Early-life stress induces EAAC1 expression reduction and attention-deficit and depressive behaviors in adolescent rats

Neonatal maternal separation (NMS), as an early-life stress (ELS), is a risk factor to develop emotional disorders. However, the exact mechanisms remain to be defined. In the present study, we investigated the mechanisms involved in developing emotional disorders caused by NMS. First, we confirmed that NMS provoked impulsive behavior, orienting and nonselective attention-deficit, abnormal grooming, and depressive-like behaviors in adolescence. Excitatory amino acid carrier 1 (EAAC1) is an excitatory amino acid transporter expressed specifically by neurons and is the route for the neuronal uptake of glutamate/aspartate/cysteine. Compared with that in the normal control group, EAAC1 expression was remarkably reduced in the ventral hippocampus and cerebral cortex in the NMS group. Additionally, EAAC1 expression was reduced in parvalbumin-positive hippocampal GABAergic neurons in the NMS group. We also found that EAAC1-knockout (EAAC1−/−) mice exhibited impulsive-like, nonselective attention-deficit, and depressive-like behaviors compared with WT mice in adolescence, characteristics similar to those of the NMS behavior phenotype. Taken together, our results revealed that ELS induced a reduction in EAAC1 expression, suggesting that reduced EAAC1 expression is involved in the pathophysiology of attention-deficit and depressive behaviors in adolescence caused by NMS.


Introduction
Early-life stressful events have detrimental effects on the brain and are risk factors for behaviors that are associated with the etiology of several psychiatric disorders. Childhood adversities are associated with maladaptive family functioning that has been related to one-third of adult psychopathology 1 . Elucidating the mechanism of the correlation between early-life manipulations and subsequent disease is difficult to be proven in humans 2 . In animal models, early developmental manipulations involving maternal care has been identified as an ethologically relevant stressor that induces cognitive and emotional dysfunction throughout life 3 . Neonatal maternal separation (NMS) enhances responses to aversive and appetitive stimuli more cautiously toward fear signals in the environment 4,5 . The postnatal period is accompanied by significant maturation of neuronal systems. Experiencing stressful events results in increased neurochemical, neurobehavioral, and immune-inflammatory abnormalities in subsequent years 6 .
Glutamate is the primary excitatory neurotransmitter in the central nervous system and plays a central role in the neurotransmission of~80% of synapses 6,7 . Dysregulation of glutamatergic neurotransmission is related to stressand depressive-like behaviors 8,9 . Excitatory amino acid carrier 1 (EAAC1, also referred to as EAAT3) is one subtype of the excitatory amino acid transporter (EAAT) family 10 . Compared with other subtypes, including the glutamate aspartate transporter (GLAST, also referred to as EAAT1), glutamate transporter-1 (GLT1, also referred to as EAAT2), EAAT4 and EAAT5, EAAC1 plays a more significant role in cysteine transport in the brain 11 . EAAC1 was first described as a neuronal glutamate transporter 10 , although it has now been shown to play only a minor role in glutamate removal from the extracellular space, as this task is primarily performed by astrocyte glutamate transporters such as GLT1 and GLAST 12,13 . EAAC1 is expressed at presynaptic GABAergic terminals, where the uptake of glutamate could contribute to GABA synthesis [14][15][16] . The loss of brain EAAC1 expression interferes with GABA synthesis and results in epilepsy 17,18 . EAAC1 expression is altered under pathological conditions and stimuli, such as epilepsy, hypoxia, multiple sclerosis, bipolar, schizophrenia, H 2 O 2 , retinoids, and neuregulin-1 16,[19][20][21][22][23] . Moreover, genetic studies implicate Slc1a1, a gene encoding EAAC1, in obsessive-compulsive disorder 24,25 .
The expression of EAAC1 varies during brain development 26 , and the transporter is expressed before both EAAT1 and EAAT2 expression in vitro 27,28 and in vivo 29 . Early expression of EAAC1 suggests a role for EAAC1 in the neuroprotection of CNS cells during brain development.
Here, we show that EAAC1 is reduced or lost in NMS and EAAC1 (−/−) rats, respectively, in which nonselective attention-deficit and depressive behaviors were shown in adolescence. These results suggest that EAAC1 exerts a regulatory role in neuromodulation and that the reduction in EAAC1 expression contributes to the pathogenesis of depression.

NMS rats exhibit impulsive behaviors in adolescence
Several previous studies have shown that early-life stress alters emotional behavior in adults and adolescents. However, whether NMS leads to increased or decreased fearful/anxiety-like behaviors in adolescence remains unclear. In this study, we used an NMS model of daily separation from mother and siblings for 3 h/day individually ( Fig. 1a).
In the first step, all the rats were investigated for their developmental milestones from PND1 to PND21 (additional file 2: Table S2a). When a certain developmental milestone (i.e., pinna detachment, incisor eruption, eye opening and surface right reflex) was measured, no significant difference was found between the NMS and CON groups (additional file 2: Table S2b, additional file 3: Fig.  S1b). However, NMS rats showed a progressive decrease in body weight compared with CON rats (Additional file 3: Fig. S1a).
The open field apparatus comprises a square arena (5 × 5) divided into the edge zone, side zone, and inner zone on the captured image (Fig. 1b). When anxious, the natural tendency of rats is to prefer to stay closer to the wall or edge area. However, impulsive-related behavior is measured by the degree to which the rat prefers entering the inner zone of the open field. Comparison of the multiple parameters (such as the time and distance in the edge zone, side zone and inner zone) demonstrated that NMS rats exhibited increased impulsive-like behavior compared with CON rats (Fig. 1c-h). The NMS rats showed more exploration than CON in the inner zone (Fig. 1e, h). Furthermore, parameters such as the number of zone transitions, velocity of movement and total distance traveled were comparable between CON and NMS rats ( Fig. 1i-k), revealing that NMS stress does not impair the general motor ability of adolescent rats.

EAAC1 expression is altered in NMS rats of various lengths
The glutamatergic circuit is involved in the pathogenesis of emotional behavior 30,31 . However, it remains unclear whether glutamate transporters are involved in Fig. 1 Schematic representation of the experimental schedule, exploratory, and ethological behavior. a Schematic representation of the experimental protocol. Maternal separation as a model of ELS was performed for 19 days (NMS21) from PND2 to 21. Pups were daily separated from the mother and siblings for 3 h/day and were placed individually in cages. Behavior tests were performed on PND 35 (OFT) and PND 50 (FST). b A square box on the left illustrates a schematic of the tested arena in an open field apparatus and shows areas that are designated as the edge, side, and inner zones. Representative track plots of a CON rat and NMS rat (on the left). Bar graph displaying the mean time spent in the edge zone (c), side zone (d), and inner zone (e). Bar graph displaying the distance of the edge zone (f), side zone (g), inner zone (h), and total traveled (k). Bar graph displaying the mean velocity (j) and zone transition number (i). l Number of rearings in OFT. m Number of wall-leanings in OFT. n Number of stretches in OFT. o Grooming time in OFT. NMS rats showed reduced rearing and wall-leaning within 5-min. The data on locomotion behavior quantification are expressed as the means ± S.E.M. The number of investigated rats is marked on each bar. **P < 0.01 and ***P < 0.001. impaired emotional behavior. Next, we performed NMS using three length models from PND 2 until PND 7, PND 14, or PND 21 (Fig. 3a). Some variability was revealed in the NMS lengths used. Western blotting was performed to examine the expression of EAAC1, EAAT1, and EAAT2 in the hippocampus (Fig. 3b). Interestingly, the level of the EAAC1 protein was significantly reduced in the length models of PND 14 and PND 21 (Fig. 3b, c). However, we observed no significant difference in the protein levels of EAAT1 and EAAT2 in all three length models ( Fig. 3b-e). These results indicate that the expression of EAAC1 protein is easily affected by NMS stress from an early stage.

NMS induces a decrease in PV-positive GABAergic neurons and EAAC1 expression in the hippocampus of adolescence rats
We checked the immunoreactivity of EAAC1 and PV in NMS rats to investigate the involvement of the GABAergic pathway. Previously, we reported that EAAC1 is expressed in GABAergic neurons in the prefrontal cortex 16 . In further agreement, EAAC1 was found in PVpositive hippocampal interneurons of 3-week-old mice 15 .
To determine the in vivo subcellular localization of EAAC1 in PV-positive neurons, we stained hippocampal sections of adolescent NMS rats. EAAC1 was detected in puncta-ring-like structures and neuropils, colocalizing with PV (Fig. 6a). Additionally, 72.34% of PV clusters were EAAC1-positive, suggesting EAAC1 localization at specific subsets of GABA terminals (Fig. 6c). However, 47.28% of EAAC1 clusters colocalized with PV (Fig. 6c), consistent with the notion that EAAC1 is also localized at non-GABAergic synapses 15 . Interestingly, NMS rats exhibit decreased expression of EAAC1 (CON, 1.00 ± 0.05; NMS, 0.36 ± 0.04; t(30) = 9.777, P < 0.001; Fig. 6d) and PV (CON, 1.00 ± 0.07; NMS, 0.30 ± 0.04; t(30) = 8.238, P < 0.001; Fig. 6e) at neuronal terminals. However, western blotting revealed that GAD65, GAD67, and PV did not significantly differ between the animal groups ( Fig.  6f, g). Taken together, these results indicate that decreased EAAC1 expression in the PV-positive interneuron terminal may affect GABAergic dysfunction.  Maternal separation as a model of ELS was performed for 7 days (NMS7), 14 days (NMS14), and 21 days (NMS21). b Brain sampling was performed on each of the last days (7, 14, and 21 days). Immunoblot analysis of EAAC1, EAAT1, EAAT2, and β-actin expression in hippocampal lysates. c NMS attenuated the reduction in EAAC1 expression, as shown by the densitometric values, which are shown as ratios relative to the values of the CON group; CON n = 12, NMS n = 10 *P < 0.05, ## P < 0.01. d Quantification analysis of EAAT1 expression in (b). Ratios relative to the values of the CON group; CON n = 13, NMS n = 10. e Quantification analysis of EAAT2 expression in (b). Ratios relative to the values of the CON group; n = 10.

Discussion
Early life is a sensitive period of brain development where experiencing any insult (such as stressful events) can cause far-reaching consequences. ELS has been established as a major risk factor for major depression and suicidal behavior along with other psychiatric illnesses in adolescence and adulthood 1,34-36 . The NMS model is a well-established animal model of early-life stress. However, different behavioral phenotypes have been reported depending on various NMS protocols, such as a single or sibling's separation, the duration of separation, and the number of separations. In adolescence, increased and decreased fearful/anxiety-related behaviors, increased and no effect on depressive-like behaviors were all reported 37,38 . We used transient NMS during the first three postnatal weeks as well as characterized the model of early-life stress in adolescence.
In the present study, we found that EAAC1 protein expression was significantly decreased in the hippocampus and cerebral cortex of adolescent NMS rats. However, NMS did not affect the mRNA level of EAAC1 in the hippocampus in adolescence. These results may indicate regulation of the EAAC1 protein degradation pathway. Additionally, we did not observe significantly reduced EAAT1 and EAAT2 expression in the hippocampus of adolescent NMS rats. Previous human postmortem studies reported reduced expression of EAAC1 (EAAT3) and EAAT4 mRNA in the striatum in bipolar disorder 39 . Prenatal restraint stress offspring rats showed significantly induced depressive-like behavior and decreased EAAT2 and EAAT3 mRNA expression in the hippocampus 40 . In contrast, the transcript levels of EAAC1 were not significantly altered in learned helpless rats, an animal model of depression 41 .
Our data revealed that male NMS rats showed increased impulsive behavior in adolescence. During the open field task, male NMS rats spent a significantly longer time exploring a longer distance in the inner zone (Fig. 1). These findings are in agreement with the previously reported decreased anxiety-like behavior after experiencing maternal separation during the first two postnatal weeks 42 . Male NMS rats spent less time in the periphery (edge and side zone), suggesting that they were more impulsive than CON rats 43 . Moreover, we found for the  first time that NMS rats showed hyperactivity only in the inner zone. Rats normally prefer to stay beside the wall where no predators may attack from the back. Furthermore, we found that EAAC1 (−/−) mice exhibited impulsive-like behavior as NMS rats. The initially reported Slclal/EAAT3-null mice showed decreased activity in the open field 44 , whereas other investigators have reported no significant changes in EAAC1 (−/−) mice 45,46 .
Rearing consists of stopping ambulation and standing on the hindlimbs and is used as a measure of orienting behavior because orienting or nonselective attention is associated with the duration of rearing episodes, where longer rearing indicates more orienting behavior 38 . Both rearing and leaning (against the walls) to novelty have been associated with the distribution of hippocampal mossy fibers in rats 47,48 . Additionally, rearing activity was differentially affected by hippocampal cholinergic and dynorphinergic systems, inhibition of nitric oxide synthesis, and blockade of NMDA glutamate receptors [49][50][51] . The NMS rats showed significantly less rearing and wallleaning episodes than CON rats, indicating that they may have a deficit in nonselective attention and orienting. Here, we report that EAAC1 (−/−) mice also showed decreased rearing and wall-leaning activity.
Moreover, stretch attend postures occur a riskassessment behavior indicating that the rodent is hesitant to move from its present location to a new position 52,53 ; thus, a high frequency of these postures indicates a higher level of anxiety. We observed that NMS rats showed decreased stretching numbers in adolescence. These findings correlate with an increase in spending time in the inner zone; NMS rats showed a decrease in spending time in the edge zone.
Acute stressors, including exposure to a novel environment, can potently modulate self-grooming, often increasing the frequency or total duration of episodes 54-56 . Our data revealed increased self-grooming behavior in NMS rats and EAAC1 (−/−) mice in adolescence. Interestingly, several independent genetic studies consistently implicate the Slc1al gene, which encodes EAAC1, a candidate gene for obsessive-compulsive disorder in both males and females [57][58][59] . The sequential patterned strokes performed during grooming are reminiscent of the excessive and repetitive hand-washing behaviors of patients with OCD 60 . Regarding the different grooming behavior phenotypes, Slc1al/EAAC1-null mice demonstrated no significant changes in 8-to 16-wk-old mice 45 , and others have reported increased grooming activity in EAAC1 knock-out mice (P14-P35) 33 . Additionally, other reports have published evidence for a role of EAAC1 in schizophrenia. Increased Slc1al/EAAC1 transcripts and proteins were reported in schizophrenic subjects 21 . Interestingly, we previously found that NRG1/ ErbB4 (susceptibility genes of schizophrenia) signaling is critical for EAAC1 expression and function 16 .
Our data showed that NMS rats provoke depressive-like behavior in adolescence. Notably, EAAC1 (−/−) mice also showed depressive-like behavior in adolescence. Adolescence correlates with developmental changes in the brain circuits in the limbic areas, such as the hippocampus 61 . Some investigations have demonstrated that applying pharmacological treatments during adolescence has beneficial effects on the mitigation of depressive-like behaviors induced by early-life stress 62 . However, no previous studies have investigated EAAC1 in early-life stress circuits or concerning depressive-like behaviors. Here, we reported that EAAC1 (−/−) mice significantly show depressive-like behavior in adolescence. Moreover, we found that EAAC1 expression was reduced in PVpositive hippocampal GABAergic neurons in the NMS group. Depression and stress decrease both glutamate and GABA neurotransmitter circuits in the limbic and cortical regions 63,64 . The loss of EAAC1 expression may cause an abnormality in glutamate conversion to GABA in presynaptic GABAergic terminals. Further research is needed to clarify the underlying mechanism of EAAC1 in these phenotypes. EAAC1 can transport not only extracellular glutamate but also cysteine into the neurons. Our data clarified that early-life stress, such as NMS, can provoke impulsive-like behavior, orienting and nonselective attention deficit, abnormal grooming behavior, and depressive-like behaviors in adolescence. Furthermore, we identified a novel function of EAAC1: EAAC1 (−/−) mice show similar abnormal behaviors to NMS in adolescence, which may be associated with decreased EAAC1 expression. These results suggest that early-life stress can cause changes in EAAC1 expression, which may participate in the pathogenesis of nonselective attention deficit and depression in adolescence. These results may provide new insights into EAAC1 signaling in the pathophysiology of depression and shed light on the development strategies of new antidepressants.

Animals and maternal separation procedure
Experiments with animals were performed following the institutional and Eulji University guidelines. Pregnant Sprague-Dawley (SD) rats were obtained at 15-17 days of gestation from a laboratory animal supplier (Samtako bio Korea) and were housed individually in cages under standard laboratory conditions with a 12-h light/12-h dark cycle. On postnatal day (PND) 1, the litters were culled to 6-10 pups at a constant sex ratio of 1:1 to avoid sex-based maternal behavioral biases 65,42 . During separation sessions, the pups were removed from the mother for periods of 3 h per day. From PND1, the pups were weighed daily, and their overall health was assessed (i.e., milk visible in pup's stomach and pups age-appropriately active). Daily, each pup was assessed for the achievement of species-specific developmental milestones as described in Additional file 2: Table S2 66 . EAAC1 (−/−) mice were also housed in a regulated environment. EAAC1 (−/−) mice were descendants of the strain established by Peghinni et al. 44 , in which exon 1 was disrupted by a neomycin resistance cassette. EAAC1 (−/−) mice were outbred to wild-type (WT) CD1 mice for more than 10 generations prior to these studies. A WT colony was maintained using the WT offspring from the latter outcrosses. We used only males for this experiment.

Statistical analysis
The data are presented as the means ± SEM of three or more independent experiments. Student's paired t-test was used for comparisons of the means between two groups of cells in a single experiment. P < 0.05 was considered statistically significant.

Supplemental materials and methods
The Supplemental Materials and Methods are provided in additional file 1: Table S1.