Effect of a Nonsedative Dose of Propofol on Memory for Aversively Loaded Information in Rats

We used 198 male Long Evans rats (Janvier, France) that weighed 330–350 g. They were housed two per cage in a colony room maintained on a 14-h light, 10-h dark cycle (light on at 7:00 am) with food and water provided ad libitum . Procedures and care of the animals were performed in accordance with the national and international guides (council directive #87848, October 19, 1989, Ministère de l'Agriculture et de la Forêt, Service Vétérinaire de la Santé et de la Protection Animales—NIH publication, N°86–23, revised 1985).

Propofol (5 mg/ml Diprivan, Zeneca, London, UK) was dissolved in 5%“intralipid.” Midazolam (5 mg/ml Hypnovel, Roche, Basel, Switzerland) was dissolved in 0.9% sodium chloride. All drugs were prepared immediately before use and injected intraperitoneally in a volume of 2 ml/kg.

Passive avoidance apparatus consisted of two adjacent compartments, a bright large one and a small dark one, separated by a guillotine door. The bright compartment had white Plexiglas walls (46 cm long, 42 cm large, and 44 cm high) and was illuminated by a 25 W bulb fixed on the wall opposite to the entrance to the dark one. This latter compartment had black Plexiglas walls and roof (30 cm long, 15 cm large, and 15 cm high). The roof could be opened to place directly and observe the animal in this compartment. The floor of the two compartments was made of stainless steel bars (0.6 cm diameter) regularly spaced apart by 1.8 cm. Scrambled foot shocks could be delivered through the bars of the dark compartment. The scrambler was set to deliver two moderate foot shocks at t¹= 2:00 and t²= 3:30 min during a 5-min session (effective intensity, 0.1 mA; 5 s: duration; 20 ms ON-140 ms OFF).

In this apparatus, naïve undrugged rats exhibit a spontaneous fear for the bright large compartment and escape quickly (within 10 s) from it to enter the adjacent dark small one.

The effect of propofol on forced locomotion was studied using a “Rotarod” apparatus as previously described. 17Briefly, the Rotarod apparatus used here consisted of an elevated (40 cm high) rotating (2 rotations/min) bar (40 cm long; diameter, 6 cm), divided into three 11-cm sections by four perpendicular disks (diameter, 50 cm).

In this apparatus, naïve rats were shaped daily, until each of them reached a criterion of stability on the rotating cylinder for at least 60 s.

The principles behind the behavioral situations performed using the passive avoidance apparatus are described here and separately for clarity.

Spontaneous anxiety was assessed by a “light-dark” situation in rats. Undrugged rats exhibit a spontaneous fear for bright large open spaces. Therefore, naïve undrugged rats, when initially placed in the large bright compartment, escape rapidly from it to enter the small dark secure-feeling one. The fact that a drugged animal stays in the large bright compartment while its spontaneous locomotion is not impaired indicates an anxiolytic effect of the drug. 17 

To assess the memory for an aversively loaded event, we used a “classic fear-conditioning preparation” for the acquisition phase of learning. 14,15Each rat was placed (via  opening the roof) and confined in the small dark compartment where it experienced two mild electrical foot shocks. The following day, the animal was tested for memory of the aversive experience by assessing the passive avoidance of the small dark compartment. When initially shocked in the small dark compartment, the rat has to remember this event to refrain from its natural tendency to enter this compartment when subsequently exposed to the bright large compartment. The fact that an animal enters the small dark compartment while he has been previously drugged during the conditioning session indicates no memory of the initial training.

To assess the memory of a nonaversive experience, we used the “CS-preexposure effect” based on previous works showing that an initial exposure to a CS alone will debilitate the subsequent association between this CS and an US. 18,20We showed in our preparation that an initial preexposure to the small dark compartment in the absence of shock, followed later by an association between this compartment and shocks, resulted in rats avoiding entering this compartment but to a lesser extent than when not initially preexposed. 17The fact that a debilitating effect of the preexposure on fear conditioning is suppressed when the rat is drugged during preexposure indicates no memory of this nonaversive experience.

Behavioral testing started after three daily handling sessions. All experiments were conducted between 10:00 am and 4:00 pm.

Comparative effect of propofol (0; 9 mg/kg) and midazolam (0; 3 mg/kg) on spontaneous anxiety was observed. Each rat was injected with the drug or its solvent 15 min before being placed in the bright large compartment, with the door open. We measured the latency to enter the small dark compartment (s). In this and subsequent experiments, the rat must have its four paws in the small dark compartment to be considered being in this compartment.

Thirty-six rats were randomly assigned to one of four groups (n = 9) according to the dose of midazolam or propofol they received (0 mg/kg midazolam = sodium chloride, 3 mg/kg midazolam, 0 mg/kg propofol = intralipid, 9 mg/kg propofol).

Comparative effect of propofol (0; 9 mg/kg) and midazolam (0; 3 mg/kg) on memory of an aversive event was noted. On day 1, each rat was injected with the drug or its solvent 15 min before being confined into the small dark compartment, with the door closed, where they received two moderate foot shocks during a 5-min session. On day 2, each rat was placed in the bright large compartment, with the door open, without receiving any drug. We measured the latency to enter the dark small compartment (s). Thirty-six rats were randomly assigned to one of four groups (n = 9) according to the dose of midazolam or propofol they received (0 mg/kg midazolam = sodium chloride, 3 mg/kg midazolam, 0 mg/kg propofol = intralipid, 9 mg/kg propofol).

Effect of propofol (0; 9 mg/kg) and midazolam (0; 3 mg/kg) on memory of a nonaversive event was evaluated. On day 1, rats were injected with the drug or its solvent. Fifteen minutes later, they were placed into the small dark compartment for 30 min with no shock (preexposure). A control group was injected with the solvent and remained in the home cage (no preexposure). On day 2, each rat of the three groups was confined in the small dark compartment where it experienced two moderate foot shocks during a 5-min session, without receiving any drug. On day 3, each rat was placed in the bright large compartment, with the door open, without receiving any drug. We measured the latency to enter the dark small compartment (s). In Experiment 3a (midazolam), 33 rats were randomly assigned to three groups (n = 11). Rats in two of the groups were injected either with midazolam, 0 mg/kg (sodium chloride), or midazolam, 3 mg/kg, and then placed in the small dark compartment. Rats of the control group were injected with sodium chloride and remained in their home cage. In Experiment 3b (propofol), 33 rats were randomly assigned to three groups (n = 11). Rats in two of the groups were injected either with propofol, 0 mg/kg (intralipid), or propofol, 9 mg/kg, and then placed in the small dark compartment (preexposure). Rats of the control group were injected with intralipid and remained in their home cage.

In Experiment 4a, behavioral response to electrical shocks after propofol (0; 9 mg/kg) administration was observed. Each rat was injected with the drug or its solvent. Fifteen minutes later, each rat was placed in the small dark compartment where it experienced electrical foot shock. After an initial period of 20 s, the scrambler was set to deliver one foot shock (5 s: duration; 20 ms ON-140 ms OFF) every 10 s. To detect the threshold of shock perception, the effective intensity of the shock was gradually increased from 0.04 to 0.14 mA, by 0.02-mA increments. Two experimenters unaware of the treatment recorded the smallest intensity for which they observed a behavioral response (cry) to the shock.

Thirty rats were randomly assigned to one of three groups (n = 10): propofol, 0 mg/kg (intralipid); propofol, 9 mg/kg; and control group (same volume of 0.9% sodium chloride).

In Experiment 4b, forced locomotion after propofol (0; 9 mg/kg) administration was evaluated. Rats were shaped during two daily sessions until each reached a criterion of stability on the rotating cylinder for at least 60 s. On the third day, rats were given propofol or intralipid 15 min before testing. The number of animals falling during the 60-s test session was recorded.

Thirty rats were randomly assigned to one of three groups (n = 10): propofol, 0 mg/kg (intralipid); propofol, 9 mg/kg; and control group (same volume of 0.9% sodium chloride).

For Experiments 1–3, the dependent variable was the latency to enter the small dark compartment (cutoff, 900 s). Statistical analysis 23was performed on the Log10 of this variable (Log latency) to better approximate the normal distribution necessary for the use of parametric statistics. Statistical analyses were realized using analysis of variance (ANOVA) followed, if necessary, by a pair-wise test with Bonferroni correction adjusted to the number of comparisons (two in Experiments 1 and 2 and three in Experiments 3a and 3b). For Experiment 4a, the dependent variable was the smallest intensity of shock inducing the behavioral response. Statistical analysis was realized using a one-way ANOVA.

Comparative effect of propofol (0; 9 mg/kg) and midazolam (0; 3 mg/kg) on spontaneous anxiety was observed. Figure 1Adepicts the effect of propofol and midazolam on the latency to escape from the large bright compartment to enter the small dark compartment. In the control groups (propofol, 0 mg/kg, and midazolam, 0 mg/kg), rats entered the small dark compartment quickly, which indicates the anxiogenic effect of the large bright compartment. Propofol, 9 mg/kg, and midazolam, 3 mg/kg, similarly increased the latency to enter the small dark compartment and appeared to be equivalent doses for inducing anxiolysis. Two-way ANOVA using treatment (propofol vs . midazolam) and dose (solvent vs . drug) as the between factors showed a significant main effect for the dose (P < 0.0001) but no effect for the treatment and no interaction between dose and treatment. The pair-wise test (with Bonferroni correction for two comparisons) showed that the latency was increased significantly for the propofol, 9 mg/kg, and midazolam, 3 mg/kg, groups compared with their respective control groups (propofol, 0 mg/kg, and midazolam, 0 mg/kg).

Comparative effect of propofol (0; 9 mg/kg) and midazolam (0; 3 mg/kg) on memory for an aversive event was evaluated. Figure 1Bdepicts the effect of propofol and midazolam on the latency to enter the small dark compartment. In the control groups (propofol, 0 mg/kg, and midazolam, 0 mg/kg), rats avoided the small dark compartment, which indicates an effective fear conditioning. None of the rats reached the time cutoff of 900 s. Propofol, 9 mg/kg, and midazolam, 3 mg/kg, decreased the latency to enter the small dark compartment. Two-way ANOVA using treatment (propofol vs . midazolam) and dose (solvent vs . drug) as the between factors showed a significant main effect for the dose (P < 0.0001) but no effect for the treatment and no interaction between dose and treatment. The pair-wise test (with Bonferroni correction for two comparisons) showed that the latency was decreased significantly for the propofol, 9 mg/kg, and midazolam, 3 mg/kg, groups compared with their respective control groups (propofol, 0 mg/kg, and midazolam, 0 mg/kg).

Effect of propofol (0; 9 mg/kg) on memory of a nonaversive event was examined. Figure 2adepicts the effect of propofol, 0 and 9 mg/kg, administered before preexposure on the latency to enter the small dark compartment. A control group of rats received 0 mg/kg of propofol and was not preexposed. ANOVA showed a significant difference between the three groups (P < 0.01). The pair-wise test (with Bonferroni correction for three comparisons) showed that the latency to enter the small dark compartment was significantly reduced in the group of rats preexposed after propofol, 0 mg/kg, compared with the control group (significant preexposure effect). The latency to enter the small dark compartment was similar in the group of rats preexposed after propofol, 9 mg/kg, and in the control group, showing that propofol, 9 mg/kg, totally suppressed the effect of preexposure on the subsequent fear conditioning.

Effect of midazolam (0; 3 mg/kg) on memory of a nonaversive event was observed. Figure 2bdepicts the effect of midazolam, 0 and 3 mg/kg, administered before preexposure on the latency to enter the small dark compartment. A control group of rats received 0 mg/kg of midazolam and was not preexposed. ANOVA showed a significant difference between the three groups (P < 0.01). The pair-wise test (with a Bonferroni correction for three comparisons) showed that the latency to enter the small dark compartment was significantly reduced in the group of rats preexposed after midazolam, 0 mg/kg, compared with the control group (P < 0.05; preexposure effect). The latency to enter the small dark compartment was similar in the group of rats preexposed after midazolam, 3 mg/kg, and in the control group, which showed that midazolam totally suppressed the effect of preexposure on the subsequent fear conditioning.

Response to electrical shock after propofol (0; 9 mg/kg) was evaluated. The mean effective intensity (and SEM) inducing the behavioral response (cry) was 0.0815 (0.0040), 0.0800 (0.0038), and 0.0784 (0.0027) mA for sodium chloride; propofol, 0 mg/kg; and propofol, 9 mg/kg, respectively. A one-way ANOVA found no difference between the three groups.

Forced locomotion after propofol (0; 9 mg/kg) was examined. Rats given propofol exhibited no alteration of performance on the Rotarod test. All the animals stayed on the bar during the 60-s test; the number of falls was 0, 0, and 0 for sodium chloride; propofol, 0 mg/kg; and propofol, 9 mg/kg, respectively.

Propofol, 9 mg/kg, injected before fear conditioning decreased the subsequent latency to enter the compartment where rats had previously been shocked, indicating a reduction of fear elicited by this compartment compared with the fear exhibited by rats that had not received propofol (Experiment 2). We observed that equipotent anxiolytic doses of propofol and midazolam display similar effects on fear conditioning during the same experiment (Experiment 1). The administration of propofol, 9 mg/kg, before preexposure to the compartment prevented preexposure from impairing subsequent fear conditioning to this compartment (Experiment 3). These results show clearly that propofol has impaired the memory of the nonaversive event (preexposure to the compartment with no shocks) and the memory of the aversive event (shocks in the compartment), which suggest an amnesic effect (propofol-induced anterograde amnesia).

To clearly assess the effects of propofol on memory, different factors that may be involved in the observed amnesic effect must be controlled. The dose of propofol used here has anxiolytic effect as previously described. 16During fear conditioning, the aversive stimulus (here, foot shocks) induces an immediate emotional state that is responsible for the subsequent fear of the environment associated with it. Any anxiolytic effect could interact with fear conditioning by blunting the emotional state induced by the foot shocks. 24We observed that the same dose of propofol impaired the memory of an aversive event and the memory of a nonaversive event (preexposure to the environment). If an anxiolytic effect may reduce the anxiety-producing aspects of an aversive stimulus, it probably plays a negligible role when compared with the main effect of propofol on memory processes. Propofol is an anesthetic drug. Its effect on awareness may interact with the perception of a nonaversive or aversive event because of reduced perception of the stimuli to be acquired. 12,25We previously showed that midazolam-induced amnesia for an aversive event could be observed at a low dose totally devoid of sensorimotor impairment. 17The amnesic effect of propofol was also observed at a low dose (about 10% of the anesthetic dose for the strain of rats we used), for which we did not previously observe an impairment of spontaneous locomotor activity as assessed by the spontaneous movements between the sectors of an exploratory test cage. 16Results of Experiment 4b emphasize and extend this previous finding, showing that this low dose of propofol has no deleterious effect on forced locomotor activity when rats were tested for their ability to stay on a rotating bar. At this minimal dose, it is unlikely that propofol has reduced the perception of the different components of the event by decreasing awareness. One could argue that propofol has reduced the perception of the aversive stimulus (shock) via  an analgesic effect. We did not observe any modification of the behavioral response (cry) to the shock after propofol administration (Experiment 4a), showing the lack of analgesic effect of the minimal dose of propofol used here.

After Veselis et al .’s findings on emotionally unloaded information, our experimental data further illuminate a specific effect of propofol per se  on memory processes. This effect was observed whatever the emotional component of the event (aversive or not) and appeared not to rely on sedation or anxiolysis. Insofar, the expression of a memory trace relies on numerous processes: acquisition of multiple components of an event, memory trace formation, and retrieval of it. 11,12From our results, it is not possible to determine what mnemonic process(es) is(are) impaired by propofol. It remains to further characterize the memory process(es) that is(are) involved in propofol-induced amnesia.

Studies were performed in humans to compare the amnesic effects of propofol and midazolam. 26–30Some found that propofol has equivalent amnesic effects to those of midazolam, whereas others found more profound and reliable amnesic effects of midazolam. Veselis et al . found that at equal light sedation, midazolam and propofol provided equivalent memory impairment for emotionally unloaded information. 6From our results, low anxiolytic doses of propofol and midazolam display equivalent memory impairment of an aversively loaded event in a rodent model of contextual fear learning.

Propofol and midazolam can be used for conscious sedation (where verbal contact is maintained) or sedation. 31–33One of the main goals of sedation is preventing memory of scary events in patients. 34,35Fear conditioning is a fundamental associative learning observed in humans and animals that has relevance to clinical practice. This particular form of automatic learning may lead to an emotional memory trace in the absence of conscious thought 36and appears more involved in the mechanism of anxiety disorders after a stressful experience. 37In the present study, we used a Pavlovian fear-conditioning procedure in which the subject passively experienced information in a specific context. This form of learning has appeared to us particularly relevant for clinical practice as patients usually experience passively scary information in a specific environment, during sedation. Our results suggest that propofol and midazolam could prevent the subsequent conditioned anxiety resulting from an aversive event occurring during conscious sedation (where verbal contact is maintained in humans).

In conclusion, propofol impaired memory of an aversive event and of a nonaversive event. This amnesic effect may be demonstrated in rodents at a low dose of 9 mg/kg, which is devoid of locomotor impairment; the anesthetic dose of propofol being approximately 100 mg/kg for the strain of rats we used. From our results, propofol has specific effect on memory processes that did not rely on its sedative or anxiolytic properties. Such an observed amnesic effect was comparable with that observed with midazolam at a low anxiolytic and nonsedative dose in rats.

The authors thank Raymond Wilhelm, Animal Care Staff, for the care he lavished on the rats, and Christophe Mittelhauser, Technician, (both from the Institut of Physiology, Faculté de Médecine, Strasbourg, France) for his technical assistance in Experiment 4.