Baixiangdan Capsule and Shuyu Capsule Regulate Anger-out and Anger-in: GB1 Mediated GABA Can Regulate 5-ht Level in the Dorsal Raphe Nucleus

Background: To explore the intervention mechanism of Baixiangdan Capsule(BXD) and Shuyu Capsule(SY) in the treatment of anger-out and anger-in. A kind of GABABR1(GB1) mediated GABA in the dorsal raphe nucleus(DRN) regulating serotonin(5-HT) levels in the Prefrontal Cortex(PFC), Hippocampus and Hypothalamus in Anger-out and Anger-in male rats. To further explore the difference of Baixiangdan Capsule(BXD) and Shuyu Capsule(SY) in the treatment of anger-out and anger-in. Methods: The anger rat model was established by social isolation combined with resident-intruder paradigm, and behavioral evaluation was used to screen and distinguish anger-out and anger-in model rats. BXD and SY were intervention drugs of anger–out and anger-in rats respectively. On this basis, ELISA was used to detect GABA content in DRN and 5-HT contents in PFC, hippocampus and hypothalamus after different time course (0,1,3,5,7 days) treated with BXD and SY. The co expression of 5-HT and GB1 in DRN was detected by immunouorescence double labeling technique. Finally, brain stereotactic localization was performed after baclofen, the GB1 specic agonist, and CGP35348, the GB1 specic inhibitor, were injected into the DRN, the 5-HT contents in PFC, hippocampus and hypothalamus was detected by ELISA. Results: After drug treatment, ABT scores and OFT total distance in BXD group were signicantly decreased to the level of control group (P<0.05). On the contrary, that in SY group were signicantly increased to the level of control group (P<0.05). The sugar water preference coecient of BXD group and SY group was signicantly increased to the normal level (P<0.05). With the increase of the medication duration, 5-HT levels in PFC, hypothalamus and hippocampus increased, gradually corrected its abnormal decline, and returned to


Background
Anger is probably one of the mostly debated basic emotions, owing to di culties in detecting its appearance during development, its functional and affective meaning, especially in human beings 1 . The enormous impact that anger has had on people and their social interactions has been recorded in many ways and in many places throughout history 2 . In fact,the emotion of anger is de ned as a negative emotional response to goal-blockage and unfair behavior by others 3 . Based on the differences in personal characteristics and enduring propensity 4 , anger usually has two different opposite ways of expression, anger-in and anger-out 5 . Anger-out is de ned as subjective feelings classi ed as"anger directed outward away from the self", however anger-in is de ned as subjective feelings classi ed as "anger directed toward the self 6 . Anger is a normal human emotion response when encountering a range of unacceptable situations 7 . After all, anger is a subjectively intolerable emotion 8 . In extreme or long-time suffering, anger may lead to aggressive behavior, hostility, violence, anxiety and adverse health consequences 9 . A wide range of psychiatric disorders has been associated with anger, such as major depression disorder 10 , post-traumatic stress disorder 11 , and premenstrual dysphoric disorder 12 . Episodes of anger are also associated with physical diseases for instance of a transiently higher risk of cardiovascular diseases 13 . Therefore, it is of great signi cance to study the mechanism of anger and in the prevention of anger-induced diseases.
Baixiangdan capsule (BXD) and Shuyu capsule (SY) are the novel capsule formulation combining several plant extracts that have been usedin traditional Chinese medicine to treat anger-out and anger-in.
Analytical studies have shown that the main active components of BXD are paeoni orin, paeonol, and alpha-cyperone [14][15][16] , which may have antipyretic, anti-in ammatory, analgesic, and neuroprotective functions 17,18 . SY, a commercially available herbal prescription of traditional Chinese medicine, is composed of four herbal ingredients: Radix Bupleuri (Bupleurum chinense DC.), Radix Paeoniae Alba (Paeonia lacti ora Pall.), Rhizoma Cyperi (Cyperus rotundus Linn.), and Radix Glycyrrhizae (Glycyrrhiza uralensis Fisch.). Studies have shown that SY can mitigate PMS depression symptoms and that its action mechanism is concentrated in speci c cerebral areas 19,20 .Paeoni orin, the main component of Shuyu and paeony extract, has many biological effects, including enhancement of cognitive ability, improvements in learning disabilities, and nerve protection 21,22 . BXD and SY can respectively treat angerout and anger-in stably and effectively, but the neural effects of BXD and SY that underlie these actions are unclear.
Human emotion regulation circuit includes several regions of the brain such as the prefrontal cortex (PFC), the amygdala, hippocampus, hypothalamus, anterior cingulate cortex, insular cortex, ventral striatum, and other interconnected structures 23 . Study showed that defensive rage or aggression can typically be elicited by electrical stimulation of sites within hypothalamus in cats and rats 24,25 . The degree of expression of anger in pictures or sounds was positively correlated with the signal intensity of bloodstreams and glucose metabolism in the prefrontal and hippocampal 26 , the ventromedial prefrontal cortex activity responded preferentially to anger expressions oriented to self 27 . Thus, PFC, hippocampus and hypothalamus is of great signi cant in participating the anger regulation [28][29][30] .
Numerous studies have shown that various of neurotransmitters and hormones such as serotonin(5-HT), norepinephrine, dopamine, androgens and estrogens all participate in anger regulation, among these the serotonin is the most important 31,32 . Major serotonergic populations contained in the midbrain dorsal raphe and median raphe nuclei, and then projecting to various regions of the brain such as PFC, hippocampus and hypothalamus 33 , participate in regulating cognition, intuition and emotion 34 . Study showed depressed patients with anger attacks may have a relatively greater serotonergic dysregulation than depressed patients without these attacks 23 . But to different anger expression, these is still little known about how 5-HT system was taken part in. GABA is the most important inhibitory neurotransmitter in the central nervous system 35 . About 50% of synapses in the central nervous system use GABA as the neurotransmitter, which can seriously affect the function of prefrontal cortex 36 . GABABR(GB) is a metabolic G protein coupled receptor, which is composed of GB1 and GABABR2(GB2). GABABR1(GB1) is responsible for binding to GABA 37,38 , while GB2 is responsible for G protein coupling 39 . GB1 plays an important role in maintaining normal brain function. GABABR1a knockout mice are more likely to suffer from stress and pleasure loss and social escape behavior 40 . The up regulation of GABABR1a expression in dorsal raphe nucleus of socially isolated mice is closely related to abnormal behavior caused by social stress 41 . Thus, it is suggested that 5-HT, GABA and GB1 play an important role in anger and depression, and the speci c mechanism needs to be further studied.
To sum up, we speculate that BXD and SY can effectively relieve anger and depression like emotions, which may be through this pathway: GB1 mediated GABA in the dorsal raphe nucleus(DRN) regulate the 5-HT levels in PFC, hypothalamus and hippocampus.To clarify this question, this paper adopts a variety of modern biological techniques to carry out research in a rat model of anger-out and anger-in based on the widely recognized social isolation stress combined with resident-intruder paradigm 42,43 . Materials And Methods 2.1 Animals, groups and treatments. 84 male Wistar rats weighing 180-220g, and 18 male SD rats weighing 120-150g were obtained from Charles River Laboratories (Beijing, China, License No.: SCXK (Beijing) 2012-0001). The wistar rats and SD rats were housed separately with the same condition, temperature at 22 ± 2 ˚C, humidity at 55 ± 10%, noise ≤60dB and 12 h light/dark cycle (lights on at 20:00) with free access to water and food. All behavior tests were performed during the dark phase under a dim red light (<2lux). All procedures performed in this study were in strict accordance with NIH Guide for the Care and Use of Laboratory Animals.
Firstly, we make the model, which were established according to resident-intruder stress combined with social isolation stress protocol that published previously 43 . The wistar rats were used as the resident and normal control, the SD rats were used as the intruders. All procedures carried on in 4 weeks. All animals were kept 1 week for adaption. From the 2 nd week, 12 wistar rats were selected randomly into normal group with 4 rats per cages and no stress till the end of the study, the rest of the wistar rats began to be isolated as model group. During the 3 rd week, rats in model group were treated with social isolation and resident-intruder stress. The wistar rats in model group was suffered from the induced intruder SD rat into his cages for 15min every day. Intruders were confronted each day with a different resident in a Latin square design. At the end of the 3 rd week, 48 rats in model group were divided into anger-out model (AOM) group and anger-in model(AIM) group according to the composite aggression scores from the median analysis. 24 rats in model group with middle scores would be excluded. At the 4 th week, the stress was continued. Besides, anger-out Baixiangdan(BXD) group was given Baixiangdan capsule suspension at the dosage of 1.33g/kg (equivalent to 8 times of clinical dose of human), continuing for 7 days. Angerin Shuyu(SY) group was given Shuyu capsule suspension with the dosage of 1.33g/kg (equivalent to 8 times of clinical dose of human), continuing for 7 days. The control group, AOM group and AIM group were given equal volume of sterilized water, continuing for 7 days. At each end of the weeks, sucrose preference test, open eld test and aggressive behavior test (only in model group at the 2 nd to 4 th week) were performed.

Aggressive behavior test (ABT)
Aggressive behavior test has become the most commonly used method for studying depression and aggressive behavior in animals 44 . Although the change of aggressive behavior induced by social isolation combined with resident-intruder paradigm is not de ned as anger, it is most similar to human aggressive behavior 45 . In a sense, the generation of human anger is the preparation response to attack 46 , and the typical behavior and core symptom of anger emotion 47 -aggression behavior is considered as the standard of anger emotion induction 48 .
During the performing of resident-intruder stress, the rst 10 min were recorded by camera. After the test, the observers watched the video and note the duration and frequency of the behaviors. Next, "rodent aggression analysis method and device", which has obtained the national invention patent(Chinese patent:CN106472348A,2017-03-08), was used to conduct behavioral evaluation on rats, which could effectively improve the objectivity and accuracy of rodent aggression analysis study and vigorously promote the relevant research level. Composite aggression score = number of attacks + 0.2 × attack duration (s) + number of bites + 0.2 × on-top duration (s) + piloerection 45 . The aggressive behavior was evaluated by blind method, and the video was played back by three persons who were trained uniformly.
The results were recorded. The consistency test showed that kappa > 0.95.

Open eld test (OFT)
The rats were adapted to the test room for more than 10 minutes 49

Sucrose preference test (SPT)
Sucrose Preference Test allowed the rat two bottles of water for freely chosen for 24h 50 , one of which contains tap water and the other contains 0.8% sucrose solution. The position of two bottles would be switched to reduce the effect of side bias after 12h of the test. Before and after test, bottles were weighted. Sucrose preference scores were calculated as a percentage of the volume of sucrose intake over the total volume of uid intake.

Immuno uorescence 2.3.1 Tissue processing
Animals were administered 75 mg/kg of sodium pentobarbital(Sigma-Aldrich, UK) i.p. After deep anesthesia, the rats were xed in the supine position with cotton thread Lift the skin of the chest, cut the abdominal cavity along the xiphoid process, cut the diaphragm, cut the thoracic cavity upward, turn over the thymus, fully expose the heart and aorta. The perfusing needle with attened tip was inserted into the root of aorta from 45° to the right from the left apex. The needle was immediately xed with an artery clamp to prevent loosening. Gently push the syringe to expand the heart and cut a small opening on the right atrial appendage. Then the normal saline was extracted from the syringe and quickly perfused to the whole body through the perfusion needle until the liquid from the right atrial appendage was clear, about 200ml. After that, 4% paraformaldehyde was infused, rst fast and then slowly. Peripheral nerve stimulation reactions such as limb tremor and tail cocking can be seen, and the perfusion can be stopped until the whole body is stiff, about 200ml. The brain was removed en bloc and post-xed in 4 % PFA for 2h. The brains were left in 10%, 20% and 30% sucrose solution containing 0. Pick the slices out from 6 well plates and place on glass slides then cover with anti-fade uorescence mounting medium. After air-dried away from light, the slides were then viewed under confocal microscope (Zeiss, Germany).

Immuno uorescence intense analysis
The pictures were opened with Zeiss ZEN Lite 2012. Set the area of DRN as "Region of Interest". Record the value of "Arithmetic Mean Intensity" as the IF intense. Analysis the same position according to The Rat Brain Stereotaxic Coordinates (Paxinos and Watson, 1996) of different specimen.

Baclofen and CGP35348 microinjection
After behavioral testing, normal saline, baclofen and CGP35348 were injected into DRN respectively. Rats were anesthetized with 75mg/kg sodium pentobarbital i.p. (Sigma-Aldrich, UK) and placed in a stereotaxic apparatus. The method of at cranial head xation was used to x the rat's bilateral inner ear foramen and incisors at three points, adjust the position of the denture, and make the vertical position of the denture 3.3± 0.4mm below the horizontal plane connecting the two ear rods, and the anterior and posterior fontanelles at the same horizontal height. The skin was cut about 1cm along the midline above skull, and holes were drilled on the dorsal skull above the target regions. Then, microinjector was inserted and 0.2ml normal saline or 0.2ml baclofen(1.5mg/ml) or 1ml CGP35348(20mg/ml) were microinjected respectively into DRN(AP -7.8mm, L -2.0mm, V -6.3mm, 20°) of each group for 2 minutes. Stay for 2 minutes and then withdraw the microinjector. After injection, suture the skin and apply gentamycin sulfate. Brain tissues(PFC, Hippocampus, Hypothalamus) were collected respectively 10minutes after surgery.
2.5 ELISA kits 2.5.1 Tissue processing Fresh brain tissues were collected at 0, 1, 3, 5 and 7 days after treatment, and the PFC, hypothalamus, hippocampus and DRN were rapidly separated, weighted and stored at -20°C. In addition, 10 minutes after microinjection of normal saline or baclofen or CGRP35348 into DRN, brain removed on ice, then PFC, hippocampus and hypothalamus were quickly separated, weighted and stored at -20°C.
Brain tissues was rinsed with 1x PBS, homogenized in PBS and stored overnight at -20°C. After two freeze-thaw cycles to break the cell membranes, the homogenates were centrifuged at 5000×g for 5 min 4°C. The supernatant was removed and assayed immediately.

Statistical Analysis
Statistical analyses were carried out using Graphpad Prism 6.0 and results were expressed as Means ± SEM (standard error of the mean). Data were analyzed by independent t test between two groups, or oneway ANOVA when more than two groups. Differences were considered to be statistically signi cant when P < 0.05.

Experiment1:Behavior tests results
ABT score is the main indication of anger-out and anger-in rats model. As shown in Figure 1A and 1B, there was no signi cant difference in ABT score and ABT latency between groups at baseline (P > 0.05).
After modeling, rats in model group were divided into anger-out model group(AOM) and anger-in model group(AIM) according to ABT score and latency. After the medication, compared with AOM, ABT score of anger-out BXD(BXD) group was signi cantly reduced (P<0.05) and ABT latency was signi cantly increased (P<0.05). Compared with AIM, ABT score of anger-in SY(SY) group was signi cantly increased (P<0.05) and ABT latency was signi cantly decreased (P<0.05). At the same time, there was still a signi cant difference between AOM and AIM(P<0.05). OFT total distances ( Figure 1C) showed that after modeling, compared with the control group, OFT total distances of AOM signi cantly increased (P<0.05), while that of AIM signi cantly decreased (P<0.05), and was signi cantly lower than that of AOM. After the medication, compared with the control group, the signi cant difference of AOM and AIM was still maintained (P<0.05). But, compared with AOM, OFT total distances of BXD group decreased signi cantly (P<0.05) and returned to the level of the control group. Compared with AIM, OFT total distances of SY group was signi cantly increased (P<0.05) and returned to the level of the control group. SPT scores ( Figure 1D) of both AOM and AIM were signi cantly lower than those of control group after modeling (P<0.05). After the medication, compared with the control group, the signi cant difference of AOM and AIM was still maintained (P<0.05). SPT scores of BXD group and SY group were signi cantly increased (P<0.05) and returned to the level of the control group. This study con rms that social isolation combined with resident-intruder paradigm can successfully prepare scienti c and stable anger-out and anger-in rats model, which lays the foundation for subsequent experiments. Besides, BXD and SY can effectively improve the behavior of anger-out and anger-in rats.
3.2 Experiment2: The basis for the selection of drug treatment days 5-HT levels in PFC, hypothalamus and hippocampus of both AOM and AIM were signi cantly lower than that of the control group, and remained at a low level; after the drug intervention, with the increase of the medication duration, the 5-HT levels in PFC, hypothalamus and hippocampus increased, gradually corrected its abnormal decline, and returned to the normal level on the 7th day (Figure2.A,B,C;Figure3.A,B,C). Besides, the GABA level in DRN of both AOM and AIM was signi cantly higher than that of the control group, and remained at a high level. After the drug intervention, GABA level in DRN decreased with the increase of medication duration, and gradually corrected its abnormal increase, and returned to normal level on the 7th day( Figure 2D; Figure 3D).

GB1 and 5-HT neurons in DRN immuno uorescence double labeling
Previous studies have found that 5-HT neurons in DRN have nerve ber connections with PFC, hypothalamus and hippocampus. So, in order to observe the expression of 5-HT neurons and GB1 in DRN and their connections with nerves and the nerve ber connection between BXD and SY in regulating anger-out and anger-in, we examined the level of tissue cells by immuno uorescence double labeling method. As shown in Figure 4, 5-HT immunostaining positive cells(red) (5-HT neurons) in DRN of each group rats can also show GB1 immunostaining positive cells(green).

5-HT and GB1 immuno uorescence intensity
Afterwards we semi-quantitatively analyzed the intensity of immuno uorescence. Results showed that intensity in both model groups decreased signi cantly than that in control group (P<0.05). Besides, after the drug intervention, 5-HT content in DRN was elevated to normal level in BXD group and SY group(P<0.05) ( Figure 5A). However, the intensity in both model groups increased signi cantly than that in control group (P<0.05). Besides, AIM increased signi cantly than AOM(P<0.05). And, after the drug intervention, GB1 content in DRN was decreased to normal level in BXD group and SY group (P<0.05) ( Figure 5B).

Experiment4: 5-HT levels in PFC, Hippocampus and Hypothalamus were changed after injection of GB1 speci c agonist and GB1 speci c inhibitors into DRN
Finally, the change of 5-HT levels in PFC, hypothalamus and hippocampus were observed by microinjection of baclofen and CGRP35348 into DRN respectively. As shown in Figure 7, 5-HT concentrations in PFC, Hippocampus and Hypothalamus of rats in each group( Figure 7A, B, C) decreased signi cantly after injection of Baclofen, a speci c agonist of GB1, into DRN(P < 0.05), compared with the injection of normal saline, whereas 5-HT concentrations in PFC, Hippocampus and Hypothalamus of rats in each group (Figure 7a, b, c) increased signi cantly after injection of CGP35348, a speci c inhibitor of GB1, into DRN(P < 0.05).

Anger-out and Anger-in Rat Model
Social isolation combined with resident-intruder paradigm is a classic modeling method in the rat model of anger-out and anger-in 42,43 . Resident rats are raised alone for a long time, which makes them irritable and aggressive. They can attack the invaders instinctively, so as to protect their territory and generate angry psychology and behavior. Social isolation can induce aggressive behavior in rats 51 . Here we have con rmed that social isolation combined with resident-intruder paradigm could successfully prepare a scienti c and stable rat model of anger-out and anger-in, which laid the foundation for subsequent experiments.
Aggression or aggressive behavior, which might be triggered by anger 52 , has various associations with anger. Aggression may also be viewed as the behavioral expression of anger 53 . Both anger and aggression could be seen as a strong and lasting personal trait 54,55 , individuals with characteristics of aggression are often accompanied with trait anger and are more likely to generate anger and enmity in the face of offensive or irritating events 56 . Anger may include increased physiological arousal, and an increased predisposition toward aggressive behavior 57,58 . In the close association between anger and aggression, many neurochemical and genomic studies have included Anger-Hostility-Aggression de ned by Spielberger as a research object, but not just anger itself 59 .
ABT is the main indication of anger-out and anger-in rat model. OFT is often used to measure anxiety-like behavior in rats 60 . OFT and SPT testi ed the emotional changes that occurred with rats induced aggressive behavior and the complexity of emotions. The generation of anger is often accompanied with a certain degree of anxiety and depression, previous studies also illustrated this 61,62 .These three evaluation methods are also used in many researches on attack behavior evaluation 62,63 . The results of this study also con rmed that these three evaluation methods can effectively evaluate the rat model of anger-out and anger-in, and BXD and SY can effectively correct abnormal anger tendency by changing the score of ABT, OFT and SPT. This is consistent with the results of relevant reports 64,65 .

Possible Action Targets and Mechanism of BXD and SY Against Anger-out and Anger-in
Many studies have shown that 5-HT is a key neurotransmitter that regulates aggressive behavior. Decreased levels of 5-HT or 5-HIAA in brain tissue can increase the occurrence of aggressive behavior and violence 66 . Activity of 5-HT neurons in DRN increased aggressive behaviors among male mice 67 . On the contrary, studies have also found that overexpression of 5-HT1A receptors weakens the activity of 5-HT neurons in DRN and enhances the aggressive behavior of mice 68 . Studies in knockout mice also showed that increased aggressive behavior was accompanied by decreased 5-HT levels or reduced activity [69][70][71] . Some scholars think that different basal level of 5-HT and phasic chages may have different role in different types of aggression 72 . In our study( Figure 5A), both anger-out and anger-in rats have a lower level of 5-HT expression in DRN than normal rats, and anger-in rats have a lower level of 5-HT expression than anger-out rats. At the same time, BXD and SY can correct the abnormal decreased 5-HT level in DRN.
DRN contains the largest accumulation of 5-HT neuronal cell bodies in the brain, and it sends afferent projections to several distinct targets including PFC, hippocampus and hypothalamus 73 , which was also testi ed by our preliminary study. The results showed that in anger-out and anger-in rats, 5-HT levels in PFC, hypothalamus and hippocampus decreased in step with that in DRN. Similarly, BXD and SY can also correct the abnormal decrease of 5-HT levels in PFC, hypothalamus and hippocampus ( Figure 3A,B,C) , and then regulate the anger-out and anger-in , which need to take 7 days to achieve curative effect( Figure  2A,B,C). However, it has been found that increasing the level of 5-HT in PFC of mice can lead to more anger and aggression 74 , which is contrary to this study. The reason may be that the tissue size of the prefrontal region used to detect 5-HT and the selection of the left and right prefrontal regions are different. The dorsolateral area is mainly involved in cognitive and executive functions, while the ventromedial area is mainly related to positive or negative emotions 75,76 .
It can be seen from the above that the release of 5-HT from 5-HT neurons in DRN was signi cantly reduced in anger-out and anger-in model rats, and the levels of 5-HT projected to PFC, hypothalamus and hippocampus via the bers was also signi cantly decreased. BXD and SY can effectively relieve angerout and anger-in by correcting the above abnormal changes.
GABA is the most important inhibitory neurotransmitter in the central nervous system 35 . About 50% of synapses in the central nervous system use GABA as the neurotransmitter, which can seriously affect the function of prefrontal cortex 77 83 , and then regulate the activity of 5-HT neurons 84 . The increase of GABA increased its inhibition on 5-HT synthesis 85 , and eventually resulted in the decrease of 5-HT content and function. In this study, we found that GABA concentration in DRN increased, BXD and SY can effectively reduce GABA concentration ( Figure 3D) and then regulate the anger-out and anger-in, which need to take 7 days to achieve curative effect( Figure 2D).
In addition, GB can regulate the release of 5-HT 87 . Various pathological processes, including chronic pain, epilepsy and schizophrenia, are often accompanied by changes in GB1 expression and function 88,89 .
GB1a knockout mice were more prone to stress pleasure loss and social escape behavior 40 . The expression of GB1a in DRN was up-regulated in socially isolated mice 41 . It will be seen from these studies that GB1 play an important role in maintaining normal brain function. This study found that GB1 expression in DRN was up-regulated in model rats and BXD and SY can signi cantly down regulate the expression of GB1 ( Figure 5B). However, it has been reported that the expression of GB1 in hippocampal neurons of anger model rats is decreased 90 , which is negatively correlated with the results of this study. It may be due to the different brain regions detected, so the expression of GB1 is different.
The GABA terminal is in close contact with GABABR/5-HT double-standard neurons 91 .We also found that the expression of GB1 in 5-HT neurons in DRN of rats in each group was con rmed by immuno uorescence double labeling technique (Figure 4). Objective conditions for GB1-mediated GABA regulation of the existence of 5-HT neurons in DRN were determined.
Therefore, we can infer that GABA increased signi cantly in DRN of angry-out and anger-in rats, which further mediated the synaptic effect and exerted inhibitory function by binding with GB1 on 5-HT neurons in DRN, thus reducing the 5-HT level of 5-HT neurons in DRN and leading to a signi cant decrease 5-HT levels in PFC, hypothalamus and hippocampus. This is also the intervention target and way of BXD and SY.
GB1 subunit has an agonist binding site 38 , that is, GB1 can bind to agonists 92 . Because Baclofen is a GB speci c agonist, Baclofen only binds to GB1 and becomes a GB1 speci c agonist 93 to activate GB1 signaling pathway. CGP35348 is a speci c inhibitor of GB. This study showed that the levels of 5-HT in PFC, hippocampus and hypothalamus of rats in each group decreased signi cantly after injection of baclofen into the DRN, while the levels of 5-HT in PFC, hippocampus and hypothalamus increased signi cantly after injection of CGP35348 into the DRN. Therefore, from both positive and negative aspects, it was con rmed that GB1 mediates GABA in DRN to regulate the levels of 5-HT in PFC, hippocampus and hypothalamus, which plays an important role in anger-out and anger-in. And it is the corresponding target of BXD and SY. Both systemic and DRN injection of baclofen enhanced aggressive behavior 94 . Baclofen can promote the expression of GB1, improve the binding ability of GABA to its receptors, and enhance the inhibitory effect of GABA 95 , so as to inhibit the release of 5-HT from DRN.
These reports are consistent with the results of this study, but other reports are contrary: microinjection of baclofen into DRN can increase the extracellular 5-HT content of medial prefrontal cortex 94 , thereby enhancing aggressive behavior, suggesting that the enhancement of 5-HT release in medial prefrontal cortex is synchronous with the increase of aggressive behavior 67 . This may be due to the asynchrony between the total 5-HT contents and extracellular 5-HT contents.
In conclusion, BXD and SY can act on GB1 in DRN, and then affect the effect of GABA, thus affecting the levels of 5-HT in PFC, hypothalamus and hippocampus of anger-out and anger-in rats. That is, GB1 in DRN mediates GABA regulation of 5-HT levels in PFC, hypothalamus and hippocampus, which is the target of BXD and SY in regulating anger-out and anger-in.
Interestingly, BXD and SY can correct the abnormal GB1 mediated GABA regulation of 5-HT levels in PFC, hypothalamus and hippocampus in anger-out and anger-in respectively, and the correction direction is the same. However, BXD and SY can correct the abnormal behavior of anger-out and anger-in rats respectively, but the correction direction is opposite. BXD is composed of Paeoni orin, volatile oil of but also signi cantly improve learning memory and anxiety 100 . Saikosaponin can increase the 5-HT level in rat brain, and effectively improve depression like behavior 101 . It can be inferred that Paeoni orin and Cyperus volatile oil may be inclined to correct the micro mechanism of action; Paeonol may be inclined to behavioral correction of anger; Bupleurum saponin may be inclined to behavioral correction of depression. Further research is needed to con rm.

Conclusions
BXD and SY can effectively improve the abnormal behavior changes of anger-out and anger-in rats, and the optimal duration of action is 7 days. The improvement way is to correct the following abnormal changes: The signi cantly increased GABA in DRN combined with a signi cantly increased GB1 on 5-HT neurons in DRN, which further mediated the synaptic inhibition effect, thereby reducing 5-HT level of 5-HT neurons in DRN, resulting in a signi cant decrease in 5-HT levels in PFC, hypothalamus and hippocampus. Therefore, GB1 mediated GABA in DRN can regulate the 5-HT level in DRN, and regulate 5-HT levels in the PFC, hypothalamus and hippocampus, which may be one of the ways that BXD and SY treat anger-out and anger-in.