MicroRNA Regulates Early-Life Stress–Induced Depressive Behavior via Serotonin Signaling in a Sex-Dependent Manner in the Prefrontal Cortex of Rats

Abstract

Background

The underlying neurobiology of early-life stress (ELS)-induced major depressive disorder is not clearly understood.

Methods

In this study, we used maternal separation (MS) as a rodent model of ELS and tested whether microRNAs (miRNAs) target serotonin genes to regulate ELS-induced depression-like behavior and whether this effect is sex dependent. We also examined whether environmental enrichment prevents susceptibility to depression- and anxiety-like behavior following MS and whether enrichment effects are mediated through serotonin genes and their corresponding miRNAs.

Results

MS decreased sucrose preference, which was reversed by enrichment. Males also exhibited greater changes in forced swim climbing and escape latency tests only following enrichment. Slc6a4 and Htr1a were upregulated in the frontal cortex following MS. In male MS rats, enrichment slightly reversed Htr1a expression to levels similar to control rats. miR-200a-3p and miR-322-5p, which target SLC6A4, were decreased by MS, but not significantly. An HTR1A-targeting miRNA, miR-320-5p, was also downregulated by MS and showed slight reversal by enrichment in male animals. miR-320-5p targeting of Htr1a was validated in vitro using SHSY neuroblastoma cell lines.

Conclusions

Altogether, this study implicates miRNA interaction with the serotonin pathway in ELS-induced susceptibility to depression-related reward deficits. Furthermore, because of its recovery by enrichment in males, miR-320 may represent a viable sex-specific target for reward-related deficits in major depressive disorder.

Various forms of early-life stress (ELS), including neglect, have been shown to increase susceptibility to major depressive disorder (MDD) (1), which also carries an increased risk of suicide, among other warning signs (2). MDD is the most prevalent psychiatric illness and has a particularly heavy financial and societal burden in the United States (3), yet a large percentage of patients do not respond to pharmacologic antidepressant treatment (4). MDD pharmacotherapies [such as duloxetine (5)] have classically targeted the serotonin system and been proven moderately successful at managing symptoms. However, individuals who report at least one previous occurrence of ELS are less responsive to antidepressant drugs, especially those that act on the serotonin system (i.e., selective serotonin reuptake inhibitors, serotonin antagonist and reuptake inhibitors, serotonin and norepinephrine reuptake inhibitors) and may be more responsive to psychotherapy (6). Furthermore, while there is little evidence of gender differences in the experience of ELS (7), women are more likely to experience MDD (8). A clearer understanding of the serotonergic neurobiology underlying depression susceptibility, including the biological basis of effective behavioral or environmental treatments, will be crucial to improving outcomes for those with MDD.

Both MDD and ELS are associated with disruptions in serotonin signaling [reviewed in (9)]. For example, individuals who possessed the s/s genotype for 5-HTTLPR and were exposed to ELS had the highest probability of developing MDD, followed by the s/l genotype (10). Early maternal separation (MS) increased serotonin receptor 1A (Htr1a) expression in the amygdala and decreased its expression in the dorsal raphe nucleus; the serotonin transporter (Slc6a4) was also decreased in the dorsal raphe (11). Prolonged MS also decreased serotonin receptor 2A (Htr2a) expression in the prefrontal cortex (PFC) during the MS period, as early as postnatal day 7 (12). The PFC is responsible for executive functions and plays a role in emotion regulation by appraising signals from other regions, detecting threats, and moderating the reactivity of limbic regions such as the amygdala (13). It comprises mostly postsynaptic serotonergic neurons projected by the dorsal raphe (14) as opposed to presynaptic neurons. Modern antidepressants such as agomelatine target serotonin receptor 2C (HTR2C) and have been shown to improve neuroplasticity in the PFC (15). There is also a burgeoning field of research around the treatment and prevention of psychiatric disorders using nonpharmaceutical interventions. In a study comparing environmentally enriched animals to those in standard housing, the levels of serotonin in the PFC were significantly increased (16).

MicroRNAs (miRNAs), a subclass of noncoding RNAs, are estimated to regulate 20%–90% of genes in the genome (17,18). miRNAs are short sequences of nucleotides synthesized in the nucleus and then processed into mature miRNAs in the cytoplasm (19). Typically, miRNAs bind to messenger RNA and repress its ability to undergo translation into protein. While there are a few reports of miRNA expression changes in animal models of ELS (20, 21, 22, 23), there is limited research on a direct link between changes in serotonin signaling and miRNAs in response to ELS. Also, several studies have demonstrated positive behavioral outcomes and changes in serotonin signaling following enrichment (16,24); there are no studies implicating miRNAs in serotonin gene regulation in the context of ELS.

In this study, we used MS as a rodent model of ELS and tested whether miRNAs target serotonin genes to regulate ELS-induced depression-like behavior and whether this effect is sex dependent. We also examined whether environmental enrichment prevents susceptibility to depression-like behavior following MS and whether enrichment effects are mediated through serotonin genes and their corresponding miRNAs. We also validated gene targets of significantly altered miRNA in vitro. Based on previous reports of serotonin signaling aberrations in MDD and preclinical stress models, we hypothesized that genes in the serotonin signaling pathway would be altered by ELS and potentially recovered by enrichment. We also hypothesized that miRNAs targeting these genes may be partially responsible for gene expression changes. These effects will be sex specific.

Methods and Materials

Animals

Pregnant female Holtzman rats were purchased from Envigo and housed in standard conditions according to the Institutional Animal Care and Use Committee at the University of Alabama at Birmingham. A timeline of the experimental design is shown in Figure 1A. When the pups were born, the litter was assigned to the control or MS groups. Pups in the control litters were handled 5 minutes daily from postnatal day (PND) 1–14. Daily, MS animals were separated from the dam and individually housed for 180 minutes from PND 1–14. At PND 21, MS animals were assigned to enrichment (MS + environmental enrichment [Enr]) or no-Enr groups and then weaned and group housed. MS+Enr animals received 4–5 additional pieces of enrichment each week beginning on PND 21. Environmental enrichment included chewable wooden sticks, cotton or paper shreddable objects, colorful plastic toys, polyvinyl chloride tubing, and metal platforms for climbing; objects were sanitized weekly and randomly replaced to maintain novelty. Each group consisted of 6 males and 6 females (N = 36), each of which were tested for depression- and anxiety-related behaviors as follows: sucrose preference test on PND 80–84, elevated plus maze (EPM) on PND 85, forced swim test (FST) on PND 87–88, and shuttle box escape test on PND 89. At 24 hours following behavior testing, the animals were sacrificed and tissues collected. Behavior testing and tissue collection were initiated between 8:00 am and 11:00 am.

Figure 1.

Behavior changes associated with MS and Enr. Mean behavior resulting from MS and MS+Enr are shown as bar diagrams. Data are mean ± SEM. Dark gray bars indicate male animals and light gray bars indicate female animals. (A) A schematic diagram shows the timeline of behavioral experiments. (B) Sucrose preference was significantly reduced by MS and recovered by enrichment. (C) Females exhibited decreased anxiety-like behavior in the EPM, and (D) FST climb score was reduced by enrichment compared with control and MS animals. Females climbed less than males (as shown by ∗ in the key). (E) Enrichment also reduced escape latency compared with control and MS animals, and females showed reduced escape latency compared with males. (F) MS and enrichment significantly reduced adrenal weight in males but had no effect on female animals; dotted lines indicate a significant interaction effect. Significance was determined by two-way analysis of variance and follow-up independent samples t tests (n = 6 per group, N = 36); ∗p < .05. a.u., arbitrary units; Enr, environmental enrichment; EPM, elevated plus maze; FST, forced swim test; MS, maternal separation; PND, postnatal day.

Behavior Testing

Sucrose Preference Test

Rats were tested for reward-related deficits (i.e., anhedonia) using the sucrose preference test, adapted from Dwivedi et al. (25). Initially, the animals were acclimated to two 500-mL bottles of 1% sucrose water for 24 hours, followed by a bottle each of sucrose water and tap water. Food and water were then removed for 24 hours. Finally, each animal was individually housed and given food and tightly sealed 500-mL bottles of tap water and 1% sucrose water. After 24 hours, the volume of sucrose and tap water consumed was measured. Sucrose preference was calculated as the percent of sucrose water consumed out of the total liquid consumed (i.e., mL sucrose water/[mL sucrose + mL tap water] × 100).

Elevated Plus Maze

After 30-minute acclimation to the testing room, each animal was individually placed in the center of a 100-cm EPM. The animal’s movement was tracked for 5 minutes (EthoVision XT 11.5; Noldus). A single anxiety-related value (called Anxiety Index) was calculated based on the number of open arm entries, time spent in the open arm, total arm entries, and total test time (26).

Forced Swim Test

At 24 hours before testing, each animal was acclimated in a Plexiglas (Rohm and Haas) cylinder (~30 cm diameter × ~40 cm tall) filled ~20 cm deep with room temperature for 15 minutes. At testing, the subjects swam under the same conditions as the previous day for 6 minutes, which was video recorded and analyzed by a rater blinded to conditions using Kinoscope software (27); climb, swim, and immobility scores were extracted.

Shuttle Box Escape Testing

At 24 hours after FST, escape testing was conducted as described earlier (28) in a shuttle box (70 × 20 × 20 cm) (Med Associates) containing an electrified grid floor. Footshock was delivered through the grid floor by a shock generator (0.6 mA on a variable interval schedule) (model no. ENV-413; Med Associates). The testing montage began with 5 trials during which a single crossing terminated the shocks, followed by 25 trials in which crossing and then returning to the initial side was required to terminate the shocks; shocks terminated automatically after 30 seconds. Escape latency was recorded.

Tissue Collection

At 24 hours after escape testing, female animals were tested for estrous timing via vaginal lavage and then returned to their home cage for at least 1 hour. Then, each animal was anesthetized using isoflurane, and blood was collected in EDTA tubes via thoracotomy and cardiac puncture. Whole blood was centrifuged (1400 rpm, 4 °C for 15 min) to collect plasma. Adrenal glands were collected and weighed. The PFC was dissected and frozen on liquid nitrogen for storage.

Plasma-Based Corticosterone Enzyme-Linked Immunosorbent Assay and Rat Brain RNA Isolation, Complementary DNA Synthesis, and Relative Gene Expression

Corticosterone was measured in rat plasma collected at the time of sacrifice using an enzyme-linked immunosorbent assay (Enzo Life Sciences).

Total RNA was isolated from frozen PFC tissue using Trizol (Invitrogen) as previously described (29). After phase separation with chloroform, the aqueous phase was collected, and RNA was precipitated using isopropanol and 20 μg of glycogen. Following overnight incubation, the RNA pellet was collected by centrifugation, washed with 70% ethanol, and dried before resuspension in water. The RNA purity ratios (260/280 nm and 260/230 nm) were assessed by NanoDrop (Thermo Scientific, Waltham, MA).

For quantitative polymerase chain reaction (qPCR)–based gene expression testing, complementary DNA (cDNA) was synthesized following an oligonucleotide dT–based priming method using 1 μg of total RNA. For miRNA-specific expression testing, cDNA was synthesized by poly A tailing and oligonucleotide dT adapter primer as previously described (29).

qPCR-Based Gene Expression

Relative gene expression was tested by qPCR using BrightGreen chemistry. We selected serotonin signaling genes, including Slc6a4, Htr1a, Htr1b, Htr2a, Htr2b, Htr2c, Tph1, Tph2, Maoa, and Maob. Gene primer sequences are provided in Table 1. Gene expression values were normalized to the geometric mean of GAPDH, β-actin, and 18S ribosomal RNA. Relative expression was determined following Livak’s ΔΔCT calculation.

Table 1.

Primer Sequences for Real-Time Polymerase Chain Reaction Experiments

Gene/miRNA	Primer Sequence
rno_5HTR1A_F	TTTCTCATCTCCATCCCGCC
rno_5HTR1A_R	AGCGGGATATAGAAAGCGCC
rno_5HTR2A_F	GGTTTCCTTGTCATGCCTGTG
rno_5HTR2A_R	TGGATGCCGTAGAAAAGAGCA
rno_GAPDH_F	CACTGAGCATCTCCCTCACAA
rno_GAPDH_R	TGGTATTCGAGAGAAGGGAGG
rno_5HTR2B_F	GACGCAAGCAAACCAAGGAA
rno_5HTR2B_R	GGAACTACTATACCCGTGCGT
rno_Slc6a4_F	GCATACGTGGTGACTCTGCT
rno_Slc6a4_R	AAGCCCAGCATCTCCTTCAC
rno_HTR1B_F	CGCCTTTGTAATCGCTACGG
rno_HTR1B_R	GATGGACACGAGCAGGTCAG
rno_HTR2C_F	CCCCTCTAATGTGGTCAGCG
rno_HTR2C_R	TCACAGGAAAATCACGGGGG
rno_TPH1_F	TCATGCTTCTTTTCAAGCCCAC
rno_TPH1_R	TTACACTCTCTCGACCGGCT
rno_TPH2_F	TAGAGGATGTGCCGTGGTTC
rno_TPH2_R	CCTTGAATCCTGGGTGGTCG
rno_MAOB_F	AGTGCCATCCCACCTGTTTT
rno_MAOB_R	AACCCAAAGGCACACGAGTA
rno_MAOA_F	AGAGATCCCCGTTGATGCAC
rno_MAOA_R	TACGCAAATTCCCGAGCAGT
rno_mir-320-5p	GCCTTCTCTTCCCGGTTCTTCC
rno_mir-320-3p	AAAAGCTGGGTTGAGAGGGCGA
rno_mir-181a-2-5p	AACATTCAACGCTGTCGGTGAGT
rno_mir-181c	ATTCAACCTGTCGGTGAGT
rno_mir-200a-3p	TAACACTGTCTGGTAACGATGT
rno_mir-322-5p	AGCAGCAATTCATGTTTTGGA
rno_mir-322-3p	CATGAAGCGCTGCAACAA
rno_mir-185-3p	TTTCCTCTGGTCCTTCTCTAAAA
rno_mir-224-5p	AAGTCACTAGTGGTTCCGTTT
rno_mir-3572	CTTGCCCTTTTTTCCCCAG
rno_mir-16-5p	GCAGCACGTAAATATTGGCG
rno_mir-135	TATGGCTTTTCATTCCTATGTGAAAAAA
U6_F	CTCGCTTCGGCAGCACA
U6_R	AACGCTTCACGAATTTGCGT
hsa_GAPDH_F	CCACATCGCTGAGACACCAT
hsa_GAPDH_R	AGTTAAAAGCAGCCCTGGTGA
hsa_SLC6A4_F	CAGTTGGAATGTCAGGTGGT
hsa_SLC6A4_R	ACTTGCATTGGTGGTAGAGCA
hsa_HTR1A_F	TGGCTCCCCCAGTAAAACCT
hsa_HTR1A_R	AAATGCAGCGAGTGTTGGGA
hsa_miR-200a-3p	AACACTGTCTGGTAACGATGT
hsa_miR-322-5p	CAGCAGCAATTCATGTTTTGAAAA
hsa_miR-320-3p	GGTTGAGAGGGCGAAAAAAA
hsa_miR-320-5p	TTCTCTTCCCGGTTCTTCC
Universal Reverse	GCGAGCACAGAATTAATACGAC

Forward and reverse primer sequences are listed for genes and microRNAs tested using real-time quantitative polymerase chain reaction.

F, forward; hsa, Homo sapiens; R, reverse; rno, Rattus norvegicus.

miRNA Target Prediction

miRGate (30) and miRWalk (31) were used to identify targeting miRNAs. miRGate included results based on other databases, such as Targetscan, miRanda, and Rnahybrid. The predicted target list was narrowed to include miRNAs predicted by more than 1 database (e.g., miRWalk, Targetscan, miRanda) or those predicted to target both genes of interest. Top candidate miRNAs were selected for expression testing based on a literature search focused on stress, depression, or psychiatric disorders. Although miR-135, miR-200a-3p, and miR-320-3p were not detected in target prediction, they were included in expression testing based on literature search findings. To assess the potential for common regulatory mechanisms due to proximity (by long noncoding RNA, for example), miRNA and their gene targets were visualized on the rat genome using a phenogram (http://visualization.ritchielab.org/phenograms/plot). miRNA expression was tested using qPCR (primers are listed in Table 1) and normalized to U6.

SHSY-5Y Cell Transfection

To validate miRNA regulation of target genes, SHSY-5Y neuroblastoma cells were transfected with mimic and antisense fluoroarabinonucleic acid antisense oligonucleotides for miR-200a-3p, miR-320-3p, miR-320-5p, and miR-322-5p (AUM BioTech) in triplicate. A third group with no oligonucleotide was used as a control. After 72-hour incubation, total RNA was isolated from each lysate following the Trizol method. Gene- and miRNA-specific cDNA were synthesized, and qPCR was used to test for miRNA and gene expression changes using BrightGreen chemistry (primers listed in Table 1). Gene expression was normalized to GAPDH and miRNA to U6.

Statistical Analysis

A two-way analysis of variance (ANOVA) between sex and group was used to determine behavior and gene/miRNA expression differences in vivo. Tukey’s post hoc tests were used to parse significant group differences.

Results

Animal Behavior

Two-way ANOVA showed a significant effect of group on sucrose preference (F_2,30 = 5.786, p < .01), and Tukey’s post hoc tests revealed that sucrose preference was significantly decreased in MS animals compared with control animals (p < .01) or animals with enrichment (p < .05) (Figure 1B). There were no significant differences in latency to enter the open arm of the EPM or the number of entries to the open arms (p > .05) (Supplement). However, there was a significant effect of sex on EPM anxiety index (F_1,30 = 5.244, p < .05), which incorporates several EPM values into a single metric; females showed lower scores (Figure 1C) than males. There were significant main effects of sex (F_1,30 = 5.192, p < .05) and group (F_2,30 = 4.551, p < .05) on climbing in the FST (Figure 1D). Male animals showed greater climb scores than females, and a Tukey’s post hoc test revealed that enrichment was associated with significantly reduced climbing compared with control and MS animals (p < .05). There were no significant main effects of either sex or group in FST swim or immobility scores (p > .05) (Supplement). There were significant main effects of both sex (F_1,30 = 18.27, p < .001) and MS (F_2,30 = 8.954, p < .005) on escape latency in the shuttle box escape test (Figure 1E). Male animals exhibited greater latency to escape than females. Post hoc tests revealed that enrichment significantly reduced escape latency compared with both control (p < .005) and MS (p < .05) animals. There were no significant differences in animal behavior based on estrous timing (p > .05); therefore, we did not use this as a variable of interest for subsequent assays.

Adrenal Weight and Corticosterone Level

There was a significant interaction between group (control, MS, and MS+Enr) and sex on adrenal weight (F_2,30 = 6.045, p < .01), with adrenal weight decreasing after MS only in male animals (Figure 1F). There was no significant effect of MS on corticosterone levels (p > .05; not shown).

Expression Analysis of Serotonin Genes

qPCR was used to test the expression of genes associated with serotonin signaling in the PFC (Figure 2A). Two-way ANOVA revealed a main effect of group on Slc6a4 expression (F_2,30 = 5.82, p < .01). MS significantly increased Slc6a4 expression, which was not recovered by enrichment (Tukey’s post hoc test, p < .05). Htr1a was also increased by MS but it did not reach significance (F_2,30 = 2.76, p = .08). There was a significant interaction between sex and group on Tph1 expression (F_2,30 = 3.98, p < .05). In male rats, Tph1 was increased only in the MS+Enr group compared with the control group, while in females, Tph1 was decreased in both the MS and MS+Enr groups compared with the control group. MS and enrichment did not significantly alter Maoa expression, yet females exhibited significantly increased expression compared with males regardless of group (F_1,30 = 4.48, p < .05). There were no main or interaction effects on expression of Htr1b, Htr2a, Htr2b, Htr2c, Tph2, or Maob (data not shown).

Figure 2.

Serotonin pathway genes altered by MS and Enr in the prefrontal cortex and miRNA target prediction. (A) Relative rat gene expression of Slc6a4, Htr1a, Tph2, and Maoa is shown as mean ± SEM. MS and MS+Enr expression are shown relative to control. Data were normalized to the geometric mean of GAPDH,β-Actin, and 18S ribosomal RNA. Dotted lines on Tph1 indicate a significant interaction. miRWalk was used to predict miRNAs targeting Slc6a4 and Htr1a. (B) Targeting miRNAs are shown as a table with quantitative polymerase chain reaction–tested miRNAs bolded. Starred miRNAs were selected based on literature search but did not appear in miRWalk prediction. (C) Chromosomal localization of genes and miRNAs were extracted from Rat Genome Database and miRbase, respectively. Localization was visualized using PhenoGram. miRNAs in green target Slc6a4, magenta target Htr1a, and blue target both genes. Significance was determined using two-way analysis of variance (n = 6 per group, N = 36); ∗p < .05; ^#p < .10. Enr, environmental enrichment; miRNA, microRNA; MS, maternal separation.

Predicted Targeting miRNAs and Their Expression Levels

Because Slc6a4 expression was altered by MS, its targeting miRNAs were predicted using mirWalk and miRGate; we also predicted targeting miRNAs for Htr1a because its expression was increased regardless of sex, although nonsignificantly. Figure 2B lists all targeting miRNAs; bolded miRNAs were tested for relative expression changes using qPCR. A total of 26 miRNAs were found to target Slc6a4, while 1 miRNA was predicted to target Htr1a; these two genes shared five targeting miRNAs. A phenogram of the targeting miRNAs’ chromosomal localization (Figure 2C) revealed that seven Slc6a4-targeting miRNAs grouped onto chromosome 1, while another five grouped tightly on the X chromosome. Htr1a, located on chromosome 2, was not closely localized near any of its targeting miRNAs. Slc6a4 localized closely with miR-195-5p and miR-497-5p on chromosome 10.

Expression of miRNAs miR-16, -181a, -181c, -200a, -322-3p, -322-5p, -135, -185, -224, -3572, -320-3p, and -320-5p were tested in MS and MS+Enr rat PFCs (Figure 3A–L). No significant changes were detected in miRNAs miR-16 or -322-3p. Using two-way ANOVA, miR-181a was shown to be significantly altered by group (F_2,30 = 7.76, p < .005), with expression in the MS and MS+Enr groups increased compared with the control group (p < .05). miR-181c was upregulated in MS and MS+Enr animals, but it did not reach significance (F_2,30 = 2.53, p = .097). miR-200a-3p expression was reduced but not significant in MS and MS+Enr male animals compared with control animals (F_2,30 = −2.51, p = .098). Similarly, miR-322-5p expression was reduced by MS and MS+Enr compared with control animals but did not reach the threshold for significance (F_2,30 = −2.86, p = .078).

Figure 3.

Slc6a4- and Htr1a-targeting miRNAs altered by MS and Enr. Relative mature rat miRNA expression targeting Slc6a4: (A) miR-16-5p, (B) miR-181a, (C) miR-181c, (D) miR-200a-3p, (E) miR-322-3p, and (F) miR-322-5p. miRNAs targeting both Slc6a4 and Htr1a: (G) miR-135, (H) miR-185, (I) miR-224, and (J) miR-3572. miRNAs targeting Htr1a: (K) miR-320-3p and (L) miR-320-5p. Data are mean ± SEM; MS and MS+Enr expression are shown relative to control; data were normalized to U6 expression; significance was determined using two-way analysis of variance (n = 6 per group, N = 36); ∗p < .05; ^#p < .10. Enr, environmental enrichment; miRNA, microRNA; MS, maternal separation.

There were no significant group differences in expression of miR-135, -185, -224, or -3572 (Figure 3G–J). miR-320-3p expression was decreased in females compared with males but was not significant (F_1,30 = 2.90, p = .099). Similarly, miR-320-5p expression was decreased in MS animals compared with control animals, but it did not reach significance (F_2,30 = 3.00, p = .065).

In Vitro Validation of miRNAs and Target Genes

miRNAs were selected for transfection if their pattern of regulation was opposite of their target gene (Slc6a4 or Htr1a), which suggested that these miRNAs might be responsible for gene expression changes after MS. Following transfection (schematic diagram in Figure 4A) with miR-200a-3p mimic, miR-200a-3p expression was increased 10.17-fold compared with nontransfected control (t₇ = 8.29, p < .001); however, its target, Slc6a4, was not increased (p > .05) (Figure 4B). Transfection with anti–miR-200a-3p decreased miR-200a-3p expression, but it did not reach significance (t₇ = −2.04, p = .08). Conversely, anti–miR-200a-3p increased Slc6a4 expression 1.55-fold, but it also did not reach significance (t₇ = 2.17, p = .06) (Figure 4B). SHSY cells transfected with miR-322-5p mimic and anti–miR-322-5p did not show changes in miR-322-5p or SLC6A4 expression (p > .05) (Figure 4C).

Figure 4.

miRNA and gene expression changes following miRNA transfection in vitro. (A) A schematic diagram of transfection using SHSY cells. miRNA oligonucleotides (200 nM) were transfected on adherent SHSY cells at 50% density in triplicate (n = 3). After 72-hour incubation at 37 °C, the cells were collected in Trizol and RNA was isolated. miRNA and gene expression were tested using qPCR. Expression of (B) miR-200a-3p and (C) mir-322-5p were tested in miR-200a-3p– and miR-322-5p–transfected cells, respectively, along with their target, SLC6A4. Bars represent fold change relative to control and SEM. Expression of (D) miR-320-3p and (E) miR-320-5p were tested in miR-320-3p– and miR-320-5p–transfected cells, as well as their target, HTR1A. Significance was determined by independent samples t test comparing mimic or antisense to the two control groups (control and scramble); ∗p < .05. anti, antisense; cDNA, complementary DNA; miRNA, microRNA; oligo, oligonucleotide; qPCR, quantitative polymerase chain reaction.

Transfection with miR-320-3p mimic and anti–miR-320-3p did not significantly alter expression of miR-320-3p or its target, Htr1a (Figure 4D). miR-320-5p mimic significantly increased miR-320-5p expression by 1.73-fold (t₇ = 3.57, p < .01) and decreased Htr1a expression (t₇ = −6.02, p < .001) (Figure 4E). Anti–miR-320-5p significantly reduced expression of miR-320-3p (t₇ = −3.57, p < .01) but also reduced expression of Htr1a (t₇ = −6.90, p < .001) (Figure 4E).

Discussion

This study aimed to assess PFC expression changes in serotonin pathway genes and their targeting miRNAs following MS and environmental enrichment. Enrichment showed some promise as a preventive measure in that enriched MS animals showed sucrose preference levels similar to control animals; however, MS and enrichment did not clearly affect anxiety-like behaviors. Slc6a4 was significantly upregulated by MS but not recovered by enrichment, while Htr1a was also upregulated (at trend level, p = .08) by MS and subsequently recovered by enrichment, most notably in male animals. Both miR-200a and miR-322-5p were decreased by MS regardless of enrichment, although it did not reach significance; miR-320-5p, which targets Htr1a, was significantly reduced by MS. To test if their reduction might have enabled the translation of Slc6a4 and Htr1a after ELS, SHSY cells were transfected with miRNA oligonucleotides. HTR1A was validated as a target of miR-320-5p. Transfection did not confirm miR-200a-3p and miR-322-5p as targeting SLC6A4. Based on these data, miR-320-5p is an intriguing miRNA that may confer resilience to reward-related behavioral deficits after ELS through its effects on expression of HTR1A. Although a direct connection between SLC6A4 and miR-200a-3p and miR-322-5p was not established, their expression was affected by MS. Future studies might explore other gene and miRNA regulatory mechanisms that might explain changes.

Serotonin Receptor 1A and Stress Susceptibility

We are cautious to interpret changes to Htr1a expression after MS and enrichment because these findings did not reach significance. However, direct manipulation of miR-320-5p did affect Htr1a expression in vitro. In support of our findings, previous MS (32) and stress-related (33) studies have found increased Htr1a. The serotonergic system has been heavily implicated in MDD pathophysiology, yet MS only induced anhedonia-related behavior changes. Thus, Htr1a may not have been as significantly affected as animals exhibiting behavioral despair or learned helplessness. Moreover, anxiety has been associated with reductions in Htr1a expression (34). Thus, it is possible that we did not observe overt anxiety-related behavior because MS increased Htr1a expression. In male animals, enrichment returned Htr1a expression to that of control animals without the use of pharmacological agents. Considering that even recovered patients with MDD can exhibit increased Htr1a expression, indicating this phenotype as a trait of MDD susceptibility (35), environmental enrichment shows even greater promise for its ability to protect against this susceptibility phenotype. A different study of rodent ELS found that social enrichment in adulthood did not reflect in changes to Htr1a expression (36), highlighting the importance of intervention prior to adulthood. In adolescents, reduced activation of reward-related brain regions predicts later onset of depression symptoms (37). Similarly, our findings reflect coexisting reward-related deficits and Htr1a expression increases, which may promote further susceptibility to depression. Furthermore, enrichment prior to adulthood might negate this susceptibility through miR-320-5p regulation of Htr1a.

The Serotonin Transporter and Stress Susceptibility

In MS animals who also exhibited anhedonic behavior, Slc6a4 expression was significantly increased in the PFC. Generally, increased Slc6a4 binding is associated with increases in MDD symptomology (38). While our findings are consistent with this pattern of Slc6a4 expression in MDD, we were unable to confirm regulation by miRNAs. Various other epigenetic changes to Slc6a4 have been associated with ELS. In humans, prenatal exposure to maternal MDD has been shown to alter methylation of Slc6a4 in a sex-dependent manner, with males exhibiting increased methylation compared with females (39); other loci in the Slc6a4 promoter showed increased methylation regardless of sex (39). There is also the potential that Slc6a4 is responsive to synaptic serotonin concentration changes (38), which are also affected by Htr1a.

miRNA Regulation of Serotonergic Genes in ELS

Following the canonical regulatory relationship between miRNAs and their gene targets, miRNAs miR-200a-3p, -322-5p, and -320-5p were significantly reduced in MS animals. A few previous studies have explored miRNA regulation of the serotonergic system, including miR-135 effects on Slc6a4 and Htr1a (40), and miR-26a targeting of Htr1a and its upregulation by antidepressants in male mice (41). Only one study has implicated miR-322 in ELS. O’Connor et al. (42) found that miR-322 was decreased in MS rodents following treatment with electroconvulsive therapy. Our findings showed that miR-322 was also decreased in MS+Enr animals compared with control animals; however, MS-only animals also exhibited this same decreased expression. miRNA-322 is implicated in several cellular processes that are relevant to stress, such as plasticity and cell differentiation (43). While some studies have elucidated miR-200 expression changes in ELS models (23,44), to our knowledge, this study represents the first to connect miR-200 with serotonergic changes following ELS. Although we found opposite directions of expression change for miR-200a and Slc6a4 in our rat MS model, we could not confirm miR-200a inhibition of SLC6A4 in human neuroblastoma SHSY cells. A previous report validated that miR-200a inhibits Slc6a4 in the rat colon (45), which could indicate that this targeting occurs outside of the brain.

miR-320-5p was previously unidentified relative to ELS, yet in vitro transfection with miR-320-5p mimic significantly increased miR-320-5p expression and decreased Htr1a expression. Rodent chronic mild stress increased expression of miR-320-3p in the ventral tegmental area but only in animals that exhibited resilience (no change relative to control animals) in the sucrose preference test (46); Slc6a4 was subsequently reduced in the ventral tegmental area of resilient animals. We validated that miR-320-5p targets Htr1a and showed slight expression reduction after MS. Reduced endogenous expression of miR-320-5p allows more Htr1a mRNA to be translated into functional receptors without its interference. Tentatively, we speculate that MS leads to reduced miR-320-5p, which increases postsynaptic Htr1a expression (Figure 5A), thus reducing serotonin availability (Figure 5B). While it is not clear how Slc6a4 is increased, it may upregulate in response to reduced extracellular serotonin availability. In the PFC, Htr1a is a heteroreceptor primarily responsible for inhibition of downstream activity (47), which may partially explain behavior changes subsequent to increased Htr1a expression. In male animals, Htr1a and miR-320-5p expression changes are reversed by enrichment, although Htr1a changes were not significant. miR-320-5p is one of many X-linked miRNAs that may partially explain sex-based differences in its response following enrichment. There have been a few studies on X-linked miRNAs in other disorders (48,49); however, no studies have explored this potential mode of sex differences and epigenetic heritability in MDD or ELS. This is the first study, to our knowledge, to validate miR-320-5p targeting of Htr1a in a central nervous system cell line.

Figure 5.

Maternal separation reprograms miRNA expression and its regulation of Slc6a4 and Htr1a in the prefrontal cortex. (A) Early-life stress reduces the expression of miR-320 in the prefrontal cortex such that it cannot inhibit the translation of Htr1a into proteins. Increased HTR1A protein is likely reflected as (B) a higher density of postsynaptic HTR1A in maternal separation animals. Once serotonin is released into the synapse, HTR1A quickly expends serotonin, thus reducing its availability. Putatively, Slc6a4 expression is increased presynaptically to compensate for limited availability of serotonin. In male animals, enrichment reverses these Htr1a expression changes. Created with BioRender.com. miRNA, microRNA.

Timing of neurobiological assays and stress interventions, such as MS, also contribute to differences in gene expression. One study of MS found that adolescent male animals who received MS exhibited increased Slc6a4 expression immediately following the MS period, but this expression was reduced in adulthood (50). It has been frequently reported that male rodents exhibit more stress-induced behavior change (51); yet, women are considered more susceptible to MDD with an almost twofold incidence compared with men (8). Differences across species may underlie some disparities in sex-based stress susceptibility between humans and rodents. For example, distinct from humans, rodents do not exhibit different alleles of the Slc6a4 gene (52).

We have provided the first evidence that miRNAs are involved in MS-induced PFC gene expression changes and reversal following environmental enrichment. Future studies should explore other brain regions and enrichment paradigms. Our enrichment montage was relatively simple yet produced robust changes in behavior and miRNA expression changes in MS animals that were not observed with MS alone. As our study did not include a control group that received enrichment, we cannot attribute these changes solely to enrichment. Studies of enrichment in control animals may reveal unique mechanisms of effect or additional positive benefits beyond prevention of disease. It will also be of particular interest to compare enrichment to antidepressants that block SLC6A4 from binding serotonin, such as duloxetine (5), in addition to their combination. This information will provide critical insights for the development of pharmacological treatments or preventions for depression. Future studies should consider altering the expression of these miRNAs in vivo to determine if their direct manipulation conveys resilience to stress.

Limitations

Because of the complex targeting networks formed by miRNAs in which a single miRNA may target hundreds of genes (19), expression changes in a single miRNA may not translate to opposite changes in its gene target. Other targeting miRNAs may exhibit an opposite change, counteracting a single miRNA’s effects. This may partially explain why we did not find consistent depression-related behaviors such as learned helplessness (increased escape latency), behavioral despair (increased FST immobility), or anxiety. It is also clear that small differences in stress paradigms can significantly alter animal behavior at a later time point (53). It is critical to the study of ELS to develop models that more closely replicate their intended human psychiatric disorders.

Conclusions

Serotonin signaling and metabolism have long been associated with psychiatric disorders such as MDD. While a few studies have explored miRNA regulation of the serotonergic system related to stress, none have considered it in the context of MS or its treatment with enrichment. Not only does MS induce significant behavior phenotypes resembling anhedonia, but we also showed a reduction of miR-320-5p following MS, with sex differences in its response to enrichment. Its predicted target, Htr1a, showed opposite expression changes, and we validated this targeting in vitro. We also found significant increases in Slc6a4 expression in the PFC. Altogether, this study showed that enrichment may be a viable tool for prevention of anhedonia following ELS via miRNA and serotonergic gene targets.