Psychopharmacology

, Volume 184, Issue 1, pp 1–12

Effects of morphine on the plus-maze discriminative avoidance task: role of state-dependent learning

Authors

  • C. L. Patti
    • Department of PharmacologyUniversidade Federal de São Paulo
  • S. R. Kameda
    • Department of PharmacologyUniversidade Federal de São Paulo
  • R. C. Carvalho
    • Department of PharmacologyUniversidade Federal de São Paulo
  • A. L. Takatsu-Coleman
    • Department of PharmacologyUniversidade Federal de São Paulo
  • G. B. Lopez
    • Department of PharmacologyUniversidade Federal de São Paulo
  • S. T. Niigaki
    • Department of PharmacologyUniversidade Federal de São Paulo
  • V. C. Abílio
    • Department of PharmacologyUniversidade Federal de São Paulo
  • R. Frussa-Filho
    • Department of PharmacologyUniversidade Federal de São Paulo
    • Department of PharmacologyUniversidade Federal de São Paulo
Original Investigation

DOI: 10.1007/s00213-005-0238-6

Cite this article as:
Patti, C.L., Kameda, S.R., Carvalho, R.C. et al. Psychopharmacology (2006) 184: 1. doi:10.1007/s00213-005-0238-6

Abstract

Rationale

The amnesic effects of morphine may be related to its action on nociception, anxiety, or locomotion. This effect is also suggested to be related to state dependency.

Objectives

The aims of this study were to verify the effects of morphine on mice tested in the plus-maze discriminative avoidance task (DAT) that uses light and noise as aversive stimuli and allows the concomitant evaluation of learning, memory, anxiety, and locomotion and also to verify the possible role of state-dependent learning in the effects of morphine.

Methods and results

The DAT was conducted in a modified elevated plus-maze. In the training, the aversive stimuli were applied when mice entered in one of the enclosed arms, whereas in the test, no stimuli were applied. The main results showed that (1) pretraining morphine (5–20 mg/kg i.p.) induced retrieval deficits (evaluated by the time spent in the aversive arm in the test) but not acquisition deficits (evaluated by the decrease in aversive arm exploration along the training); (2) pretest morphine (5–10 but not 20 mg/kg) counteracted this deficit; (3) morphine induced hypolocomotion (decreased number of entries in the arms), irrespective of memory alterations; and (4) morphine did not alter anxiety-like behavior (evaluated by the time spent in the open arms) during the training.

Conclusions

Morphine given before training induces retrieval deficits in mice tested in the DAT, and these deficits could be related to morphine-induced state-dependent learning. Neither the memory deficit induced by pretraining morphine nor the reversal of this deficit by pretest morphine seems to be related to anxiety levels or locomotor alterations.

Keywords

Avoidance paradigmsMorphineState dependenceMemoryAnxietyMice

Introduction

Several studies have shown the important role of opioid transmission in learning and memory (Izquierdo et al. 1980; McGaugh and Baratti 1985; Ragozzino and Gold 1994; Vaccarino et al. 1998), and several studies have focused on the effects of morphine on memory. In most of these studies, pretraining administration of this opiate induced learning/memory deficits in rodents tested in avoidance paradigms that use foot shocks as unconditioned stimuli (Izquierdo 1979; Ragozzino et al. 1992; Ragozzino and Gold 1994, 1995; McNay and Gold 1998; Aguilar et al. 2004). Although a question should be raised of whether pretraining morphine-induced analgesia would account for the effects of this drug on memory performance of tasks that involve painful stimuli, other studies have shown that the amnesic action of morphine is also present after posttraining or pretesting administration (Cestari and Castellano 1997; Saha et al. 2001; Mohammad-Reza et al. 2002; Costanzi et al. 2004). In this respect, when the drug is given after the acquisition, the effects observed in the test session are due to an effect of the drug on consolidation/retention processes. However, it is still unclear if pretraining morphine administration induces learning impairments, i.e., alterations in the acquisition of avoidance tasks. Aguilar et al. (2004), for example, have shown that morphine induced a behavioral deficit in the acquisition of a conditioned avoidance response in mice. This deficit, however, was suggested to be a consequence of morphine-induced hyperactivity (Aguilar et al. 2004) because the task applied was dependent on motor behavior (Vinader-Caerols et al. 1996; Aguilar et al. 1998, 2000).

Another important feature of pretraining morphine-induced amnesia in avoidance tasks is that it is abolished by pretesting administration of this drug (Izquierdo and Dias 1983; Bruins Slot and Colpaert 1999a; Khavandgar et al. 2002, 2003; Jafari et al. 2004; Mohammad-Reza and Rezayof 2004; Vakili et al. 2004). This phenomenon is called state dependence, i.e., the retrieval of learned information requires that the animal be in a state similar to that in which the memory for this information was acquired (Izquierdo et al. 1981; Bruins Slot and Colpaert 1999a), and has been demonstrated after administration of several psychoactive drugs (Colpaert 1990; Jackson et al. 1992; Bruins Slot et al. 1999; Colpaert et al. 2001). In addition, it has been shown that state dependence may act separately on acquisition, retention, and retrieval (Colpaert et al. 2001). Because of the failure to respond in the absence of the drug, morphine state-dependent learning has been proposed as a mechanism for opiate dependence (Colpaert 1990, 1996; Spanagel 1995). In addition, it has been shown to be dose-specific and related to time of administration (Nishimura et al. 1990; Bruins Slot and Colpaert 1999a; Khavandgar et al. 2003), suggesting that the same strength of morphine effects should be present for retrieval to occur. In this respect, one of the effects of this drug that could provide cues for the conditioned learning could be the analgesic state (Bruins Slot and Colpaert 1999a). This possibility is supported by the fact that morphine state-dependent learning occurs at times and at doses at which morphine also produces analgesia (Bruins Slot and Colpaert 1999b). This issue seems to be of relevance regarding studies on morphine-induced state dependence performed in avoidance paradigms that use painful stimuli as unconditioned stimuli.

Finally, another factor that should be taken into account is the anxiolytic effect of morphine, usually observed in animal models of anxiety Motta and Brandão 1993; Anseloni et al. 1999; Kõks et al. 1999; Sasaki et al. 2002; Shin et al. 2003). The emotional alterations induced by morphine seem to be important not only because the anxiety level can be related to memory deficits found in behavioral animal models (Silva and Frussa-Filho 2000, 2002; Silva et al. 2002a, 2004a; Calzavara et al. 2004) but also because of a possible role of anxiety alterations (as relevant features of the general state induced by morphine) in state-dependent learning.

The aims of the present study are (1) to verify the effects of morphine on learning and memory performance of mice tested in the plus-maze discriminative avoidance task (DAT), (2) to verify the possible role of state-dependent learning in the effects of morphine on this task, and (3) to investigate the participation of anxiety-like behavior and motor alterations on the effects of morphine on learning and memory. In the plus-maze DAT, light and noise, instead of electric shocks, are used as unconditioned stimuli. In addition, this behavioral paradigm allows the concomitant evaluation of learning, memory, anxiety, and motor parameters (Silva et al. 1997, 2002b, 2004a; Silva and Frussa-Filho 2000).

Methods

Subjects

Three-month-old Swiss EPM-M1 male mice (outbred, raised, and maintained in the Center for Development of Experimental Models in Medicine and Biology of Universidade Federal de São Paulo, Brazil, since 1985; Festing 1993) were used. Animals weighing 30–35 g were housed under conditions of controlled temperature (22–23°C) and lighting (12-h light, 12-h dark; lights on at 7 a.m.). Food and water were available ad libitum throughout the experiments. Animals used in this study were maintained in accordance with the guidelines of the Committee on Care and Use of Laboratory Animal Resources, National Research Council, USA.

Drug

Morphine sulfate (Dimorf, Innovatec) was diluted in saline solution that was used as control solution. Both morphine and saline solutions were given intraperitoneally at a dose of 10 ml/kg of body weight.

Discriminative avoidance task procedure and statistical analysis

The apparatus employed is a modified elevated plus-maze, made of wood, containing two enclosed arms with sidewalls, and no top (28.5×7×14 cm, 03 lx at the floor level), opposite to two open arms (28.5×7 cm, 09 lx at the floor level). A 100-W lamp was placed exactly over the middle of one of the enclosed arms (aversive enclosed arm, 660 lx at the floor level). In the training session, each mouse was placed in the center of the apparatus and, over a period of 10 min, every time the animal entered the enclosed arm containing the lamp, an aversive situation was produced until the animal left the arm. The aversive stimuli were the 100-W light and an 80-dB noise produced by a small machine placed under the aversive enclosed arm. In the test session (performed in the same room with the observer in the same position), the mice were again placed in the center of the apparatus and observed for 3 min without receiving the aversive stimulation. In all experiments, the animals were observed in a random order and in a blind manner, and the apparatus was cleaned with a 5% alcohol solution after each behavioral session. Total number of entries in any of the arms, number of entries in both enclosed arms, percent time spent in the aversive enclosed arm (time spent in aversive enclosed arm/time spent in both enclosed arms), and percent time spent in open arms (time spent in open arms/time spent in both open and enclosed arms) were calculated and compared by the one- or two-way analysis of variance (ANOVA) followed by Duncan’s test or by Student’s t test. The decrease in percent time spent in the aversive arm throughout the training session was used to evaluate learning of the task (experiment II) and was compared by ANOVA with repeated measures. Memory was evaluated by the time spent in the aversive vs nonaversive enclosed arms (compared by two- or three-way ANOVA followed by Duncan’s test) and by percent time in aversive enclosed arm in the test session. The animal was considered to be in a certain arm when the four paws passed over its entrance. Anxiety-like behavior and locomotor activity were evaluated by the percent of time spent in the open arms and total number of entries in all the arms or in both enclosed arms of the apparatus, respectively.

Hot plate procedure and statistical analysis

The antinociceptive effects of the doses of morphine used in the above experiments were evaluated in the hot plate test. Mice were placed on a heated zinc plate (55°C), and the latency to hind paw withdrawal was recorded throughout 60 min at 10-min intervals (Woolfe and MacDonald 1944). A maximal latency value of 30 s was set. The data were analyzed by ANOVA with repeated measures followed by Duncan’s test.

Experimental design

Experiment I

Groups of ten mice received saline, 5, 10, or 20 mg/kg morphine. Thirty minutes after the injection, all the animals were submitted to the plus-maze discriminative avoidance training session, and a test session was performed 24 h later, 30 min after the saline injection.

Experiment II

Groups of 20 mice received saline or 10 mg/kg morphine. Thirty minutes after the injection, all the animals were submitted to the plus-maze discriminative avoidance training session, and the behavioral parameters of this paradigm were registered minute by minute.

Experiment III

Twenty-four hours after the training session of experiment II, ten animals treated with saline and ten animals treated with morphine (before the training session, experiment II) received saline, and the other 20 animals pretraining treated with saline (n=10) or morphine (n=10) received 10 mg/kg morphine. A test session was performed 30 min after this injection.

Experiment IV

Groups of mice received saline (n=28), 5 (n=10), or 20 (n=10) mg/kg morphine. Thirty minutes after the injection, all the animals were submitted to the plus-maze discriminative avoidance training session. Twenty-four hours after the training session, the animals treated with saline (before the training session) received saline (n=8), 5 (n=10), or 20 mg/kg morphine (n=10). The animals treated pretraining with 5 or 20 mg/kg morphine received the same dose of morphine before the test session. Thirty minutes after the injection, the animals were submitted to the plus-maze discriminative avoidance test session.

Experiment V

The subjects were placed on the hot plate to obtain baseline latencies. Immediately after this measurement, groups of animals (n=5) received saline, 5, 10, or 20 mg/kg morphine (each animal were tested only at a single dose) and placed again on the hot plate 10, 20, 30, 40, 50, and 60 min after the injections.

Results

Experiment I: effects of several doses of pretraining morphine on memory, anxiety-like behavior, and locomotor activity of mice tested in the plus-maze discriminative avoidance task

Two-way ANOVA with treatment as between-subjects factor and arm type (aversive vs nonaversive) as within-subject factor revealed significant effects of arm type [F(1,72)=2241.37, p<0.001] and treatment × arm type interaction [F(3,72)=2.82, p<0.05]. Post hoc analysis by Duncan’s test revealed that all the groups spent significantly less time in the aversive than in the nonaversive enclosed arm (Fig. 1a). In this session, one-way ANOVA followed by Duncan’s test showed that all the groups treated with morphine presented significantly decreased percent time in the aversive enclosed arm (Fig. 1c) when compared to saline-treated animals [F(3,36)=3.76, p<0.05].
https://static-content.springer.com/image/art%3A10.1007%2Fs00213-005-0238-6/MediaObjects/213_2005_238_Fig1_HTML.gif
Fig. 1

Effects of pretraining morphine on mice tested in the plus-maze DAT. Mice were treated with saline (0), 5, 10, or 20 mg/kg morphine 30 min before a 10-min-long training session and tested during 3 min, 24 h later, in the absence of the drug. Results are presented as mean±SE of time (seconds) spent in the nonaversive (Nav) and in the aversive (Av) enclosed arms (a, b) and percent time spent in the aversive enclosed arm (c, d) of a plus-maze discriminative avoidance apparatus in the training and test sessions, respectively. *p<0.05 compared to the time spent in the nonaversive enclosed arm (two-way ANOVA and Duncan’s test). °p<0.05 compared to saline-treated group (one-way ANOVA and Duncan’s test)

In the test session, two-way ANOVA revealed only a significant treatment × arm type interaction effect [F(3,72)=4.16, p<0.01]. Post hoc analysis by Duncan’s test revealed that only saline-treated animals presented significantly less time in the aversive enclosed arm than in the nonaversive one (Fig. 1b). Accordingly, in this session, one-way ANOVA followed by Duncan’s test showed that all the groups treated with morphine presented significantly increased percent time in the aversive enclosed arm (Fig. 1d) when compared to saline-treated animals [F(3,36)=4.77, p<0.01].

No differences were found in percent time spent in the open arms in the training or test sessions (one-way ANOVA followed by Duncan’s test; Fig. 2a,b).
https://static-content.springer.com/image/art%3A10.1007%2Fs00213-005-0238-6/MediaObjects/213_2005_238_Fig2_HTML.gif
Fig. 2

Effects of pretraining morphine on mice tested in the plus-maze DAT (see legend for Fig. 1). Results are presented as mean±SE of percent time spent in the open arms (% TO) (a, b) and total number of entries and number of entries in both enclosed arms (c, d) of a plus-maze discriminative avoidance apparatus in the training and test sessions, respectively, presented by mice treated with saline (0), 5, 10, or 20 mg/kg morphine 30 min before the training session. *p<0.05 compared to saline-treated group (one-way ANOVA and Duncan’s test)

In the training session, when locomotor-activity-related parameters were compared, no differences were found between saline controls and mice treated with 5 mg/kg morphine, whereas groups treated with 10 or 20 mg/kg morphine presented total number of entries [F(3,36)=5.66, p<0.005], and number of entries in the enclosed arms [F(3,36)=4.80, p<0.01] significantly decreased (one-way ANOVA followed by Duncan’s test; Fig. 2c). In the test session, no differences were found between mice treated with 5 or 20 mg/kg morphine and saline controls, whereas groups treated with 10 mg/kg morphine presented number of entries in the enclosed arms [F(3,36)=4.02, p<0.05] significantly increased (one-way ANOVA followed by Duncan’s test; Fig. 2d).

No significant effect of morphine doses was found in any of the parameters analyzed.

Experiment II: effects of pretraining morphine on learning of mice during the plus-maze discriminative avoidance training

ANOVA for the percent time spent in the aversive arm with treatment as a between-subjects factor and time (minutes of observation) as a repeated-measures factor revealed significant effects of treatment [F(1,37)=6.91, p<0.05] and time [F(9,333)=17.21, p<0.001]. Indeed, although the percent time spent in the aversive arm was decreased in both groups from the second minute onward, this decrease had a greater magnitude in morphine-treated mice (Fig. 3a). Similarly to experiment 1, the percent time spent in the aversive arm (t=3.27, p<0.005), the total number of entries (t=4.73, p<0.001), and the number of entries in both enclosed arms (t=5.56, p<0.001) were significantly decreased in morphine-treated animals, while percent time in the open arms was not modified (Student’s test; Table 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs00213-005-0238-6/MediaObjects/213_2005_238_Fig3_HTML.gif
Fig. 3

Effects of morphine on training performance of mice in the plus-maze DAT. Mice were treated with saline (Sal) or 10 mg/kg morphine (M10) 30 min before a 10-min training session. Data are presented as mean±SE of percent time spent in the aversive enclosed arm (% TAv) of a plus-maze discriminative avoidance apparatus in each of the 10 min of the training session. ANOVA with repeated measures revealed treatment and time significant effects

Table 1

% TAv, % TO, TE, and EE of a plus-maze discriminative avoidance apparatus (mean±SE) in the entire 10 min of the training session presented by mice treated with saline or 10 mg/kg morphine 30 min before the session (see legend for Fig. 3)

 

Saline

Morphine (10 mg/kg)

% TAv

08.41±8.71

1.98±1.22*

% TO

07.51±9.30

3.04±6.46

TE

20.60±9.40

9.15±5.30*

EE

17.40±6.74

7.75±9.85*

% TAv Percent time spent in the aversive enclosed arm, % TO percent time spent in the open arms, TE total number of entries, EE number of entries in both enclosed arms

*p<0.05 compared to saline-treated animals (Student’s t test)

Experiment III: effects of pretest 10 mg/kg morphine on memory of pretraining morphine-treated mice tested in the plus-maze discriminative avoidance task

Three-way ANOVA with pretraining and pretest treatments as between-subjects factors and arm type (aversive vs nonaversive) as within-subject factor revealed significant effects of arm type [F(1,72)=79.38, p<0.001], arm type × pretraining treatment interaction [F(1,72)=12.51, p<0.001], arm type × pretest treatment interaction [F(1,72)=21.09, p<0.001], and arm type × pretraining treatment × pretest treatment interaction [F(1,72)=5.03, p<0.05]. Post hoc analysis by Duncan’s test revealed that both groups treated with saline before training and mice treated with morphine before both sessions spent significantly less time in the aversive than in the nonaversive enclosed arm, whereas mice treated with morphine before training and with saline before testing showed similar times spent in aversive and nonaversive enclosed arms (Fig. 4a).
https://static-content.springer.com/image/art%3A10.1007%2Fs00213-005-0238-6/MediaObjects/213_2005_238_Fig4_HTML.gif
Fig. 4

Effects of morphine on testing performance of pretraining-treated mice in the plus-maze DAT. Mice were treated with saline (S-) or 10 mg/kg morphine (M10-) 30 min before a 10-min training session and tested during 3 min, 24 h later, 30 min after an injection of saline (-S) or 10 mg/kg morphine (-M10). Results are presented as mean±SE of time (seconds) spent in the nonaversive (NAv) and in the aversive (Av) enclosed arms (a), percent time in the aversive arm (% TAv) (b), percent time in the open arms (% TO) (c), total number of entries and number of entries in the enclosed arms (d) of a plus-maze discriminative avoidance apparatus in the test session. *p<0.05 compared to the time spent in the nonaversive enclosed arm (three-way ANOVA and Duncan’s test). °p<0.05 compared to S–S group (two-way ANOVA and Duncan’s test)

Two-way (pretraining × pretest treatments) ANOVA for percent time spent in the aversive arm revealed significant effects of both pretraining [F(1,36)=16.58, p<0.001] and pretest [F(1,36)=6.36, p<0.05]. Post hoc analysis by Duncan’s test revealed that (1) mice treated with saline before training and with morphine before testing did not differ significantly from control (saline + saline) group, (2) mice treated with morphine before training and saline before testing presented significantly increased percent time spent in the aversive enclosed arm when compared to control group, (3) mice treated with morphine before both sessions were not significantly different from the control group, and (4) mice treated with morphine before both sessions presented significantly decreased percent time spent in the aversive arm when compared to mice that received morphine only before training (Fig. 4b).

Two-way ANOVA for total number of entries and number of entries in the enclosed arms revealed pretest treatment [F(1,36)=11.69 and 7.21, respectively, p<0.05] and pretraining × pretest treatment interaction [F(1,36)=6.43 and 4.77, respectively, p<0.05] effects. Post hoc analysis by Duncan’s test revealed that only mice that received morphine before training and saline before testing presented increased locomotor activity parameters when compared to animals that were treated with saline before both sessions (Fig. 4c).

Two-way ANOVA for percent time spent in the open arms revealed only a significant effect of pretest treatment [F(1,36)=10.03, p<0.005]. Indeed, post hoc analysis by Duncan’s test revealed that both groups treated with morphine before testing presented decreased percent time in the open arms when compared to the respective controls (Fig. 4d).

Experiment IV: effects of pretest 5 and 20 mg/kg morphine on memory of pretraining morphine-treated mice tested in the plus-maze discriminative avoidance task

Two-way ANOVA with treatment as between-subjects factor and arm type (aversive vs nonaversive) as within-subject factor revealed significant effects of arm type [F(1,90)=470.67, p<0.001]. Post hoc analysis by Duncan’s test revealed that all the groups spent significantly less time in the aversive than in the nonaversive enclosed arm (Fig. 5a). In this session (training), one-way ANOVA followed by Duncan’s test revealed that there was no significant difference in any of the groups for percent time in the aversive enclosed arm (Table 2).
https://static-content.springer.com/image/art%3A10.1007%2Fs00213-005-0238-6/MediaObjects/213_2005_238_Fig5_HTML.gif
Fig. 5

Effects of morphine on testing performance of pretraining-treated mice in the plus-maze DAT. Mice were treated with saline (Sal), 5 (M5), or 20 (M20) mg/kg morphine 30 min before a 10-min training session and tested during 3 min, 24 h later, 30 min after an injection of saline (-Sal), 5 (-M5), or 20 (-M20) mg/kg morphine. Results are presented as mean±SE of time (seconds) spent in the aversive enclosed arm (Av) and in the nonaversive enclosed arm (Nav) of a plus-maze discriminative avoidance apparatus in the training (a) and test (b) sessions. *p<0.05 compared to the time spent in the nonaversive enclosed arm

Table 2

% TAv, % TO, TE, and EE of a plus-maze discriminative avoidance apparatus (mean±SE) presented by mice treated with saline, 5, or 20 mg/kg morphine 30 min before training or test sessions (see legend for Fig. 5)

 

Training

Test

Sal

M5

M20

Sal–Sal

Sal–M5

Sal–M20

M5–M5

M20–M20

% TAv

4.24±0.60

7.47±3.85

3.42±4.74

39.16±11.15

0.68±0.68a,b

0.34±0.34a,b

33.66±12.47

56.09±14.72

% TO

15.93±3.69

13.92±7.94

6.00±1.39

17.66±4.56

0.74±0.74a

1.70±1.70a

2.42±1.40a

2.92±2.44a

TE

17.64±1.60

14.20±2.79

14.10±2.34

7.63±1.35

1.20±0.25a

2.00±1.12a

3.60±1.17a

2.90±1.38a

EE

14.07±1.36

11.2±2.03

11.2±1.91

5.00±0.82

1.10±0.18a

1.60±0.72a

3.00±0.93

2.10±0.81a

% TAv Percent time spent in the aversive enclosed arm, % TO percent time spent in the open arms, TE total number of entries, EE number of entries in both enclosed arms

ap<0.05 compared to Sal–Sal group in the test session

bp<0.05 compared to M5–M5 and M20–M20 groups in the test session (ANOVA followed by Duncan’s test)

In the test session, two-way ANOVA revealed only a significant arm-type effect [F(1,90)=23.195, p<0.001]. Post hoc analysis by Duncan’s test revealed that all the groups except M20–M20 (animals that received 20 mg/kg morphine before both sessions) presented significantly less time in the aversive enclosed arm than in the nonaversive one (Fig. 5b). In this session, one-way ANOVA followed by Duncan’s test showed that the groups treated with saline before training and 5 or 20 mg/kg morphine before test presented significantly decreased percent time in the aversive enclosed arm (Table 2) when compared to all the other groups [F(4,43)=5.8421, p<0.001].

No differences were found in percent time spent in the open arms in the training session (one-way ANOVA followed by Duncan’s test; Table 2). In the test session, all the animals that received morphine before testing presented a decrease in percent time spent in the open arms [F(4,43)=4.0917, p<0.01].

In the training session, no differences were found when mice treated with 5 or 20 mg/kg morphine were compared to saline controls for total number of entries [F(2,45)=1.0154, p>0.05] or number of entries in the enclosed arms [F(2,45)=1.0291, p>0.05] (one-way ANOVA followed by Duncan’s test; Table 2). In the test session, one-way ANOVA followed by Duncan’s test showed a decrease in the total number of entries [F(4,43)=2.9584, p<0.05] and number of entries in the enclosed arms [F(4,43)=2.9443, p<0.05] presented by all the animals treated with morphine before the test session, except by the animals treated with 5 mg/kg morphine before both sessions (Table 2).

Experiment V: effects of morphine in nociception evaluated by the hot plate procedure

ANOVA for latency to hind paw withdrawal with treatment as between-subjects factor and time (minutes of observation) as a repeated measure revealed significant effects of treatment [F(3,16)=8.67, p<0.01], time [F(6,96)=5.39, p<0.001], and treatment × time interaction [F(18,96)=2.84, p>0.01] (Fig. 6). All the animals presented a similar baseline latency. After treatment, the dose of 5 mg/kg morphine did not modify latency to hind paw withdrawal. The dose of 10 mg/kg morphine induced an analgesic effect from 30 min onward (increased latency when compared to saline-treated animals). In addition, the dose of 20 mg/kg morphine was effective in inducing an analgesic effect from 20 min onward. Finally, at 40 min of observation, the group treated with 20 mg/kg morphine was significantly different from the group treated with 10 mg/kg morphine, suggesting a dose-dependent antinociceptive effect of morphine.
https://static-content.springer.com/image/art%3A10.1007%2Fs00213-005-0238-6/MediaObjects/213_2005_238_Fig6_HTML.gif
Fig. 6

Effects of morphine on hind paw withdrawal from a hot plate. Mice were treated with saline (Sal), 5 (M5), 10 (M10), or 20 (M20) mg/kg morphine. Latency for hind paw withdrawal was measured upon exposure to a hot plate immediately before (0) and 10–60 min after treatment. Data are presented as mean±SE latency to hind paw withdrawal (s).*p<0.05 compared to Sal group. °p<0.05 compared to the M5 group. Hp<0.05 compared to M10 group (ANOVA with repeated measures followed by Duncan’s test)

Discussion

In the plus-maze discriminative avoidance paradigm, learning is indicated by a decrease in the exploration of the aversive enclosed arm throughout the training session (Silva et al. 2004a), while retrieval of the task is indicated by decreased exploration of this arm in the test session when the aversive stimuli are no longer present. Indeed, amnesic procedures increase (Claro et al. 1999; Silva et al. 1999, 2002a,b, 2004a; Silva and Frussa-Filho 2000, 2002), while memory-improving manipulations decrease (Silva et al. 1997, 1999, 2000; Claro et al. 1999) aversive arm exploration in the test session. In this respect, results from experiment I showed that all doses of morphine administered before the training induced retrieval deficits in the test session. Indeed, besides spending the same amount of time in the aversive and in the nonaversive enclosed arms (see Fig. 1b), all the groups treated with morphine presented significantly increased percent time in the aversive enclosed arm when compared to saline-treated animals (Fig. 1d). These data are in agreement with other studies reporting a memory deficit in avoidance tasks after pretraining morphine administration (Izquierdo 1979; Ragozzino et al. 1992; Ragozzino and Gold 1994, 1995; McNay and Gold 1998; Aguilar et al. 2004).

The possibility that the retrieval deficit shown by morphine-treated animals in experiment I was the result of some effect of the drug in the training session—which would modify the learning process and therefore impair retention and/or retrieval—could be raised. Regarding motor behavior, it was observed that 10 and 20 mg/kg morphine decreased total number of entries and entries in the enclosed arms of the apparatus (Fig. 2c; Table 1). Although morphine has been mostly reported as a stimulant of motor behavior in mice (Hynes and Berkowitz 1983; Stevens et al. 1986; Kuribata 1995; Kuzmin et al. 2000) and rats (Vivian and Miczek 1999; Kalinichev et al. 2004), there are some studies showing that morphine, depending on the dose and route of administration, can decrease locomotion in rats (Havemann et al. 1982; Boyer et al. 1998; Vivian and Miczek 1998; Timar et al. 2005). In addition, our data are in accordance with the fact that systemic administration of morphine, especially at doses over 10 mg/kg, is associated with a biphasic effect on locomotor activity. Indeed, when morphine is administered to rats, an initial period of hypolocomotion (lasting 30–60 min) and a subsequent period of hyperactivity are observed (Babbini and Davis 1972; Vasko and Domino 1978; Roberts et al. 1978; Walter and Kuschinsky 1989; Johnson and Glick 1993; Kosten and Bombace 2000). This biphasic pattern of locomotor alterations has also been demonstrated in hamsters (Schunr et al. 1983) and mice (Saito 1990). Notwithstanding, the alterations induced by morphine in the locomotor parameters did not seem to be related to the amnesic effect of the drug. In this respect, 5 mg/kg morphine, despite inducing retrieval deficits, did not modify locomotor activity in the training session.

Pretraining morphine administration (especially at the dose of 10 mg/kg) induced an increase in the parameters related to locomotor activity in the test session (Figs. 2d and 4c). This increase does not seem to be related to a stimulant effect of the previously given (24 h before) morphine treatment, since the stimulant phase of morphine effect on locomotor activity in mice ends approximately 240 min after the injection (Kosten and Bombace 2000). It is possible that the animals that received pretraining morphine were not able to remember not only the presentation of aversive stimuli in a specific arm but also the general information about the apparatus. In other words, the impairment effect of morphine may not have been specifically related to the associative learning induced by pairing light and noise to the aversive arm. Habituation (i.e., the decrement of exploratory activity when a rodent is exposed to a new environment; Conceição et al. 1994; Silva et al. 1996) deficit may have also occurred in pretraining-morphine-treated mice.

Another modification induced by morphine administration in the training session was the decrease in percent time spent in the aversive enclosed arm (Fig. 1c; Table 1). Although controversial to a decreased retrieval, this result could reflect an improvement in learning of the task; that is, the animals treated with morphine could have spent less time in the aversive arm because they learned faster to avoid it. Indeed, the analysis of the aversive enclosed arm exploration throughout the training session does provide an indication of better learning by morphine-treated animals (Fig. 3). In fact, although both saline- and morphine-treated mice presented decreased percent time in the aversive arm from the second minute onward, this decrease had a greater magnitude in morphine-treated mice. This increased learning of the task by morphine-treated mice could be related to the aversive properties of this drug, which can induce increased responses to aversive situations. Indeed, Anseloni et al. (1999) demonstrated that although low doses of systemically administered morphine (0.1–0.3 mg/kg) induced an anxiolytic-like behavior, a high dose of this drug injected into the dorsal periaqueductal gray increased anxiety-like behaviors in an elevated plus-maze.

The analysis of percent time spent in the open arms (currently used as a measurement of anxiety-like behavior in rodents; Pellow and File 1986; Lister 1987; Frussa-Filho et al. 1991, 1999; Goto et al. 1993; Silva et al. 2004b) showed that morphine had no effect on anxiety-like behavior of mice in the training session of the plus-maze discriminative avoidance apparatus (Fig. 2a; Table 1). This result is not in accordance with other studies that show an anxiolytic effect of this drug in the conventional elevated plus-maze (Motta and Brandão 1993; Anseloni et al. 1999; Kõks et al. 1999; Sasaki et al. 2002; Shin et al. 2003). These discrepant results could be due to qualitative differences between the exploration of the open arms of the conventional elevated plus-maze apparatus and the plus-maze apparatus used here. However, percent time spent in the open arms of the apparatus employed here has been extensively used for evaluation of anxiety-related behaviors in mice, presenting similar results to those obtained in the conventional elevated plus-maze (Silva et al. 1997, 2002a, 2004b; Silva and Frussa-Filho 2000, 2002; Calzavara et al. 2004). This concern, notwithstanding the above-mentioned aversive properties of morphine, may increase responses to aversive situations (Anseloni et al. 1999). Accordingly, morphine administration induced an anxiogenic-like behavior in the plus-maze apparatus when administered at high doses in the dorsal periaqueductal gray (Anseloni et al. 1999). Furthermore, the opioid antagonist naltrexone potentiated the anxiolytic effect of chlordiazepoxide in rats and mice submitted to this and other animal models of anxiety (Belzung and Agmo 1997; Frussa-Filho et al. 1999). Interestingly, the facilitation of the anxiogenic component of morphine produced by an aversive environment could also explain the observed decrease in the locomotor activity parameters. Indeed, open-arm exploration and locomotor activity in the plus-maze apparatus are closely related (Moser 1989; Silva et al. 2004b).

In experiment III, we attempted to verify whether the amnesic effect reported in the plus-maze DAT would be related to state-dependent learning. The results were in accordance with the first experiment, showing that 10 mg/kg morphine induced a retention deficit in mice tested in the plus-maze apparatus (Fig. 4). Corroborating previous studies of morphine effects in avoidance paradigms (Izquierdo 1979; Ragozzino et al. 1992; Ragozzino and Gold 1994, 1995; McNay and Gold 1998; Aguilar et al. 2004), the pretest administration of 5 or 10 mg/kg morphine completely counteracted the amnesic effect of pretraining administration of this drug. In this respect, it has been suggested that the improvement in performance induced by pretest morphine would be related to a direct effect of this drug on memory retrieval mechanisms (Nishimura et al. 1990; Shiigi and Kaneto 1990). Accordingly, in the present study, mice that received morphine only before the test session presented better retrieval when compared to control mice (decreased percent time in the aversive arm in the test session; Table 2).

The present results lend support to the state-dependent learning induced by morphine hypothesis. Indeed, mice that received morphine (5 or 10 mg/kg) before both sessions (training and test) were not different from controls, showing the same amount of retrieval (experiments III and IV). In contrast, the present results do not corroborate the observation that state-dependent learning is bidirectional, i.e., animals that are trained after a saline injection should present retention deficits if morphine is administered before the test session. In the present study, mice treated with morphine before testing showed similar or even better retention levels than that presented by the control group (Figs. 4 and 5; Table 2). In this respect, previous data have shown that drug-to-vehicle state alterations often exert greater effects than vehicle-to-drug changes do (Colpaert 1990; Jackson et al. 1992).

The impaired retrieval induced by the dose of 20 mg/kg was not counteracted by pretest administration of this dose (Fig. 5b). This finding suggests that pretraining administration of morphine, depending on the dose, can induce a memory deficit that is not related to state-dependent learning. Indeed, opioid transmission was suggested to exert an inhibitory modulation of memory consolidation, since posttraining administration of opioid agonists and antagonists impair and improve retention, respectively (see Izquierdo 1982 for a review; Castellano et al. 1994; Cestari and Castellano 1997).

Although no effect of morphine on anxiety levels in the training session was detected, this drug was able to induce an anxiogenic-like effect in the test session. In this respect, there are several studies showing that the anxiety-like behavior related to the open arms of a plus-maze is qualitatively different in animals that had been previously exposed to the maze (Rodgers et al.  1997; Holmes and Rodgers 1999). Indeed, benzodiazepines—classical anxiolytic drugs—do not modify open-arm exploration in plus-maze-experienced rodents (File 1990; Pereira et al. 1999; Frussa-Filho and Ribeiro 2002; Calzavara et al. 2005). Anxiogenic substances, however, usually maintain their effects (decreased open-arm exploration) upon retesting (File 1993). In this context, morphine, which has been shown to induce aversion (Anseloni et al. 1999), could be potentiating the already increased preference for the enclosed arms in a second exposition to the plus-maze apparatus. This concern, notwithstanding the anxiety state generated by morphine, does not seem to be related to the state-dependent learning induced by this drug, since this effect was not present in the training session.

As mentioned in the “Introduction,” morphine-induced analgesia should be considered when evaluating the effects of this drug on learning and memory. In this respect, in the present study, we have used nonnociceptive stimuli (light and noise) to engender avoidance behavior. In addition, results from experiment V (morphine-induced analgesia) seem to corroborate the absence of a relationship between the effects of this opiate on nociception and memory. Indeed, while the three doses induced retrieval deficits, a significant antinociceptive effect was found only at 10 and 20 mg/kg. Additionally, while the analgesic effect was dose-dependent, no dose relation was found in the plus-maze DAT test session. Finally, the results depicted in Fig. 6 also suggest dissociation between antinociceptive effects and state-dependent learning, since the range of doses that induced these effects was not the same.

In conclusion, the results reported here indicate that morphine given before the training session induces retrieval deficits in mice tested in the plus-maze DAT, an avoidance paradigm that does not involve painful stimulation. This deficit was shown to be related, in part, to a morphine-induced state-dependent learning, since pretest morphine administration restored retention performance to control levels. Neither the memory deficit induced by pretraining morphine nor the reversion of this deficit by pretest morphine seems to be related to anxiety-related, locomotor, or antinociceptive effects of this opiate.

Acknowledgements

This research was supported by grants from Fundação de Amparo a Pesquisa do Estado de São Paulo (proc. 01/10713-7 and FAPESP/CEPID proc. 98/14303-3), Fundo de Auxílio ao Docente e Aluno da UNIFESP (FADA), Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), and Associação Fundo de Incentivo à Psicofarmacologia (AFIP).

The authors would like to thank Ms. Teotila R. Amaral, Mr. Cleomar S. Ferreira, and Mr. Antônio Rodrigues dos Santos for capable technical assistance. The experiments performed in the present study comply with the current Brazilian laws and are in accordance with the Ethics Committee of Universidade Federal de São Paulo.

Copyright information

© Springer-Verlag 2005