Introduction

Animals exposed to inescapable stressors exhibit a variety of behaviors that are thought to mimic some aspects of depression in humans (Maier 1984). The three most common stress-based models of depression used to validate antidepressant efficacy in rodents are the tail suspension test (TST) (Steru et al. 1985, 1987), forced swim test (FST) (Porsolt et al. 1977) and learned helplessness (LH) paradigm (Sherman et al. 1979; Shanks and Anisman 1989). In the TST and FST paradigms, animals will alternate between active escape responses and periods of immobility. In the LH model, animals exposed to inescapable shock will subsequently show impaired learning of an escape response (Maier and Seligman 1976). A wide range of antidepressants including tricyclic antidepressants (TCA), monoamine oxidase (MAO) inhibitors, selective serotonin reuptake inhibitors (SSRIs), and atypical antidepressants, have been shown to increase active escape responses in both rat (Sherman et al. 1979, 1982) and mouse (Porsolt et al. 1977; Steru et al. 1985; Shanks and Anisman 1989) models.

One approach to understanding the mechanism of action of antidepressants is to examine specific genes that have been implicated in antidepressant action. A number of genetically altered mice have been generated that have targeted genes that influence neurotransmitter systems (Xu et al. 1994; Baik et al. 1995; Picciotto et al. 1995; Tecott et al. 1995; Link et al. 1996; MacMillan et al. 1996; Rubinstein et al. 1997; Parks et al. 1998), second messengers systems (Bourtchuladze et al. 1994) and neurotrophic factors (Ernfors et al. 1995) that have been implicated in the mechanisms of action of antidepressants (Vaidya and Duman 2001).

One of the most common genetic backgrounds on which knockout mice are maintained is the C57BL/6J (B6) inbred strain. B6 mice have been reported to be resilient with respect to their hormonal (Anisman et al. 2001) and behavioral (Tannenbaum and Anisman 2003) responses to stress as well as to excitotoxic cell death (Schauwecker 2002). Despite this resistance to stressors, antidepressant-like responses have been reported previously for the B6 strain in the FST (Dalvi and Lucki 1999; Bai et al. 2001; Lucki et al. 2001), TST (Bai et al. 2001; Liu and Gershenfeld 2001) and LH (Shanks and Anisman 1989) stress-based models of depresion. Given the limitations of the available animal models (Dalvi and Lucki 1999; Cryan et al. 2002), it is important to establish consistent antidepressant effects across several models before assessing antidepressant effects in knockout and transgenic mice.

One potential difficulty in interpreting studies assessing the efficacy of antidepressants in mice is that the behavioral response to antidepressants in the FST and TST is often assessed following an acute injection of the drug (Porsolt et al. 1977; Steru et al. 1985). In humans, however, the therapeutic effects of antidepressant are not seen for weeks, or even months after initiation of treatment (Wong and Licinio 2001). A second potential problem is that few mouse behavioral studies have compared antidepressant efficacy in both males and females, although it is well established that the rate of depression is higher in women than in men (Weissman and Klerman 1985). Finally, in B6 mice, no study has compared different lengths of antidepressant treatment across sexes in multiple behavioral models. Therefore, in the present study, we examined the effects of chronic and subchronic oral administration of the TCA amitriptyline (AMI) in male and female B6 mice across three models of depression: the TST, FST, and LH paradigm.

Materials and methods

Animals

Experimentally naive male and female C57BL/6J (B6) mice were obtained from Jackson Laboratory (Bar Harbor, Maine, USA). Mice, ranging in age from 2 to 4 months, were group housed in cages with a maximum of five mice per cage. Mice were kept in a colony room maintained at 22°C on a 12-h light:dark cycle, with lights on at 7:00 a.m. Food and water were available at all times. All behavioral testing was carried out during the light portion of the light dark cycle. All animal procedures were in strict accordance with NIH Care and Use of Laboratory Animals Guidelines and were approved by the Yale Animal Care and Use Committee.

Oral amitriptyline administration

Amitriptyline HCl (AMI) (Sigma, St Louis, Mo., USA) was administered in the drinking water as the sole source of fluid. The concentration of AMI used was based on a previous study of LH in rats (Sherman et al. 1979). AMI (200 μg/ml, free base) was dissolved in 2% saccharin to increase palatability and control mice received 2% saccharin alone. Mice were given AMI or saccharin either chronically (21 days) or subchronically (4 days) before the start of behavioral testing and solutions were changed twice a week. Mice continued to drink AMI or saccharin throughout the duration of behavioral testing.

Behavioral testing

Tail suspension test

Approximately 1 cm from the end, each mouse's tail was taped (Scotch 35 Vinyl Electrical Tape) to a piece of Tygon tubing, 36 cm in length. Mice were suspended by the tail approximately 120 cm above the floor and the duration of immobility, defined as the absence of all movement except for those required for respiration, was recorded by an observer for 6 min. If a mouse climbed up on its tail during the testing session, it was gently pulled down and testing continued. Mice that continued to climb their tails were eliminated from the study.

Forced swim test

The FST was carried out in a glass cylinder (18.5 cm diameter, 25 cm height) filled with water to the height of 17 cm. Water was maintained at 23–25°C. Mice were placed into the water and immobility times were recorded by an observer for 15 min. Immobility was defined as absence of all movement except motions required to keep the mouse's head above the water.

Learned helplessness

Apparatus

Learned helplessness training was carried out in two shuttle boxes (Med Associates, St Albans, Vt., USA) in which the front, back and ceiling were clear Plexiglas, and the sides were aluminum. The shuttle boxes were enclosed in sound attenuating melamine chambers in which the inside dimensions were 59.5×55.9×35.6 cm. Sound attenuating chambers were equipped with houselights and fans that remained on during training and testing. Scrambled shock (0.30 mA) was delivered by a shock source to a grid floor made of stainless steel bars spaced 0.50 cm apart. A gate with an archway measuring 11.4×8.9 cm was inserted into the chamber, dividing it into two equal compartments (20.3×15.9×21.3 cm). The motor-driven gate opened automatically at the start of each trial. A series of photocells on both sides of the chamber monitored the position of the animal, measured escape latencies, and triggered the closing of the gate after an escape response was made.

Training

Procedures were based on those reported previously (Shanks and Anisman 1988; Caldarone et al. 2000). Learned helplessness was induced by administering 120 inescapable 4-s footshocks with a random intertrial interval (range 3–50 s) over a 1-h session. Training was given in two sessions that were spaced approximately 24 h apart. Mice were placed on either side of the shock chamber, so that either one or two mice were administered shock simultaneously. Mice of the same gender were always shocked together, but drug treatment varied. A control group that did not receive shock was exposed to the apparatus for an equivalent period of time.

Shuttle escape testing

Approximately 24 h after the second LH training session, mice were tested on the shuttle escape task. The side of the chamber on which each mouse was placed at the start of the test session was alternated. Mice were given 30 shuttle escape trials with 30-s intervals between the start of each trial. The gate opened when the shock turned on and the trial terminated when the mouse crossed through the gate into the adjacent compartment. Shock termination was delayed by 1 s after crossing. If an escape response was not made, the trial was terminated 24 s after shock onset.

Locomotor activity

One day prior to the FST, mice were tested for locomotor activity. The locomotor apparatus consisted of eight photocells, spaced 4 cm apart, which was connected to a computer that collected data through Windows compatible software. Mice were tested in clear plastic chambers (47×25×21 cm) that were covered with an opaque lid. Locomotor activity, as assessed by the number of beam breaks, was measured for 20 min.

Serum levels of AMI/NOR and measurement of AMI intake

A subset of mice that were tested in the FST or LH paradigm was killed by rapid decapitation. Trunk blood was collected, centrifuged, and stored at −80°C. Serum levels of AMI and its major metabolite nortriptyline (NOR) were measured by fluorescence polarization immunoassay (FPIA) (Abbott Laboratories, Abbott Park, Ill., USA). The assay is a competitive binding procedure that uses a flouroscein-labeled ligand. When the ligand is displaced from the antibody by the unlabeled drug in the sample, fluorescence polarization is reduced. The assay provides semiquantitative levels of AMI plus NOR levels (AMI/NOR). The assay was carried out on an AxSYM analyzer (Abbott Laboratories).

An independent group of mice was individually housed and intake of AMI was measured for 3 weeks. Mice were weighed each week and intake of AMI was calculated by averaging intake (mg AMI/kg body) across days 1–7 (subchronic) and days 8–21 (chronic).

Data analyses

In the TST, time immobile was averaged over the 6 min test. Data were analyzed by analysis of variance (ANOVA) with drug (saccharin, AMI), duration (subchronic, chronic) and sex as the between group variables. In the FST, time immobile was averaged in three 5-min blocks and in the locomotor activity test, data were assessed by the number of horizontal beam breaks averaged in four 5-min blocks. FST and locomotor activity data were analyzed by an ANOVA with drug (saccharin, AMI), duration (subchronic, chronic), and sex as the between group variables and block as the within group variable.

For the LH data, the 30 escape trials were divided into six blocks with latencies averaged across five trials. For mice that received inescapable shock training, an ANOVA was carried out with block as the within subject variable, and drug (saccharin, AMI), duration (subchronic, chronic), and sex as the between subject variables. No differences were found between non-shocked mice that received either chronic or acute saccharin or AMI so these data were also collapsed over duration to increase statistical power. A separate ANOVA was carried out on these mice with block as the within-subject variable, and drug (saccharin, AMI) and sex and the between-group variables.

Significant effects were followed up by the post-hoc Tukey honestly significant difference (HSD) test (α=0.05).

Results

Tail suspension test

The mean duration of immobility in the TST for female and male mice treated subchronically and chronically with AMI or saccharin is shown in Fig. 1. Treatment with AMI reduced immobility times, but the magnitude of the reduction was not influenced by duration of treatment or sex. ANOVA revealed a significant main effect of drug [F(1,60)=17.34, P<0.001], but no main effect of sex, duration, or any interaction of drug with sex or duration of treatment.

Fig. 1.
figure 1

Effects of subchronic and chronic oral AMI treatment on duration of immobility in the TST in female and male C57BL/6J mice (female subchronic n=7–8/group; female chronic, n=10/group; male subchronic, n=7–8/group; male chronic, n=9/group). Data are presented as mean (±SEM) immobility time across a 6-min test. Both suchronic and chronic AMI reduced immobility, but this reduction did not depend on duration of treatment and did not differ between females and males. †P<0.001, ANOVA, main effect of drug, saccharin versus AMI

Full size image

Forced swim test

The mean duration of immobility in the FST for female and male mice treated subchronically and chronically with AMI or saccharin is shown in Fig. 2. In general, treatment with AMI reduced immobility times, but the magnitude of the reduction varied depending on the duration of treatment. ANOVA revealed a significant block×drug×duration×sex interaction [F(2,158)=3.227, P<0.05]. Post hoc tests showed that females receiving subchronic AMI spent less time immobile than saccharin controls, but only in the second 5-min block. Females receiving chronic AMI spent less time immobile than saccharin controls across all three 5-min blocks. Males that received chronic AMI spent less time immobile than saccharin controls in the second and third 5-min block.

Fig. 2.
figure 2

The effects of subchronic and chronic oral AMI treatment on duration of immobility in the FST in female and male C57BL/6J mice (female subchronic, n=10/group; female chronic, n=10/group; male subchronic, n=15/group; male chronic, n=8–9/group). Data are presented as mean (±SEM) immobility time across three 5-min blocks. Chronic AMI was more effective than subchronic treatment in reducing immobility. *P<0.05, Tukey, saccharin versus AMI

Full size image

A secondary analysis was conducted to examine sex differences in baseline immobility in saccharin treated mice. An ANOVA with block (1–3) as the within subjects variable and sex as the between group variable revealed a significant block×sex interaction [F(2,84)=3.727, P<0.05]. Post hoc analyses revealed that females spent more time immobile than males in the second 5-min block.

Learned helplessness

Escape latencies for female and male mice receiving LH training and treated either subchronically or chronically with AMI are shown in Fig. 3A. In general, treatment with AMI reduced escape latencies, but the magnitude of the reduction varied according to duration of treatment and sex. ANOVA revealed a significant block×drug×sex interaction [F(5,84)=2.4, P<0.05] and a significant block×duration×sex interaction [F(5,84)=3.4, P<0.01]. Further subanalyses were conducted to determine the nature of the drug effect on these interactions. Repeated measures ANOVAs with block (1–6) as the within-subjects measure and drug (saccharin, AMI) as the between-subjects factor were carried out. Female mice that received subchronic AMI treatment showed a reduction in escape latencies compared to saccharin controls [F(1,19)=6.0, P<0.05]. Female mice that received chronic AMI treatment showed a reduction in escape latencies that varied according to block [F(5,22)=2.7, P<0.05]. Post hoc tests showed that females treated chronically with AMI had shorter escape latencies than saccharin controls in blocks 2–6. Escape latencies were not altered in males treated subchronically with AMI, but males treated chronically with AMI showed a reduction in escape latencies that varied according to block [F(5,19)=2.4, P<0.05]. Post hoc testing revealed that males treated chronically with AMI had shorter escape latencies than saccharin controls in blocks 3 and 4.

Fig. 3.
figure 3

A The effects of subchronic and chronic oral AMI treatment on escape latencies in female and male C57BL/6J mice that received inescapable shock in the LH training phase (female subchronic, n=10–11/group; female chronic, n=12/group; male subchronic, n=13/group; male chronic, n=10–11/group). Data are presented as mean (±SEM) escape latencies averaged over blocks of five trials. Subchronic AMI treatment decreased escape latencies only in females, whereas chronic AMI decreased escape latencies in both sexes. #P<0.05, ANOVA, main effect of drug, saccharin versus AMI; *P<0.05, Tukey, saccharin versus AMI. B Effects of subchronic and chronic oral AMI treatment on escape latencies in female and male C57BL/6J mice that did not receive shock during training in the LH test (female, n=8/group; male, n=8/group). No differences were found between subchronic and chronic treatment, so data for these treatments were combined. Data are presented as mean (±SEM) escape latencies averaged over blocks of five trials. AMI had no effect on escape latencies in either sex

Full size image

Escape latencies for female and male mice that did not receive inescapable shock during the LH training phase are shown in Fig. 3B. Treatment with AMI had no effect on escape latencies in these non-shocked control mice. ANOVA revealed no main effect of drug or sex or interaction of block with these variables.

An additional analysis was conducted to examine sex differences in baseline LH in saccharin treated mice. A repeated measures ANOVA was carried out with block (1–6) as the within subjects variable and sex and shock pre-treatment (shock, no shock) as the between group variable. Escape latencies were increased in mice that received inescapable shock pretraining [F(1,57)=26.5, P<0.0001], but did not vary according to sex.

Locomotor activity

The mean number of horizontal beam breaks for female and male mice treated subchronically or chronically with AMI or saccharin is shown in Fig. 4. Locomotor activity decreased over the 20-min test [F(3,79)=85.1, P<0.001], but did not vary according to drug treatment or sex.

Fig. 4.
figure 4

The effects of subchronic and chronic oral AMI treatment on locomotor activity in female and male C57BL/6J mice (female subchronic, n=10/group; female chronic n=10/group; male subchronic, n=15/group; male chronic, n=8–9/group). Data are presented as mean (±SEM) number of beam breaks over four 5-min blocks. AMI had no effect on locomotion in either sex

Full size image

Serum levels of AMI/NOR and measurement of AMI intake

Table 1 shows serum levels of AMI plus nortriptyline (AMI/NOR) in female and male mice following subchronic and chronic administration. The major purpose of this semiquantitative assay was to assure that adequate drug exposure was achieved in the mice using the oral route. The saccharin treated mice showed a background immunoreactivity of 54±12.6 (mean±SEM) (n=6), and are not included in the table. The total levels in the treated mice were in the approximate range observed in humans treated with amitriptyline (Ziegler et al. 1976; Breyer-Pfaff et al. 1982, 1989; Ulrich et al. 2001). An ANOVA on serum levels of AMI/NOR was carried out with duration (subchronic, chronic) and sex as the between group variables. AMI/NOR levels were not different between mice treated subchronically or chronically with AMI, although males had significantly higher AMI/NOR levels than females [F(1,33)=7.2, P<0.05]. An ANOVA on intake with time (subchronic, chronic) as the within subjects factor and sex as the between group factor was also carried out. These results showed that females consumed greater amounts of AMI corrected for body weight than males [F(1,7)=6.4, P<0.05].

Table 1. Mean (±SEM) serum levels of amitriptyline plus nortriptyline (AMI/NOR) (ng/ml), mean (±SEM) intake of AMI (mg/kg per day), and mean body weight in female and male C57BL/6J mice following subchronic and chronic treatment with AMI or saccharin. Males had higher serum levels of AMI/NOR (P<0.05), but females consumed more AMI per body weight than males (AMI/NOR serum levels: n=9–10/group; AMI intake: n=4–5/group). Females drinking AMI weighed less than female saccharin drinkers (P<0.01), but male AMI drinkers did not differ from male controls (n=12–17/group). Separate groups of mice were used to access serum levels, intake and body weight
Full size table

Body weights of mice treated subchronically and chronically with AMI or saccharin were assessed for mice following LH testing (Table 1). Separate ANOVAs on body weights of females and males were carried out with duration (subchronic, chronic) and drug (saccharin, AMI) as the between-group measures. There was a small, but significant difference in weight between AMI and saccharin treated females [F(1,53)=13.8, P<0.01], but male AMI treated mice were not different from saccharin controls. Although females AMI treated mice weight slightly less than controls, the general health of both males and females was good and animals did not appear sick from the oral AMI treatment.

Discussion

The behavioral performance of mice was evaluated in the TST, FST, and LH models of depression following treatment with the tricyclic antidepressant AMI. The main purpose of the study was to identify sex differences in antidepressant responses and to develop a standard treatment, sensitive across multiple models, which could be used to assess the mechanism of action of antidepressants in knockout and transgenic mice. The B6 mouse strain was tested because it is one of the most common background strains used for the development of genetically altered mice. The present study is the first to directly compare whether antidepressant efficacy varies according to duration of treatment and sex across three animal models of depression in B6 mice.

Although it is well established that antidepressants are effective only following chronic treatment in humans (Wong and Licinio 2001), this time course has been difficult to reproduce in rodent models of depression in which antidepressant-like responses are often seen following acute injection (Porsolt et al. 1977; Steru et al. 1985; Shanks and Anisman 1988). In the TST, both subchronic and chronic treatment with AMI reduced immobility in mice. In the FST, however, mice treated subchronically with AMI showed less of a reduction in immobility than mice treated chronically, suggesting that chronic AMI treatment resulted in a greater antidepressant-like response than subchronic exposure. Data from the LH study also suggest that chronic oral AMI may be more effective than subchronic treatment in male mice. Although females that received subchronic and chronic AMI showed an equivalent antidepressant-like effect, males that received chronic AMI showed a greater reduction in escape latencies than males that received subchronic treatment. Neither chronic nor subchronic AMI treatment altered escape latencies in control mice that did not receive inescapable shock training, suggesting that AMI does not affect general escape learning. In addition, the antidepressant-like effects of AMI observed in the three depression models are not due to general increases in locomotor activity because locomotion was not affected by either subchronic or chronic AMI treatment. These data suggest that both the FST and LH paradigm (in males) show face validity for evaluation of long-term antidepressant treatment in mice, but that the TST may not be as sensitive to the alterations in second messenger systems that may be critical for antidepressant action in humans (Vaidya and Duman 2001).

Administration of antidepressants in the drinking water in mice has several advantages over injection. First, the time course of action of AMI in the FST and LH paradigm (in males), when administered in the drinking water, closely models the delayed onset of action seen in humans (Wong and Licinio 2001). In the FST, many studies have reported that a single injection of an antidepressant can reduce immobility in mice (Porsolt et al. 1977; Bai et al. 2001; Lucki et al. 2001). Similarly, an acute injection of AMI was found to attenuate the escape deficit in LH in B6 mice, but chronic injections of AMI had no effect on escape latencies (Shanks and Anisman 1989). With oral administration, however, the present study demonstrated that chronic AMI was more effective than subchronic AMI in the LH paradigm in males and in FST in both sexes. A second advantage of oral administration is that the stress associated with handling from chronic injections is reduced. Although administration of antidepressants has been reported in many studies to prevent or reverse LH in rats (Leshner et al. 1979; Petty and Sherman 1979; Sherman et al. 1979, 1982; Telner and Singhal 1981), the same effects have been difficult to reproduce in mice. It is possible that the repeated stress of injections in mice is a stressor that cannot be overcome by the antidepressant treatment. Support for this idea comes from a study showing that repeated saline injections in rats produced a "depressive-like" state, including a prolonged immobility in the TST (Izumi et al. 1997). Finally, because antidepressants are administered orally in humans, oral administration of antidepressants in the drinking water in mice more closely models the metabolic pathway of antidepressants in humans.

In humans, it has been reported that men are more responsive to TCAs and women show a more favorable response to SSRIs (Kornstein et al. 2000; Yonkers and Brawman-Mintzer 2002). In the mouse, sex differences in responses to TCAs versus SSRIs appear to depend on the behavioral test being studied. In the mouse FST, both the TCA imipramine and the SSRI paroxetine reduced immobility at lower doses in females than males (David et al. 2001). In contrast, in the TST, both imipramine and paroxetine reduced immobility at lower does in males than females (David et al. 2001). We found no sex differences in responsiveness to oral AMI in the TST, but in the FST and LH paradigm, females tended to show a greater response to AMI treatment than males. This difference may be due to altered sensitivity to the drug, or variations in metabolism or consumption. Intake of AMI showed that females consumed a higher dose of AMI per body weight than males, but had lower serum levels of AMI/nortriptyline. It is likely that this discrepancy between intake and blood levels is due to higher levels of metabolism in female mice (Shapiro et al. 1995). These data suggest that the greater responsiveness of TCAs in men compared to women is not consistent across mouse models and may depend on factors such as the behavioral test, dose and mode of administration.

The effects of antidepressants in the FST and TST have already been examined in knockout mice lacking cAMP response element binding protein (CREB) (Conti et al. 2002), α2A-adrenergic receptors (Schramm et al. 2001), dopamine-β-hydroxylase (Cryan et al. 2001), 5HT1A and 5HT1B receptors and the 5HT transporter (Mayorga et al. 2001; Holmes et al. 2002), and DARPP-32 (Svenningsson et al. 2002). In these studies, however, the antidepressant effects were assessed after acute or subchronic administration. Because chronic antidepressant treatment results in changes in receptor densities, electrophysiological properties, and levels of neurotransmitters in serotonergic and noradrenergic neurons (Blier and de Montigny 1994), as well as changes in second messenger systems that are not seen after acute or subchronic treatment (Vaidya and Duman 2001), it is important to assess effects of these genes after chronic drug treatment. Furthermore, drug metabolism may differ between acute and chronic treatment and antidepressants such as amitriptyline can have active metabolites (nortriptyline) that increase in response to repeated administration (Coudore et al. 1996).

The results of the present study demonstrate that continuous administration of AMI in the drinking water results in antidepressant-like responses in B6 mice across three models of depression. The delayed onset of antidepressant action in the LH paradigm in males and the FST in both sexes may model the delayed onset of action of antidepressants observed in humans more closely than the immediate antidepressant action seen with acute injections in rodents. These results demonstrate that the B6 strain is appropriate for use in studying antidepressant responses in genetically altered mice and suggest that chronic oral treatment may be a preferred mode of administration over injections.