Long‐term stability and characteristics of behavioral, biochemical, and molecular markers of three different rodent models for depression

Abstract Objective The present study was designed to explore the long‐term differences between three mouse models for depression. Method In the present study, the unpredictable chronic mild stress (UCMS) model, the glucocorticoid/corticosterone model, and the olfactory bulbectomy model were compared at two, three, and five weeks after model induction. Behavioral testing performed included forced‐swimming, tail suspension, open‐field and elevated plus‐maze tests. In addition, 5‐hydroxytryptamine (5‐HT) and dopamine levels, and mRNA and protein expressions related to 5‐HT synthesis, transport, and signaling were analyzed in the hippocampus of tested animals. Results Our results revealed that each model demonstrated a specific profile of markers, whereas the stability of them differed over testing time. Conclusions Each model provided a unique set of advantages that can be considered depending on the context and aims of each study. Among the three models, the UCMS model was mostly stable and appeared to the best model for testing long‐term depression‐like state.


| INTRODUC TI ON
Depression is a common mental illness that affects more than 300 million people all over the world (Organization, 2018). In the last decades, depression has become a leading cause of disability worldwide and a major contributor to the global disease burden (Organization, 2018). Despite the high prevalence of the disease, its etiology remains unclear; the current treatments are only moderately effective. One reason for the lack of effective treatment is the missing understanding of the underlying neurobiological mechanisms of depression, but also results from the heterogeneity and comorbidity of depression (Krishnan & Nestler, 2008). To improve our understanding, animal models of depression can provide insights into the pathogenic mechanisms of the disease and allow researchers to better tease apart the details of the inner workings of the brain (Nestler & Hyman, 2010).
Based on the limited data regarding the etiology of depression, the available animal models predominantly rely on chronic exposure to stressful experiences such as the unpredictable chronic mild stress (UCMS) model and the high-dose administration glucocorticoid/corticosterone (CORT) model. In addition, the rodent olfactory bulbectomy (OB) models behavioral and neurochemical changes resembling some of the symptoms observed in depressed patients (Song & Leonard, 2005). Each model has its own set of advantages and disadvantages. In all cases, the depression-related behavior and neurochemical changes continue throughout the length of the experiment (Mucignat-Caretta, Bondi, & Caretta, 2006).
The UCMS model includes chronic depression by presenting unpredictable stress stimuli (Gupta et al., 2014) such as multiple, randomly scheduled food and water deprivation, overnight illumination, cage tilting, and other similar stressors that may be associated with human patients (Boyle et al., 2005). In most published studies, the UCMS model induced depression for 2-4 weeks, but some have been tested for up to eight weeks (He et al., 2016;Wu et al., 2017;Xia et al., 2016;Yan et al., 2017;N. Zhang et al., 2018). High levels of serum cortisol, related to elevated hypothalamo-pituitary-adrenal (HPA) axis activity, can be associated with depression. The CORT model tests the pharmacologically evoked anxio-depressive state in animals by application of corticosterone, the rodent functional analogue of cortisol (Krugers, Lucassen, Karst, & Joels, 2010;Rosa et al., 2014) where corticosterone are subcutaneously applied on daily basis for two, three, five, or eight weeks (Gupta, Radhakrishnan, & Kurhe, 2015;Kv et al., 2018;Schloesser et al., 2015;Zhang et al., 2016). The OB models chronic agitated hypo-serotonergic depression in mice (Lumia, Teicher, Salchli, Ayers, & Possidente, 1992).
Although animals in these models show several phenotypes of depression-related behavior, including prolonged immobility time in a forced-swimming test (FST) or tail suspension test (TST), as well as altered neurotransmitter activity and reduced neurotransmitter-related gene expression, it has not been confirmed whether these characteristics persist for long periods or whether the models show varied qualitative and quantitative similarity to human depression.
Comprehensive comparative data are still lacking. In the present study, we analyzed and compared the above three animal models over two, three, and five weeks to investigate the systemic impact on functions ranging from behavior to molecular expression over time. The aim of the study was to gain a better understanding of the pathophysiological processes involved in these animal models. Our results will help researchers choose the most suitable model and determine which time parameters are most useful in each model when evaluating the novel treatment for depression.

| Animals
In total, 180 male C57BL/6 mice aged five weeks were provided by Animals.). They were housed under a 12-hr dark/ light cycle (light period from 07:00 to 19:00) at room temperature (25 ± 1°C) and provided food and water ad libitum. Unless mentioned otherwise, all mice were randomly housed in groups of four per cage upon arrival for one week to allow habituation until experimental use.

| Reagents
Isoflurane was purchased from HeBeiJiuPai Company.

| Establishment of depression models
For each model, mice were weighed weekly. Behavioral testing began two, three, or five weeks after induction of the model as described below.

| UCMS model
The modeling method was designed according to literature (Forbes, Stewart, Matthews, & Reid, 1996) with minor modifications. Sixty mice were randomly divided into control (housed four mice per cage) and UCMS group (housed individually). The mice of UCMS group were exposed to a variety of unpredictable mild stressors including the following: body restraint (3 hr), inverted light/dark cycle (48 hr), food deprivation (24 hr), water deprivation (12 hr), electric shock (0.5 A, shock/min, 5 min), damp sawdust (12 hr), tail pinch (1 min), empty cage (no litter in the cage, 24 hr), cage tilting of 45° (24 hr), and soiled cage (24 hr). These stress procedures were randomly scheduled during the experimental period with at least one stressor every day. The mice of control group were housed under normal conditions without specific stimulus.

| CORT model
The model was established based on previous glucocorticoid-induced depression (Kv et al., 2018;Mesripour, Alhimma, & Hajhashemi, 2019) and had been slightly modified. Sixty mice were randomly divided into control and CORT groups. The mice of CORT group mice were intraperitoneally (i.p.) administered 40 mg/kg methylprednisolone consecutively for two, three, and five weeks. The mice of control group were given i.p. saline injections.

| OB model
Sixty mice were randomly divided into control and OB groups. The mice in the OB group underwent olfactory bulbectomy while the mice in the control group underwent sham surgery. For bulbectomy, the bilateral olfactory bulbs were removed as described previously (Almeida et al., 2017) with minor modifications. The mice were anaesthetized with isoflurane mixed with oxygen, and an incision made in the overlying skin to expose the skull. Then, a burr hole was drilled on each side, 2 mm from the midline of the frontal bone overlying the olfactory bulbs (2 mm in diameter; 4 mm anterior to bregma). The olfactory bulbs were aspirated using a blunt hypodermic needle and a vacuum pump. Special care was taken to avoid damaging the frontal cortex. Sham-operated controls underwent all of the same surgical procedures, but the olfactory bulbs were left intact. After the surgery, all incisions were closed with 4-0 vicryl suture. The animals were allowed to recover for two weeks.

| Behavioral testing
Unless mentioned otherwise, all the tests were performed between 09:00 and 12:00. The experimental schedule is shown in Figure 1.

| Forced-swimming test
Mice were individually placed in a glass cylinder (height 30 cm, diameter 20 cm) filled with 20°C water to a height of 20 cm. The animals were allowed to swim for 6 min, and the duration of immobility was quantified from video recordings of the last 4 min. Immobility was defined as floating passively with no active movements.

| Tail suspension test
Mice were individually suspended 50 cm above the floor by taping the tail to a horizontal bar for 6 min. The duration of immobility was quantified from video recordings of the last 4 min. Immobility was defined as no active movements except normal respiration.

| Open-field test (OFT)
Mice were individually placed in the center of an open field (50 × 50 × 50 cm). Time spent in the central area and total distance traveled by mice in 5 min were recorded and analyzed by a videotracking system (Mobiledatum Inc.). The OFT was used to evaluate anxiety-related behavior.

| LC-MS/MS analysis
Twelve hours after the final behavioral tests, the mice were anesthetized with an overdose of 1% pentobarbital sodium. Hippocampi were dissected on ice, and one half of each was processed for LC-MS/MS analysis. Briefly, the hippocampus was homogenized in a disposable glass tube after adding of 400 μl ice-cold methanol with 0.1% formic acid and 10 μl internal standard. The homogenate was vortexed for 1 min and then centrifuged at 18,000 × g for 10 min at 4°C. The supernatant was transferred and evaporated to dryness under a nitrogen stream. The resultant dry residue was reconstituted in 100 μl of initial mobile phase (0.1% formic acid in water/acetonitrile, 98:2, v/v), and a 10 μl aliquot was injected into the LC-MS system for analysis under conditions described previously (Huang et al., 2014).

| Real-time PCR analysis
Total RNA was extracted from the second half of the dissected hippocampi of five mice per group using Trizol according to the manufacturer's instructions (Life Technologies). After eliminating trace amounts of DNA contamination with DNase I, the RNA extracts were reverse transcribed into cDNA with a Revert Aid First Strand cDNA Synthesis kit (Fermentas). The synthesized cDNAs were used as templates for real-time PCR as described previously (Cao et al., 2017). The primer sequences used in PCR are shown in

| Statistical analysis
The data of behavioral tests were processed by Animal behavior analysis system version 2.20 software (Mobiledatum, Shanghai, China). The data are expressed as mean ± standard error of the mean (SEM) and analyzed by two-way analyses of variance or Student's t test using GraphPad Prism 6 software. p values less than .05 were regarded as statistically significant.

| Body weight gain
As can be seen in Figure

| Depression-like behaviors
Hopeless behavior of three model mice was determined using FST and TST. The changes of immobility times in FST and TST were different across the three models. For the CORT and OB

| Anxiety-like behavior
The

| Locomotor activity
In OFT, the UCMS model mice had significantly longer total traveling distances than controls at week 3 and week 5 (Figure 5a

| Hippocampal neurotransmitter level
As shown in Table 2

| Hippocampal mRNA expression
As shown in Figure

| Hippocampal protein expression
As shown in Figure 7a,d, the UCMS model mice had increased 5-HTR2C protein expression (p < .05) and reduced 5-HTR1A, BDNF and TrkB expression (p < .1, p < .01 and p < .05, respectively) compared to their controls. The CORT model mice had significant lower 5-HTR1A and 5-HTR2C expression (Figure 7b,e, p < .05 and p < .001), and increased the level of BDNF (p < .01) than the controls. The OB model mice showed no differences in 5-HTR1A and 5-HTR2C expressions, but displayed significantly higher BDNF and TrkB expression than controls (Figure 7cF, p < .001 and p < .05). ior at the second week, but anxiolytic-like behavior accompanied by hyperactivity at the fifth week (Zueger et al., 2005). The present data indicated that different behavioral changes emerged at various times after modeling. Notably, there were clear differences between the OB model group and the UCMS/CORT model groups.

| D ISCUSS I ON
Body weight disorder can indicate depression (Blashill & Wilhelm, 2014). In our study, the mice of three models showed decreased body weight gain over time relative to controls; similar as previous literature (Lucca et al., 2008) suggested that decreased body weight gain represents a symptom of depression. The immobility time in the FST and TST are frequently used to represent "behavioral despair" as a marker of depression. The FST is the best-validated method for assessing depressive symptoms and predicting antidepressant medications' efficacy (Nestler et al., 2002). We found longer immobility times in the UCMS and CORT models, but not in OB model. It has been demonstrated that 85% of patients with depression also experience significant symptoms of anxiety, whereas comorbid depression occurs in up to 90% of patients with anxiety disorders F I G U R E 6 Hippocampal mRNA expressions of 5-hydroxytryptamine receptor 1A (5-HTR1A), 5-hydroxytryptamine receptor 2C (5-HTR2C), serotonin transporter (SERT) and tryptophan 5-hydroxylase 2 (TPH2) in the UCMS, CORT and OB model mice. All tissues are from the mice at 5 weeks. Data are expressed as mean ± SEM. N = 5 for each group. *, p < .05; **, p < .01 versus Control group (Gorman, 1996). The OFT and the EMPT are well validated for use in investigating anxiety-related activity (Belzung & Griebel, 2001;Ramamoorthy, Radhakrishnan, & Borah, 2008;Walf & Frye, 2007).
The UCMS model mice showed anxiety-like behavior at week 5. The total distance moved in both the UCMS and OB models was also higher at week 5, but not in the CORT model. Based on behavioral performance and neurotransmitter changes, we chose to study hippocampal gene and protein expression at week 5. We found reduced SERT mRNA expression in all three models, and reduced TPH2 expression only in the UCMS model mice. SERT mediates the clearance and reuptake of 5-HT during synaptic transmission (S. Ramamoorthy et al., 1993), its reduction in the three models, as showed in the current study, may be related to the impaired release of 5-HT in these models. As the synthesis of neuronal 5-HT in the CNS is regulated by TPH2, the decrease of TPH2 may contribute to the decrease of 5-HT in the UCMS model mice. Among all the 5-HT receptor subtypes, the 5-HTR1A and 5-HTR2C seem to play key roles in depressive neuropathology (Drevets et al., 2000;Hirvonen et al., 2008;Iwamoto, Kakiuchi, Bundo, Ikeda, & Kato, 2004;Martin, Hamon, Lanfumey, & Mongeau, 2014). It has been reviewed that the stimulation of postsynaptic 5-HTR1A might result in antidepressant-like effect; but for 5-HTR2C subtype the data are ambiguous, as both agonists and antagonists induce antidepressant-like activity (Zmudzka, Salaciak, Sapa, & Pytka, 2018 (Gourley, Kiraly, Howell, Olausson, & Taylor, 2008;Lopes et al., 2018). Though the common hypothesis of depression predicts decreased BDNF expression in depression-related brain areas, our results showed increased BDNF protein in CORT and OB mice at week 5. In addition, some reports indicate that OB modeling in mice leads to increased BDNF levels (Hellweg, Zueger, Fink, Hortnagl, & Gass, 2007). For the CORT model, the decreased BDNF level in the mouse hippocampus could be attributed to the time course of changes in BDNF after chronic stress as reported (Hashikawa et al., 2015).
Although development of convincing, useful animal models for depression represents a major challenge, these three models appear quite useful to further understand disease pathophysiology and for developing treatments based on new molecular targets (Nestler & Hyman, 2010). The UCMS model has high validity but can produce different profiles of behavioral and neuropharmacological effects based on experimental design and end-point selection (Willner, 2005). In our experiment, the UCMS model was established in single-caged fed mice stimulated with chronic unpredictable mild stress. According to our experimental results, the UCMS model was more suitable for modeling depression with anxiety and was stable for more than three weeks. The CORT model was more suitable for the study of depression caused by 5-HT disorders, but it was not stable and might accumulate difficulties after long study duration. The OB model was more potentially better characterized by manic-depressive symptoms, but it was very unstable at five weeks, which may be related to the unusual plasticity of olfactory system (Mucignat-Caretta et al., 2006). The OB model's behavioral characterization may suitable, in particular, for testing the presence of depression-like behaviors in transgenic mice.
To the best of our knowledge, this is the first study comparing the UCMS, the CORT, and the OB models in terms of stability over longer experimental times. Our study emphasized that the awareness of the advantage and disadvantage of the three depression models is essential for ensuring that one's desirable experimental aim and durations can be addressed appropriately.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data available on request from the authors.