineHCl, sodium nitroprusside and verapamil HCl were purchased fromSigma-Aldrich (Poole, UK) and were dissolved in normal saline(0.9% w/v NaCl in water).
3. Results
3.1. Comparison of anaesthetics
Pentobarbital, either given alone or after fentanyl, provided stableanaesthesia whereas the anaesthesia obtained with Hypnorm/Hypnovel tended to be less stable over time and isoflurane was unsuit-able. Induction of anaesthesiawith 5% isoflurane resulted in profuse sal-ivation and in the first animal, regurgitation of food. In the twofollowing animals food was withdrawn the night before the experi-ments but there was still marked salivation, spontaneous respirationwas rather erratic and the concentration of isoflurane had to be adjust-ed frequently to try tomaintain a stable heart rate within normal limits.In the subsequent batch of guinea pigs given isoflurane, induction with2–4% caused less salivation but the necessity for frequent adjustmentsof the concentration of isoflurane remained. In addition, mean arterialblood pressure and heart rate were significantly lower (Fig. 1) andthere were some minor differences in PR and QRS intervals (Table 2).
The time taken to reach surgical anaesthesia with Hypnorm/Hypnovel was 12±1 min (individual data: 9, 11, 13, 15 min), withpentobarbital alone was 18±9 min (individual data: 6, 7, 15,43 min) and with pentobarbital given 5 min after fentanyl was 6±1 min* (individual data: 5, 5, 5, 7 min); *Pb0.05 compared to othergroups, Kruskal–Wallis test. After completion of the surgical prepara-tion and a 20 minute stabilisation period, baseline values for heartrate were similar with these three anaesthetic regimes (Fig. 1a) butthere was more variation in the baseline mean arterial blood pres-sures (Fig. 1b). With Hypnorm/Hypnovel mean arterial blood pres-sure tended to be lower and the heart rate had declinedsignificantly by the end of the recording period (Fig. 1a). In contrast,there were no significant differences between the initial and finalvalues for heart rate or mean arterial blood pressure in either pento-barbital group (Fig. 1). In the guinea pigs anaesthetized withHypnorm/Hypnovel, pentobarbital or fentanyl/pentobarbital therewere no differences in ECG intervals within groups or betweengroups at any time point (Table 2).
Fig. 1. a) Heart rate and b) mean arterial blood pressure (BP) with time in groups of an-imals anaesthetized with either Hypnorm/Hypnovel, isoflurane, pentobarbital alone orgiven after pre-medication with fentanyl. Baseline data were recorded after completionof a minimum 20 min stabilisation period. The experiments with Hypnorm/Hypnovel,isoflurane and pentobarbital were performed at AstraZeneca whereas those with fenta-nyl/pentobarbital were carried out at the University of Strathclyde. Data are shown asmean±s.e. mean, n=4 per group. *Pb0.05 compared with time 0 min within group,paired t test. #Pb0.05 compared with fentanyl/pentobarbital group at that time,+Pb0.05 compared with pentobarbital group at that time, one way ANOVA withTukey's test.
46 L. Mooney et al. / Journal of Pharmacological and Toxicological Methods 66 (2012) 43–51
3.2. Drug responses during unilateral and bilateral carotid occlusion
Occlusion of the second carotid artery did not alter baseline valuesfor heart rate, systolic or diastolic arterial blood pressure (Fig. 2). Iso-prenaline increased heart rate in a dose-dependent manner but had
Table 2ECG intervals in guinea pigs anaesthetized with isoflurane, Hypnorm/Hypnovel, sodium pe
The experiments with Hypnorm/Hypnovel, isoflurane and pentobarbital were performedUniversity of Strathclyde. Data are shown as mean±s.e. mean, n=4 per group. aPb0.05one way ANOVA with Tukey's test.
limited effects on systolic and diastolic arterial blood pressure(Fig. 2a–c). In contrast, angiotensin II caused significant dose-dependent increases in systolic and diastolic pressure but had verylittle effect on heart rate (Fig. 2d–f). When the data obtained duringunilateral and bilateral carotid artery occlusion are compared visuallyit can be seen that the patterns of responses are virtually identical(Fig. 2). Statistical analysis of the areas under the curve of the dose/time-response relationships of the data expressed as % change frombaseline indicated that there were no significant differences in theheart rate or blood pressure responses to either isoprenaline or angio-tensin II during bilateral compared to unilateral carotid occlusion.
3.3. Reproducibility and reflex responses
Repeated administration of the same dose of sodium nitroprussidecaused decreases in mean arterial blood pressure of similar magni-tude which were accompanied by increases in heart rate whereasphenylephrine had opposite effects; increased blood pressure and de-creased heart rate (Fig. 3a–d). The experiments with increasing dosesof sodium nitroprusside and phenylephrine resulted in a linear rela-tionship between the change in mean BP and the change in heartrate, where y=−0.989x+0.378 (r2=0.953). This gives a mean bar-oreflex sensitivity which would result in a change of 9.9 beats min−1
in heart rate in response to a change in blood pressure of 10 mm Hg(Fig. 3e).
3.4. Effects of drugs on LVdP/dtmax and QA interval
Isoprenaline caused rapid dose-dependent increases in LVdP/dtmax
and decreases in the QA interval which were sustained during contin-uous infusion of the drug (Fig. 4a). Infusion of increasing doses ofouabain resulted in a gradual rise in LVdP/dtmax which reached signif-icance during the latter stages of infusion of the highest dose. Theeffects of ouabain on the QA interval were a mirror image of thoseon LVdP/dtmax with the reduction in the QA interval also reachingstatistical significance during infusion of the highest dose (Fig. 4b).Although verapamil did not reduce LVdP/dtmax significantly, it onlyjust failed to do so by the smallest margin with the ANOVA indicatinga P value of 0.051. The effect of verapamil on the QA interval wasclearer as significant changes were detected at several time pointsduring administration of the final dose (Fig. 4c). Infusion of the samevolumes of vehicle did not alter either LVdP/dtmax or the QA interval(Fig. 4d).
ntobarbital or fentanyl plus sodium pentobarbital.
Pentobarbital Fentanyl plus pentobarbital
(ms)56±2 57±256±2 56±156±1 56±257±2 56±2
l (ms)24±2 24±125±2 23±125±2 23±125±2 24±1
l (ms)154±10 151±11152±12 150±13159±12 150±14165±10 148±16
at AstraZeneca whereas those with fentanyl/pentobarbital were carried out at thecompared with all other groups, bPb0.05 compared with Hypnorm/Hypnovel group,
Time (min)-5 0 5 10 15 20 25 30 35
Dia
stol
ic B
P (
mm
Hg)
0
10
20
30
40
50
60
70
Time (min)-5 0 5 10 15 20 25 30 35
0
10
20
30
40
50
60
70
0.03
Sys
tolic
BP
(m
mH
g)
40
50
60
70
80
90
100
110
0.31.0
0.1
a
b
c
Hea
rt R
ate
(bea
ts m
in-1
)
240
260
280
300
320
340
360
Unilateral
Bilateral
240
260
280
300
320
340
360
40
50
60
70
80
90
100
110
0.10.3
d
e
f
Isoprenaline Angiotensin II
Fig. 2. The effects of isoprenaline on a) heart rate, b) systolic arterial blood pressure (BP), c) diastolic BP and of angiotensin II on d) heart rate, e) systolic BP, f) diastolic BP duringunilateral and bilateral carotid artery occlusion. The grey bars indicate the dose of the drugs in nmol kg−1 min−1. Mean±s.e. mean, n=4. *Pb0.05 compared to value within groupat time 0 min, one way ANOVA with Dunnett's test.
47L. Mooney et al. / Journal of Pharmacological and Toxicological Methods 66 (2012) 43–51
The observation that the graphs of LVdP/dtmax and QA interval dataseem to be mirror images of each other suggests that there may be in-verse correlations between these two indices of contractility. To exam-ine this further, Fig. 4 shows the relationships between LVdP/dtmax andQA interval. Plotting the QA interval against LVdP/dtmax revealed acurvilinear relationship (Fig. 5a) which could be converted to a linearcorrelation by plotting the log10 values for each of these parameters(Fig. 5b).
The effects of isoprenaline on heart rate and arterial blood pressurein the group of animals used here for the contractility experiments(data not shown) were virtually identical to those detailed in Fig. 2 forthe animals used to study carotid occlusion; i.e. a significant increasein heart rate but no significant effects on blood pressure. The highestdose of ouabain increased heart rate significantly (Fig. 6a) but the ap-parent increases in systolic and diastolic arterial blood pressure didnot reach significance (Fig. 6b). Verapamil reduced heart rate (Fig. 6c)and systolic blood pressure but did not reduce diastolic blood pressuresignificantly (Fig. 6d). Heart rate, systolic, diastolic and mean bloodpressure remained steady throughout the experiments in the group re-ceiving vehicle (see Fig. 1 for heart rate andmean blood pressure data).
4. Discussion
This work demonstrates, for the first time, the detection of drug-induced changes in two indices of cardiac contractility (LVdP/dtmax
and the QA interval) measured simultaneously in anaesthetized guineapigs. In addition, the results prove that cannulation of both carotidarteries in this species does not have any detrimental effects on re-sponses to drugs, thus validating this approach and contributing to op-timisation of the experimental conditions.
4.1. Anaesthesia
As the main aim of this work was to optimise conditions for study-ing the effects of drugs on haemodynamics and cardiac contractility inanaesthetized guinea pigs, we felt that it was important first of all toreconsider the suitability of sodium pentobarbital as our anaestheticof choice (Batey et al., 1997; Michael et al., 2008). Previous studieson ECG intervals in guinea pigs have employed a variety of anaes-thetics. Although Hamlin et al. (2003) used ketamine/xylazine, theirexperiments used non-invasive methods for assessing the ECG and
Time (min)
0 10 20 30 40 50
Mea
n B
P
(mm
Hg)
10
30
50
70
Mea
n B
P
(mm
Hg)
30
50
70
90
a
b
c
Hea
rt R
ate
(bea
ts m
in-1
)
240
260
280
300
Phenylephrine
Hea
rt R
ate
(bea
ts m
in-1
)
260
280
300
320
Sodium Nitroprussided
Change in Mean BP (mmHg)-150 -100 -50 0 50 100 150
Cha
nge
in H
eart
Rat
e (b
eats
min
-1)
-150
-100
-50
0
50
100
150e
Fig. 3. a) Heart rate and b) mean arterial blood pressure (BP) responses to repeated ad-ministration of phenylephrine (75 nmol kg−1) and c) heart rate and d) mean BP re-sponses to repeated administration of sodium nitroprusside (336 nmol kg−1) at thetimes indicated by the arrows. Data are presented as mean±s.e. mean, n=4 foreach drug. e) The linear relationships for the change in heart rate in response to achange in mean arterial blood pressure in individual animals (dotted lines) extrapolat-ed from responses to increasing doses of sodium nitroprusside (33.6–336 nmol kg−1)and phenylephrine (7.5–75 nmol kg−1). The solid line represents the mean linear rela-tionship where y=−0.989x+0.378 (r2=0.953).
48 L. Mooney et al. / Journal of Pharmacological and Toxicological Methods 66 (2012) 43–51
others have reported that this anaesthetic regime was not sufficientfor surgical interventions in guinea pigs (Buchanan et al., 1998).Hauser et al. (2005) used urethane but as this is carcinogenic (Field& Lang, 1988), for the sake of operator safety, this was also excludedfrom our list of possible anaesthetics.
Inhalational anaesthetics offer advantages in terms of speed of in-duction and rapid responses to altering the dose during procedures tomaintain steady anaesthesia. Unfortunately we found that the guineapigs tended to hold their breath for some time when they smelt theisoflurane, which meant that induction was not as smooth as it is inother species such as the rat. They also salivated profusely and onceplaced on their backs on the operating table we were concerned
that this excess fluid could impede respiration. As we wanted an an-imal model with intact baroreceptor reflexes we did not want to giveantimuscarinic premedication to reduce the secretion of saliva. Usinga lower concentration of isoflurane for induction of anaesthesia in thesecond batch of experiments did reduce salivation but heart rate andmean arterial blood pressure were lower than with other anaestheticsand there were still difficulties in maintaining stable haemodynamics.Thus experiments with this anaesthetic were discontinued.
In our experience, rats anaesthetized with pentobarbital orHypnorm/Hypnovel normally breathe spontaneously whereas artifi-cial respiration is almost always necessary in pentobarbital-anaesthetized guinea pigs. At the doses used in the present study ar-tificial ventilation was required in Hypnorm/Hypnovel anaesthetizedguinea pigs thus offering no advantage over pentobarbital in this re-spect. We also found that heart rate declined significantly in theHypnorm/Hypnovel group but not in those given pentobarbital indi-cating that the latter was more suitable for our purposes.
Although pentobarbital is used commonly for terminal anaesthesiaof laboratory animals, veterinary surgeons prefer to use anaesthetics,such as Hypnorm/Hypnovel, that also have analgesic properties. Forthis reason we decided to investigate whether premedication with fen-tanyl prior to pentobarbital had any advantages or disadvantages. Thedose of fentanyl used here (50 μg kg−1 s.c.) seemed to relax the guineapigs and thismay have contributed to the faster induction time after thesubsequent administration of pentobarbital. However, as fentanyl wasgiven 5 min before pentobarbital, the overall time from first injectionto surgical anaesthesia was very similar. During induction of anaesthe-sia with pentobarbital, animals can go through a hyperalgesic phase(Ewen, Archer, Samanani, & Roth, 1995) and there is evidence of in-creased nociception after i.p. administration of pentobarbital(Svendsen, Kok, & Lauritzen, 2007). Premedication with an analgesicsuch as fentanyl will offset these actions of pentobarbital, thus improv-ing animal welfare. Fentanyl had no adverse effects on haemodynamicsor the ECG.
4.2. Bilateral carotid occlusion
In general, guinea pigs are considered to be more difficult to workwith than other similarly sized species such as the rat. As well asbeing more difficult to anaesthetize (Buchanan et al., 1998) theblood vessels of the guinea pig are much smaller (Librizzi, Biella,Cimino, & De Curtis, 1999) and very fragile. As a consequence it canbe technically challenging to successfully isolate and cannulate femo-ral arteries (or veins) in the guinea pig. Previously, it was noticed thatwhen both carotid arteries were cannulated in the anaesthetized rat,the heart rate and blood pressure responses to drugs such as isopren-aline and angiotensin II were unusual because they lacked the normalreflex components (S.J. Coker, personal observation). The data pres-ented above demonstrate that this does not apply in the guinea pig,as the responses to isoprenaline and angiotensin II were virtuallyidentical when either one or both carotid arteries were cannulated.This suggests that the baroreceptors in the aortic arch are muchmore important than those in the carotid sinuses in guinea pigs. Un-like the rat, two‐thirds of the blood supply to the brain in the guineapig is delivered through the vertebrobasilar arterial system (Librizziet al., 1999; Majewska-Michalska, 1998). Thus bilateral occlusion ofthe carotid arteries is unlikely to seriously compromise blood flowto the brain in the guinea pig.
As well as demonstrating the reproducibility of responses withtime, the experiments with sodium nitroprusside and phenylephrinealso confirmed the existence of functional baroreceptor responses inboth directions. The ratio for the change in heart rate to the changein blood pressure of 0.99 beats min−1 mmHg−1 is lower than thevalue of 1.59 reported for responses to phenylephrine in consciousrats (Coleman, 1980) but in rabbits it has been shown that anaesthe-sia depressed baroreflex sensitivity (Blake & Korner, 1981). The data
Time (min)0 10 20 30 40 50
QA
inte
rval
(m
s)
20
30
40
50
60
a
b
c
LVdP
/dt m
ax (
mm
Hg
s-1)
LVdP
/dt m
ax (
mm
Hg
s-1)
2000
4000
6000
8000
10000 Isoprenaline
* * * * * *
Time (min)0 10 20 30 40 50
20
30
40
50
60
2000
4000
6000
8000
10000 Ouabain
Time (min)0 10 20 30 40 50
QA
inte
rval
(m
s)
30
40
50
60
70
0
2000
4000
6000
8000 Verapamil
Time (min)0 10 20 30 40 50
30
40
50
60
70
0
2000
4000
6000
8000 Vehicle
d
#
Fig. 4. The effects of a) isoprenaline (0.1, 0.3 and 1.0 nmol kg−1 min−1), b) verapamil (14, 42 and 140 nmol kg−1 min−1), c) ouabain (1, 3, and 10 nmol kg−1 min−1) and d) ve-hicle (equal volumes of normal saline) on left ventricular (LV) dP/dtmax and the QA interval. The grey bars indicate the increasing doses of the drugs infused. Mean±s.e. mean,n=4. #P=0.051 one‐way ANOVA; *Pb0.05 compared to value within group at time 0 min, one way ANOVA with Dunnett's test.
49L. Mooney et al. / Journal of Pharmacological and Toxicological Methods 66 (2012) 43–51
presented here suggest that in this anaesthetized guinea pig modeladequate reflex sensitivity is retained.
4.3. Assessing cardiac contractility via changes in LVdP/dtmax and theQA interval
For many years LVdP/dtmax has been used as an index of cardiaccontractility despite the limitation that it can be influenced bychanges in heart rate, preload and afterload (Wallace et al., 1963).For safety pharmacology purposes this is not necessarily a disadvan-tage as any undesired effect on heart rate or blood pressure wouldalso be of concern. The QA interval could be an even simpler methodof assessing cardiac contractility as it does not require insertion of acatheter into the lumen of the left ventricle. However, since it wasfirst described as an index of cardiac contractility (Cambridge &Whiting, 1986) only about a dozen other full papers using the QA in-terval have been published. Just two of these papers have looked atrelationships between the QA interval and LVdP/dtmax (Adeyemi etal., 2009; Norton et al., 2009).
In conscious rats, Adeyemi et al. (2009) found an inverse linearcorrelation between log10 LVdP/dtmax and log10 QA interval whereasin conscious dogs, Norton et al. (2009) reported inverse linear
correlations between drug-induced changes in the absolute valuesof these parameters. In the latter report on the experiments in dogsthe correlations found with a positive inotrope (pimobendan) and anegative inotrope (atenolol) were shown in separate graphs. Al-though reasonable linear correlations were found for each drug indi-vidually it is not possible to tell whether or not there would still be alinear relationship if the data had all been plotted on the samegraph. Adeyemi et al. (2009) plotted separate straight lines (on thesame graph) through the correlations between the absolute valuesof LVdP/dtmax and the QA interval for each of the three drugs theyhad used (verapamil, milrinone and salmeterol). The slope of theline through their verapamil data is steeper than the slopes of thelines through the data obtained with the positive inotropes, milrinoneand salmeterol, suggesting that when all of their data are consideredtogether there is a curvilinear relationship between the absolutevalues of LVdP/dtmax and the QA interval. The results detailed above(Fig. 5) indicate that in the anaesthetized guinea pig there is also acurvilinear relationship between the absolute values for LVdP/dtmax
and the QA interval which becomes a linear correlation when thelog10 values of both parameters are plotted. Thus these results inthe anaesthetized guinea pig agree well with those found in consciousrats (Adeyemi et al., 2009).
Log10 LVdP/dtmax
3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 4.0
Log 1
0 Q
A in
terv
al
1.4
1.5
1.6
1.7
1.8
1.9Isoprenaline
Ouabain
Verapamil
1000 2000 3000 4000 5000 6000 7000 8000 9000
QA
Inte
rval
(m
s)
20
30
40
50
60
70 Isoprenaline
Ouabain
Verapamil
b
a
r2 = 0.959
LVdP/dtmax (mmHg s-1)
Fig. 5. Correlations between a) LVdP/dtmax and the QA interval and b) log10 LVdP/dtmax
and log10 QA interval in guinea pigs which received isoprenaline, ouabain or verapamil.Dotted lines indicate the 95% confidence limits.
50 L. Mooney et al. / Journal of Pharmacological and Toxicological Methods 66 (2012) 43–51
As the QA interval covers the period of time from the initiation ofdepolarization in the ventricles until the time when the aortic valveopens, thus starting the rise in arterial pressure, there are a numberof possible sites where drugs devoid of inotropic activity could stillalter the QA interval (Hamlin & Del Rio, 2010). For example, drugsthat slow conduction velocity through the heart could prolong theQA interval. The data for verapamil in Figs. 4b and 5b suggest that itmay have had a slightly greater effect on the QA interval than onLVdP/dtmax. If confirmed, this could simply be a consequence of thereduction in blood pressure caused by verapamil since a reduced arte-rial pressure, and a reduced arterial stiffness due to dilatation, wouldtend to slow the propagation of the pulse wave thus increasing theQA interval (Hamlin & Del Rio, 2010). Further studies with other va-soactive drugs would help to clarify this point.
Another potential issue when considering the use of the QA inter-val as an index of myocardial contractility instead of LVdP/dtmax is theopposite effects of inotropes on these parameters. Since positiveinotropes decrease the QA interval there is a limit to the extent ofchanges that can be seen; larger doses of drugs may produce progres-sively smaller changes as the values get nearer to zero. In contrast,
values of LVdP/dtmax can increase indefinitely with positive inotropes.Obviously, the opposite situation occurs with negative inotropes.Thus LVdP/dtmax may have greater potential to detect positive inotro-pic effects whereas the QA interval may be more powerful indetecting negative inotropy.
4.4. Limitations
The recognised “gold standard” method for estimating cardiaccontractility is the LV pressure–volume loop; see Hamlin & Del Rio(2010) for references and discussion. However, this technique is chal-lenging, the equipment required is expensive and calibration of thecatheters can be complicated. As our main aim was to develop a sim-ple model, we decided to measure both LVdP/dtmax values and QA in-tervals to allow us to evaluate whether the QA interval offered anyadvantages. Although good correlations between the two parameterswere found, the relationship between LVdP/dtmax and the QA intervalmay differ for positive and negative inotropes or drugs with addition-al electrophysiological or haemodynamic effects and thus further ex-periments with other positive and negative inotropes and withadditional compounds are required to explore this in more detail.
Small group sizes have been used in these studies for two reasons.First, UK legislation on the use of animals for scientific procedures re-quires licence holders to reduce the number of animals used wherev-er possible. Thus if significant differences can be obtained with n=4it is difficult to justify using more animals per group. Second, themodel is intended for use in safety pharmacology studies wheretime and costs are important factors. However, if certain test com-pounds have effects of lower magnitude or there is wider variationbetween the responses in individual animals within a group, in-creased numbers may be required to see statistically significant ef-fects. The present results indicate that simultaneous screening forproarrhythmia and adverse effects on cardiac contractility is feasibleand should lead to reductions in the number of animals required forthese processes.
4.5. Conclusions
The results presented above indicate that pentobarbital, eithergiven alone or after fentanyl, provides anaesthesia suitable for the na-ture and duration of these experiments. Occlusion of both carotid ar-teries (by cannulation) did not alter drug-induced changes in heartrate or blood pressure thus indicating that this approach is valid inthe guinea pig. Both LVdP/dtmax and the QA interval can be measuredin the anaesthetized guinea pig and strong inverse correlations be-tween these indices of cardiac contractility were found. Taken togeth-er these results indicate that the anaesthetized guinea pig has greatpotential as an early pre-clinical model in safety pharmacology forthe simultaneous detection of drug-induced changes in the ECG,haemodynamics and cardiac contractility.
Authors' contributions
All authors contributed to the study design, data interpretationand critical revision of the manuscript. The experimental work anddata collation at Strathclyde University was performed by L Mooneyand at AstraZeneca by L Marks and KLP. Statistical analysis of datawas performed by L Mooney, MS and SJC. The manuscript wasplanned by MS, SJC and SC and written mainly by SJC.
Acknowledgements
The work carried out at the University of Strathclyde was fundedby the Medical Research Council (MRC) and AstraZeneca throughthe MRC Working with Industry Scheme (Award No. G0800135).
Time (min)0 10 20 30 40 50
0
20
40
60
80
100 SystolicDiastolic
Time (min)0 10 20 30 40 50
Art
eria
l BP
(m
mH
g)
0
20
40
60
80
100
a
b
Hea
rt R
ate
(bea
ts m
in-1
)
220
240
260
280
300
320
220
240
260
280
300
320Ouabain Verapamilc
dSystolicDiastolic
Fig. 6. The effects of ouabain on a) heart rate and b) systolic and diastolic arterial blood pressure (BP) and the effects of verapamil on c) heart rate and d) systolic and diastolic BP inanaesthetized guinea pigs. Mean±s.e. mean, n=4. *Pb0.05 compared to value within group at time 0 min, one way ANOVA with Dunnett's test.
51L. Mooney et al. / Journal of Pharmacological and Toxicological Methods 66 (2012) 43–51
The authors would like to thank Dr. Michael Wilkinson and Pascal vanTroys for their veterinary input and advice on anaesthesia.
References
Adeyemi, O., Roberts, S., Harris, J., West, H., Shome, S., & Dewhurst, M. (2009). QA inter-val as an indirect measure of cardiac contractility in the conscious telemeterisedrat: Model optimisation and evaluation. Journal of Pharmacological and Toxicologi-cal Methods, 60, 159–166.
Batey, A. J., Lightbown, I. D., Lambert, J. P., Edwards, G., & Coker, S. J. (1997). Compar-ison of the acute cardiotoxicity of the antimalarial drug halofantrine in vitro andin vivo in anaesthetized guinea-pigs. British Journal of Pharmacology, 122,563–569.
Blake, D. W., & Korner, P. I. (1981). Role of baroreceptor reflexes in the hemodynamicand heart rate responses to althesin, ketamine and thiopentone anesthesia. Journalof the Autonomic Nervous System, 3, 55–70.
Brown, J. N., Thorne, P. R., & Nuttall, A. L. (1989). Blood pressure and other physiolog-ical responses in awake and anesthetized guinea pigs. Laboratory Animal Science,39, 142–148.
Buchanan, K. C., Burge, R. R., & Ruble, G. R. (1998). Evaluation of injectable anestheticsfor major surgical procedures in guinea pigs. Contempory Topics in Laboratory Ani-mal Science, 37, 58–63.
Cambridge, D., & Whiting, M. V. (1986). Evaluation of the QA interval as an index ofcardiac contractility in anaesthetised dogs: Responses to changes in cardiac load-ing and heart rate. Cardiovascular Research, 20, 444–450.
Coleman, T. G. (1980). Arterial baroreflex control of heart rate in the conscious rat.American Journal of Physiology, 238, H515–H520.
Ewen, A., Archer, D. P., Samanani, N., & Roth, S. H. (1995). Hyperalgesia during seda-tion: effects of barbiturates and propofol in the rat. Canadian Journal of Anaesthesia,42, 532–540.
Field, K. J., & Lang, C. M. (1988). Hazards of urethane (ethyl carbamate): A review of theliterature. Laboratory Animals, 22, 255–262.
Force, T., & Kerkelä, R. (2008). Cardiotoxicity of the new cancer therapeutics —
mechanisms of, and approaches to, the problem. Drug Discovery Today, 13,778–784.
Hamlin, R. L., & Del Rio, C. (2010). An approach to the assessment of drug-inducedchanges in non-electrophysiological properties of cardiovascular function. Journalof Pharmacological and Toxicological Methods, 62, 20–29.
Hamlin, R. L., Kijtawornrat, A., Keene, B. W., & Hamlin, D. M. (2003). QT and RR inter-vals in conscious and anesthetized guinea pigs with highly varying RR intervalsand given QTc-lengthening test articles. Toxicological Sciences, 76, 437–442.
Hauser, D. S., Stade, M., Schmidt, A., & Hanauer, G. (2005). Cardiovascular parametersin anaesthetized guinea pigs: A safety pharmacology screening model. Journal ofPharmacological and Toxicological Methods, 52, 106–114.
Librizzi, L., Biella, G., Cimino, C., & De Curtis, M. (1999). Arterial supply of limbic struc-tures in the guinea pig. Journal of Comparative Neurology, 411, 674–682.
Majewska-Michalska, E. (1998). The vertebrobasilar arterial system in guinea pig ascompared with dog and human. Folia Morphol (Warsz), 57, 121–131.
Michael, G., Kane, K. A., & Coker, S. J. (2008). Adrenaline reveals the torsadogenic effectof combined blockade of potassium channels in anaesthetized guinea pigs. BritishJournal of Pharmacology, 154, 1414–1426.
Norton, K., Iacono, G., & Vezina, M. (2009). Assessment of the pharmacological effectsof inotropic drugs on left ventricular pressure and contractility: An evaluation ofthe QA interval as an indirect indicator of cardiac inotropism. Journal of Pharmaco-logical and Toxicological Methods, 60, 193–197.
Schwenke, D. O., & Cragg, P. A. (2004). Comparison of the depressive effects of four an-esthetic regimens on ventilatory and cardiovascular variables in the guinea pig.Comparative Medicine, 54, 77–85.
Svendsen, O., Kok, L., & Lauritzen, B. (2007). Nociception after intraperitoneal injectionof a sodium pentobarbitone formulation with and without lidocaine in rats quan-tified by expression of neuronal c-fos in the spinal cord—A preliminary study. Lab-oratory Animals, 41, 197–203.
Testai, L., Calderone, V., Salvadori, A., Breschi, M. C., Nieri, P., & Martinotti, E. (2004). QTprolongation in anaesthetized guinea-pigs: An experimental approach for prelim-inary screening of torsadogenicity of drugs and drug candidates. Journal of AppliedToxicology, 24, 217–222.
Wallace, A. G., Skinner, N. S., Jr., & Mitchell, J. H. (1963). Hemodynamic determinants ofthe maximal rate of rise of left ventricular pressure. American Journal of Physiology,205, 30–36.
Yao, X., Anderson, D. L., Ross, S. A., Lang, D. G., Desai, B. Z., Cooper, D. C., et al. (2008).Predicting QT prolongation in humans during early drug development usinghERG inhibition and an anaesthetized guinea-pig model. British Journal of Pharma-cology, 154, 1446–1456.
