Seediscussions,stats,andauthorprofilesforthispublicationat:https://www.researchgate.net/publication/10842530

IsofluraneAntagonizestheCapacityofFlurothylor1,2-DichlorohexafluorocyclobutanetoImpairFearConditioningtoContextandTone

ArticleinAnesthesia&Analgesia·May2003

DOI:10.1213/01.ANE.0000055360.30078.FF·Source:PubMed

CITATIONS

32

READS

19

8authors,including:

RobertAPearce

UniversityofWisconsin–Madison

92PUBLICATIONS4,357CITATIONS

SEEPROFILE

MichaelFanselow

UniversityofCalifornia,LosAngeles

281PUBLICATIONS23,952CITATIONS

SEEPROFILE

JamesMSonner

UniversityofCalifornia,SanFrancisco

118PUBLICATIONS3,492CITATIONS

SEEPROFILE

AllcontentfollowingthispagewasuploadedbyMichaelFanselowon30November2016.

Theuserhasrequestedenhancementofthedownloadedfile.Allin-textreferencesunderlinedinblue

arelinkedtopublicationsonResearchGate,lettingyouaccessandreadthemimmediately.

Isoflurane Antagonizes the Capacity of Flurothyl or1,2-Dichlorohexafluorocyclobutane to Impair FearConditioning to Context and ToneEdmond I Eger II, MD*, Yilei Xing, MD*, Robert Pearce, MD†, Steven Shafer, MD‡,Michael J. Laster, DVM*, Yi Zhang, MD*, Michael S. Fanselow, PhD§, andJames M. Sonner, MD*

*Department of Anesthesia and Perioperative Care, University of California, San Francisco, California; †Department ofAnesthesiology, University of Wisconsin, Madison, Wisconsin; ‡Department of Anesthesiology, Stanford University,Stanford, California; and §Department of Psychology, University of California, Los Angeles, California

In animals, the conventional inhaled anesthetic, isoflu-rane, impairs learning fear to context and fear to tone,doing so at concentrations that produce amnesia in hu-mans. Nonimmobilizers are inhaled compounds thatdo not produce immobility in response to noxious stim-ulation, nor do they decrease the requirement for con-ventional inhaled anesthetics. Like isoflurane, the non-immobilizer 1,2-dichlorohexafluorocyclobutane (2N)impairs learning at concentrations less than thosepredicted from its lipophilicity to produce anesthe-sia. The capacity of the nonimmobilizer di-(2,2,2,-trifluoroethyl) ether (flurothyl) to affect learning andmemory has not been studied. Both nonimmobilizerscan cause convulsions. We hypothesized that ifisoflurane, 2N, and flurothyl act by the same mecha-nism to impair learning and memory, their effectsshould be additive. We found that isoflurane, 2N,

and flurothyl (each, alone) impaired learning fear tocontext and fear to tone in rats, with the nonimmobi-lizers doing so at concentrations less than those thatcause convulsions. (Fear was defined by freezing [vo-litional immobility] in the presence of the condi-tioned stimulus [context or tone].) However, the com-bination of isoflurane and 2N or flurothyl producedan antagonistic rather than an additive effect onlearning, a finding in conflict with our hypothesis.And flurothyl was no less potent than 2N (at least noless potent relative to the concentration of each thatproduced convulsions) in its capacity to impair learn-ing. We conclude that conventional inhaled anesthet-ics and nonimmobilizers impair learning and mem-ory by different mechanisms. The basis for thisimpairment remains unknown.

(Anesth Analg 2003;96:1010 –8)

A lthough they have a lipophilicity that indicates acapacity to produce anesthesia (1,2), certain gasesand vapors do not prevent movement in response

to noxious stimulation. They do not prevent movementwhen given alone, nor do they add to the capacity ofknown anesthetics to prevent movement (3). While thesecompounds, known as nonimmobilizers, do not produceimmobility, they can impair learning and memory at

concentrations close to those predicted from their li-pophilicity to produce such effects (4,5). Thus, the non-immobilizer 2N (1,2-dichlorohexafluorocyclobutane)impairs learning at concentrations roughly a third ofthose predicted to produce anesthesia (4,5). The con-centrations predicted to produce anesthesia actuallycause convulsions.

Most nonimmobilizers, such as 2N, differ from con-ventional inhaled anesthetics by having a lower polar-ity, manifested by a lower affinity to water (3). Thishas led to the suggestion that the capacity of conven-tional anesthetics to produce immobility required thatthey be amphipathic (have an affinity to both polarand nonpolar phases) and that their capacity to pro-duce immobility resulted from an action at an inter-face between polar and nonpolar phases (e.g., at thecell membrane surface) (6). In contrast, it was sug-gested that amnesia (impaired learning and memory)

This work was supported by National Institutes of Health Grant1PO1GM47818-08.

Dr. Eger is a paid consultant to Baxter Healthcare Corp. Dr. Shaferhas received research grants from and/or has been a consultant toAbbott Laboratories, Baxter Healthcare, and Zeneca.

Accepted for publication December 17, 2002.Address correspondence and reprint requests to Dr. Edmond I

Eger II, Department of Anesthesia, S-455, University of California,San Francisco, CA 94143-0464. Address e-mail to egere@anesthesia.ucsf.edu.

DOI: 10.1213/01.ANE.0000055360.30078.FF

©2003 by the International Anesthesia Research Society1010 Anesth Analg 2003;96:1010–8 0003-2999/03

resulted from an action confined to a nonpolar site(e.g., one within the membrane).

This reasoning presumes that anesthetics and mostnonimmobilizers impair learning and memory by acommon mechanism. If correct, then the effects ofinhaled anesthetics and nonimmobilizers on learningand memory should be additive. However, at leastone nonimmobilizer does not fit this reasoning: flu-rothyl [di-(2,2,2,-ethyl) ether] has a solubility in watersimilar to that of conventional inhaled anesthetics (7).Flurothyl and 2N may differ in one other respect: aninitial report (but not a more recent report) (8) sug-gested that 2N does not antagonize the effect ofgamma aminobutyric acid (GABA) (9), whereas flu-rothyl does (10). Thus, the mechanism by which flu-rothyl impairs learning and memory might differ fromthe mechanism by which 2N impairs learning andmemory (11). If correct, this implies that the mecha-nism of the capacity of 2N and flurothyl to impairlearning and memory might differ. The present inves-tigation tested these hypotheses, finding that bothwere incorrect. We used two measures of learning andmemory: fear conditioning to context and fear condi-tioning to tone. Fear conditioning to context testsmemory of the surroundings associated with applica-tion of an unconditioned stimulus (a footshock),whereas fear conditioning to tone tests memory of atone associated with an unconditioned stimulus. Fearconditioning to tone requires processing by the amyg-dala but not hippocampus; fear conditioning to con-text requires processing by both the amygdala andhippocampus (12–18). Both measures of learning ledto the same conclusions.

MethodsThe Committee on Animal Research of the Universityof California, San Francisco, approved our study ofmale [Crl:CD(SD)BR] Sprague-Dawley rats weighing275 to 325 g obtained from Charles River Laboratories(Hollister, CA). Animals were housed in our animalcare facility for a week before study under 12-h cyclesof light and dark, two per cage, and had continuousaccess to standard rat chow and tap water before thestudy. Rats were handled on each of at least 3 daysbefore training to accustom them to being handled.We purchased the nonimmobilizers 2N and flurothylfrom SynQuest Laboratories, Inc. (Alachua, FL).

We assessed the effect of 2N and of flurothyl (sep-arately) on fear conditioning to context, and to tone, inthe absence and presence of various concentrations ofisoflurane (Tables 1 and 2). Fear was defined by freez-ing (volitional immobility) in the presence of the con-ditioned stimulus (context or tone). A rat was givenone dose of one nonimmobilizer alone or in combina-tion with one concentration of isoflurane (i.e., wasstudied only once).

On the training day, the rats (two/cage) werebrought to the training area and their tails markedwith a permanent felt-tipped pen for identification.They were placed in clear plastic containers coveredwith a clear plastic plate surmounted by a chimneybroad enough to allow us to easily reach into thecontainer and extract a rat (equilibration chamber, Fig.1). The clear plastic plate was pierced to allow deliveryof a �2 L/min inflow of oxygen with or without 2N,flurothyl, and/or isoflurane. The concentrations of2N, flurothyl, and/or isoflurane were established atpredetermined target values before placement of therats in the containers. These concentrations were thesame as those provided in the training chambers (seebelow). Isoflurane, 2N, and flurothyl concentrationswere separately measured by gas chromatography.The gas chromatograph was calibrated with second-ary standards from tanks. Rat temperatures were notmeasured because we previously found that 2N doesnot affect temperature regulation (19).

The rats were continuously observed during theperiod of equilibration before training. Ten to 15 min-utes after placement of the rats in the containers, eachanimal was rapidly (�10 s) transferred from the equil-ibration chamber via the chimney to a training cham-ber (Fig. 1) via an 8-cm port in the chamber top oth-erwise sealed with a rubber cork. (We separatelydetermined that the introduction of the rats did notmaterially change the chamber concentrations of thedelivered gases.) Animals were allowed to explore thechamber for 3 min before training began.

The training chambers were 4 identically shapedchambers (25 � 20 � 17 cm) constructed of clearacrylic and located in a well lit room. Inlet and outletports allowed continuous ventilation through thechambers. A circular flow through the 4 chambers wasmaintained by a fan producing a background noise of70 dB (A-scale) (Sound Level Meter; Radio Shack, Ft.Worth, TX). A 5 L/min oxygen fresh gas inflow con-veyed the target concentrations of isoflurane, 2N,and/or flurothyl. Carbon dioxide was removed with asoda lime canister, and gas concentrations were sam-pled from a port in the circle system. A shock gridformed the floor of each training chamber. The shockgrid consisted of 14 stainless steel rods (6-mm diam-eter), spaced 1.8 cm center to center and wired to ashock scrambler (Gemini Avoidance System; San Di-ego Instruments, San Diego, CA). A speaker wasmounted on the rear wall of each training chamber.Training chambers were cleaned with 2% ammoniumhydroxide before and after each animal occupiedthem.

For tone conditioning, animals received three tone-shock pairs consisting of a 30-s tone (90 dB, A-scale,2000 Hz) co-terminating with a 2-s electric shock(11-Hz bipolar square waves, 2 mA for all groups);shock pairs were 90 s apart. Animals were returned to

ANESTH ANALG ANESTHETIC PHARMACOLOGY EGER ET AL. 10112003;96:1010–8 NONIMMOBILIZERS, ANESTHETICS, AND AMNESIA

their home cages (free of isoflurane) within 60 s afterthe last shock. For context conditioning, an identicalprocedure was followed except that no tone wasdelivered.

The next day, we tested freezing to tone or to con-text. The chambers for tone testing (tone testing cham-ber, Fig. 1) provided a different environment from thatprovided by the training chambers. The clear acrylic

Table 1. Isoflurane, 1,2-Dichlorohexafluorocyclobutane (2N), and Di-(2,2,2-Trifluoroethyl) Ether (Flurothyl) Effects onFreezing to Context

Concentration (%)

n

Average %freezing to

contextStandard

errorIsoflurane 2N Flurothyl

0 0 0 20 65.8 6.40.05 0 0 4 60.6 7.90.09 0 0 20 62.1 7.00.11 0 0 13 68.8 7.00.19 0 0 17 41.7 6.50.28 0 0 24 22.0 6.30.38 0 0 12 3.1 1.40.19 0.4 0 8 45.6 7.60.28 0.4 0 8 30.1 6.00.38 0.4 0 8 13.6 3.60.00 0.8 0 12 25.0 5.50.11 0.8 0 11 64.9 4.90.19 0.8 0 8 48.3 9.30.28 0.8 0 8 32.4 10.30.38 0.8 0 8 8.7 2.40.0 1.5 0 12 9.3 2.80.11 1.5 0 12 24.6 3.90.19 1.5 0 12 28.4 6.30.28 1.5 0 4 14.0 6.10.0 3.0 0 8 0.2 0.20.11 3.0 0 12 11.2 2.20.19 3.0 0 12 19.6 7.60.28 3.0 0 8 11.7 3.10 0 0.033 8 55.9 10.80 0 0.066 8 22.0 6.90.11 0 0.066 8 29.8 7.50 0 0.10 24 12.9 3.90.11 0 0.10 20 31.0 3.90.19 0 0.10 4 41.9 11.3

Table 2. Isoflurane, 1,2-Dichlorohexafluorocyclobutane (2N), and Di-(2,2,2,-Trifluoroethyl) Ether (Flurothyl) Effects onFreezing to Tone

Concentration (%)

n

Average %freezing to

toneStandard

errorIsoflurane

concentration2N

concentration Flurothyl

0 0 0 22 58.9 7.30.19 0 0 12 60.3 10.40.38 0 0 10 32.0 10.30 3.0 0 4 15.3 7.40.11 3.0 0 8 25.9 4.90.19 3.0 0 8 51.5 9.90 0 0.066 8 64.4 9.20 0 0.10 8 54.3 11.10 0 0.15 4 33.1 16.20 0 0.20 4 8.1 1.80.25 0 0.20 8 42.1 10.80 0 0.40 7 12.8 11.70.35 0 0.40 4 13.2 9.3

1012 ANESTHETIC PHARMACOLOGY EGER ET AL. ANESTH ANALGNONIMMOBILIZERS, ANESTHETICS, AND AMNESIA 2003;96:1010–8

testing chambers for tone testing had an A-frame roof.Each had a 25 � 28 cm base and 21 � 28 cm sides andhad a smooth floor. They were in a different roomfrom the training chambers. The tone test room was litwith a 25-W red lightbulb, whereas the training roomwas lit with conventional white-light fluorescent ceil-ing lamps. The testing cages were cleaned with a

pine-scented solution, and there was no backgroundnoise. A speaker was mounted on the rear wall of eachtesting chamber.

For context testing, each animal was returned to thechamber in which it was trained (the training cham-ber) and allowed to explore the chamber for 8 min;neither tone nor shock was administered. Four ani-mals were observed simultaneously, one in each of thefour chambers.

For tone testing, each animal that had received tonetraining was placed in the A-frame testing chamber inthe different room (see above), and, after 3 min ofexploration, a tone (90 dB, A-scale, 2000 Hz) wascontinuously sounded for 8 min; shocks were notadministered. Again, four animals were observed si-multaneously, one in each of the four chambers. Ani-mal observations for both the test of freezing to con-text and freezing to tone were via a video camera thatallowed observation of all rats simultaneously. Nopersonnel were in the test rooms during this period.

Freezing was assessed by a trained observer. Toscore freezing 24 h after training, an observation ofone of the four animals was made every 2 s (20).Therefore, each animal was scored once every 8 s.Behavior was judged as freezing if there was no visiblemovement except for breathing. The observation pe-riods were also videorecorded for scoring at a latertime, if needed.

We studied a minimum of 4 and a maximum of 24 ratsat each concentration or combination of concentrations.

For each group, the gas concentrations were calcu-lated as the mean and se of the concentrations mea-sured in the clear plastic containers and in the trainingchambers before and after training of that group. Thepercentage of time an animal froze during the 8-minobservation periods 24 h after training was calculatedas the number of observations judged to be freezingdivided by the total number of observations in 8 min,i.e., 60 observations (20). For each group score forconditioning to context and to tone, the mean and seof the mean were calculated.

The effects of isoflurane, 2N, and flurothyl on freez-ing to context and freezing to tone were modeledusing a sigmoidal curve, described by the relationship

Effect � E0 � �E0 � Emin�C�

C50� � C�

,

where Effect is freezing to context or freezing to tone,E0 is the baseline effect when no drug is present, Eminis the lowest score (i.e., maximal effect; 0 for freezingto context and about 10 for freezing to tone), C is thedrug concentration, and � is the Hill coefficient indi-cating the steepness of the relationship.

Figure 1. Three chambers were used in this study. The equilibrationchamber was used to preequilibrate rats to the concentration(s) ofisoflurane and/or nonimmobilizers for 30 min before insertion intothe training chamber (which also served as the test chamber forcontext). The rats were taken from the equilibration chamber via thechimney and immediately (within a few seconds) placed through ahole (not shown) in the test chamber that otherwise was closed witha large rubber stopper (not shown). The training chamber had ashock grid as the floor on which the rats stood and had a speakerthat could be activated for training to tone. The tone testing chamberwas of a different shape from the training chamber. It smelleddifferent, had different lighting, and had a flat smooth floor surfacerather than a grid.

ANESTH ANALG ANESTHETIC PHARMACOLOGY EGER ET AL. 10132003;96:1010–8 NONIMMOBILIZERS, ANESTHETICS, AND AMNESIA

The interaction between isoflurane and each nonim-mobilizer was analyzed with the response-surface in-teraction model of Minto et al. (21). The response-surface approach follows closely the approach forsingle drugs, in that the drug effect is still described bya sigmoidal relationship:

Effect � E0 � �E0 � Emin�

� C

C50���� ����

1 � � C

C50���� ���� ,

where effect, E0, and Emin are as previously defined. Cis the sum of normalized concentrations. In otherwords,

C �Iso

C50,Iso

�2N

C50,2N

or

C �Iso

C50,Iso

�flurothyl

C50,flurothyl

,

depending on which nonimmobilizer was being givenin conjunction with isoflurane. � is the fraction of the(normalized) nonimmobilizer in the total drug given:

� �

2N

C50,2N

Iso

C50,Iso

�2N

C50,2N

or

� �

flurothyl

C50,flurothyl

Iso

C50,Iso

�flurothyl

C50,flurothyl

,

depending on which nonimmobilizer was being stud-ied. The drug interaction is captured by C50(�), whichacts to either magnify or reduce the concentrationbeyond simple additivity for the drugs in combina-tion. We chose a simple quadratic model for the inter-action shape, as is commonly the practice in interac-tion modeling:

C50��� � 1 � �2N � � � �2N � �2 or

C50��� � 1 � �flurothyl � � � �flurothyl � �2 ,

depending on which nonimmobilizer was being stud-ied. The parameter � captures the shape of the inter-action. If � � 0, then C50(�) is always 1, and theinteraction is additive. If � � 0, then C50(�) becomes�1 except when � is 0 or 1, which acts to magnify theeffect of the drug, producing a supraadditive (e.g.,

synergistic) relationship. If � � 0, then C50(�) becomesgreater than 1 except when � is 0 or 1, which acts todiminish the effect of the two drugs, producing aninfraadditive (e.g., antagonistic) relationship.

The function �(�) permitted each drug to have itsown steepness for the concentration versus responserelationship. �(�) varied linearly with � (i.e., variedlinearly depending on the fraction of total drug madeup by the nonimmobilizer):

���� � �iso � �1 � �� � �2N � � or

���� � �iso � �1 � �� � �flurothyl�� ,

depending on which nonimmobilizer was being studied.Parameter estimation was performed with NON-

MEM (University of California, San Francisco, CA).We used the NONMEM objective function to testwhether the relationship between isoflurane and thenonimmobilizers significantly differed from additivityand whether the interaction term, �, and the Hillcoefficient, �, were statistically distinguishable be-tween 2N and flurothyl. Specifically, we examined thedecrease in the NONMEM objective function with theaddition of each parameter, since this decrease ap-proximately follows the �2 distribution for the degreesof freedom (i.e., number of parameters) added to themodel. For example, a decrease of 3.84 with the addi-tion of a single parameter indicates that the expandedmodel is significantly better, with a probability of�0.05.

ResultsIsoflurane, 2N, and flurothyl each suppressed freezingto context in a dose-related manner (Table 1), butisoflurane antagonized the effect of 2N (Fig. 2, topgraph) and flurothyl (Fig. 2, bottom graph).

The response-surface modeling for fear to context(Fig. 3) demonstrated a highly significant interaction(P � 0.001). Flurothyl and 2N did not differ in (fearwas defined by freezing [volitional immobility] in thepresence of the conditioned stimulus [context or tone])either � or �, suggesting that these drugs act via thesame mechanism. The parameters of the final inter-action model were as follows: C50,isoflurane � 0.22,C50,2N � 0.90, C50,flurothyl � 0.059, �isoflurane � 3.2,�nonimmobilizer � 2.3, and � � �4.8. In Figure 3, theblack line is the “isobole” showing 50% drug effect.The upper projection demonstrates the shape of thesurface, while the lower projection, a view fromdirectly above the surface, shows the extreme an-tagonism evident in the 50% isobole.

As with freezing to context, isoflurane, 2N, and flu-rothyl each suppressed freezing to tone in a dose-relatedmanner (Table 2), but isoflurane antagonized the effectof 2N (Fig. 4, top graph) and flurothyl (Fig. 4, bottom

1014 ANESTHETIC PHARMACOLOGY EGER ET AL. ANESTH ANALGNONIMMOBILIZERS, ANESTHETICS, AND AMNESIA 2003;96:1010–8

graph). Isoflurane appeared to enhance learning fear totone in the presence of a nonimmobilizer.

The response-surface modeling for fear to tone (Fig.5) demonstrated a highly significant interaction (P �0.001). Flurothyl and 2N did not differ in either � or �,suggesting that these drugs act via the same mecha-nism. The parameters of the final interaction modelwere as follows: C50,isoflurane � 0.37, C50,2N � 1.9,C50,flurothyl � 0.138, �isoflurane � 4.8, �nonimmobilizer �8.8, and � � �6.4. As observed for freezing to context,the relationship was profoundly antagonistic, a find-ing particularly evident in the Figure 5, bottom graph,viewing the surface from directly above, where theextreme shape of the 50% isobole can be directlyobserved.

Suppression of freezing to tone required larger con-centrations of all vapors than were required to pro-duce freezing to context. The minimum average offreezing to tone was 12% and was 0% with freezing tocontext. These results suggest that tone creates a moreprofound association than context with the noxious

electrical shock; i.e., it requires larger concentrations toblock the learning process.

None of the rats gave evidence of convulsive activ-ity in the presence of isoflurane. In the absence ofisoflurane, rats convulsed at flurothyl concentrationsof 0.15% and larger.

Figure 2. In the absence of a nonimmobilizer, increasing concentra-tions of isoflurane applied during training fear to context decreasepercent freezing in a dose-related manner. Similarly, increasingconcentrations of each nonimmobilizer in the absence of isoflurane(first point for each graph) decrease percent freezing in a dose-related manner. However, the addition of isoflurane to a givenconcentration of 1,2-dichlorohexafluorocyclobutane (2N) (upperpanel, A) or flurothyl (lower panel, B) increases the percent freezing(i.e., learning fear to context), indicating an antagonism of the effectsof the two compounds. Points for the percent isoflurane in this andsubsequent graphs are offset slightly to improve clarity.

Figure 3. The antagonism between isoflurane and nonimmobilizerson learning fear to context can be visualized as the bowing of theinteraction surface away from the origin and the vertical axis. In aninteraction surface, synergy would appear as bowing towards theorigin and vertical axis. Simple additivity would appear as astraight line on any horizontal slice through the vertical axis. Theprofound antagonism (bowing away from the origin and verticalaxis) is most easily visualized in the lower projection, where thesurface is viewed from directly above. Isoflurane concentration isexpressed in percent partial pressure. The nonimmobilizerconcentrations have been normalized by dividing each concen-tration by the C50 for learning fear to context, allowing 1,2-dichlorohexafluorocyclobutane (2N) and flurothyl to share acommon axis.

ANESTH ANALG ANESTHETIC PHARMACOLOGY EGER ET AL. 10152003;96:1010–8 NONIMMOBILIZERS, ANESTHETICS, AND AMNESIA

DiscussionOur results indicate that the effects of isoflurane onlearning antagonize rather than add to the effects ofthe two nonimmobilizers, 2N and flurothyl. This is nota subtle result. The outward bowing of the isobolo-grams seen in Figures 3 and 5 shows tremendousantagonism. The log-likelihood ratios for antagonismof fear to context and fear to tone were 29 and 7,respectively. Since twice the log-likelihood ratio ap-proximates the �2 distribution, these correspond to Pvalues for antagonism approximately equal to 2 �10�13 and 9 � 10�4, respectively. Because these areapproximations, in our results we have listed the sig-nificance as P � 0.001.

The results from this study are not consistent with ourhypothesis that conventional anesthetics and nonimmo-bilizers act in the same manner to impair learning andmemory. It suggests that at least two sites of action areimportant to this impairment. It undermines the argu-ment that conventional inhaled anesthetics and nonim-mobilizers both must act in a nonpolar environment toimpair learning and memory. The finding that flurothyl,a polar compound, also potently impairs learning and

memory also suggests that an action limited to a nonpo-lar environment may not be essential to the impairmentof learning and memory.

Our results demonstrate that both 2N and flurothylimpair learning and memory. The effect of 2N hadbeen demonstrated previously (4,5), but no such dem-onstration had been made for flurothyl. For both non-immobilizers, the concentrations that produce convul-sions [approximately 4%–5% for 2N (3) and 0.12%–

Figure 4. As in Figure 2, in the absence of a nonimmobilizer, increasingconcentrations of isoflurane applied during training fear to tone de-crease the percent freezing in a dose-related manner. Similarly, increas-ing the concentrations of 1,2-dichlorohexafluorocyclobutane (2N)(upper panel, A) or flurothyl (lower panel, B) applied during train-ing fear to tone decreases the percent freezing in a dose-relatedmanner (see the first points for each nonimmobilizer’s concentration).As in Figure 2, isoflurane reverses the effects of each nonimmobilizer,indicating an antagonism of the effects of the compounds.

Figure 5. As in Figure 3, the outward bowing of the interactionsurface away from the origin and vertical axis demonstrates antag-onism between isoflurane and nonimmobilizers, seen here for learn-ing fear to tone. The profound antagonism is most readily visual-ized in the lower graph, where the surface is viewed from above.Isoflurane concentration is expressed as percent partial pressure.The nonimmobilizer concentrations have been normalized by divid-ing each concentration by the C50 for learning fear to tone, allowing1,2-dichlorohexafluorocyclobutane (2N) and flurothyl to share acommon axis.

1016 ANESTHETIC PHARMACOLOGY EGER ET AL. ANESTH ANALGNONIMMOBILIZERS, ANESTHETICS, AND AMNESIA 2003;96:1010–8

0.15% for flurothyl (22)] exceed, by a factor of 3 to 6,the concentrations that impair learning fear to context(Table 1, Figs. 2 and 3). They also exceed the concen-trations that impair learning fear to tone, but the dif-ference is much less (Table 2, Figs. 4 and 5).

These results do not indicate that 2N and flurothyldiffer in their capacities to impair learning and mem-ory. Clearly, the capacity of nonimmobilizers to im-pair learning and memory does not correlate withtheir lipophilicity. That is, the product of the concen-tration of nonimmobilizer producing a 50% impair-ment of learning fear to context (perhaps 0.6% 2N and0.05% flurothyl; Tables 1 and 2, Figs. 2 and 4) times theoil/gas partition coefficient [43.5 for 2N (3) and 46.9for flurothyl (7)] differs by an order of magnitude. Itmay correlate with the capacity of nonimmobilizers toinhibit neuronal nicotinic acetylcholine receptors. Bothflurothyl and 2N inhibit such receptors, doing so atconcentrations close to those that impair learning (23).They share this capacity with that of conventionalinhaled anesthetics (24,25), and one might be temptedto consider that this underlies the capacity of bothconventional anesthetics and nonimmobilizers to im-pair learning and memory. However, as noted above,the absence of an additive effect of isoflurane with 2Nor flurothyl makes a common mechanism unlikely.

Does the impairment of learning and memory bythe nonimmobilizers correlate with their capacity toproduce convulsions? None of the rats gave overtevidence of convulsive activity at concentrations thatcompromised learning fear to context. We previouslyfound that concentrations of up to 4.2% 2N do notproduce evidence of epileptiform activity in sponta-neous (unstimulated) electroencephalographic record-ings (26). Indeed, in that study we found no effect of2N on the middle latency auditory-evoked potential.

Another explanation for the capacity of the nonim-mobilizers to impair learning and memory is that thenonimmobilizers make the rat anxious. An inhaledanesthetic such as isoflurane might decrease this anx-iety and thereby antagonize the effect of the nonim-mobilizers. Other investigators have suggested thatthe blockade of the effect of GABA by the anxiogenicdrug DMCM induces a state of panic and that thisblocks fear conditioning (27). Like 2N and flurothyl,DMCM can produce convulsions. Both flurothyl (10)and DMCM (28) can inhibit the action of GABA onGABAA receptors. One report suggests that 2N doesnot produce inhibition (9), but another report (8) findsan effect.

Inhaled anesthetics can produce antianalgesia at 0.1to 0.2 minimum alveolar anesthetic concentration(MAC) (29). This might partly explain the antagonisticeffects seen at these concentrations. However, concen-trations larger than 0.2 MAC have analgesic effects(29), and yet such concentrations also antagonized the

impairment of learning and memory produced byflurothyl and 2N.

We conclude that conventional inhaled anestheticsand nonimmobilizers can, at least in part, impairlearning and memory by different mechanisms. If theydo partly share a common mechanism, one or bothhave additional mechanisms that come into play.

References1. Meyer HH. Theorie der Alkoholnarkose. Arch Exp Pathol Phar-

makol 1899;42:109–18.2. Overton E. Studien uber die Narkose, Zugleich ein Beitrag zur

allgemeinen Pharmakologie. Jena, Germany: Gustav Fischer,1901:1–195.

3. Koblin DD, Chortkoff BS, Laster MJ, et al. Polyhalogenated andperfluorinated compounds that disobey the Meyer-Overton hy-pothesis. Anesth Analg 1994;79:1043–8.

4. Kandel L, Chortkoff BS, Sonner J, et al. Nonanesthetics cansuppress learning. Anesth Analg 1996;82:321–6.

5. Sonner J, Li J, Eger EI II. Desflurane and the nonimmobilizer1,2-dichlorohexafluorocyclobutane suppress learning by amechanism independent of the level of unconditioned stimula-tion. Anesth Analg 1998;87:200–5.

6. Eger EI II, Koblin DD, Harris A, et al. Hypothesis: inhaledanesthetics produce immobility and amnesia by different mech-anisms at different sites. Anesth Analg 1997;84:915–8.

7. Koblin DD, Eger EI II, Johnson BH, et al. Are convulsant gasesalso anesthetics? Anesth Analg 1981;60:464–70.

8. Perouansky M, Chesney MA, Banks MI, Pearce RA. The pro-convulsant non-immobilizer F6 affects GABAA receptor-mediated currents in hippocampal neurons [abstract]. Anesthe-siology 2001;95:A86.

9. Mihic SJ, McQuilkin SJ, Eger EI II, et al. Potentiation of gamma-aminobutyric acid type A receptor-mediated chloride currentsby novel halogenated compounds correlates with their abilitiesto induce general anesthesia. Mol Pharmacol 1994;46:851–7.

10. Wakamori M, Ikemoto Y, Akaike N. Effects of two volatileanesthetics and a volatile convulsant on the excitatory andinhibitory amino acid responses in dissociated CNS neurons ofthe rat. J Neurophysiol 1991;66:2014–21.

11. Eger EI II, Koblin DD, Sonner J, et al. Nonimmobilizers andtransitional compounds may produce convulsions by twomechanisms. Anesth Analg 1999;88:884–92.

12. Kim JJ, Fanselow MS. Modality specific retrograde amnesia offear following hippocampal lesions. Science 1992;256:675–7.

13. Romanski LM, Clugnet MC, LeDoux JE. Somatosensory andauditory convergence in the lateral nucleus of the amygdala.Behav Neurosci 1993;107:444–50.

14. Phillips RG, LeDoux JE. Differential contribution of amygdalaand hippocampus to cued and contextual fear conditioning.Behav Neurosci 1992;106:274–85.

15. Maren S, Fanselow MS. The amygdala and fear conditioning:has the nut been cracked? Neuron 1996;16:237–40.

16. Fendt M, Fanselow MS. The neuroanatomical and neurochem-ical basis of conditioned fear. Neurosci Biobehav Rev 1999;23:743–60.

17. Fanselow MS, LeDoux JE. Why we think plasticity underlyingPavlovian fear conditioning occurs in the basolateral amygdala.Neuron 1999;23:229–32.

18. LeDoux JE. Fear and the brain: where have we been, and whereare we going? Biol Psychiatry 1998;44:1229–38.

19. Maurer AJ, Sessler DI, Eger EI II, Sonner JM. The nonimmobi-lizer 1,2-dichlorohexafluorocyclobutane does not affect thermo-regulation in the rat. Anesth Analg 2000;91:1013–6.

20. Fanselow MS. Conditioned and unconditional components ofpost-shock freezing. Pavlov J Biol Sci 1980;15:177–82.

21. Minto CF, Schnider TW, Short TG, et al. Response surface modelfor anesthetic drug interactions. Anesthesiology 2000;92:1603–16.

ANESTH ANALG ANESTHETIC PHARMACOLOGY EGER ET AL. 10172003;96:1010–8 NONIMMOBILIZERS, ANESTHETICS, AND AMNESIA

22. Fang Z, Laster MJ, Gong D, et al. Convulsant activity of non-anesthetic gas combinations. Anesth Analg 1997;84:634–40.

23. Raines DE, Claycomb RJ, Forman SA. Nonhalogenated anes-thetic alkanes and perhalogenated nonimmobilizing alkanes in-hibit �4�2 neuronal nicotinic acetylcholine receptors. AnesthAnalg 2002;95:573–7.

24. Flood P, Ramirez-Latorre J, Role L. �4�2 Neuronal nicotinicacetylcholine receptors in the central nervous system are inhib-ited by isoflurane and propofol, but �7-type nicotinic acetylcho-line receptors are unaffected. Anesthesiology 1997;86:859–65.

25. Violet JM, Downie DL, Nakisa RC, et al. Differential sensitivitiesof mammalian neuronal and muscle nicotinic acetylcholine re-ceptors to general anesthetics. Anesthesiology 1997;86:866–74.

26. Dutton R, Rampil I, Eger EI II. Inhaled nonimmobilizers do notalter the middle latency auditory-evoked response of rats.Anesth Analg 2000;90:213–7.

27. Sieve AN, King TE, Ferguson AR, et al. Pain and negative affect:evidence the inverse benzodiazepine agonist DMCM inhibitspain and learning in rats. Psychopharmacology 2001;153:180–90.

28. Kleingoor C, Ewert M, von Blankenfeld G, et al. Inverse but notfull benzodiazepine agonists modulate recombinant alpha 6beta 2 gamma 2 GABAA receptors in transfected human embry-onic kidney cells. Neurosci Lett 1991;130:169–72.

29. Zhang Y, Eger EI II, Dutton RC, Sonner JM. Inhaled anestheticshave hyperalgesic effects at 0.1 minimum alveolar anestheticconcentration. Anesth Analg 2000;91:462–6.

Attention Authors!

Submit Your Papers Online

You can now have your paper processed and reviewed faster by sending it to us through our new,web-based Rapid Review System.

Submitting your manuscript online will mean that the time and expense of sending papers through themail can be eliminated. Moreover, because our reviewers will also be working online, the entire reviewprocess will be significantly faster.

You can submit manuscripts electronically via www.rapidreview.com. There are links to this site from theAnesthesia & Analgesia website (www.anesthesia-analgesia.org), and the IARS website (www.iars.org).

To find out more about Rapid Review, go to www.rapidreview.com and click on “About Rapid Review.”

1018 ANESTHETIC PHARMACOLOGY EGER ET AL. ANESTH ANALGNONIMMOBILIZERS, ANESTHETICS, AND AMNESIA 2003;96:1010–8