Exploration of Molecular Mechanism of Traditional Chinese Tuina Therapy for Insomnia by Proteomics Analysis

Insomnia, one of the most common mental disorder, not only affects the quality of life, but also damages physical and mental health. Therefore, it is very necessary to explore the molecular mechanism of insomnia and nd some suitable treatments. At present, methods of using drugs to treat insomnia is not satisfactory due to lack of evidences and side effects. Hence, development of non-drug treatments is particularly important. Tuina, a Chinese massage method, has achieved certain therapeutic effects on injuries, rheumatism, neurological diseases and other types of diseases. We have treated patient with insomnia by Tuina, and obtained therapeutic effects indeed. In the current study, we used iTRAQ (isobaric Tags for Relative and Absolute Quantitation) quantitative proteomics to analyze plasma samples taken from the healthy control (HC) group, insomnia patients group (before Tuina treatment, BTT) and insomnia-therapy group (after Tuina treatment, ATT) to identify the molecular correlation of insomnia. In BTT vs HC, we found that the expression of many immune-related and stress-related proteins were out of control, and it revealed that Tuina had the capacities to regulate expression of immune-related and stress-related proteins in ATT vs BTT, suggesting that Tuina may improve insomnia by regulating immune-related and stress-related proteins. The proteomics verication results had been veried by commercial ELISA(Enzyme linked immunosorbent assay. All in all,Our study not only found a good way to treat insomnia, but also provided a research foundation for improving insomnia.


Introduction
Insomnia is one of the most common mental disorders in the world [1], usually manifested as di culty in starting or maintaining sleep or early morning wake-up related to impaired daytime functioning [2].
According to the de nition of the National Institutes of health, insomnia disorder includes sleep de ciency, sleep homeostasis, sleep fragmentation, insu cient sleep or impairment of sleep quality or quantity induced by a sleep disorder [3]. Sleep is a biological process that is important for optimal neurologic function, as well as systematic biology, including appetite regulation, metabolism, hormonal balance, immunity, and cardiovascular system [4,5]. The insu cient sleep will not only affect the life quality of the patients, but also do harm to psychological and physical health, which may further cause other conditions, such as anxiety, fatigue, depression, and cognitive decline, etc [6][7][8]. Previous research data indicates that the prevalence of sleep disturbance ranges between 36% and 50% [9,10], and approximately 15% of people meet the criteria for insomnia in China [11]. Sleep disorders have been affecting a large number of people in the world and increasing in prevalence [12]. In addition, the prevalence of insomnia has imposed a huge economic burden on individuals and society [13][14][15]. Thus, it is really necessary to explore the molecular mechanism of insomnia and nd suitable and effective treatments.
The treatments for insomnia disorder mainly consist of pharmacological and non-pharmacological therapies currently [16], among which hypnotics are the most commonly used pharmacological treatment methods as we know. However,long-term use of hypnotics is generally not recommended, due to lack of evidences and kinds of side effects [2]. These side effects mainly include tolerance, dependence, impaired cognitive function and poor quality of life [16,17]. Generally, non-drug therapies have a lower risk of drug-related side effects than most of drug therapies. Hence, patients are more inclined to choose non-drug therapies to improve their sleep conditions [18]. Moreover, the cognitive behavioral therapy for insomnia (CBT-I) which is a standard of non-pharmacological therapy for insomnia, has been underutilized for the scarcity of limited insurance coverage, CBT-I providers, non-responsiveness, and poor compliance [2,19,20]. Therefore, new complementary and alternative therapies with favorable bene t to risk ratio have been considered to be potential options for insomnia patients.
Tuina, a kind of massage methods, is used as an independent method and as an alternative to traditional and Western treatments [21,22]. Previous studies have indicated that Tuina has been successfully applied to various diseases such as neurological diseases, injuries, rheumatism and other types of diseases [23]. Tuina improves some common health problems by stimulating the acupressure points eliminate meridian obstruction and balance the ow of Qi and blood. Tuina has been used to assist in the treatment of muscle pain or stiffness for thousands of years [24][25][26], but there are not wide applications in insomnia. Studies have shown that the insomnia induced by joint de cit of the heart and spleen, treated with acupuncture, moxibustion and Chinese Tuina, provided signi cantly better results than that of patients treated with acupuncture and moxibustion alone [24]. It implies that Tuina may have a therapeutic effect on insomnia.
We tried to treat insomnia through Tuina treatment, and found that it does have a therapeutic effect. In this study, in order to explore the molecular mechanism of Tuina therapy for insomnia, we recruited a group of insomnia patients and healthy people, and collected plasma samples of healthy control group (HC) and insomnia patients group (before Tuina treatment, BTT) and insomnia treatment group (after Tuina treatment, ATT). Subsequently, iTRAQ (isobaric Tags for Relative and Absolute Quantitation) quantitative proteomics and analyses were carried out. We found that Tuina may improve insomnia by regulating immune-related proteins and stress-related proteins. The proteomics veri cation results have been veri ed by ELISA(Enzyme linked immunosorbent assay. All in all, our research not only found a good way to treat insomnia, but also provided a research foundation for improving insomnia. Here we report the results.

Participant characteristics
Our team found that Tuina is effective for treating insomnia in clinical trials. In order to explore the relevant mechanisms deeply, we recruited 27 healthy people and 29 patients with insomnia. The standard deviation of the mean age of the control group participants was 6.3 ± 2.0 months, and the insomnia patients was 6.8 ± 2.1 months.
Afterwards, we performed Tuina therapy for the patients with insomnia, and the effect was also achieved. For patients with insomnia, we took plasma sample before and after the Tuina treatment (BTT, ATT). For the HC group, we took plasma sample without taking any treatment measures, and carried out subsequent experiments using samples from the three groups.
The table lists gender distribution, average age and weight (data not shown). In the iTRAQ analysis, the genders were matched. In the ELISA analysis, the gender distribution and average age between the two groups were similar.
The summery of iTRAQ-based proteomics analysis In this part of the study, we randomly selected 9 BTT and 9 ATT plasma samples from 29 recruited insomnia patients, and 9 plasma samples from 27 recruited healthy controls. Subsequently, the plasma from three persons in the same group was randomly mixed together at a ratio of 1:1:1, and the mixed 9 standard samples (HC-1, HC-2, HC-3; BTT-1, BTT-2, BTT-3; ATT-1, ATT-2, ATT-3) were identi ed by mass spectrometry.
Before analyzing the proteins identi ed from the high throughput assay, we rst looked at the quality of the data that obtained from the mass spectrometry. For the same we looked at these features such as the distribution of unique peptide number, peptide length, the distribution of coverage identi ed, and repeatability using parameters such as coe cient of variation. Unique peptides are de ned as the peptides that are found only for one protein. From the presence of this type of peptide, the existence of the corresponding protein can be uniquely determined. It shows the coordinate distribution of the number of unique peptides contained in all the proteins identi ed in this assay (Fig. 1A). The x-axis is the number of unique peptides that contained in the protein, the left y-axis depicts the number of proteins corresponding to the x-axis and the right y-axis corresponds to the ratio of total protein. For example, there are 325 proteins with 2 as the unique number of peptides, which is 80.65 % of the total number of proteins obtained. To increase the number of unique peptides, this inference could be further made. Next, the distribution of peptide length was analyzed (Fig. 1B). As it could be seen from the gure, the average length of the polypeptide identi ed in the assay was 14.43, which was within a reasonable range of the peptide length. The gure also showed that the length of peptide was mainly concentrated between 7 and 19 with a length of 11 peptides number showing maximum number of peptides. For an identi ed protein, the more peptides that support the protein, the higher the con dence of the protein. Therefore, the identi cation coverage of the protein indirectly re ected the overall accuracy of the identi cation results (Fig. 1C). The different colored sectors in the pie chart represented the percentage of proteins with different ranges of determined coverage. It was clear from the gure that 73.20% of the identi ed proteins had equal to or more than 10% of the peptide coverage, and 58.06% had equal to or more than 20% of the peptide coverage. The 520 proteins were classi ed through Blast2go to assess gene ontology (GO) enrichment. Based on different biological process, these grouped proteins were as follows: metabolic process (8.18%), immune system process (4.24%), cellular component organization or biogenesis (5.72%), response to stimulus (7.66%), biological regulation (8.29%). The common proteins are mainly involved in metabolic, structural, and regulating processes (Fig. 1D).

Exploration of differentially expression proteins (DEPs)
We consider the cut-off value of all iTRAQ ratios as a 1.2-fold change, that is, ratios > 1.2 or < 0.80, and classify proteins as up-regulated or down-regulated, respectively.
In BTT vs HC, 97 DEPs were screened successfully, of which 55 were up-regulated and 42 were downregulated ( Fig. 2) (Table S1). It suggested that in ATT vs HC, there were 46 up-regulated and 46 downregulated DEPs (Table S2). In ATT vs BTT, 63 DEPs were identi ed, including 23 up-regulated and 40 down-regulated DEPs (Table S3). In addition, there were 17 common DEPs were identi ed between all three comparisons (Fig. 2).

Validation of protein identi cation and quanti cation by ELISA
We performed ELISA to analyze the expression levels of C4, C5 and C5b-9 to con rm the results of iTRAQlabeled LC-MS/MS analysis. Compared with the HC group, the expression levels of C4, C5 and C5b-9 in the BTT group and ATT group were decreased. Compared with the BTT group, the expression levels of C4, C5 and C5b-9 were increased in the ATT group (Fig. 4). The result of ELISA supported the result of iTRAQ, which proved that our experiment was very reliable.

Discussion
Sleep is an important part of our daily life and one of the important predictors of people's health [27]. It is a complex biological process which is necessary for optimal neurologic function, as well as systematic biology including metabolism, appetite regulation, immunity, hormonal balance and cardiovascular system [4,5]. Good sleep quality has the capacities to enhance the immune defense of the bodies,while poor sleep status does harm to physical and mental health [28]. Sleep disorders have affected many people in China and even in the world [12]. For example, surveys conducted in Tianjin in 2019 and Liaoning Province in 2013 showed that the incidence of sleep disorders was 13.2% and 11.59%, respectively [29]. Insomnia is a huge problem of the public health, so people have been seeking complementary and alternative therapies all the time, due to the side effects of drug therapy [2,18]. Due to its wide availability, practicality and high compliance, Tuina therapy has been an important treatment in China for thousands of years to help treat muscle pain or stiffness in the body [24,26]. In this study, we used Tuina therapy to treat patients with insomnia and achieved good results. In order to explore the molecular mechanism of Tuina therapy for insomnia, we used iTRAQ quantitative proteomics technology for the rst time and found that the immune-related proteins and stress-related proteins of patients with insomnia have undergone signi cant changes through functional analysis of DEPs.
The sleep and circadian systems have strong regulatory effects on immune functions [30]. Innate immunity is the rst line of defense against tissue damage and microbial infection in the animals [31]. The discovery of reciprocal connections between the central nervous system, sleep and the immune system has suggested that sleep could promote immune defences and the afferent signals from the immune cells might improve sleep. One proposed mechanism by which sleep provides a vital advantage for survival is to support the neurointegrated immune system, which may anticipate the threat of injury and infection [32]. The interaction between sleep and the biological mechanisms of in ammation highlights the impact of sleep dysfunction on the risk of in ammatory diseases. Deciphering the mechanisms is good for development of treatments which regulate in ammation, and improve sleep health [33]. Moreover, the chronic sleep disturbances might be the cause or the consequence of other known triggers of low-grade in ammation, including obesity [34], circadian disruption [35], detrimental lifestyle habits [36], physical inactivity [37], psychosocial in uences [38], stress [39], and low socioeconomic status [40]. These are all factors that should be controlled for in studies linking short or disturbed sleep and in ammatory processes. In BTT vs HC, ATT vs HC and ATT vs BTT, multiple DEPs were associated with immunity. On the one hand, it supports the relationship between insomnia and immune disorders. On the other hand, it suggests that Tuina may affect the immunity and thus the therapeutic effect.
Naturalistic stress exposure, a precipitant of insomnia [41], has long been suggested that insomnia might begin with an organic predisposition towards poor sleep and wake-promoting hyperarousal [42]. Truely, insomnia disorder is always induced by the stressed events, the hyperarousal could interfere with sleep and lead to chronic insomnia in some persons [43]. The literature on stress and sleep is very vast and rich, and its comprehensive analysis deserves further review. In BTT vs HC, ATT vs HC and ATT vs BTT, multiple DEPs were associated with stress. On the one hand, these ndings support that insomnia is related to response to stress. On the other hand, it suggests that Tuina might in uence response to stress, thereby treatment effect.
All in all, we show here that the expressions of the immune-related proteins and stress-related proteins of patients with insomnia are different between the insomnia group and the insomnia treatment group. It is speculated that Tuina may regulate the immune-related proteins and stress-related proteins of patients with insomnia, thereby improving insomnia. These results indicate that Tuina may have the effect of treating insomnia, which will provide a new idea for the future drug-free treatment of insomnia.

Ethical certi cation
This study was approved by the Hubei TCM Regional Ethics Review Committee (TCM), and the ethical approval identi er is HBZY2019-C23-01. All subjects' participation was voluntary and informed consent was obtained. We con rm that all experiments were performed in accordance with relevant guidelines and regulations.

Recruitment of participants
According to the diagnostic criteria of insomnia in previous study [18], 29 participants with a diagnosis of insomnia disorder and 27 control participants with good sleep quantity were recruited. They were between ages 18 and 50 years old without any co-morbid medical or psychiatric disorders. Participants would be excluded if they took any drugs other than contraceptives or received any medication during the study during the study. In addition, each subject needed a statement from a primary care physician, which shows that they had no medical conditions interfering with insomnia disorder. Before the beginning of the experiment, all participants had enough time to decide whether they were like to participate in the experiment. They would be told the details of the study and the bene ts and risks of participating in the trial. Insomnia patients received Tuina treatment, while the healthy control group did not receive any treatment.

Tuina treatment
We treat patients with insomnia based on the previous Tuina method [25,44]. Tuina treatment was performed by a specialized researcher with professional physician certi cate.
Operation method: (1) Operating position: the subject should be asked to relax and take a sitting position, and expose the upper nape line of occiput to the bilateral scapular spine, cervical spinous process to the front edge of the sternocleidomastoid muscle. (2) Manipulation: the operator should stand behind the subject and rub evenly the muscles of the neck and shoulders with the kneading method. The operation time is 5 minutes, and the muscles could be relaxed accordingly. The muscles of the back and neck area should be operated with thumbs and rub, and the plucking should be used to focus on the positive reaction points of abdominal or intermuscular contractures. The operation time is 15 minutes while the time of Tuina operation is 20 minutes. (3) Frequency, course of treatment and time: every day from 8:00 to 12:00, once a day, 20 minutes each time, within 10 days. According to the patient's feedback, and insomnia of the all patients had been improved accordingly.

Sample collection
There were three groups of experimental samples including plasma samples of healthy controls (HC), plasma samples of insomnia patients before Tuina treatment (BTT), and plasma samples of insomnia patients after Tuina treatment (ATT). The fasting plasma samples was taken at 11:00, 1 h after participants relaxed in a temperature-controlled (25 °C) room. Centrifuge immediately after collecting blood on ice in tubes with K3-EDTA or Corvac. The plasma samples were obtained after centrifugation, and stored at -80 °C until further use.

Protein enrichment
The frozen plasma samples from the three groups (BTT (n = 9), ATT (n = 9) and HC (n = 9)) were thawed. Subsequently, the plasma from three persons in the same group was randomly mixed together at a ratio of 1:1:1, and the three experimental groups were mixed in this way to avoid individual errors.
According to the instructions of manufacturer, the pooled plasma samples were pretreated by Proteo-Miner™ protein enrichment kit (Bio-Rad, Hercules, CA, USA) to remove high-abundance proteins and collect the nal eluate. The collected protein solution after depletion was transferred to another centrifuge tube, to which a 4-fold volume of cold acetone was added and kept at -20 °C for 10 h. The acetone protein precipitant was collected through centrifugation, then dried in air and re-dissolved in 8 M urea/100 mM triethylamonium bicarbonate (TEAB) (pH 8.0). The dithiothreitol (DTT) was added at a nal concentration of 10 mM for the reduction reaction for 30 min in at 56 °C, and then the iodoacetamide was added to a nal concentration of 55 mM for an alkylation reaction for 30 min in the dark. A Pierce BCA protein assay kit (Thermo Scienti c, Rockford, IL, USA) was used to measure the concentration of protein samples. In order to evaluate the effect of pretreatment, quality control was performed by SDS-PAGE (sodium dodecyl sulfate-polyacrylamide gel electrophoresis).

iTRAQ quantitative proteomics analysis
We mainly referred to previous experimental methods and made some adjustments [45]. Approximately 100 µg protein from each sample was digested by trypsin. We diluted the protein solution 5 times by 100 mM tetradecyl trimethylammonium bromide (MTEAB), next added trypsin at a ratio of trypsin: protein of 100:1 for enzymatic hydrolysis at 37 °C for 12-16 h. After trypsin digestion, the resulting peptide was Next, the polypeptide solution was added to Durashell C18 column (5 µm, 100 Å, 4.6 x 250 mm) by an Ultimate 3000 HPLC system (Thermo DINOEX, USA). The peptides segment were separated by increasing the concentration of Acetonitrile (CAN) at the ow rate of 1 ml/min under alkaline conditions. The fractions were collected every 1 min, and a total of 42 secondary fractions were collected. Finally, the collected fractions were combined into 15 fractions and dried by vacuum centrifugation.

LC-MS/MS analysis
The TripleTOF 5600 plus liquid chromatography-mass spectrometry (SCIEX, USA) in combination with an Eksigent nanoLC system (SCIEX, USA) was used to capture the mass spectrometry (MS) data. These polypeptide samples were dissolved in the solution comprising 2% acetonitrile / 0.1% formic acid, and subsequently transferred to the C18 capture columns (5 µm, 100 µm x 20 mm). Gradient elution was implemented in a C18 analytical column (3 µm, 75 µm x 150 mm) with a 90 min gradient at a ow rate of 300 nl/min (mobile phase A comprising 0.1% (v/v) formic acid and 2% (v/v) acetonitrile; mobile phase B comprising 0.1% (v/v) formic acid and 80% (v/v) acetonitrile). The Information Dependent Acquisition (IDA) was used to scan the 1st-order mass spectra (MS1) with an ion accumulation time of 250 ms, while the 2nd-order mass spectra (MS2) of 30 precursor ions were collected with an ion accumulation time being 50 ms. The spectra of MS1 and MS2 were captured in the range 350-1,500 m/z and 50 − 2,000 m/z, respectively. In this iTRAQ project, we set the energy of the ion fragmentation at 35 ± 5 eV, while the parent ion dynamic exclusion set was set to half of the peak time (about 15 s).

Data analysis
ProteinPilot V4.5 software was employed for proteome identi cation and iTRAQ quanti cation. For proteome identi cation, Uniprot/Swiss-Prot database of human species (downloaded in June 2020) was used. Other parameters were set as follows: instrument, TripleTOF 5600; cysteine alkylation, iodoacetamide; digestion, trypsin; biological quali cations including ID focus and trypsin digestion; the Quantitate, Bias Correction and Background Correction was checked for protein quanti cation and normalization. Only the proteins with at least one unique peptide and unused value ≥ 1.3 (credibility ≥ 95%) was used for the further analyses.
The pairwise comparisons method between biological replicates was standardized as the ratios, and the smallest ratio was used as p-values to explore the differentially expressed proteins (DEPs) under student's t test (two-tailed or unpaired). For the determination of DEPs, the fold changes (FC) were calculated as the average comparison pairs among biological replicates, and the proteins with FC 1.2 and p < 0.05 were considered to be the DEPs .

Functional analysis
Identi ed DEPs were submitted to Gene Ontology (GO) [46] Terms (http://geneontology.org/) for classi cation, by which the DEPs could be assigned into three branches: molecular function (MF), biological process (BP), and cellular components (CC). In addition, the Pathway enrichment analysis was implemented by the Kyoto Encyclopedia of Genes and Genomes (KEGG) [47] (https://www.kegg.jp/). The enrichments were checked statistically with Fisher's exact test, and those with p-values ≤ 0.01 were considered to be statistically signi cant.

Declarations Author Contribution
Yan Zhao conceived of the study, and participated in its design. Jing Zhou participated in the design of the study and coordination and drafted the manuscript. Bi wei Cao participated in the design of the study assessed treatment effectiveness and collected clinical data. Meng Wei participated in the design of the study and performed the statistical analysis. Yuan Xiong interviewed patients before and after treatment and collected clinical data. Wan Liu recruited and screened eligible participants from outpatient department, and assigned patients to either massage group or control group. Li Zhu participated in trial design and helped to prepare the manuscript. All authors read and approved the nal manuscript.

Funding
The experiment was funded by the Hubei Provincial Science and Technology Project (approval number 2017ACA192). The funding resource had no role in study design; data collection, analysis and interpretation of data; writing of the report; or in the decision to submit the article for publication.