Stress-induced changes in respiratory quotient, energy ... - Springer Link

Psychopharmacology (I 994) 116:475-482

Psychopharmacology © Springer-Verlag 1994

Stress-induced changes in respiratory quotient, energy expenditure and locomotor activity in rats: effects of midazolam Iain S. McGregor 1, Andrew M. Lee 1, R.F. Westbrook 2 lDepartment of Psychology,Universityof Sydney,NSW 2006, Australia 2School of Psychology,University of New South Wales, NSW 2033, Australia Received: 23 April 1993 / Final version: 24 March 1994

Abstract. Changes in 02 consumption, CO2 production and locomotor activity were examined in rats exposed to (1) brief footshock, (2) an aversive conditioned stimulus (CS) predicting footshock, or (3) the anxiogenic drug FG-7142. Respiratory quotient (RQ = CO2 produced/O2 consumed) and energy expenditure [EE = 02 consumed (364 + 113RQ)] were derived to give an estimate of the energy substrate (fat, carbohydrate or protein) being utilised and total substrate oxidation respectively. In experiment 1, footshock (4 x 5 s 0.6 mA shocks over 2 min) produced an immediate increase in RQ, EE and activity. The RQ and EE effects were attenuated by the benzodiazepine midazolam (1 mg/kg). In experiment 2, an aversive CS, consisting of flashing light and buzzer that had 24 h earlier been repeatedly paired with footshock (20 x 5 s 0.6 mA shocks) caused a pronounced drop in RQ, an increase in EE and locomotor activity suppression. The effects of the aversive CS on RQ and EE were reversed by midazolam (1 mg/kg). In experiment 3, FG-7142 (10 mg/kg) produced a steep drop in RQ that persisted for at least 2 h and which was reversed by midazolam (1 mg/kg) and delayed by the benzodiazepine antagonist RO 15-1788 (10 mg/kg). FG-7142 also tended to inhibit EE and locomotor activity, but these effects did not reach statistical significance. Overall, these data show that stress causes profound alterations in RQ, EE and activity and that the pattern of change in these parameters differs with the nature of the stressor involved. Key words: Stress - Anxiety - Rat - Metabolism Energy expenditure - Energy substrate utilisation - Respiratory quotient - Midazolam - FG-7142 - RO 15-1788 Panic - Ityperventilation - Sympathetic nervous system Benzodiazepine - Conditioned stimulus - Respiration.

Parts of this paper were presented at the joint meetingof the British Psychopharmacology Society and European Society for Behavioural Pharmacology.Cambridge, U.K. August 1992. Correspondence to." I.S. McGregor

One of the most reliable physiological changes occurring with emotion is altered respiration. In an article published in the first volume of the Journal of Experimental Psychology, Feleky (1916) noted with regard to emotion that "the respiratory muscles speak as clear a language as do the muscles of the larynx". Feleky documented changes in the pattern and amplitude of respiration during the recall of emotional states, finding that the recall of fear produced the greatest changes in respiration. Such respiratory changes have been conceptualised as a vital part of the stress response whereby the increased mobilisation and oxidation of fuel reserves primes an animal for "fight or flight". Selye (1936) proposed that acute stress elicits a fundamental redirection of energy resources from storage to release with stress-induced activation of the sympathetic nervous system and the hypothatamicpituitary-adrenal axis leading to an inhibition of the storage of glucose, fats and proteins and an enhancement of their mobilisation. Such stress-induced changes in availability of energy substrates are thought to be paralleled by cardiovascular and respiratory changes allowing these newly available energy substrates to be rapidly oxidised (e.g. Johnson et al. 1992). This account suggests that the measurement of respiration and/or energy substrate oxidation might provide a very sensitive index of emotional states. However, with the exception of Feleky's (1916) study mentioned above, this idea has been subjected to surprisingly little experimental scrutiny. The present series of experiments therefore set out to characterise the basic nature of changes in respiration and energy substrate utilisation induced by stress. To this end, 02 consumption and CO2 production were measured in freely moving rats subjected to various stressful procedures. Basic respiratory measurements were used to derive two parameters, namely energy expenditure (EE) and respiratory quotient (RQ). EE, expressed in Joules/g, is very closely related to 02 consumption, but employs corrections for body weight and substrate utilisation to give a more accurate and uniform estimation of total energy substrate oxidation (Kleiber 1975). RQ, on the other hand, is an estimate of the type of energy substrate (carbohydrate, protein or fat) being utilised by an animal. An

476 R Q o f a r o u n d 1.0 indicates the exclusive use o f c a r b o h y d r a t e while an R Q o f 0.7 indicate the exclusive use o f lipids. I n t e r m e d i a t e R Q s ( o f a r o u n d 0.8) can indicate either exclusive p r o t e i n utilisation or the utilisation o f a m i x e d c a r b o h y d r a t e a n d lipid s u b s t r a t e ( K l e i b e r 1975; Le M a g n e n 1985). It was a s s u m e d f r o m classical theories o f the stress response t h a t by virtue o f increasing energy s u b s t r a t e o x i d a t i o n , stress w o u l d increase 0 2 c o n s u m p tion a n d hence EE. H o w e v e r , it was u n c l e a r w h e t h e r stress w o u l d alter R Q . Selye's t h e o r y predicts t h a t stressors m i g h t increase the utilisation o f all s u b s t r a t e types, n a m e ly c a r b o h y d r a t e , a m i n o acids a n d lipids, which w o u l d l e a d to u n p r e d i c t a b l e effects on R Q . T h e p r e s e n t s t u d y also m e a s u r e d l o c o m o t o r activity d u r i n g stress. U n d e r n o r m a l c i r c u m s t a n c e s 0 2 c o n s u m p tion, a n d hence EE, covaries closely with activity, a n d high levels o f activity can a c c o u n t for m o r e t h a n h a l f o f t o t a l EE. T h e a m o u n t o f E E a t t r i b u t a b l e to l o c o m o t o r activity is sometimes called voluntary EE, to distinguish it f r o m o t h e r f o r m s o f E E which reflect p u r e l y i n v o l u n tary factors such as b a s a l m e t a b o l i s m , d i e t - i n d u c e d thermogenesis a n d c o l d - i n d u c e d thermogenesis. Since it has been d e m o n s t r a t e d t h a t stressors m a y p r o v o k e r o b u s t changes in l o c o m o t o r activity (e.g. F a n s e l o w 1982), it was i m p o r t a n t to d e t e r m i n e the extent to which a n y stressi n d u c e d increases in E E reflected v o l u n t a r y c o m p a r e d to i n v o l u n t a r y factors. This was achieved t h r o u g h c o m p a r ing E E with activity a n d d e t e r m i n i n g the extent to which changes in the f o r m e r reflect the changes in the latter (see M c G r e g o r et al. t990, 1991). A l s o o f interest in the p r e s e n t study was the w a y in which p h y s i o l o g i c a l responses m i g h t v a r y a c c o r d i n g to the stressor involved. I n p a r t i c u l a r it was o f interest to c o m p a r e the response to a p r o t o t y p i c a l physical stressor (brief f o o t s h o c k ) w i t h the response to a m o r e " p s y c h o logical" f o r m o f stress, n a m e l y a n t i c i p a t i o n o f an aversive event. This f o r m o f stress was i m p l e m e n t e d t h r o u g h the p r e s e n t a t i o n o f a c o n d i t i o n e d stimulus (CS) t h a t predicts f o o t s h o c k . A s a t h i r d m a n i p u l a t i o n , the effects o f the ~3-carboline d r u g F G - 7 1 4 2 were assessed. This drug, which acts as a p a r t i a l inverse a g o n i s t at the G A B A / b e n z o d i a z e p i n e r e c e p t o r c o m p l e x , induces stressor-like effects when a d m i n i s t e r e d to rats ( T h i e b o t et al. 1988; M c G r e g o r a n d A t r e n s 1990; F a n s e l o w 1991) o r h u m a n s ( D o r o w et al. I983). I t was therefore o f interest to c o m p a r e the R Q , E E a n d activity changes i n d u c e d by physical stress o r " p s y c h o l o g i c a l " stress with those i n d u c e d p h a r m a c o logically by F G - 7 1 4 2 . Finally, the present study investigated the ability o f the anxiolytic b e n z o d i a z e p i n e d r u g m i d a z o l a m ( M D Z ) to reverse the changes in EE, R Q a n d activity p r o d u c e d b y the three stressors d e s c r i b e d above. T h e r e is c u r r e n t l y m u c h c o n c e r n a b o u t the reliability a n d v a l i d i t y o f the anim a l m o d e l s c o m m o n l y used to assess the efficacy o f anxiolytic drugs, causing a p r o n o u n c e d interest in develo p i n g m o r e valid m e a s u r e s o f anxiety in a n i m a l s ( G r e e n I991). I t was r e a s o n e d t h a t s h o u l d stress p r o d u c e p r o n o u n c e d effects on E E a n d R Q t h a t are b e n z o d i a z e p i n e reversible then the p a r a d i g m s used here m a y be o f use in d e t e r m i n i n g the efficacy o f o t h e r anxiolytic drugs.

Materials and methods

Subjects The three experiments used a total of 52 experimentally naive male Wistar rats weighing between 300 and 450 g. They were group housed in large plastic tubs (six to eight rats per cage) in a temperature controIled vivarium (22 2 _+ 1° C) with a 14:10 h light dark cycle. All testing took place during the light cycle between the hours of 0830 and 1800 hours. The rats had ad tib access to standard lab chow and water, except when inside the metabolic test cage or the conditioning chamber. The rats were randomly assigned to treatment groups with each rat receiving only one treatment and participating in only one experiment.

Drugs Midazolam ("Hypnovel", Roche Australia) was dissolved in a vehicle of 0.9% sterile saline to a volume of l mg/ml which was injected SC at a dose of 1 mg/kg. Control rats received equivalent injections of vehicle. FG-7142 (Research Biochemicals, USA) (10 mg/kg) and RO 15-1788 (10 mg/kg) (a gift from ttoffman-LaRoche, Ltd., Australia) were suspended in 0.9% sterile saline containing 2% Tween 80 and were injected IP at a concentration of I0 mg/ml.

Apparatus RespiratoIT exchange. 02 consumption and CO2 production were measured in an open circuit calorimeter (McGregor et al. 1990, 1991). Individual rats were placed in a cylindrical clear acrylic test cage (volume 6.2 t) with a stainless steel grid floor made of 16 metal bars. Compressed atmospheric air was drawn from a large storage cylinder and passed through the test cage at slight positive pressure. The flow rate was continuously monitored by computer and maintained at 1.5 1.6 1/min by a series of flow controllers and pressure regulators. After passing through the test cage, the air was directed through a Perma Pure (PD-750-12PP) permeation drier. The air stream was then split and a sample of 110 mI/min passed at slight negative pressure through an Ametek CD-3A CO2 analyser and an Ametek S-3A 02 analyser. The analysers were calibrated daily with primary gravimetric standards obtained from Commonwealth Industrial Gases (Sydney). The test cage was placed on a Mettler PE-2000 electronic balance and the unintegrated signal from the strain gauge of the balance recorded every minute by computer. This produced a measure of locomotor activity,. Test sessions were always of t25 rain duration, with the first 5 min used for calibration and settling of the equipment. The first respiratory measures were therefore obtained 5 rain into the test session. In experiment 1, the metal bars of the test cage floor were connected to a Coulbourn Instruments Test Grid Shocker to allow footshock to be introduced into the test cage by the experimenter. In experiment 2, a 60 W clear light globe and buzzer were placed just outside the test cage for presentation of a CS. Aversive conditioning. The aversive conditioning used in experiment 2 took place in a standard Coulbourn test chamber placed inside a wooden sound attenuation box. The Coulbourn chamber had a stainless steel grid floor, the 16 bars of which were connected to a Coulbourn Instruments Test Grid Shocker. A buzzer and 60 W clear globe were placed in the attenuation box just outside the operant chamber. A fan inside the attenuation box produced masking noise throughout the conditioning session. The onset and offset of the globe, buzzer and shocker were controlled by an IBM PC compatible computer programmed using Analog Connection Workbench software (Strawberry Tree, USA).

Procedure In all three experiments, each rat was pre-exposed to the metabolic test cage for 60 rain on each of the 4 days preceding testing. This

477 was to ensure that any stress induced by novelty of the apparatus had dissipated prior to testing.

Experiment 1. Three groups were used in this experiment, namely (1) Shock (n = 5), (2) Control (n = 5), and (3) Shock/MDZ (n = 5). The Shock and Control groups received saline injection immediately before being placed in the test cage while the Shock/MDZ group received M D Z (1 mg/kg). Rats in the Shock and Shock/MDZ groups received four 5-s duration 0.6 mA experimenter-delivered shocks over a 105-s period (shocks starting at 0, 60, 80 and 100 s) inside the test cage. The Control group received no footshock. In the Shock group, the footshocks were delivered only after activity and EE had stabilised. To this end, the experimenter waited until a 10-min period had passed in which activity counts approached zero before delivering shock. This low activity ensured a relatively low and stable baseline EE against which the effects of shock could be compared. The actual time of shock onset in the Shock group was between the 50th and 60th min of testing for individual subjects (mean 52 rain). In the Shock/MDZ group, the experimenter always delivered shock after the 30th rain of testing, regardless of the behaviour of the rat. This was done for two reasons, namely (1) to attempt to ensure a homogenous drug effect in all rats tested, and (2) to ensure an adequate drug effect given that the serum halfqife of M D Z (1.5 mg/kg) in the rat has been shown to be 32.2 min (Tang et al. 1988). The sedative effect of M D Z ensured that baseline EE and activity stabilised much earlier in this group. Experiment 2. Three groups were used in this experiment, (I) CS (n = 7), (2) Control (n = 71) and (3) CS/MDZ (n = 6). On the day before metabolic testing, rats in the CS and CS/MDZ groups were subjected to an aversive conditioning procedure. This involved the presentation of a synchronised flashing light and buzzer (0.5 s on/ 0.5 s off) conditioned stimulus (CS) immediately prior to the delivery of a footshock unconditioned stimulus (US) (5 s, 0.6 mA). Rats were given a single 2-h conditioning session in which 20 such CS-US pairings were presented. The length of the CS was progressively increased over the 20 trials from 30 s to 10 min (actual CS durations (s) 4 x 30, 4 x 60, 4 × 120, 4 x 240, 3 x 360, 1 x 600). This ascending CS length was used to establish a prolonged playsiological response to a single 600-s CS during testing. The intertrial interval (ITI) was varied randomly between 60 and 360 s but the same sequence of ITI values were used for each rat [mean ITI = 194.21 _+21.93 (SEM) seconds]. The Control group received identical exposure to the CSs during the conditioning session but in the absence of the US. Aversive conditioning was performed in a different room to metabolic testing. The test phase occurred 24 h later. Immediately before being placed in the metabolic chamber, rats in the CS/MDZ group received M D Z (1 mg/kg) while the rats in the CS and Control groups received saline. During the test session all rats were presented with a single 600 s CS without any shock delivery. For rats in the CS and Control groups the CS was presented once baseline EE and activity had stabilised (using the same stability criterion as in experiment 1 above). The range of times for CS onset was 55 77 min in the CS group (mean 61 min) and 55-78 min in the Control group (mean 62 min). As in experiment 1, every rat in the CS/MDZ group was presented with the CS 30 min into the test session to ensure an optimal and uniform drug effect.

Experiment 3, In experiment 3, rats were randomly allocated to one of four drug treatment groups (n = 4 per group). The treatment groups were as follows: VEH & VEH, FG-7142 (10 mg/kg) & VEH, FG-7142 (10 mg/kg) & M D Z (1 mg/kg), FG-7142 (10 mg/kg) & RO 15-1788 (10 mg/kg). Rats were placed in the metabolic test cage immediately following injection and their respiratory exchange and locomotor activity monitored for 120 min. Ethics of the animal experimentation These experiments were undertaken according to the principle that studies involving aversive manipulations should employ the bare

minimum number of subjects and conditions necessary to produce reliable effects. Reliable effects were achieved with approximately five to seven rats per group in experiments 1 and 2 and four rats per group in experiment 3. In addition, only single dose determinations of drug effects were performed in order to minimise the number of subjectts exposed to aversive procedures or anxiogenic drugs. The tbotshock used was kept at the minimum level required to produce reliable shock-induced metabolic effects (experiment 1) or aversive CS-induced metabolic effects (experiment 2). All procedures used were approved by the University of Sydney Animal Care and Ethics Committee.

Dam Analysis For each minute of testing, mean 02 consumption, mean COa production and total activity counts were recorded. From these data respiratory quotient (RQ = vol CO2 produced/vol 02 consumed) and energy expenditure [EE (k J) = vol 02 consumed (364 + 113RQ)] were computed. EE was expressed in Joules/g to account for difference in body weight across subjects. In all three experiments, RQ data were transformed to percent of baseline to minimise variance due to slight differences in nutritional status and time of day of testing across rats. The absolute mean RQ values for each group are given in figure captions. In experiment 1, data for RQ, EE and activity were analysed by separate contrasts involving two-way ANOVAs (group x time) comparing (t) Shock versus Control groups, and (2) Shock versus Shock/MDZ groups. The increased family-wise error rate resulting from the use of these two separate contrasts was controlled for by use of the Bonferroni adjustment. This means, in effect, that the threshold of statistical significance was P < 0.025 (e.g. Howell 1992). Statistical analyses were carried out for the pre-shock (baseline) and post-shock periods. Since the pre-shock period was always 30 min in the M D Z group, this provided the maximum length of the baseline period for between-group comparisons. However, since the first few minutes in the metabolic chamber are typically marked by high activity and EE (e.g. McGregor et al. 1991) it was decided that only the 21 min of data preceding shock would be used for baseline. Otherwise the pre-shock data of the Shock/MDZ group would have been qualitatively different by virtue of containing the data from the first 10 min of testing. Data for RQ were expressed as percentage of the RQ seen in the 22nd min before shock. The length of the post-shock period varied across rats due to differences in the time of shock onset. Visual inspection of the data suggested that most effects of stress had dissipated at 30 rain post-shock so this was judged to be an acceptable post-shock period for purposes of group comparisons. In experiment 2, separate contrasts involving two-way A N O V A with Bonferroni adjustment were used to compare the (1) CS versus Control groups, and (2) CS versus CS/MDZ groups on RQ, EE and activity. Separate contrasts were conducted for the pre-CS, CS and post CS periods. The same length of baseline was used as in experiment 1 (i.e. the 21 min prior to CS onset for each subject) and data for RQ were expressed as the percentage of the RQ measured in the 22nd min before CS onset. The CS period was 10 min for each group and the post-CS period was taken as the 20 min following CS offset. In experiment 3, RQ, EE and activity data for the 120-rain test period were compared across the four groups using ANOVA followed by post hoc Newman-Keuls tests for evaluation of differences between individual group means.

Results Experiment 1 T h e c h a n g e s in R Q , E E a n d a c t i v i t y o c c u r r i n g t o f o o t s h o c k a r e i l l u s t r a t e d in F i g . 1. S u m m a r y statistics are

478 115.

i~il !i!i

Table 1. Results of ANOVAs with Bonferroni adjustment in experiment 1

[ - 4 - s.ocK ] i - o - s.oc~,~,~oz/

Shock vs. Control RQ RQ x Time EE EE × Time ACT ACT x Time

105

-~T(iiS) =-7~353 ....

F(29,232) = 18.841"** F(1,8) = 46.588*** F(29,232) = 14.420"** F(I,8) = 23.887** F(29,232) = 2.507***

Shock vs. Shock/MDZ F(I,8) = 2.t14

F(29,232) = 2.684*** F(1,8) = 18.338"* F(29,232) = 5.063*** F(1,8) = 14.704"** F(29,232) = 1.36

* F i 0 . 0 i i / ; c p < 0 . 0 i , ***e < 0.001 90, -20

-15

-10

' '~ 0

-5 TIME

o,6o

:o.,5~

-~. os0. .

10

15

20

2.5

30

(MINUTES)

,,

,

5

,..

ili

;

::il

.

snocx

1-.o- sHoc~z ]

I

gi!:~

cONTROL i

ill L,

in activity that continued for at least 25 min (Fig. 1). The increased EE during and after shock showed a clear dissociation from locomotor activity, since while EE showed a gradual decline over the 30 min following shock, activity showed a steady increase. It can be surmised from this that stress produced a large increase in involuntary EE that dissipated slowly over time. M D Z significantly attenuated both the RQ and the EE increases to shock and, while not affecting the immediate activity burst to shock, significantly attenuated the postshock activity increase (Table 1).

Experiment 2 0.20

.... -20

, ......... -15 -10

~oek-----I~i !!! , .... -5

', ;;;. , . , . . . . 0 5

, .... 10

, .... 15

, .... 20

, • .'"~""',-" 25 30

250 shock---'w~i

SHOCK

! i!~

snoc~,loz CONTROL

200

~oo 50

-20

-15

-10

-5

0 TIME

5

10

15

20

25

30

(MINUTES)

Fig. I. Changes in RQ (upper), EE (middle) and activity (lower) in the Shock, Shock/MDZ and Control groups (n = 5). Data for RQ are expressed as percentage of RQ on the 22nd min prior to shock. Actual baseline values for RQ (+_SEM) were: Shock: 0.944 _+ 0.023, Shock/MDZ: 0.938 + 0.010, No Shock: 0.965 _+ 0.009. Error bars are omitted to enhance clarity

presented in Table 1. Comparison of groups over the baseline period showed no significant differences between groups in the 21-min pre-shock period in RQ, EE or activity (Fs < 1.2) Footshock caused immediate and highly significant increases in RQ, EE and activity. The increase in RQ peaked at about 3 min following shock onset and had dissipated by 10 min post-shock. EE also peaked at about 3 min following shock and remained elevated for at least 30 min post-shock. Shock caused an immediate increase in locomotor activity, followed by a short period of lowered activity and then a gradual and sustained increase

The changes in RQ, EE and activity occurring in experiment 2 are illustrated in Fig. 2. Summary statistics are presented in Tables 2A and 2B for the CS and post-CS periods, respectively. Comparison of groups over the pre-CS (baseline) period showed that there were no significant differences between groups in RQ. However, the CS group showed a significant activity x time [-/7(20,240) = 1.967, P < 0.01] and EE x time effect [F(20,240) = 4.783, P < 0.001] relative to the Control group and a significant EE x time effect relative to the CS/MDZ group [F(20,220) = 2.763, P < 0.001]. These effects reflect the higher activity (and corresponding EE) in the CS group relative to the Control group and in the CS/MDZ group relative to the CS group. However, by the time of CS onset the mean RQ, EE and activity values were broadly similar across groups (see Fig. 2). CS presentation caused an initial rise in RQ in the CS group for approximately 3 min, followed by a precipitous drop in RQ that continued until the time of CS offset. This pattern was reflected in a significant group x time effect in RQ in the CS group relative to the Control group (Table 2A). The CS group also showed a large rise in EE that peaked at CS offset, and which occurred despite very low levels of activity. The Control group showed a similar magnitude increase in EE to the CS group but showed significantly greater activity (Table 2A). The large increase in EE in the Control group can therefore be almost certainly attributed to this locomotor activity. It was observed that CS presentation invariably causing a pronounced orienting response followed by high levels of activity and investigatory behaviour in control rats. This is in stark contrast to rats in the CS group who showed a profound freezing response to the CS. M D Z significantly attenuated the RQ changes seen to the CS (Fig. 2) and also significantly attenuated the

479 102 t

0

1

Table 2. A Results of ANOVAs with Bonferroni adjustment for CS period in experiment 2

~ [ ......

100

CONTROL

CS vs. Control RQ F(1,12) = 0.583 RQ x Time F(9,108) = 5.853*** EE F(1,12) = 0.013 EE x Time F(9,108) = 5.540*** ACT F(1,12) = 4.886 ACT × Time /7(9,108)= 2.206 *P

80.

7

60. 40 20.

v ',,,,,-~" 10

20

30

/ "'--40

50 T~IE

60

70

80

90

100

110

120

(Mlff~,~UTES)

Fig. 3. Changes in RQ (upper), EE (middle) and activity (lower) following saline or FG-7142 alone or in combination with MDZ or RO 15-1788. Data or RQ are expressed as percentagge of RQ on the 1st rain of testing. ActuaI baseline values or RQ (+_SEM) were: saline: 0.964 _+0.012, FG-7142:0.992 + 0.005, FG-7142/MDZ: 0.912 _+0.011; FG-7142/RO 15-1788:0.968 _+0.015. For clarity of exposition, data are averaged over each 5 min of testing ing increases in respiratory rate during fear and anxiety states which would be likely to be associated with higher 02 consumption (e.g. Feleky 1916; Hofer 1970). Less predictable on the basis of previous research is the profound increase in R Q that was observed. The mean increase in R Q to shock was from approximately 0.92 to 1.04, suggesting a shift in substrate utilisation from a mixture of fat and carbohydrates (0.92) to the exclusive use of carbohydrate (1.00 and above). The fact that the R Q actually increases above 1.00 m a y indicate the net syn-

thesis of fat from carbohydrate Kleiber 1975; McGregor et al. 1990, 1991). However, such an interpretation must proceed with caution since RQ can only be claimed to be an accurate measure of energy substrate utilisation when measured for sufficiently long periods (Kleiber 1975; Le Magnen 1985). Large transient changes in R Q may merely reflect short-lived respiratory phenomena such as hyperventilation rather than underlying changes in energy substrate utilisation (Kleiber 1975). Since hyperventilation is sometimes seen with negative emotional states (e.g. Van den Hour et al. 1992), it is important to determine whether the R Q results might merely reflect this factor. A clear argument can be made against this case since hyperventilation would eventually lead to a relative lowering of CO2 production (hypocapnia) relative to 02 consumption, as more and more CO2 is "blown off" through overbreathing (Fried 1987). This would cause a decrease in RQ, an effect that is not evident in the data from experiment 1 (Fig. 1). This suggests that the R Q increase to shock m a y well reflect a transient increase in carbohydrate utilisation, perhaps involving sympathetically-mediated glycogenolysis in the liver. The large increase in activity seen to footshock in experiment t corresponds well with previous reports of a "post-shock activity burst" (Fanselow 1982). Less predictable was the large increase in activity seen following shock offset. On the basis of previous reports (e.g. Fanselow 1982; Fanselow and Helmstetter 1988) one might predict that decreased activity (i.e. freezing) would occur in an environment in which shock had been experienced. However, the extensive (4 × 1 hour) pre-exposure to the test environment given in the absence of shock m a y have produced latent inhibition which attenuated aversive conditioning to that environment. Clearly M D Z was able to significantly attenuate both the R Q and EE increases produced by footshock. Given that benzodiazepines can have analgesic effects, it might be suggested that this merely reflects decreased nociceptive sensitivity in rats given MDZ. However, we have recently shown that M D Z at a dose of 1.25 mg/kg has no analgesic effect in rats in the hotplate test at 20 rain following injection (Harris et al. 1993). It therefore seems that M D Z exerts a more specific inhibitory effect on the neuroendocrine mechanisms through which stress responses are expressed. Previous studies have documented the ability of benzodiazepines to attenuate autonomic and endocrine indices of stress through a centrally mediated process (e.g. Le Fur et al. 1979; Breier et al. 1991).

481 The results of experiments 2 and 3 indicate that two anxiogenic stressors (an aversive CS and F G 7142) produced decreased RQ, an effect that is opposite to that produced by a physical stressor in experiment 1. The pattern of RQ change seen with the aversive CS in experiment 2 is consistent with hyperventilation, with an initial rise in RQ occurring as rapid breathing leads to elevated CO2 production relative to 02 consumption. This may then be followed by a longer term reduction in CO2 production relative to 02 consumption as hyperventilation leads to hypocapnia. RQ may then slowly return to baseline as ventilation returns to normal. Similarly, the extended reduction in RQ seen with FG-7142 might reflect chronic hyperventilation leading to a prolonged state of hypocapnia. An initial transient rise in RQ, typical of hyperventilation, is also suggested with FG-7142 since the mean absolute R Q value was elevated in the FG-7142 group relative to the vehicle at the start of testing (see Fig. 3 legend). Interestingly, chronic hypocapnia is regularly seen in humans suffering from panic disorder and other anxiety states (Van den Hout et al. 1992). This suggests that administration of FG-7t42 to rats might provide an intriguing animal model of panic disorder. The fact that the decreased RQ produced by FG-7t42 or an aversive CS were completely reversed by the anxiolytic drug M D Z clearly suggests that the effect is anxiety related. The benzodiazepine antagonist RO 15 1788 only delayed the RQ effect (Fig. 3) probably reflecting the short half life of RO 15-1788 in rats, reported to be in the order of 16 min following 10 mg/kg IP (Lister et al. 1984; Mandema et al. 1991). It remains possible, however, that the reduced RQ seen with an aversive CS and FG-7142 in experiments 2 and 3 might not reflect hyperventilation but rather an alteration in energy substrate utilisation involving increased reliance on proteins or lipids as an energy substrate (Kleiber 1975). Conclusive determination of this must await further research in which hyperventilation is specifically measured in combination with urinary nitrogen (an index of protein utilisation) and free fatty acid levels (an index of lipid utilisation) to allow an unambiguous assessment of the role of these different factors in the observed RQ decrease. Nonetheless, it is worth noting here that there is a large literature showing increased release of free fatty acids during "psychological stress" in humans (Niaura et al. 1992), an effect that is thought to be mediated, at least partly, by direct sympathetic nervous system innervation of white adipose tissue. Conceivably such a mechanism might underlie the lowered RQ induced by the arguably more "psychological" stressors used in experiments 2 and 3. Finally it is worth noting that FG-7142, in contrast to physical stress or an aversive CS did not increase EE (Fig. 3). In fact there was strong, though non-significant, tendency towards an inhibition of EE with this treatment. This suggests that the physiological response to this "pharmacological stressor" is in some ways very different to that provoked by other more conventional forms of stress. The lowered EE seen with FG-7142 may in part be a reflection of the tendency towards lowered locomotor activity with this treatment, an effect that has been

reported in other studies (Taylor et al. 1985; Jaskiw and Weinberger 1990). In addition, previous studies have shown that FG-7142 induces hypothermia (Taylor et al. 1985; Jackson and Nutt 1991), suggesting that inhibition of basal metabolism may result from administration of this drug. The relatively small sample size used in experiment 3 precludes any definite conclusions being made regarding this issue at present. In conclusion, the major finding of this study is that the measurement of energy expenditure and respiratory quotient can provide a novel and apparently sensitive index of stress. Further studies will hopefully verify the utility of these measures. Two other findings are of clear interest. Firstly, it is evident that different types of stress produce different, sometimes opposite physiological responses. This is evident from the opposite effects on RQ produced by a physical stressor in experiment 1 relative to a "psychological" stressor in experiment 2 and a "pharmacological" stressor in experiment 3. This fact is also evident from the opposite effects on EE produced by the "pharmacological" stressor. This finding suggests that theories which posit a unitary physiological response to diverse stressors are in need of revision. A second interesting finding is that the model outlined here appears to have utility in the assessment of the efficacy of anxiolytic drugs. The ability of midazolam to reverse the physiological changes induced by disparate stressors in the present study indicates that the measurement of respiratory exchange during stress may provide a novel means tbr assessing drug-induced anxiolysis.

Acknowledgements. Research supported by a University of Sydney Research Grant to Iain S. McGregor. Sincere thanks are due to Dale Atrens for allowing generous access to his indirect calorimeter. Peter Home and Jose Menendez are thanked for their expert technical assistance. The authors are grateful to the members of the Australian Learning Group and to Fran Abbott, Keith Franklin, Stylio Nicolaidis and two anonymous reviewers for their valuable comments on this work.

References Breier A, Davis OR, Buchanan (1991) Alprazolam attenuates metabolic stress-induced neuroendocrine and behavioral effects in humans. Psychopharmacology 104 : 479-484 Dorow R, Horowski R, Paschelke G, Amin M, Braestrup C (1983) Severe anxiety induced by FG 7142, a /~-CarboIine ligand for benzodiazepine receptors. Lancet ii: 98-99 Fanselow MS (1982) The postshock activity burst. Anita Learn Behav 10 : 448454 Fanselow MS (1991) The midbrain periaqueductal gray as a coordinator of action in response to fear and anxiety. In: Depaulis A, Bandler R (eds) The midbrain periaqueductai gray matter: functional, anatomical and neurochemical organization. Plenum Press, New York, pp 151 174 Fanselow MS, Helmstetter FJ (1988). Conditional analgesia, defensive freezing and benzodiazepines. Behav Neurosci 102 : 233 243 Feleky A (1916) The influence of the emotions on respiration. J Exp Psychol 1 : 218-241 Fried R (1987) The hyperventilation syndrome: research and clinical treatment. John Hopkins University Press, Baltimore. Green S (1991) Benzodiazepines, putative anxiolytics and animal models of anxiety. Trends Neurosci 14:10t-103

482 Harris J, McGregor IS, Westbrook RF (I993) Low and high doses of midazolam differentially affect hypoalgesia in rats conditioned to a heat stressor. Psychopharmacology 111 : 62 68 Hofer MA (1970) Cardiac and respiratory function during sudden prolonged immobility in wild rodents. Psychosom Med 32 : 633 647 Howell DC (1992) Statistical methods in psychology, 3rd edn. Duxbury Press, US Jackson HC, Nntt DJ (1991) Comparison of the effects of benzodiazepine and ~q-carboline inverse agonists on body temperature in mice. Eur J Pharmacol 205 : 213~16 Jaskiw GE, Weinberger DR (1990) Ibotenic acid lesions of the medial prefrontal cortex potentiate FG-7142-induced attenuation of exploratory activity in the rat. Pharmacol Biochem Behav 36 : 695-697 Johnson EO, Kamilaris TC, Chrousos GP, Gold PW (1992) Mechanisms of stress: a dynamic overview of hormonal and behavioral homeostasis. Neurosci Biobehav Rev 16:115-130 Kleiber M (1975) The fire of life: an introduction to animal energetics. Kreiger, New York. Le Fur G, Guilloux F, Mitrani N, Mizoule J, Uzan A (1979) Relationships between plasma corticosteroids and benzodiazepines in stress. J Pharmacol Exp Ther 211:305 308 Le Magnen J (1985) Hunger. Cambridge University Press, UK Lister RG, Greenblatt DJ, Abernethy DR, File SE (1984) Pharmacokinetic study on RO 15-1788, a benzodiazepine receptor ligand, in the brain of the rat. Brain Res 290 : 183 186 Mandema JW, Gubbens-Stibbe JM, Danhof M (1991 ) Stability and pharmacokinetics of flumazenil in the rat. Psychopharmacology 103 : 384-387

McGregor IS, Atrens DM (1990) Stressor-like effects of FG-7142 on medial prefrontal cortex self-stimulation. Brain Res 516 : 170-174 McGregor IS, Menendez JA, Atrens DM (1990) Metabolic effects obtained from excitatory amino acid stimulation of the sulcal prefrontal cortex. Brain Res 529 : 1-6 McGregor IS, Menendez JA, Atrens DM, Lin HQ (1991 ) Prefrontal cortex ~2 adrenoceptors and energy balance. Brain Res Bull 26 : 683-691 Niaura R, Stoney CM, Herbert PN (1992) Lipids in psychological research: the last decade. Biol Psychol 34 : 143 Selye H (1936) A syndrome produced by diverse nocuous agents. Nature 138 : 32 Tang M, Lau CE, Fatk JL (1988) Midazolam and discriminative motor control: chronic administration, withdrawal and modulation by the antagonist RO 15-1788. J Pharmacol Exp Ther 246 : 1053-1060 Taylor SC, Little H J, Nutt DJ, Setlars N (1985) A benzodiazepine agonist and contragonist have hypothermic effects in rodents. Neuropharmacology 24 : 6 9 ~ 3 Thiebot MH, Soubrie P, Sanger D (1988) Anxiogenic properties of beta-CCE and FG-7142: a review of promises and pitfalls. Psychopharmacology 94 : 452463 Van den Hour MA, Hoekstra R, Arntz A, Christiaanse M, Ranschaert W, Schouten E (1992) Hyperventilation is not diagnostically specific to panic patients. Psychosom Med 53 : 182m191