Retention of nociceptor responses during deep barbiturate anesthesia in frogs

Comparative Biochemistry and Physiology Part C 124 (1999) 203 – 210 www.elsevier.com/locate/cbpc

Retention of nociceptor responses during deep barbiturate anesthesia in frogs Hall Downes *, Dennis R. Koop, Beth Klopfenstein, Nickola Lessov Department of Physiology and Pharmacology, Oregon Health Sciences Uni6ersity, 3181 SW Sam Jackson Park Road, Portland, OR 97201, USA Received 8 October 1998; received in revised form 26 July 1999; accepted 3 August 1999

Abstract Bullfrogs (Rana catesbeiana) anesthetized with a large dose of thiopental (42.8 mg/kg) retained movement responses to nociceptor stimuli despite an average plasma drug level of 51 mg/l, of which 63% was bound to plasma proteins. This concentration, when corrected to include only unbound and uncharged drug, was 2-fold greater than those reported to abolish nociceptor response (NR) during surgical anesthesia in man. The median anesthetic dose (AD50) for loss of the righting reflex was 11.2 mg/kg by s.c. injection into the abdominal lymph sac; however, at 54.0 mg/kg, all frogs retained NRs, although otherwise deeply anesthetized. The ratio of NR-blocking dose to light AD was thus \ 4.8, as compared to B2 in mammalian studies. Whole body levels of thiopental determined at 3 h after intralymphatic injection showed that about half the injected drug had been eliminated by this time and that termination of anesthesia was chiefly due to drug elimination. Even though the pharmacokinetics of thiopental appears to differ markedly in frogs and men, the poor analgesia seen in the present study frequently has been reported during clinical barbiturate anesthesia. Since this deficiency is much more pronounced in the bullfrog than in man, its neurophysiological basis might profitably be studied using the bullfrog as a model; however, the high mortality associated with deep thiopental anesthesia in the frog should preclude its use as a practical anesthetic in amphibia. © 1999 Elsevier Science Inc. All rights reserved. Keywords: Rana catesbeiana; Frog; Anesthesia; Thiopental; Barbiturate; Analgesia; Nociception; Blood levels; Protein binding

1. Introduction It is commonly thought that the effective concentration of a general anesthetic is relatively constant throughout vertebrate classes [29]. However, in a previous study in bullfrog tadpoles [10], we found that all five of the anesthetics tested required very high aqueous concentrations to produce the usual endpoints of anesthesia, such as loss of the righting reflex (RR) or loss of nociceptor response (NR). The resistance to anesthetic depression was particularly pronounced for barbiturate effect on the NR, as measured by loss of propulsive swimming movements in response to pinching of the tail fin. For thiopental, the steady-state aqueous concentration of uncharged drug in the ambient solution associated with loss of the tadpole NR was 6 – 11-fold higher than the equivalent plasma levels of unbound * Corresponding author. Tel.: +1-503-494-7808; fax: + 1-503-4944352.

and uncharged drug associated with loss of NR in mammals, as calculated (see Materials and methods) from plasma levels reported in men [2] and rats [15]. Direct measurement of unbound drugs in tadpole blood confirmed an equilibrium between uncharged drug in the bath and in the aqueous phase of blood [10]. A lack of analgesic effect — or even an anti-analgesic effect — has been noted at low doses of barbiturates in both man [7,11] and rodents [1,26]; however, in mammals, barbiturates depress the NR at doses that are only slightly higher (B 2-fold) [21] than those needed to suppress the RR and that are clearly in the sublethal range. Thiopental in the tadpole [10] had a median anesthetic concentration for loss of nociceptor response, the AC50 (NR), that was six times greater than the median anesthetic concentration for loss of righting reflex, the AC50 (RR), and the concentrations used to determine the AC50 (NR) produced death in some animals. In contrast, in the same experiments, the ratios of AC50 (NR)/AC50 (RR) for the non-barbiturate

0742-8413/99/$ - see front matter © 1999 Elsevier Science Inc. All rights reserved. PII: S 0 7 4 2 - 8 4 1 3 ( 9 9 ) 0 0 0 6 9 - 9

204

H. Downes et al. / Comparati6e Biochemistry and Physiology, Part C 124 (1999) 203–210

drugs (ether, ethanol ,and diazepam) were between 1.6 and 1.9. Thus, the lack of analgesic effect of thiopental in the tadpole was reflected both in the very high concentrations per se and in the high AC50 (NR)/AC50 (RR) ratio. Since the resistance of the tadpole NR to barbiturate depression could represent a distinctive characteristic of neural circuits in the tadpole stage, the present study was undertaken to determine whether adults of the same species also required exceptionally high drug levels to block NR or showed a large separation between the median anesthetic dose (AD50) (NR) and the AD50 (RR). Preliminary studies showed that the high injected doses of thiopental ( 65 mg/kg) needed to completely block the NR produced a moribund animal and greatly impaired collection of an adequate volume of blood by arterial or ventricular puncture. Therefore, for studies of plasma drug levels, we determined total and unbound drug in arterial blood after administration of a lower dose (42.8 mg/kg), which depressed but did not prevent the somatic response to nociceptor stimulation. Thiopental levels also were determined in homogenates of the whole animal at 3 h after drug injection to determine if the wear-off of effect evident at this time represented drug loss or pharmacodynamic tolerance.

2. Materials and methods

2.1. Animals Housing conditions and experimental protocols were approved by the Institutional Animal Care and Use Committee (IACUC). Postmetamorphic bullfrogs (Rana catesbeiana) obtained from commercial sources (Charles Sullivan, Nashville, TN) were housed in slanted-floor terraria with free access to water and were fed crickets ad libitum. Tests involved three batches of animals. Group I consisted of three large bullfrogs (260 –386 g) used to observe the effect of intravenous (i.v.) thiopental. For that purpose, on the day before receiving thiopental, they were anesthetized by intraperitoneal (i.p.) injection of tricaine methanesulfonate (180–200 mg/kg) to permit placement of a PE50 catheter in the right 6ena cutanea magna under direct vision through a small skin incision. The catheter was filled with heparinized 0.7 N saline and secured in place with sutures, the skin incision was sutured closed, and the animals were allowed to recover. Group II contained six large frogs (196 – 440 g) used to obtain arterial blood during moderately deep thiopental anesthesia. These received two consecutive i.p. injections of thiopental (21.4 mg/kg) about 20 min apart. Since these two doses were inadequate to block all somatic movement in response to nociceptive stimuli, the animals were rendered completely unresponsive by

intracranial injection of bupivacaine just prior to opening the body cavity to obtain blood by catheterization of an aortic arch or puncture of the cardiac ventricle. A 25 gauge needle was inserted cephalically through the foramen magnum and 0.25 ml of 0.75% bupivacaine HCl was deposited either in the space separating the periosteum of the skull from the meninges or, if the meninges were punctured, directly into the cerebral spinal fluid on the surface of the brain; in either event, the animals became completely unresponsive within less than 1 min after the injection. Assessment of anesthetic depth was conducted just before the intracranial injection of bupivacaine, which was at  18 min after the second injection of thiopental; preliminary studies had shown that the peak effect of thiopental occurred at about 20 min after either i.p. or subcutaneous (s.c.) injection. The blood samples were immediately centrifuged and the plasma stored at − 80°C until analysis for barbiturate levels. Group III contained small postmetamorphic bullfrogs (6–18 g) used to determine dose–response relationships and to measure the disappearance of thiopental from whole body homogenates. Since at lower doses these animals were quite active, they received thiopental by s.c. injection with precautions to prevent any possibility of drug loss through leakage from the puncture site; a number 27 needle was introduced into the saccus submaxillaris lymph space and passed through the connective tissue partitions into the adjacent saccus thoracis and then into the saccus abdominalis. After injection of drug into the saccus abdominalis, the animals were held in a supine position for 1 min as a further safeguard against leakage. Observation of behavioral effects occurred at 15-min intervals for 60 min in those not losing the RR, and until recovery of the RR in those that did. Animals in the behavioral studies (n= 12) received different doses of thiopental (5.1, 7.7, 10.2, 12.8, 15.4, 21.4, 54.0, and 64.2 mg/kg) at  2-week intervals for a total of three to four doses per animal. Five or six animals were used to test the effect of each dose. The lowest dose did not prevent righting in any animal and served as a control for possible ‘placebo’ effects of handling and injection. Frogs of similar size (n= 6) were used to determine the whole body content of thiopental at 1–2 min and at 180 min after s.c. injection of 42.8 mg/kg of thiopental; since NRs were present, these animals were sacrificed by pithing immediately before homogenization in 100 ml of 0.5 M potassium phosphate buffer (pH 5.8). Homogenization was performed using  12 short highspeed bursts (3 s) spaced over a 5-min interval. Homogenates were frozen and stored at − 20°C for determination of whole body thiopental levels later in the same week. Frogs that received thiopental doses sufficient to depress ventilatory movements were placed on moist towels in closed chambers ventilated with 100% oxygen

H. Downes et al. / Comparati6e Biochemistry and Physiology, Part C 124 (1999) 203–210

at flow rates of 0.5– 2 l/min. Care was taken to keep the skin moist at all times, but the animals were not immersed in water in order to avoid trancutaneous drug loss.

2.2. Tests of anesthetic depth The RR was rated as lost if the animal was unable to right itself within 60 s of being placed in a supine position. Nociceptor responsiveness was tested in three ways: 1) repeated pinching of the carpus with toothed forceps; 2) similar repeated pinching of the abdominal skin; 3) stroking of the cornea with a wisp of moistened, soft paper. A positive response to carpal pinch was any movement of the forearm or shoulder girdle; typically, the most drug-resistant part of this response was a shrugging motion of the shoulder girdle which was both visible and palpable. A positive response to abdominal pinch involved contraction of the abdominal muscles to produce a quick ‘jackknifing’ movement of the whole body. A positive response to corneal stimulation was blinking of the lid. The NR was rated as lost only if all three of these NRs were lost to repeated stimuli. NR stimuli were always repeated several times (  3), since a single stimulus sometimes did not elicit a response in animals that responded vigorously to a second or third stimulus. Typically, these three responses disappeared at about the same time, but individual animals might lose one or two such responses in a lighter level of anesthesia than required to suppress the third. Preliminary studies showed that if any of these three were present, subsequent surgical procedures would evoke gross movement. Such movements, although typically feeble and poorly coordinated, indicated that true surgical anesthesia had not been attained. Movement of the leg in response to pinching of the hindfoot, either manually or with forceps, was found to be a less reliable test for surgical anesthesia and potentially damaging during repeated testing. Indeed, coordinated and ‘purposeful’ movements of the legs were depressed in relatively light anesthesia.

2.3. Analytical procedure Determination of thiopental, total and unbound, were as described in our previous study [10], as modified from Burch and Stanski [5] and Ebling et al. [12]. Standard stock solutions of thiopental in 50% methanol were prepared and added to blank plasma, frog homogenate, or distilled water to produce standard solutions of thiopental in the range of 5–200 mg/ml. The internal standard, thiamylal, was prepared in 50% methanol at a concentration of 0.1 mg/ml. Plasma samples were diluted with 100 mM sodium phosphate buffer (pH 7.8) to produce a final phosphate buffer concentration of 10 mM (nine parts plasma+

205

one part 100 mM sodium phosphate buffer). The pH of the samples was verified to be in the range of 7.7–7.9. Samples were then subjected to ultrafiltration for 30 min at 2000 × g with an Amicon Centrifree Micropartition Device (Amicon, Inc., Beverly, MA) with a cutoff of 30 000 daltons [10]. Centrifugation was performed using a 28° fixed-angle rotor in a Sorvall RC-2B Superspeed temperature-controlled centrifuge maintained at 20°C. Protein concentrations in the supernatant were determined by a Coomassie Brillant Blue assay (BioRad, Hercules, CA). Aliquots of diluted whole plasma and diluted filtered plasma were placed in glass extraction tubes (16 mm × 150 mm) and the internal standard, thiamylal [12], was added to produce a concentration of 5 mg/ml. Samples were then extracted with six volumes of pentane by vortexing for 30 s and centrifuging to separate the phases. The organic phases were transferred to 13 mm× 100 mm glass tubes and dried under reduced pressure. Whole frog homogenate or fluid samples were placed in glass extraction tubes (16 mm× 150 mm) and thiamylal was added to produce a concentration of 1 mg/ml. Eight volumes of pentane were then added and the tubes were vortexed for 30 s. The tubes were centrifuged to separate the phases, and the organic phases were transferred to 13 mm× 100 mm glass tubes and dried under reduced pressure. The pentane extracts were evaporated to dryness and the residues containing thiopental and thiamylal were dissolved in 200 ml of 55% methanol and resolved on a Shimadzu HPLC system (LC-10 AS solvent delivery units and SPD-10AV UV-VIS detector) equipped with a Supelcosil LC-18 column (4.6 mm×15 cm ×5 mm). The column was eluted with a mixture of acetic acid (0.1%, adjusted to pH 4.5 with 1 N NaOH) and methanol with the methanol concentration beginning at 55% and increasing linearly to 95% over the next 15 min at a flow rate of 1 ml/min. Samples were analyzed for absorption at 290 nm. The retention times of thiopental and thiamylal were 7.3 and 8.1 min, respectively. Samples obtained from untreated frogs showed no detectable absorption at these times. Thiopental concentration was quantified by comparing peak area with the internal standard. With a fixed amount of internal standard and an increasing amount of thiopental (0–200 mg/ml), the ratio of thiopental to thiamylal was linear with thiopental concentration, with a r 2 of 0.998.

2.4. Calculations For calculations of the unbound form of thiopental from plasma levels reported in the literature, 15% of the total drug was considered to be unbound in both man [2,5] and rat [13].

206

H. Downes et al. / Comparati6e Biochemistry and Physiology, Part C 124 (1999) 203–210

For calculation of uncharged thiopental, the pKa of thiopental was taken to be 7.6 at room temperature [20,31] and 7.4 at 37°C, assuming a temperature coefficient for ionization similar to those of other barbiturates [22]. The average pH of bullfrog plasma in vivo at 20°C has been reported as 7.87 [17] and 7.96 [14] in normal breathing animals and 7.59 [14] during apnea in 100% oxygen. The pH was taken to be 7.8 for calculation of uncharged thiopental in the animals used in this study, since ventilatory movements, although usually present, were obviously depressed. Quantal dose–response curves were analyzed by the method of Litchfield and Wilcoxon [25] to determine an AD50 (RR) and its 95% confidence limits. All other statistical values are expressed as means9 standard errors (9 S.E.).

2.5. Chemicals Thiopental Na product was obtained from Sigma Chemical Co (St. Louis, MO) and was dissolved in distilled water to make a 2.5% solution for i.p. or s.c. injection. Such thiopental Na products contained calcium carbonate as a buffer in addition to the drug salt. Chemical analysis by Sigma of thiopental Na product showed that thiopental as the free acid represented 85.64% of the thiopental Na product [10]. Drug doses and concentrations are reported as the free acid. For studies with i.v. injection, a similar 2.5% solution was made from the thiopental Na product obtained from Abbott Laboratories (Chicago, IL). Bupivacaine HCl was employed as the 0.75% clinical solution marketed by Astra Pharmaceuticals (Wayne, PA). For HPLC analysis, HPLC grade methanol was obtained from Mallinckrodt Chemical Co. (Paris, KY), glacial acetic acid from Baker Chemical Co. (Phillipsburg, NJ), thiamylal and pentane (HPLC grade) from Sigma Chemical Co. (St. Louis, MO) and monobasic and dibasic sodium phosphate from Fisher Scientific (Pittsburgh, PA).

3.2. Subcutaneous injection The AD50 (RR) for thiopental was 11.2 mg/kg (9.0– 14.0 mg/kg 95% CL). At doses close to the AD50 (RR), peak effects occurred in all animals at between 5 and 40 min post injection, with peak effects in 60% at between 15 and 25 min. At 42.8 (n= 3) and 54.0 mg/kg (n= 5), all animals retained a discernable NR, although with considerable impairment of motor function; at 64.2 mg/kg, three of five lost the NR and all of these failed to recover. Occasional buccal respirations persisted after 42.8 mg/ kg, but the 54.0 and 64.2 mg/kg doses produced complete apnea in all. The average time to recover the RR was 385 min (32 min S.E.) in the five receiving 54.0 mg/kg and 544 min in the two animals that recovered after 64.2 mg/kg. The animals receiving 42.8 mg/kg were sacrificed for determination of drug levels at 3 h post injection; at this time, only one had recovered its RR (at 138 min post injection).

3.3. Intraperitoneal injection and plasma le6els At  18 min after the second of two injections of 21.4 mg/kg of thiopental, all of the frogs had lost RR, but none had lost NR. Ventilatory movements were suppressed in all animals, but only one was apneic. At this time, the frogs were given an intracranial injection of bupivacaine to permit opening of the body cavity and withdrawal of blood from an aortic arch or ventricle. In blood drawn at 20 min after the second thiopental injection, the plasma thiopental levels (Table 1) ranged from 28.8 to 66.5 mg/ml (mean 50.6 96.0 S.E., mg/ml). Of this, 63.4 ( 92.4 S.E.) mg/ml was protein-bound. In a sample of rat plasma spiked with 40.9 mg/ml of thiopental and filtered at 36°C, the thiopental was 82.8% protein-bound, which was similar to binding Table 1 Plasma levels at peak effecta of 42.8 mg/kg Thiopental free acid, mg/ml

3. Results

3.1. Intra6enous injection Injection of 18.3 mg/kg of thiopental produced loss of RR in three of three frogs, but all retained vigorous withdrawal responses. There was a marked lag in onset of action, with almost no effect apparent during the first 3 min post injection and peak depression at 4.5 – 8 min. The RR was lost at 4.5 – 6 min and returned in from 288 to 405 min (mean= 341 min) post injection.

Frog wt, g

Total

Unbound

Unbound and unchargedb

196 144 198 156 440 114

38.49 54.16 64.86 50.92 28.82 66.52

13.54 24.06 19.84 17.74 9.14 28.56

5.24 9.31 7.68 6.86 3.54 11.05

Mean( 9 S.E.)

50.6 (6.0)

18.8(2.9)

a

7.3(1.1)

Approximately 20 min after the second of two intraperitoneal doses of 21.4 mg/kg. b Assuming a plasma pH of 7.8 for bullfrogs at room temperature.

H. Downes et al. / Comparati6e Biochemistry and Physiology, Part C 124 (1999) 203–210

reported by others [13]. Plasma protein was 3.5 ( 9 0.2 S.E.) mg% in the frog samples and 5.2 mg% in the rat.

3.4. Whole body homogenates In three frogs receiving 42.8 mg/kg s.c., the concentration found in the homogenates at 3 h post injection ranged from 6.8 to 29.5 mg/g of frog (mean 19.5, S.E. 4.9 mg/kg), with the lowest concentration in the animal that had recovered its righting reflex prior to sacrifice. These levels accounted for 45.5% of the total drug injected. Another 1.1% was accounted for in the wash water used to remove expelled urine from the test chamber (total drug recovery of 46.6%). In the three control frogs that were homogenized within 2 min of injection of the same dose, drug levels ranged from 37.7 to 45.1 mg/g of frog (mean 41.4, S.E. 2.1 mg/g), indicating virtually complete recovery of the injected dose.

4. Discussion The endpoints used in this study and in the preceding study in tadpoles [10] are those customarily employed in studies of anesthetic potency in laboratory rodents, i.e. loss of righting and loss of gross movement in response to a mock surgical stimulus [21,29]. These are quantal ‘yes – no’ types of response, and it is important to recognize that there is marked motor dysfunction at doses substantially subthreshold for complete loss of the response. Loss of NR, therefore, marked true surgical anesthesia with a generalized depression of neural function rather than the selective analgesic effect typical of opiate-type drugs. Drugs such as diethyl ether, ethanol, and diazepam [10] abolished NR at concentrations less than twice those needed to block the RR and that were without lethal effect in any animal. In contrast, the high concentrations of thiopental needed to abolish NR were frequently lethal in both larval and adult frogs. For this reason, the blood samples obtained in the present study were drawn at doses that depressed, but did not completely abolish, NR. Our choice of anuran species originally was dictated by the large size of the bullfrog tadpole, which both facilitated testing of motor responses and provided sufficiently large blood samples on ventricular puncture to permit measurement of blood drug levels. The present study, for continuity with the preceding work in larval animals [10], used the same species and similar end points. Stevens et al. [27,33,34] have developed a sensitive test for the analgesic effect of opiates in frogs, that uses as its endpoint the vigorous wiping response of a

207

Table 2 Thiopental concentration (mg/ml) at loss of NR total drug/unbound and uncharged Nociceptor test Mana, Cp50 (NR)2

Rata, Cp50 (NR)15

Corneal reflex Trapezius squeeze Tail clamp Tail pressure

60.0/4.5

39.4/2.45 42.4/3.15

38.3/2.87 39.2/2.94

Tadpoleb, AC50 (NR)10

–/28.1

a

Cp (plasma concentration) at which 50% of subjects lost NR. AC (anesthetic concentration) in the bath solution at which 50% of subjects lost NR. b

hindlimb elicited by brief application of a drop of acetic acid to the skin of the dorsal thigh. This procedure yields a graded dose–response curve in which increasing the dose of an analgesic drug progressively increases the strength of acetic acid needed to elicit the response. Doses are chosen so that they do not produce motor dysfunction or marked sedative effects. We did not use this test since it was inapplicable to immersion anesthesia, as used in our preceding work [10], and not well suited for use in the present study, in which the motor impairment accompanying general anesthesia could preclude a clear-cut wiping response irrespective of the strength of acetic acid. In the tadpole study [10], we calculated steady-state AC50s using the concentrations of nonionized drug in the surrounding bath after a 4–6-h incubation in a constant concentration of thiopental. Samples of ventricular blood obtained at that time showed that the unbound and uncharged drug in blood was virtually identical to that in the bath water. We assumed an equilibration between blood and neuraxis, since there was a stable level of anesthetic depression. In the present study, in which steady-state conditions did not apply, we assumed that blood and neuraxis were in near equilibrium at the time of peak effect (20 min post i.p. injection) and during the subsequent slow decline in drug level and effect. The slow time course for both the onset of and recovery from behavioral depression contrasts dramatically with that seen in mammals [13,28], and indicates that blood and brain levels were not rapidly changing during the period in which blood samples were obtained. Therefore, these samples, which were obtained 20 min after the second of two successive injections (40 min after the first injection) were at or near the peak effect of the second injection and during the declining phase of the first, and should be reasonable estimates of the equilibrium blood levels producing the observed degree of anesthetic depression.

H. Downes et al. / Comparati6e Biochemistry and Physiology, Part C 124 (1999) 203–210

208

4.1. Plasma le6els Table 2 shows the plasma drug concentrations in man [2] and rat [15] that were associated with loss of NR in 50% of subjects, and compares them with the AC50 (NR) of uncharged and unbound drug in tadpoles, as determined in our previous study [10]. We were unable to obtain blood samples from adult bullfrogs anesthetized to loss of the NR and instead sampled plasma from frogs that received a lower dose equivalent to about two-thirds of the estimated AD50 (NR) (see below). Total plasma barbiturate in these frogs (Table 1), all of which retained a discernable NR, was in the same range as the AC50 (NR)s of total barbiturate as determined in mammalian studies (Table 2); however, since 37% of the total drug in bullfrog plasma was in the unbound form compared to 15% in man [2,5] and rat [13], the concentration of unbound drug was 2–3-fold higher in the frogs. In terms of freely diffusable drug, the lesser degree of protein binding will be partly offset by a greater ionization at the more alkaline pH of amphibian blood. Nonetheless, at a plasma pH of 7.8 (see Materials and methods), the average concentration of freely diffusable drug (Table 1, unbound and uncharged) was still 2-fold higher than mammalian AC50 (NR)s of freely diffusable drug (Table 2). Furthermore, since ventilatory movements were obviously depressed by this dose, plasma pH may have been more acidic, which would have further increased freely diffusable drug. For example, at a plasma pH of 7.59, as reported in bullfrogs at room temperature in which gas exchange was exclusively by diffusion across the skin [14], the levels of freely diffusable thiopental would have been 31% greater than shown in Table 1. The only other study of thiopental levels during surgical anesthesia in an amphibian species is that of Shim and Andersen [30] in the marine toad, Bufo marinus. They found blood thiopental concentrations of 28 mg/ml of total drug as the average concentration needed to abolish movement in response to clamp of the lower extremity; this concentration is actually less than the concentrations associated with block of mammalian NR (Table 2). The discrepancy between their

results and ours could represent differences in the type of nociceptor stimulus employed; indeed, we did not employ foot clamping as a test because in our bullfrogs it gave inconsistent results that often did not agree with other tests of nociception. Alternatively, the difference may reflect a species difference, possibly related to a difference in protein binding. Drug binding to the plasma proteins of ectothermic vertebrates has rarely been studied. The single report involving thiopental is in the nurse shark, Ginglymostoma cirratum [35], in which only 10–14% of plasma drug was bound to protein; a similar low level of protein binding in the toads used by Shim and Andersen could underlie the greater analgesic potency of thiopental in their study.

4.2. Dose response Table 3 compares the injected doses in human beings, rats, and bullfrogs required to produce light anesthesia versus anesthesia of sufficient depth to block NR. Indices of light anesthesia were loss of RR in bullfrogs and rats and loss of the eyelid reflex in human subjects. The route of injection and time course for drug effect are different in the bullfrog from those in the two mammals; however, the ratio of doses required to produce different endpoints of anesthesia (Table 3, righthand columns) ought to afford a valid cross-species comparison of analgesic versus anesthetic effect, irrespective of differences in pharmacokinetics. Since the very high doses required in the bullfrog to abolish NR were usually lethal, the true value for the AC50 (NR)/ AC50 (RR) ratio could not be obtained; however, the continued presence of the NR at the highest nonlethal dose indicates that this ratio was \ 4.8 and at least 3-fold greater than seen in mammalian studies (Table 3). Therefore, dose–response relationships also show a relative lack of analgesic potency of thiopental in the bullfrog. Although s.c. injection was employed in the frogs to obtain the above ratios, we also administered thiopental by i.v. injection. The 18.3 mg/kg dose, which we gave as an i.v. bolus, is about the AD50 (NR) in the rat [21]. This dose had little, if any, effect on the bullfrog NR, although it produced a very long period of loss of

Table 3 Thiopental doses and dose ratios for light anesthesia compared to block of NR Dose, mg/kg

Rat, i.v. Man, i.v. Bullfrog, i.l. a b

Dose ratio

Light anesthesia (LA)

Block of NR

Lethal dose

AD50 (RR)21 12.9 AD50a,8 2.9 AD50 (RR) 11.2

AD50 (NR)21 17.2 AD50 (NR)b,8 5.0 AD50 (NR) \54.0

LD50

21

57.8

LD50 64.2

Lid reflex (blink in response to brushing both eyelashes), children 5–15 years of age. Response to squeeze of the trapezius muscle, children 5–15 years.

NR dose:LA dose 1.3 1.7 \4.8

Lethal dose:NR dose 3.4 1

H. Downes et al. / Comparati6e Biochemistry and Physiology, Part C 124 (1999) 203–210

RR. The long duration of effect coupled with the marked lag in onset of effect indicates a pronounced difference in the i.v. pharmacokinetics of thiopental in the bullfrog compared to mammalian species, a difference which we did not pursue further in this study. The slow time course for the onset and offset of effect of i.v. thiopental in the frog has been noted previously [16]. Probably the major factor responsible is the low blood flow to the frog’s brain relative to its rapidly perfused viscera, such as the gastrointestinal tract, kidney, and spleen [6]. Indeed, the bullfrog brain is only slightly more rapidly perfused than muscle [6]. This could account not only for the lag in onset of effect, but also for the prolonged duration of anesthesia, since the drug would not rapidly redistribute from brain to muscle, as seen in mammals [13,28]. The blood– brain barrier in amphibia is based on the ‘tight’ junctions of the vascular endothelium [4,9], as in mammals, and should not be a significant hindrance to passage of a highly lipid soluble drug such as thiopental; however, the low body temperature of the frog will contribute to the delay in onset of effect because of the temperature dependence of the coefficient of diffusion [23]. At least in theory [24,32], a low body temperature also might affect anesthetic potency, since cold has the potential to antagonize the nonspecific changes in membrane lipids produced by anesthetic drugs. However, such a nonspecific effect (if any) should not alter the ratio of NR- to RR-blocking doses. Furthermore, in the preceding study in tadpoles [10], involving several nonbarbiturate anesthetics, the ratio of AC50 (NR)/AC50 (RR) was \2 only for the barbiturate drugs. Kaplan et al. [19] have previously reported on the high mortality produced by NR blocking doses of thiobarbiturates (thiopental and thiobarbital) in frogs (R. pipiens). The 60 – 70 mg/kg dose of thiopental Na product ( 51.4–60.0 mg/kg of the free acid) needed to produce ‘surgical anesthesia’ in their study is close to the dose needed to block NR in our study ( \54.0 but =64.2 mg/kg). The mechanism underlying the high lethality of NR-blocking doses of thiobarbiturates in frogs in unknown and does not apply to oxybarbiturates. Indeed, both hexobarbital and pentobarbital have been recommended for anesthetic use in frogs [18,19]. In our previous study in tadpoles [10], the AC50 (NR)/AC50 (RR) ratio was even higher for hexobarbital (ratio of 9.8) than for thiopental (ratio of 6.2). We chose to use thiopental rather than hexobarbital for the present study, since we were unable to find either concentration – or dose – response studies for hexobarbital effect on NR in mammals, despite the voluminous literature on hexobarbital-sleep time as measured by return of the RR.

209

4.3. Drug elimination A peculiar feature of barbiturate anesthesia in frogs is the recovery within a few hours, which is not likely to be explained on the basis of redistribution. Since amphibia are presumed to have relatively little drug metabolizing capability [3], the relatively rapid recovery of the RR in the animals used in this study suggests either a high degree of acute tolerance or an unexpectedly rapid drug elimination. To test the latter possibility, whole body levels of thiopental were determined at 3 h after injection, at which time anesthesia had noticeably lightened. At that time, about half of the thiopental had been lost from the body, which was sufficient in itself to explain the apparent lightening of anesthesia. Assuming a log-linear elimination process, the estimated half-lives in individual frogs would be 1.1, 3.5, and 5.7 h. Only one of these animals recovered RR prior to sacrifice; in this animal (t0.5  1.1 h), the estimated whole body thiopental level at the time of recovery of RR (138 min post injection) would have been 10.4 mg/g, which is in good agreement with the AD50 (RR) of 11.2 mg/kg. Therefore, contrary to expectations, thiopental elimination in the bullfrog occurs within a time frame similar to that seen in mammals, e.g. elimination half-lives of 1.4 h in the rat [36] and 12 h in man [5]. The route of drug loss is undetermined, but clearly did not involve urinary excretion or loss across the skin, since the frogs were not immersed in water and only 1% of injected drug (presumably lost in urine) could be recovered from the water used to rinse out the test chamber. It should be pointed out that Brodie and Maickel [3] also found a relatively rapid elimination of hexobarbital from bullfrogs, although elimination from R. pipiens and R. esculenta was negligible.

5. Summary Adult as well as larval bullfrogs require exceptionally high concentrations of barbiturate to abolish movement responses to nociceptor stimuli. For thiopental, the doses needed to prevent such movement responses overlapped the lethal range, and were more than 4.8 times greater than those adequate to produce loss of RR and other signs of light anesthesia. Although thiopental is obviously a poor choice for anesthetic use in the bullfrog, the marked resistance of the bullfrog NR to barbiturate-induced depression affords a potentially useful model for study of the so-called anti-analgesic effects of barbiturates.

210

H. Downes et al. / Comparati6e Biochemistry and Physiology, Part C 124 (1999) 203–210

References [1] Archer DP, Ewen A, Roth SH, Samanani N. Plasma, brain, and spinal cord concentrations of thiopental associated with hyperalgesia in the rat. Anesthesiology 1994;80:168–76. [2] Becker KE. Plasma levels of thiopental necessary for anesthesia. Anesthesiology 1978;49:192–6. [3] Brodie BB, Maickel RP. Comparative biochemistry of drug metabolism. In: Brodie BB, Erdo˜s EG, editors. Proceedings of the first international pharmacological meeting. New York: The Macmillan Company, 1962:299–324. [4] Bundgaard M. Ultrastructure of frog cerebral and pial microvessels and their impermiability to lanthanum ions. Brain Res 1982;241:57 – 65. [5] Burch PG, Stanski DR. The role of metabolism and protein binding in thiopental anesthesia. Anesthesiology 1983;58:146 – 52. [6] Burggren W, Moalli R. ‘Active’ regulation of cutaneous gas exchange by capillary recruitment in amphibians: experimental evidence and a revised model for skin respiration. Resp Physiol 1984;55:379 – 92. [7] Clutton-Brock J. Some pain threshold studies with particular reference to thiopentone. Anaesthesia 1960;15:71–2. [8] Cote´ CJ, Goudsouzian NG, Liu LMP, Dedrick DF, Rosow CE. The dose response of intravenous thiopental for the induction of general anesthesia in unpremedicated children. Anesthesiology 1981;55:703 – 5. [9] Crone C, Olesen SP. Electrical resistance of brain microvascular endothelium. Brain Res 1982;241:49–55. [10] Downes H, Courogen PM. Contrasting effects of anesthetics in tadpole bioassays. J Pharmacol Exp Ther 1996;278:284– 96. [11] Dundee JW. Alterations in response to somatic pain associated with anesthesia: II. The effect of thiopentone and pentobarbitone. Br J Anaesth 1960;32:407–14. [12] Ebling WF, Mills-Williams L, Harapat SR, Stanski DR. Highperformance liquid chromatographic method for determining thiopental concentrations in twelve rat tissues: application to physiologic modeling of disposition of barbiturate. J Chromatog 1989;490:339 – 53. [13] Goldstein A, Aronow L. The durations of action of thiopental and pentobarbital. J Pharmacol Exp Ther 1960;128:1–6. [14] Gottlieb G, Jackson DC. Importance of pulmonary ventilation in respiratory control in the bullfrog. Am J Physiol 1976;230:608 – 13. [15] Gustafsson LL, Ebling WF, Osaki E, Stanski DR. Quantitation of depth of thiopental anesthesia in the rat. Anesthesiology 1996;84:415 – 27. [16] Hong SS, Cho KO. Some observations on the reduction of thiopental sleeping time induced by fat in frogs and rabbits. Arch Int Pharmacodyn Ther 1959;118:249–57.

.

[17] Howell BJ, Baumgardner FW, Bondi K, Rahn H. Acid-base balance in cold-blooded vertebrates as a function of body temperature. Am J Physiol 1970;218:600 – 6. [18] Kaplan HM. Anesthesia in amphibians and reptiles. Fed Proc 1969;28:1541 – 6. [19] Kaplan HM, Brewer NR, Kaplan MR. Comparative value of some barbiturates for anesthesia in the frog. Proc Anim Care Panel 1962;12:141 – 8. [20] Kazimierczuk Z, Psoda A, Shugar D. Ultraviolet and infrared spectral studies on the structure and tautomeric forms of 2-thiobarbituric acids and 2-thiobarbitals. Acta Biochem Pol 1973;20:83 – 99. [21] Kissin I, McGee T, Smith LR. The indices of potency for intravenous anesthetics. Can Anaesth Soc J 1981;28:585–90. [22] Krahl ME. The effect of variation in ionic strength and temperature on the apparent dissociation constants of thirty substituted barbituric acids. J Phys Chem 1940;44:449 – 63. [23] Kutchai HC. Cellular physiology. In: Berne RM, Levy MN, editors. Physiology, 4th Edition. St. Louis: Mosby, 1998:7–8. [24] Lever MJ, Miller KW, Paton WDM, Smith EB. Pressure reversal of anesthesia. Nature 1971;231:368 – 71. [25] Litchfield JT, Wilcoxon F. A simplified method for evaluating dose-effect experiments. J Pharmacol Exp Ther 1949;96:99–113. [26] Neal MJ. The hyperalgesic action of barbiturate in mice. Br J Pharmacol 1965;24:170 – 7. [27] Pezalla PD. Morphine-induced analgesia and explosive motor behavior in an amphibian. Brain Res 1983;273:297 – 305. [28] Price HL, Kovnat PJ, Safer JN, Conner EH, Price ML. The uptake of thiopental by body tissues and its relation to the duration of narcosis. J Clin Pharmacol Ther 1960;1:16–22. [29] Quasha AL, Eger EI, Tinker JH. Determination and applications of MAC. Anesthesiology 1980;53:315 – 34. [30] Shim CY, Andersen NB. The effect of oxygen on minimal anesthetic requirements in the toad. Anesthesiology 1971;34:333 – 7. [31] Shore PA, Brodie BB, Hogben CAM. The gastric secretion of drugs: a pH partition hypothesis. J Pharmacol Exp Ther 1957;119:361 – 9. [32] Stern SA, Frisch HL. Dependence of inert gas narcosis on lipid ‘free volume’. J Appl Physiol 1973;34:366 – 73. [33] Stevens CW. Alternatives to the use of mammals for pain research. Life Sci 1992;50:901 – 12. [34] Stevens CW, Klopp AJ, Facello JA. Analgesic potency of mu and kappa opiods after systemic administration in amphibians. J Pharmacol Exp Ther 1994;269:1086 – 93. [35] Walker MD. Physiologic and pharmacologic aspects of barbiturates in elasmobranchs. Comp Biochem Physiol 1972;42A:213– 21. [36] Yu HY, Sawada Y, Sugiyama Y, Iga T, Hanano M. Effect of sulfadimethoxine on thiopental distribution and elimination in rats. J Pharm Sci 1981;70:323 – 6.