Introduction
Retinoic signaling is reportedly linked with the development of the central nervous system (CNS) and the pathogenesis of depression in adults [1][2][3][4]. Excessive consumption of vitamin A (hypervitaminosis A) has long been known to cause adverse psychiatric events [5]. A synthetic retinoid used to treat acne, 13-cis-retinoic acid (13-cis-RA; isotretinoin), has been linked to depression and suicide, since its approval in 1982 [6]. All-trans retinoic acid (ATRA), the endogenous active derivative of vitamin A, plays a role in cell growth, neural differentiation, and synaptic plasticity during development and operates exclusively by regulating gene transcription [7][8][9]. Recently, our group found that chronic ATRA administration could induce depression-like behavioral changes in adult rats [10][11][12].
The hippocampus is involved in mood disorders, such as anxiety and depression. Brain imaging and post-mortem studies provide evidence of changes in the cellular architecture and/or morphology within this brain region, including a reduction in hippocampal volume, and atrophy of hippocampal pyramidal neurons [13][14][15][16]. Recent studies have indicated that abnormal synaptic plasticity in specific areas of the CNS, particularly the hippocampus, may be a core factor in the pathophysiology of depression [17]. Studies in rodents have provided evidence in support of a reduced synapse number and decreased levels of synaptic signaling proteins in the hippocampus in a depression model [18][19]. Furthermore, antidepressants, such as the highly prescribed selective serotonin reuptake inhibitors, could enhance synaptic plasticity in the hippocampus, as demonstrated in electrophysiological studies [20][21]. All of the above studies confirm a link between altered synaptic plasticity in the hippocampus and major depression. Moreover, the hippocampus seems to be a main target of retinoids [20]. Our group found that ATRA-induced impairments in hippocampal neurogenesis correlate with depression-like behavior [11]. It has been reported that ATRA treatment enhances excitatory synaptic transmission in the hippocampus [21][22] and ATRA is mediated in synaptic plasticity via a synaptic protein synthesis-dependent mechanism [23]. These reports demonstrate the potential role of ATRA in synaptic plasticity of the hippocampus. To better understand the molecular mechanism of synaptic plasticity as influenced by ATRA, we used a whole-genome complementary DNA microarray to investigate changes of gene expression in the human neuroblastoma cell line BE2(c) cells after ATRA administration. Several genes are altered by ATRA (Table S1). We only chose these two synaptic-associated genes, discs large homolog 2 (DLG2) and synapse differentiation-inducing gene protein 1 (SynDIG1), among the genes altered by ATRA. DLG2 is thought to have vital roles in synaptic plasticity [24]. It has been reported that a reduction of DLG2 expression is found in the hippocampus in depression disorders [25], and DLG2 might be involved in depression disorders, according to results of a genome-wide association study [26]. SynDIG1 has been identified as an α-amino-3-hydroxy-5-methylisoxazole-4-propionic acid subtype glutamate receptor (AMPAR)-interacting transmembrane protein that could regulate excitatory synapse development via AMPAR content [27]. Moreover, SynDIG1 has been found to be involved in depressive symptoms in a genome-wide association study [28].
In this study, we aimed to investigate the effect of ATRA on the expression of two synaptic- associated genes, DLG2 and SynDIG1, in the hippocampus and their relationship with anxiety- or depression-like behavior in young mice.
ATRA-Induced Anxiety- and Depression-Like Behavior in Young Mice
We used the open-field test (OFT), elevated-plus maze (EPM), forced swimming test (FST), and sucrose preference test (SPT) behavioral tests to investigate the effect of ATRA on anxiety- and depression-like behavior in young mice. In the OFT, significant difference in duration (F (3, 17) = 6.558, p = 0.0038, A) and distance (F (3, 17) = 4.828, p = 0.0131, B) traveled in the central area were found among the control, 5, 10, and 20 mg/kg ATRA groups. Post-hoc analysis revealed ATRA treatment significantly decreased the duration (p = 0.0016, p = 0.0025, p = 0.0020) and distance (p = 0.0036, p = 0.0090, p = 0.0083) traveled in the central area among the groups receiving 5, 10, and 20 mg/kg doses. There was no significant difference in the velocity of movement in the center area among the four groups in the OFT (F (3, 17) = 0.9254, p = 0.4497, C). This indicates that ATRA did not affect motor function in mice, in comparison with the vehicle. The results of the EPM showed that there was no significant difference in the time (F (3, 17) = 2.713, p = 0.0773, D), distance (F (3, 17) = 2.713, p = 0.0847, E), and frequency (F (3, 17) = 2.808, p = 0.0709, F) in the open arms.
Figure 1. Effects of different doses of ATRA on anxiety-like behavior in mice. (A–C) center duration, distance and velocity in OFT in ATRA treatment and control groups; (D–F) duration, distance and frequency in open arms in EPM test. Data are expressed as mean ± SEM, with n = 5–6 in each group. ** p < 0.01 versus controls, using one way ANOVAs with least significant difference (LSD) test.
Results of the FST showed a difference was found among the four groups (F (3, 17) = 4.555, p = 0.0162, A). The result showed that only mice who were administered a dose of 10 mg/kg ATRA had a significant reduction in mobility time during the 6-min period, as compared with the control group (p = 0.0027). There was no significant difference in mobility time during the FST between the 5 and 20 mg/kg groups and the control group (p = 0.0670, p = 0.0823, respectively). In the SPT, there was no significant difference between all ATRA groups and vehicle, which suggested that anhedonia was not affected by ATRA administration (F (3, 17) = 0.3580, p = 0.7840, B). These results indicate that ATRA induced depression-like behavior in young mice.
Figure 2. Effects of different doses of ATRA on depression-like behavior and anhedonia level in mice. (A) Mobility time in the FST. (B) result of the SPT in ATRA and control groups. Data are expressed as mean ± SEM, with n = 5–6 in each group. ** p < 0.01 versus controls, using one way ANOVAs with LSD test.
ATRA-Induced Changes in mRNA Expression of DLG2, SynDIG1, and Retinoic Acid Receptors in the Hippocampus
Significantly different levels of DLG2 mRNA expression were found among the four groups (F (3, 17) = 6.391, p = 0.0043, A). Levels of expression of DLG2 mRNA in the hippocampus were significantly decreased in all of the ATRA treatment groups, compared with the control group (p = 0.0011, p = 0.0042, p = 0.0030). A difference in expression of SynDIG1 mRNA was found among the four groups (F (3, 17) = 3.108, p = 0.0469, B). Expression of SynDIG1 mRNA was significantly increased at a dose of 10 mg/kg ATRA (p = 0.0142). To investigate the possible relationship between the expression of synaptic-related genes and ATRA receptor, mRNA expression of three types of ATRA receptors in the hippocampus were simultaneously measured. A significant difference in retinoic acid receptor α (RARα) mRNA levels was found among the four groups (F (3, 17) = 6.252, p = 0.0047, C). The 20 mg/kg ATRA treatment group showed a significant reduction in RARα mRNA levels (p = 0.0008). There was a significant difference in RARβ mRNA expression among the four groups (F (3, 17) = 9.459, p = 0.0007, D). The expression of RARβ mRNA in the hippocampus was increased with administration of 10 mg/kg ATRA, compared with the control (p = 0.0014). Significantly different levels of RARγ mRNA expression were also found among the four groups (F (3, 17) = 20.54, p < 0.0001, E). The expression of RARγ mRNA in all of the ATRA treatment groups was significantly decreased (p < 0.0001).
Figure 3. Effects of different doses of ATRA on mRNA expression levels of target genes in the hippocampus. Changes in mRNA expression levels of (A) DLG2, (B) SynDIG1, (C) RARα, (D) RARβ, and (E) RARγ in mice treated with 5 mg/kg, 10 mg/kg, and 20 mg/kg RA, compared with controls. Data are expressed as mean ± SEM, with n = 5–6 in each group. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001 versus controls, using one way ANOVAs with LSD test.
Association of DLG2 and SynDIG1 mRNA Levels with Anxiety- and Depression-Like Behavior and RARs in the Hippocampus
We performed correlation analysis, to explore the association of
DLG2 and
SynDIG1 expression with anxiety- and depression-like behavior in mice. The results showed that relative
DLG2 mRNA levels in the hippocampus were significantly positively correlated with duration (
p = 0.0111, r = 0.5420, A) and distance (
p = 0.0174, r = 0.5128, B) traveled in the central area in the OFT. No significant correlation was found between
DLG2 with time (
p = 0.5778, C) and distance (
p = 0.2834, D) in the open arms in the EPM or mobility time in FST (
p = 0.7821, E).
Table S2 shows the correlation more intuitively. Furthermore, relative
DLG2 mRNA levels in the hippocampus were significantly positively correlated with relative mRNA levels of
RARα (r = 0.5091,
p = 0.0184, F) and
RARγ (r = 0.7873,
p < 0.0001, H). No significant correlation was found between
DLG2 and
RARβ mRNA (r = −0.1139,
p = 0.6231, G).
Figure 4. Correlation of DLG2 mRNA levels in the hippocampus with behavior and RARs mRNA levels in young mice. Correlation between mRNA levels of DLG2 in the hippocampus and (A) center duration, (B) center distance in OFT, (C) duration, and (D) distance in the open arms in EPM, (E) mobility time in FST. Correlation between mRNA levels of DLG2 in the hippocampus and (F) RARα, (G) RARβ, and (H) RARγ. Correlation analysis was performed using Pearson’s correlation test. * p < 0.05, **** p < 0.0001.
Interestingly, relative
SynDIG1 mRNA levels in the hippocampus were significantly negatively correlated to mobility time in FST (
p = 0.0052, r = −0.5861, E). No significant correlation was found between relative SynDIG1 mRNA levels and duration (
p = 0.0927, A), or distance (
p = 0.1120, B) traveled in the central area in the OFT, or time (
p = 0.1278, C) and distance (
p = 0.1871, D) in open arms in EPM.
Table S2 shows the correlation more intuitively. The results showed that relative
SynDIG1 mRNA levels in the hippocampus were significantly negatively correlated with
RARγ (r = −0.4728,
p = 0.0304, H). No significant correlation was found between relative mRNA levels of
SynDIG1 and
RARα (r = −0.2084,
p = 0.3647, F), or
RARβ (r = –0.1088,
p = 0.6387, G).
Figure 5. Correlation of the SynDIG1 mRNA levels in the hippocampus with behavior and RARs mRNA levels in young mice. Correlation between mRNA levels of SynDIG1 in the hippocampus and (A) center duration, (B) center distance in OFT, (C) duration, and (D) distance in the open arms in EPM, (E) mobility time in FST. Correlation between mRNA levels of DLG2 in the hippocampus and (F) RARα, (G) RARβ, and (H) RARγ. Correlation analysis was performed using Pearson’s correlation test. * p < 0.05, ** p < 0.01.