OUP user menu

We are in the dark here: induction of depression- and anxiety-like behaviours in the diurnal fat sand rat, by short daylight or melatonin injections

Tal Ashkenazy, Haim Einat, Noga Kronfeld-Schor
DOI: http://dx.doi.org/10.1017/S1461145708009115 83-93 First published online: 1 February 2009


Circadian rhythms are considered an important factor in the aetiology, expression and treatment of major affective disorders, including seasonal affective disorder (SAD). However, data on the effects of daylight length manipulation or melatonin administration are complex. It has been suggested that since diurnal and nocturnal mammals differ significantly in their physiological and behavioural responses to daylight, diurnal rodents offer a preferable model of disorders related to circadian rhythms in the diurnal human. We previously found that diurnal fat sand rats maintained under short daylight (SD), show depression-like behaviour in the forced swim test (FST). The present study was designed to test additional behaviours related to affective disorders and study the involvement of melatonin in these behaviours. Sand rats were divided into short-daylight (SD, 5 h light:19 h dark) and long-daylight (LD, 12 h light:12 h dark) groups, and received 100 µg melatonin or vehicle administration for 3 wk (5 h and 8.5 h after light onset in the LD room). Animals were then tested for reward-seeking behaviour (saccharin consumption), anxiety (elevated plus-maze), aggression (resident-intruder test), and depression-like behaviour (FST). SD or melatonin administration resulted in a depressed/anxious-like behavioural phenotype including reduced reward seeking, increased anxiety, decreased aggression and decreased activity in the FST, supporting the notion that in a diurnal animal, reduced light results in a variety of behavioural changes that may model depression and anxiety; and that melatonin may be a significant factor in these changes. We suggest that the sand rat may offer an excellent model species to explore the interactions between daylight, affective behaviour and the related underlying mechanisms.

Key words
  • Depression
  • diurnal
  • fat sand rat
  • melatonin
  • photoperiod


Ample evidence has recently been gathered that demonstrates a possible relationship between circadian rhythms and major affective disorders, including major depressive disorder (MDD), bipolar disorder (BPD) and seasonal affective disorder (SAD) (for a recent review see McClung, 2007). Melatonin, the fundamental hormone related to circadian rhythms, is secreted during the night, and the duration of the dark period influences the duration of melatonin secretion, which transduces the photoperiodic information into a physiological signal (Arendt, 2000). Hence, under short photoperiod melatonin secretion duration is long and vice versa. However, the effects of melatonin on mood, whether naturally secreted or exogenously administered, are confusing, and studies in animal models as well as in humans seem to be complex.

In human patients afflicted with affective disorders, treatments such as sleep deprivation or bright-light therapy (which result in decreased melatonin levels) were demonstrated to be effective for some diseases (Kripke, 1998; Terman and Terman, 2005), and it has been suggested that part of the therapeutic effect of antidepressants and mood stabilizers stems from their effects on the circadian clock (Goodwin et al., 1982; Nestler et al., 2002). A new antidepressant (agomelatine) has been shown to have a partial effect on the melatonin receptor (but also on serotonin receptor; Pjrek et al., 2007; Rouillon, 2006). However, melatonin administration was not demonstrated to have any antidepressant effect on treatment-resistant depression (Dalton et al., 2000), or on SAD (Wirz-Justice et al., 1990), and had only some effect on sleep quality but not on depressive symptoms in patients when used alone (Leppamaki et al., 2003), or in combination with the antidepressant fluoxetine (Dolberg et al., 1998), or with electroconvulsive therapy (Grunhaus et al., 2001).

In animal models, the effects of altering daylight length or of melatonin administration appear to be even more complex and include many apparent contradictions. For example, long daylight (LD) conditions appear to induce an anxiety-like response in meadow voles (Microtus pennsylvanicus; Ossenkopp et al., 2005), but were reported to be anxiolytic in Siberian hamsters (Phodopus sungorus; Prendergast and Nelson, 2005), and resulted in increased anxiety-like behaviour in mice that was reversed by a melatonin antagonist (Kopp et al., 1999). Continuous light exposure in rats resulted in dopaminergic supersensitivity (Abilio et al., 1999), while manipulations of the CLOCK gene in mice resulted in dopaminergic supersensitivity as well as an increased response to cocaine reward that might be relevant to manic patients (McClung et al., 2005). Melatonin administration was reported to reduce cocaine-withdrawal-related anxiety (Zhdanova and Giorgetti, 2002), but also to slightly enhance cocaine-induced behavioural sensitization in rats (Sircar, 2000). Chronic melatonin in rats was shown to have mild antidepressant-like effects in the chronic mild stress model when injected in the evening but not in the morning (Papp et al., 2003), but such antidepressant-like activity is not universal, as in the forced swim test (FST) model some studies report an antidepressant-like action (Micale et al., 2006), whereas others found no such effects (Bourin et al., 2004).

The reasons for these inconsistencies may include the use of different species (seasonal or not, laboratory strains vs. wild species and more), different laboratories, schedules of drug administration or light exposure and different testing protocols. Another important issue is that of the natural activity pattern of the studied species. Most, if not all, related animal research was performed on nocturnal animals while modelling disorders in a diurnal mammal (humans).

Whereas it is conceivable that much of the information derived from research in nocturnal animals is significant for the understanding of mechanisms, some of the behavioural effects of altering daylight cycles or of melatonin administration in nocturnal animals may differ greatly from effects in diurnal species (including humans), and it is conceivable that using models in diurnal animals may give us a much better understanding of the effects of circadian rhythms on mood in humans, and also possibly better ways to develop improved treatments in relation to affective disorders.

In line with this approach, we recently demonstrated that chronic (3 wk) short daylight (SD) exposure (5 h/19 h light/dark cycle) in the diurnal fat sand rat (Psammomys obesus) results in a depression-like behavioural change (Einat et al., 2006). We now suggest that since diurnal mammals are active when melatonin levels are low (day time), and may develop SAD under short photoperiod when melatonin levels are high, the physiological and behavioural effect of short photoperiod and melatonin may differ between nocturnal and diurnal mammals. The present project was therefore designed with two goals: (1) to test a broader range of behavioural changes related to affective disorders after SD exposure; and (2) to test the significance of melatonin in the behavioural effects of SD exposure.

Materials and methods


For the sweet-solution consumption experiment (see below), two groups (n=8 each) of male adult (aged ∼1 yr) sand rats, born in the breeding colony of the I. Meier Segals Gardens for Zoological Research at Tel-Aviv University, were individually housed in standard plastic cages (21×31×13 cm) and in separate temperature-controlled rooms (25°C). The test group was housed under a 5 h light/19 h dark cycle and the control group under a 12 h light/dark cycle. This lighting regime was designed according to these rodent's winter/summer activity in nature, where during winter they are active outside their burrows for about 5 h daily around mid-day, while during summer their activity starts in the early hours of the morning and ends in the late afternoon (about 12 h from activity onset to offset), with a pause during the mid-day hours of extreme desert heat (Ilan and Yom-Tov, 1990; Khokhlova et al., 2005), and on our previous results (Einat et al., 2006), showing that chronic (3 wk) exposure to such photoperiod results in a depression-like behavioural change. In order to avoid the development of diabetes, the sand rats were provided with special low-energy pellets ad libitum (product 1078, Koplock Ltd, Israel, containing 14% protein, 1.7% fat, 15% fibre, 11.2% ash, 1.02% calcium, 0.64% phosphorus, 1.15% sodium chloride, vitamins and minerals), Atriplex branches and water, and remained under these conditions for 3 wk prior to onset of experiments, which is sufficient for physiological acclimation (Kronfeld-Schor et al., 2000), and synchronization of circadian rhythms (Cohen and Kronfeld-Schor, 2006; Levy et al., 2007).

For all other experiments, adult male (n=40, aged 6 months) fat sand rats (Psammommys obesus) provided by Harlan Laboratories (Jerusalem, Israel) served as subjects. As described above, animals were divided into two groups (n=20/group) and were individually housed in separate rooms with different light/dark cycles. Three animals from the SD room lost significant weight during the first 10 d of the acclimatization period and were therefore eliminated from the study. As described below, the animals were then further subdivided into melatonin/non-melatonin groups.

All experimental procedures were approved by the Tel-Aviv University IACUC (protocol no. L-07-077) and were conducted in accordance with NIH guidelines for the treatment and care of laboratory animals.


After the 3-wk acclimation to either SD (n=17) or LD (n=20), each group was further divided into two subgroups that received i.p. injections of either 100 µg melatonin (Sigma, St Louis, MO, USA; diluted in 0.1 ml 10% ethanolic saline) or vehicle solution. Injections were administered twice daily, 5 h and 8.5 h after the onset of light in the LD room. These administration hours were chosen to mimic the effects of the SD cycle in the animals housed under LD conditions, as the half-life of exogenous melatonin is ∼3.5 h (Hiebert et al., 2006). The experiment now constituted four groups with daylight hours as one factor (short or long) and melatonin administration as a second factor (melatonin or vehicle). Final numbers for each group were: long daylight/vehicle (LV; n=10); long daylight/melatonin (LM; n=10); short daylight/vehicle (SV; n=8); short daylight/melatonin (SM; n=9). Following 3 wk of injections all animals underwent behavioural tests as detailed below while continuing to receive the injections until the last day of the tests.

Order of experiments

To minimize the effects of behavioural experience on results, experiments were conducted from the less to the more intrusive. The order of experiments for the second group of sand rats was, accordingly: (1) elevated plus maze; (2) resident-intruder; (3) FST.

Body weight

For the first group of animals, body weight was noted at the initiation of the experiment and was then used only to normalize saccharin consumption as detailed below. Body weight was continuously monitored with the second group of animals, which were weighed 1 wk after arrival, at the end of the 3-wk light entrainment period, and once a week thereafter for a period of 7 wk.

Behavioural testing

Sweet-solution consumption

To assess the sand rats' reward-seeking (hedonistic) responses, we used the sweet-solution (saccharin) consumption test (Mazarati et al., 2007; Qi et al., 2006). For 4 d, animals were provided continuously with a bottle containing 1% saccharin solution in addition to their regular food and water. The saccharin solution bottles were weighed daily 5 h after light onset, and daily consumption (ml) was calculated and normalized for body weight.

Elevated plus-maze test

The plus-maze was constructed from black painted aluminum, and consisted of two open arms (50×10 cm) and two enclosed arms (50×10×15 cm), elevated 50 cm above the ground and illuminated (200 lx at the maze floor level). Test and control sand rats were individually placed at the centre of the plus-maze and were allowed to explore freely for 5 min. Behaviour was videotaped using a video camera (Panasonic M9000) located above the centre of the maze. A blind observer then scored the videotapes for time spent in each section of the maze and the number of entries. To be considered in the open or closed arms of the maze, the animal had to be in it with all four legs. The test was held during the light phase of both groups.

Resident-intruder test

Aggressive behaviour can be the consequence of an affective-like change in animals and in rodents can be evaluated using the resident-intruder test (Einat, 2007; Malkesman et al., 2006).

Eight naive male sand rats from the colony of the I. Meier Segals Gardens for Zoological Research were used as intruders. The males were aged 1 yr and weighted ∼250 g (similar to the tested individuals). For technical reasons, each intruder was used four times, but care was taken to counterbalance the order of use between experimental groups. No effects were found on residents' behaviour that correlated with the order of exposure of intruder animals. All trials were performed during the light phase of both groups.

For each test, an intruder was introduced into the home cage of an experimental sand rat for a 5-min session and the behaviour was videotaped for later analysis. The experimenter was present throughout the test session and if an attack became very vicious and included significant biting, the experimenter used a plastic probe to separate the animals for a few seconds. The probe was then withdrawn and the experiment continued. In this design there is no possibility of scoring total time of attacks and, accordingly, videotapes were used to score the number of attacks within the session and the latency to first attack. This paradigm had already been demonstrated to be valid in demonstrating changes in mood (Einat, 2007).


The FST is a commonly used test for depression-like behaviour in both rats and mice, with some methodological difference in application across species (Porsolt et al., 1978). For the present study, animals underwent the swim test over two consecutive days, which comprised a training day and a test day, similar to the standard procedures for rats (Gould et al., 2004). Since fat sand rats are diurnal, in both groups testing started at 08:00 hours (1 h after lights-on) and terminated within 3–4 h, well within the light/activity phase of the two groups, and was therefore preformed in the light (200 lx). The test was performed in a white, opaque plastic cylinder, 30 cm in diameter and 45 cm high. The cylinder was filled with water (21–23°C) to a depth of 25 cm, preventing the sand rats from touching the floor or escaping. A video camera (Panasonic M9000) was mounted above the cylinder and videotapes were later used to score behaviours. Animals were tested in alternating order of the four groups, with each animal individually placed in the water by the experimenter and left there to swim. Sand rats are not as good swimmers as rats and mice and, moreover, they cannot float and therefore, when they reach the state when they do not swim or struggle, they sink into the water. Because of their different behaviour in the water, the testing procedure was modified (Einat et al., 2006). Specifically, instead of monitoring swimming/struggle vs. immobility (floating) as the procedure is with rats and mice, we monitored ‘time to sink’. The sand rats usually swim and struggle, stop, sink, start swimming and struggling again, stop and sink and so on. Eventually, they sink for more than a few seconds. To eliminate the possibility of drowning, when an animal became entirely immersed in water (above the tip of the snout) for 5 s the experimenter rescued it and the session terminated. At that time the experimenter dried the animal with a cloth, and placed it back in the home cage. A time limit of 6 min for the first and second dives were set, however, none of the animals tested reached that limit. Accordingly, the tapes were used to score the first and second times that an animal was completely immersed in water (sink 1 and sink 2) as well as the time to reach rescue criteria (when the tip of the sand rat's snout was under the water for 5 s; Einat et al., 2006).

Statistical analysis

Data from all experiments was analysed using an analysis of variance (ANOVA) with Statistica 7.0 software (Statsoft, Tulsa, OK, USA). Data for the saccharin consumption test; weights and FST were analysed with repeated-measures ANOVA with photoperiod length (saccharin) and photoperiod length and melatonin (weight and FST) as main factors and day (saccharin) week (weight) or behaviour (FST) as a repeated-measure factor. Since these ANOVAs indicated main effects and interaction, effects were analysed separately using Student's t test (saccharin) or ANOVA (weight and FST) for each session.

All other experiments were analysed using a factorial analysis with photoperiod and melatonin as main factors. When results of the ANOVA indicated significant interactions, it was followed by post-hoc Scheffé tests to explore specific effects of groups. Level of statistical significance was set at p<0.05.


Body weights

Animals exposed to a SD cycle did not gain weight at the same rate as those under LD conditions. As shown in Figure 1, the initial weights of animals from the different groups were not different; but whereas sand rats in the LD groups showed a gradual increase in weight, the animals from the SD group first demonstrated a reduction in weight and only later started gaining weight [repeated-measures ANOVA, light effect: F(1, 36)=8.02, p<0.001; day effect: F(6, 31)=6.94, p<0.001; day×light interaction: F(6, 31)=10.5, p<0.001]. Interestingly, the melatonin injections did not affect weight gain rate in the animals under any of the light conditions.

Figure 1

Body weight (weekly mean±s.e.) of fat sand rats acclimated to either short daylight or long daylight divided into two subgroups that received i.p. injections of either 100 µg melatonin or vehicle solution. X-axis represents weeks from the beginning of treatment: long daylight/vehicle (LV, n=10); long daylight/melatonin (LM, n=10); short daylight/vehicle (SV, n=8); short daylight/melatonin (SM, n=9). Repeated measures, * signifies difference between short daylight and long daylight groups (p<0.05).

Saccharin consumption

SD exposure significantly decreased saccharin consumption across the 4 d of testing, indicating reduced reward-seeking behaviour and anhedonia [Figure 2; repeated-measures ANOVA across days, group effect: F(1, 27)=30.15, p<0.001; day effect: F(3, 25)=3.14, p=0.03; group×day interaction: F(3, 25)=5.09, p=0.003. Following comparisons for separate days, day 1: t(27)=2.4, p=0.03; day 2: t(28)=8.0, p<0.001; day 3: t(28)=3.7, p=0.001; day 4: t(27)=4.0, p<0.001)].

Figure 2

Saccharin consumption/body weight ratio (mean±s.e.) across the 4 d of testing. * Indicates a significant difference between groups (p<0.05).

Elevated plus-maze

Sand rats maintained under SD or under LD and receiving melatonin spent significantly less time in the open arm sections [Figure 3; open/closed time ratio: ANOVA light effect: F(1, 52)=2.8, n.s.; melatonin effect: F(1, 52)=4.82, p=0.035; light×melatonin interaction: F(2, 51)=6.94, p=0.013; post-hoc Scheffé: LV different than SV and LM with a strong tendency for difference from SM (p=0.07)]. Similar, albeit non-significant, trends were shown for the ratio of number of entries into open/closed arms (data not shown). Moreover no differences were shown in the total number of entries into all arms, indicating that the behavioural effects are probably not related to changes in activity levels (data not shown).

Figure 3

Ratio of time spent in the open/closed arm sections of the elevated plus-maze (mean±s.e.) by fat sand rats acclimated to either short daylight or long daylight divided into two subgroups that received i.p. injections of either 100 µg melatonin or vehicle solution: long daylight/vehicle (LV, n=10); long daylight/melatonin (LM, n=10); short daylight/vehicle (SV, n=8); short daylight/melatonin (SM, n=9). * Indicates a significant difference from all other groups (p<0.05).

Resident-intruder test

Sand rats maintained under SD conditions showed a reduced tendency to attack an intruder in the resident-intruder test, with melatonin treatment having a similar effect to those of the SD conditions. These effects are demonstrated by the significant reduction in the number of attacks within the session [Figure 4; ANOVA light effect: F(1, 52)=10.37, p=0.003; melatonin effect: F(1, 52)=3.76, n.s.; light×melatonin interaction: F(2, 51)=11.72, p=0.002; post-hoc Scheffé: LV group different from all other groups] and a similar trend for the latency to first attack (data not shown).

Figure 4

Resident-intruder number of attacks (mean±s.e.) by fat sand rats acclimated to either short daylight or long daylight divided into two subgroups that received i.p. injections of either 100 µg melatonin or vehicle solution: long daylight/vehicle (LV, n=10); long daylight/melatonin (LM, n=10); short daylight/vehicle (SV, n=8); short daylight/melatonin (SM, n=9). * Indicates a significant difference from all other groups (p<0.05).


As previously demonstrated (Einat et al., 2006), animals from the two groups did not differ in their FST behaviour during the first day of testing but revealed significant differences on the second test day. Fat sand rats held under SD conditions acquired despair-like behaviours faster than those held under LD conditions. Moreover, melatonin administration during the light hours to the LD animals had a similar effect on behaviour to those of exposure to SD conditions, as demonstrated by the significant interaction between light and melatonin [Figure 5; repeated-measures ANOVA across behaviours (sink 1; sink 2; rescue) light effect: F(1, 52)=4.33, p=0.046; melatonin effect: F(1, 52)=3.18 (n.s.); light×melatonin interaction: F(1, 52)=7.34, p=0.011)]. Further analysis of the specific behaviours demonstrated that the animals maintained under the SD conditions, or those receiving melatonin treatment, had a lower latency to first sink [Figure 5; ANOVA light effect: F(1, 52)=17.9, p<0.001; melatonin effect: F(1, 52)=9.74, p=0.004; light×melatonin interaction: F(2, 51)=15.06, p<0.001; post-hoc Scheffé: group LV different from all other groups] and second sink [ANOVA light effect: F(1, 52)=6.41, p=0.017; melatonin effect: F(1, 52)=5.13, p=0.03; light×melatonin interaction: F(2, 51)=10.92, p=0.002; post-hoc Scheffé: group LV differs from all other groups] and a similar non-significant trend for reaching rescue criteria.

Figure 5

The time to reach first and second sinking and rescue criteria in the FST. Fat sand rats were acclimated to either short daylight or long daylight divided into two subgroups that received i.p. injections of either 100 µg melatonin or vehicle solution: long daylight/vehicle (LV, n=10); long daylight/melatonin (LM, n=10); short daylight/vehicle (SV, n=8}; short daylight/melatonin (SM, n=9). * Indicates a significant difference from all other groups (p<0.05).


It has previously been suggested that diurnal and nocturnal mammals may differ in their response to a short photoperiod, and that a diurnal animal maintained under SD conditions will develop depression-like behaviour (Einat et al., 2006). We now bring further support to this idea: the present data demonstrate that diurnal fat sand rats maintained under SD conditions display reduced reward-seeking behaviour, reduced aggression and increased anxiety-like behaviours. Moreover, the present data clearly show that these behavioural changes can also be induced by melatonin injections in a regimen that mimics a SD cycle, indicating that the underlying biological mechanism is related to melatonin effects.

The results show that SD or its mimicking by melatonin injection reduces active behaviours in the FST and reduce sweet-solution consumption, two behaviours that had been strongly validated in the context of depression (Cleary et al., 2008; Porsolt, 2000; Qi et al., 2006). It is therefore suggested that these manipulations result in depressed-like behaviours. Moreover, the same manipulations resulted in reduced time spent in the open arms of the elevated plus-maze a behaviour that is thought to represent increased anxiety (Bertoglio and Carobrez, 2002; Einat et al., 2005). The reduced aggression demonstrated in the resident-intruder test cannot be directly related to either depression or anxiety but changes in aggression levels had been shown to be related to depression and mania and to treatment with both antidepressant and mood-stabilizing drugs (D'Aquila et al., 1994; Einat, 2007; Holmes et al., 2002).

It is interesting to note that previous studies on the effect of day length on aggression in Siberian hamsters (Phodopus sungorus, Jasnow et al., 2000, 2002; Scotti et al., 2007) Syrian hamster (Mesocricetus auratus, Badura and Nunez, 1989; Caldwell and Albers, 2004; Fleming et al., 1988), beach mice (Peromyscus polionotus, Trainor et al., 2007a), deer mice (Peromyscus maniculatus, Trainor et al., 2007b), and laboratory mice (Trainor et al., 2008), all nocturnal rodents, found that acclimation to short photoperiod increases aggression in the resident-intruder test compared with long photoperiod. In Siberian hamsters the effect of photoperiod on aggression can be replicated by melatonin administration (Jansnow et al., 2002) and is independent of gonadal response (Demas et al., 2004; Scotti et al., 2007), and in laboratory mice melatonin injections increase aggression (Paterson and Vickers, 1981). Meadow voles (Microtus pennsylvanicus) in a long photoperiod are more anxious than voles in a short photoperiod (Ossenkopp et al., 2005). On the contrary, our results in the diurnal fat sand rat show that long photoperiod causes an increase in aggression, and a decrease in anxiety, and that melatonin injections blunt this reaction to the level observed in short photoperiod. Moreover, in (diurnal) humans, SAD increases under short photoperiod and in high latitudes (where day length is short), and melatonin has been shown to exert a sedative effect (Dollins et al., 1994), induce sleep (Arendt, 2000; Dollins et al., 1994), and impair mood and performance (Dollins et al., 1993, 1994; Lieberman et al., 1984). It has been suggested that the effect of photoperiod, and consequently the duration of melatonin secretion, may lead to changes in the hypothalamic–pituitary–adrenal (HPA) axis, which subsequently mediate changes in aggression (Jansnow et al, 2002), and other behaviours. If true, this effect should differ between the diurnal fat sand rat and the nocturnal rodents studied so far. More generally, we suggest that melatonin may exert differential physiological and behavioural effects in diurnal and nocturnal mammals, and therefore that research of the effects of day length and melatonin on behaviour in diurnal mammals may be more relevant to humans compared to nocturnal mammals. In order to test this hypothesis, further research of diurnal species is needed.

The significance of circadian rhythms in the aetiology of mood disorders has been emphasized for some time (Bunney and Bunney, 2000; Lenox et al., 2002; McClung, 2007; Nestler et al., 2002), but the results of both human and animal experimentation relating to circadian mechanisms and mood have been inconsistent, especially at the clinical or behavioural levels. When working with animals it is hard to draw far-reaching conclusions regarding the effects of daylight length or melatonin administration on depression-, manic- or anxiety-like behaviours and on the mechanisms that might be involved in any behavioural changes. This inability may stem from the ecological and behavioural background of the different species studied and the selection pressures faced in its natural habitat. For example, winter breeders are expected to respond differently to short day length then summer breeders. Moreover, related studies have been performed with nocturnal rather than diurnal animals and therefore it may not be clear how relevant the behavioural results are to humans: melatonin is secreted during the dark in both diurnal and nocturnal mammals, which means that nocturnal mammals are active when melatonin levels are high, while diurnal mammals are active when melatonin levels are low (reviewed by Challet, 2007). In nocturnal rodents a light pulse at night-time rapidly reduces activity levels, whereas light can trigger activity in diurnal mammals (reviewed by Redlin, 2001). Therefore it is reasonable to assume that melatonin will have a differential effect on nocturnal and on diurnal mammals.

The use of diurnal animals in the research of affective disorders has been quite limited, probably mostly due to technical limitations. With only a few truly non-hibernating diurnal species amongst the rodents, and hardly any diurnal rodents maintained or sold by suppliers, the access of laboratories to such animals is very limited. Nevertheless, some behavioural work was previously done in the diurnal sand rat (Khokhlova et al., 2005; Neuman et al., 2005), Nile grass rat (Novak et al., 2007; Refinetti et al., 2006), degu (Octodon degus; Smale et al., 2001), and various species of squirrels (Sharma et al., 2003), but none of these studies had a direct relevance to mood disorders.

The relationship between circadian rhythms, melatonin and affective disorders has been observed at the clinical level (for a recent review see Srinivasan et al., 2006), and received significant support at the biochemical and the molecular levels. Manipulating light exposure was demonstrated to have effects on various neurotransmitter systems related to mood (Abilio et al., 1999, 2003), including serotonin (Zawilska et al., 1997), and adrenaline (Zubidat et al., 2007). Antidepressants and mood stabilizers were demonstrated to have effects on systems related to circadian rhythms (Castro et al., 2005; Goodwin et al., 1982; Jawed et al., 2007; Nestler et al., 2002; Wirz-Justice et al., 1980), and modifications of genes related to circadian rhythms were shown to have effects on affective-like behaviours (Bunney and Bunney, 2000; Easton et al., 2003; McClung et al., 2005). The present results suggest that the utilization of a diurnal rodent as an additional animal model may offer an effective way to further explore the affective-like behavioural changes related to circadian rhythms and the underlying mechanisms of these changes.

Further establishment of the model is still needed at a number of levels. It is important to establish the predictive validity of the model for drug-screening purposes (Willner, 1984), and to test whether antidepressant, mood-stabilizer or light-exposure treatment can ameliorate the depressive- and anxiety-like behaviours that were demonstrated by the SD or the melatonin-treated groups. It is also important to explore whether the behavioural changes seen in the sand rat are unique to this species or are shared by other diurnal species. Last, but not least, it is important to examine the relationship between the behavioural changes and biochemical and molecular effects, including possible changes in those intracellular systems strongly implicated in affective disorders such as BDNF, the Erk pathway, GSK-3 PKC and AMPA receptors. Previous data suggests interactions between circadian systems and relevant intracellular systems (Cirelli et al., 2002; Payne et al., 2002; Tononi and Cirelli, 2006), and specific studies with the present model are underway.


This study was supported by the National Institute for Psychobiology in Israel, founded by the Charles E. Smith Family. The authors would like to thank Orly Barak for her technical assistance and the employees of the I. Meier Segals Gardens for Zoological Research at Tel-Aviv University for their constant and reliable assistance. We greatly appreciate the help and good advice of Professor David Eilam.

Statement of Interest



View Abstract