Histone phosphorylation analysis of two main TCM syndromes of chronic fatigue syndrome Authors: Tingting Xu, Shuo Gao, Xinxin Cheng, Wenxuan Man, Yanjie Wang, Yonghui Yin Published: 2025 ABSTRACT AbstractObjectiveThis study aimed to investigate differences in protein expression, protein function, and epigenetic modifications related to histone phosphorylation in peripheral blood mononuclear cells among patients with chronic fatigue syndrome (CFS) presenting two common Traditional Chinese Medicine (TCM) syndrome types compared with healthy controls.MethodsThree groups of participants were selected from the Medical Examination Center of the Affiliated Hospital of Shandong University of Traditional Chinese Medicine between January 2024 and December 2024. The first group consisted of 10 CFS patients with the Qi and Blood Deficiency syndrome pattern; the second group included 10 CFS patients with the Liver Depression and Spleen Deficiency syndrome pattern; and the third group comprised 10 healthy controls. Phosphopeptide enrichment analysis was employed to detect and identify differentially expressed proteins, followed by Gene Ontology (GO) enrichment analysis.ResultsDifferences were observed between both TCM syndrome types of CFS (Qi and Blood Deficiency, Liver Depression and Spleen Deficiency) and the healthy control group in various biological processes and cellular components, including intracellular signal transduction, actin filament bundling, and cortical cytoskeleton organization.ConclusionMultiple processes, structures, and functions—such as intracellular signaling, presynaptic modulation of chemical synaptic transmission, actin filament bundling, cortical cytoskeleton organization, SH3 domain binding, and profilin binding—are associated with the occurrence and development of CFS.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12967-025-07579-9.Keywords: Chronic fatigue syndrome, Histone phosphorylation, TCM syndrome patterns, GO enrichment analysis, KEGG enrichment analysis Introduction Chronic fatigue syndrome (CFS) is a multisystem condition characterized by prolonged fatigue lasting ≥ 6 months, with cognitive dysfunction (including memory deficits), widespread pain, and neurological or psychiatric symptoms. Increasing societal stressors have recently been associated with a progressive increase in CFS, establishing it as a significant global public health concern. Epidemiological studies have recognized CFS as a multifactorial disorder that requires urgent interdisciplinary attention. The epidemiological studies estimate the CFS global prevalence to be between 0.2% and 0.4%, affecting approximately 17 million individuals globally. Country-specific data reveal notable burdens, with Spain reporting over 40,000 confirmed cases [1]. U.S. epidemiological studies have revealed sex-disaggregated patterns in fatigue-related morbidity, with unremitting fatigue symptoms present in 14% of the male and 20% of the female populations. Notably, longitudinal data indicate that approximately 1 in 8 individuals with persistent fatigue progress to meet the diagnostic criteria for CFS [2, 3]. To date, no specific laboratory biomarkers have been established for CFS, which frequently presents multimorbidity and various chronic conditions. Although the precise etiology and pathological mechanisms remain unclear in current medical research, traditional Chinese medicine (TCM) and Western medicine recognize CFS as a multisystem disorder involving dysregulation across multiple organ systems. Current evidence suggests that CFS pathogenesis involves complex interactions among neuroendocrine dysfunction, immune dysregulation, and metabolic disturbances [4, 5]. Additionally, emerging evidence suggests that these pathological mechanisms may be altered by epigenetic modifications [6]. These epigenetic modifications demonstrate substantial associations with intricate pathophysiological alterations observed in CFS, suggesting a contribution to disease onset and progression. Wilfred C de Vega et al. [7]. have recently identified aberrant DNA methylation patterns in patients with CFS. These findings highlight the importance of examining histone modifications to elucidate the CFS epigenetic etiology. Recent advancements in epigenetic technologies and protein modification analyses have led to the identification of an expanding array of pathogenic factors and genetic associations in CFS. A recent study demonstrates that epigenetic mechanisms, including miRNA expression, DNA methylation, and histone modifications, are vital for modulating CFS pathogenesis by dysregulating immune responses, oxidative stress pathways, neural function, and hormonal balance. Consequently, it is vital to detect common downstream pathways that elucidate CFS pathogenesis and discover universal therapeutic targets that may alleviate fatigue-related symptoms. Given this gap in the literature, this study aimed to enhance the knowledge of CFS etiology. The diagnostic criteria issued by the U.S. National Academy of Sciences, Engineering, and Medicine in 2016 require, patients must exhibit persistent or recurrent severe fatigue lasting at least 6 months accompanied by post-exertional malaise and unrefreshing sleep, with at least one additional symptom from the following: Cognitive impairment (including “brain fog”) or orthostatic intolerance (characterized by symptom exacerbation upon standing) [8]. CFS is classified into multiple distinct patterns in TCM, each characterized by unique clinical manifestations and underlying pathological mechanisms. Considering the complex CFS etiology, it is imperative to classify these distinct TCM patterns, as we hypothesize that each pattern may exhibit distinctive epigenomic underpinnings. Chen Gang et al. [9] conducted syndrome differentiation analysis in 40 patients with CFS, identifying 16 patients with dual deficiency of Qi and blood, 13 with liver depression-spleen deficiency, and 11 with liver-kidney insufficiency. According to the Clinical Research Guidelines for New Chinese Medicines in Treating Chronic Fatigue Syndrome [10], the TCM syndrome diagnostic criteria classify CFS into six patterns: Qi-blood deficiency, spleen-kidney Yang deficiency, marrow vacuity, liver depression-spleen deficiency, liver-kidney Yin deficiency, and liver fire hyperactivity. The diagnostic criteria for syndrome differentiation were established based on the Guiding Principles for Clinical Research of New Chinese Drugs (2002) and The Clinical Research Guidelines for New Chinese Drugs in Treating Chronic Fatigue Syndrome [10], along with our supervisor’s clinical expertise in tongue and pulse diagnosis. The four TCM patterns identified were Qi-blood deficiency, liver depression-spleen deficiency, spleen-stomach Yang deficiency, and essence-marrow depletion. The diagnostic criteria for TCM syndrome differentiation are presented in Table 1. A case was defined as positive if it fulfilled one of the following criteria: Three or two major symptoms coupled with two or more minor symptoms. This clinical assessment was integrated with an examination of the tongue and pulse features. Subsequently, the two most frequently observed syndromes were compared. To minimize confounding factors from comorbid conditions, a thorough physical examination was performed on all potential subjects. This screening led to the exclusion of individuals with identifiable underlying diseases or a history of conditions known to cause fatigue. Finally, 92 eligible patients with a confirmed diagnosis of CFS were included in the study. The examination of 92 patients with CFS revealed two predominant TCM syndrome patterns: Qi-blood deficiency syndrome (54 patients, 58.7%) and liver depression-spleen deficiency syndrome (27 patients, 29.3%). Therefore, we allocated 10 patients each from healthy controls (group J), patients with CFS and Qi-blood deficiency syndrome (group Q), and patients with CFS and liver depression-spleen deficiency syndrome (group G) for histone phosphorylation analysis to study the differential histone expression patterns among these groups. Table 1Diagnostic criteria for traditional Chinese medicine syndrome differentiationQi-blood deficiencyLiver depression-spleen deficiencySpleen-stomach yang deficiencyEssence-marrow depletionMain symptomsClinical manifestations include mental fatigue, persistent dull headache (aggravated by exertion), dizziness, blurred vision, insomnia, palpitations, spontaneous sweating, and shortness of breath.Patient presents with fatigue, chest tightness, poor appetite (anorexia), muscle/joint weakness, emotional depression or irritability, and frequent sighing.Patient exhibits epigast-ric distension or dull pain (alleviated by pressure and warmth), loose stools, and fatigue.Patient presents with dizziness, tinnitus, insomnia, poor memory, difficulty concentrating, memory decline, lethargy with desire to lie down, and soreness/weakne-ss in the lumbar region and knees.Secondary symptomsPatient exhibits poor appetite, pale complexion, and numbness in the extremities.Patient exhibits epigastric/hypochondriac distension and pain, postprandial bloating, reduced food intake, loose or alternating stools, borborygmi/flatule-nce, and diarrhea with pain relief following defecation.Patient exhibits reduced food intake, shortness of breath, reluctance to speak, vomiting of clear fluid, and a bland taste in the mouth.Patient exhibits cold intolerance and cold limbs, night sweats with tidal fever, and premature graying of hair.Tongue and pulse manifestationsTongue manifestation exhibits a pale tongue body and thin white coating, while the pulse is deep, thready, and weak.Tongue presentation exhibits white or greasy coating, with a wiry or thready pulse.Tongue appears pale with teeth marks, and the pulse is thready and weak.Tongue appears pale with a white coating, and the pulse is deep and thready. Diagnostic criteria for traditional Chinese medicine syndrome differentiation This study aimed to identify potential common pathways contributing to physical fatigue through gene ontology (GO) and Kyoto encyclopedia of genes and genomes (KEGG) analyses of CFS-associated proteins and genes. Based on the identified CFS-associated genes, we identified several common pathways across different TCM-based CFS subtypes. The advancing understanding of CFS epigenetics will clarify the etiological basis of these subtypes, providing crucial information for diagnosis, disease evaluation, and development of potential therapeutic strategies. Methods TiO2 enrichment Reagents and consumables for TiO2-based phosphopeptide enrichment Trifluoroacetic acid (TFA): ACROS (Cat # 139721000, USA), acetonitrile (ACN): MERCK (Cat # UN1648, USA), lactic acid: Sigma (Cat # L6661-100ML, USA), TiO2 beads: Titansphere (Cat # 5020–75000, Japan). Experimental procedure for TiO2 phosphopeptide enrichment The experimental procedure for TiO2 phosphopeptide enrichment involved the following steps: First, TiO2 beads were prepared by suspending them in binding buffer (80% ACN/6% TFA), followed by centrifugation at 1,000 × g to remove the supernatant. This washing step was repeated twice. The peptide samples were subsequently dissolved in binding buffer and centrifuged at 14,000 × g for 15 min to remove insoluble material. Subsequently, the supernatant was transferred to the pre-washed TiO2 beads for phosphopeptide binding through vertical mixing at room temperature for 30 min. Following centrifugation at 2,000 × g for 2 min, the supernatant was then sequentially transferred to fresh TiO2 beads, and the binding process was repeated three times to ensure maximal phosphopeptide capture. The enriched TiO2 beads were thoroughly washed with binding buffer and subsequently loaded onto a C18 desalting column. The bound phosphopeptides were eluted using a stepwise ACN gradient (6%, 9%, 12%, 15%, 18%, 21%, 25%, 30%, 35%, and 50%) in alkaline buffer. The eluted fractions were pooled into three groups and lyophilized for subsequent mass spectrometry analysis. Critical quality control measures were implemented, including the use of a β-casein standard to verify > 90% phosphopeptide recovery and parallel processing of Qi-blood deficiency and liver-spleen deficiency samples to prevent cross-contamination. Data-dependent acquisition (DDA) mass spectrometry analysis Mobile phases were prepared by dissolving 0.1% formic acid in either 100% water (mobile phase A) or 80% ACN (mobile phase B). The lyophilized peptide samples were reconstituted in 10 µL of mobile phase A, centrifuged at 14,000 × g for 20 min at 4 ℃, and 400 ng of the supernatant was injected for liquid chromatography tandem mass spectrometry (LC-MS/MS analysis). Chromatographic separation was performed in DDA mode using a timsTOF_HT mass spectrometer that was equipped with a Captive Spray ion source, under the gradient conditions as indicated in Table 1. The mass spectrometer was configured with an m/z scan range of 100–1700, MS1 resolution of 60,000 (at m/z 1222), and ion mobility separation (0.6–1.6 cm2/[V·s]) during an accumulation time of 100 ms. Additional parameters comprised a capillary voltage of 1.6 kV, a total cycle time of 1.1 s, and 10 PASEF MS/MS scans per cycle. These raw mass spectrometry data (.d files) were generated for subsequent processing. The separation was performed at a flow rate of 500 nL/min with the gradient program as outlined in Table 2. Table 2Separation gradientTime (min)Mobile phase B (%)00:003.517:003218:009520:009521:00122:001 Separation gradient Data-independent acquisition (DIA) mass spectrometry analysis The mobile phases were prepared as follows: Mobile phase A consisted of 0.1% formic acid in water (100% aqueous), while mobile phase B contained 0.1% formic acid in 80% ACN. The lyophilized powder was reconstituted in 10 µL of mobile phase A and centrifuged at 14,000 × g for 20 min at 4 ℃ for sample preparation. A 400-ng aliquot of the supernatant was injected for LC-MS/MS analysis under the chromatographic conditions outlined in Table 1. Mass spectrometry was performed in DIA mode using a timsTOF_HT instrument that was equipped with a Captive Spray ion source. The mass spectrometer was operated with the following parameters: MS scan range of m/z 300–1500, MS1 resolution of 60,000 (at m/z 1222), TIMS accumulation time of 50 ms, ion mobility range of 0.70–1.30 cm2/(V·s), and capillary voltage of 1.5 kV. The total cycle time was set to 1.23 s, and raw data were acquired in .d format. The separation was conducted at a flow rate of 300 nL/min using the gradient program specified in Table 3. Table 3Separation gradientTime (min)Mobile phase B (%)00:00504:001046:002453:003654:009560:0095 Separation gradient Data collection of proteins associated with CFS The principles of database selection were as follows: for sequenced species, we directly used their dedicated databases; for non-sequenced species, we selected the most relevant higher-level proteomic databases. The database used in this study was Homo sapiens SP (protein count: 20434; database: UniProt; download date: 2024.03.07). Results Baseline demographic and clinical characteristics A total of 20 patients met the diagnostic criteria for CFS, and 10 healthy volunteers were enrolled in the study. Based on TCM syndrome differentiation, patients with CSF were categorized into two groups: Qi and blood deficiency syndrome group (Q group, n = 10) and liver depression and spleen deficiency syndrome group (G group, n = 10). The cohort of healthy controls was designated as the J group (n = 10). The comparative analysis of demographic characteristics demonstrated statistically non-significant differences in gender or age among the three groups (all p > 0.05), thus establishing comparability at baseline (Table 4). Table 4Baseline demographic and clinical characteristicsGroupAge (years), Mean ± SDGender (Male/Female), nQi and Blood Deficiency (Q)30 ± 9.234/6Liver Depression-Spleen Deficiency (G)32.2 ± 8.744/6Healthy Control (J)39.2 ± 11.15/5P value>0.05>0.05 Baseline demographic and clinical characteristics GO characterizes genes and proteins in various species across three domains: Biological processes (BP), cellular components (CC), and molecular functions (MF). This study used enrichment analysis to identify overrepresented MF, CC, and BP (significant enrichment) in CFS-associated protein lists compared to random protein datasets. We conducted an enrichment analysis of differentially expressed proteins using the Saccharomyces cerevisiae database. The enrichment fold change, p-value, and list of CFS-related genes associated with each term were presented with the significant terms associated with these proteins. The analysis was classified in descending order of enrichment fold change, with only terms with statistically significant associations (p < 0.05) being retained. The p-values reflect the probability of randomly observing proteins annotated to a specific term within the gene list. Comparisons were made among three groups: Healthy controls, patients with CSF and Qi-blood deficiency pattern, and patients with CSF and liver depression-spleen deficiency pattern. Compared to healthy controls, patients with CSF and Qi-blood deficiency pattern exhibited 223 differentially expressed proteins (132 upregulated and 91 downregulated; Fig. 1a). The liver depression-spleen deficiency group exhibited 351 dysregulated proteins (169 up- versus 182 downregulated; Fig. 1b). Typically, 188 differential proteins were identified between the two CFS subtypes (80 upregulated compared to 108 downregulated; Fig. 1c). Volcano plot of differentially expressed proteins. a between healthy controls and patients with CSF and Qi-blood deficiency syndrome. b between healthy controls and patients with CSF and liver depression-spleen deficiency syndrome. c between patients with CSF and Qi-blood deficiency syndrome and those with liver depression-spleen deficiency syndrome Comparative phosphoproteomic analysis exhibited distinct phosphorylation patterns among the CFS subtypes. The liver depression-spleen deficiency group presented the most pronounced alterations, with 406 differentially phosphorylated sites (198 upregulated and 208 downregulated) compared to healthy controls. Conversely, the Qi-blood deficiency group indicated 242 differential sites, with a predominance of upregulated phosphorylation events (148 upregulated versus 94 downregulated). The direct comparison between the two CFS subtypes demonstrated 201 differentially phosphorylated sites, with a greater number of downregulated sites (114 downregulated compared to 87 upregulated; Fig. 2). Statistics of differential sites among healthy controls, patients with CSF and liver depression-spleen deficiency syndrome, and patients with CSF and Qi-blood deficiency syndrome GO enrichment analysis of CFS-associated proteins The proteins were analyzed based on their associated BP, CC, and MF, as defined by GO classification. The comparative GO analysis between healthy controls and patients with CSF and Qi-blood deficiency pattern revealed significant molecular alterations (Fig. 3a). The most enriched BP was “intracellular signal transduction” (GO:0035556), involving protein products encoded by MAST3, PRKAR1A, MARK2, and 19 additional signaling-related genes. Among the CC, the “actin filament bundle” (GO:0032432) exhibited the highest differential expression (Fig. 3b), comprising structural proteins derived from FLNA, LCP1, and RFLNB. The top MF was “SH3 domain binding” (GO:0017124), which was mediated by adaptor proteins translated from PTPN12, HCLS1, LYN, and eight other genes involved in protein-protein interactions (Fig. 3c). GO term enrichment analysis of biological processes across the three groups The comparative GO analysis between healthy controls and patients with CSF and liver depression-spleen deficiency pattern identified distinct functional alterations (Fig. 4a). The most significantly enriched BP was “intracellular signal transduction” (GO:0035556), involving genes, including AKAP11, MAP4K1, and PAK2. “Cortical cytoskeleton” (GO:0030863) revealed the highest differential expression among CC (Fig. 4b), with key genes including CAPZB, FLNA, and ACTB. The top MF was “profilin binding” (GO:0005522), which was mediated by proteins translated from VASP, EVL, CTTN, HTT, and WIPF1 (Fig. 4c). GO term enrichment analysis of cellular components across the three groups Comparative GO analysis between patients with CSF and Qi-blood deficiency patterns and those with liver depression-spleen deficiency patterns revealed subtype-specific molecular signatures (Fig. 5a). The most enriched BP was “presynaptic modulation of chemical synaptic transmission” (GO:0099171), involving protein products encoded by GIT1, PRKCB, and ADRA2A. Among CC, “cytoskeleton” (GO:0005856) highlighted the highest differential expression (Fig. 5b), containing proteins derived from ACTB, CTTN, CAPZB, and 18 additional cytoskeletal genes. The top MF was “diacylglycerol-dependent serine/threonine kinase activity” (GO:0004697), which was mediated by kinase proteins translated from PRKCB, PRKCA, PRKCD, PKN1, and PKN3 (Fig. 5c). GO term enrichment analysis of molecular functions across the three groups Pathway enrichment analysis results KEGG enrichment analysis is a widely adopted pathway analysis method that enables the systematic annotation of differentially expressed proteins to KEGG pathways. This approach identifies and characterizes the primary metabolic and signal transduction pathways associated with these proteins. The comparative analysis between healthy controls and patients with CSF and Qi-blood deficiency pattern revealed the “Notch signaling pathway” (hsa04330) as the most significantly enriched pathway (Fig. 6a), involving the genes NUMB, HDAC1, SNW1, ADAM17, and HDAC2. KEGG pathway analysis across the three groups In the comparison of healthy controls and patients with CSF and liver depression-spleen deficiency pattern, the “Spinocerebellar ataxia pathway” (hsa05017) exhibited the highest enrichment (Fig. 6b), comprising the genes SP1, VDAC3, TRAF2, PSMD11, ITPR1, PSMA3, PUM2, PRKCB, NOP56, VDAC2, and PRKCA. The analysis between the two CFS subtypes (Qi-blood deficiency versus liver depression-spleen deficiency) identified the “serotonergic synapse pathway” (hsa04726) as the top-enriched pathway (Fig. 6c), with BRAF, ITPR1, PRKCB, PLA2G4A, and PRKCA involvement. Discussion This study evaluates CFS pathophysiology by characterizing disease-associated proteins within their biological contexts, which has potential implications for diagnostic biomarker development. The “intracellular signal transduction” pathway (GO:0035556) was identified as the primary enriched functional term in the comparison between healthy controls and patients with CSF and Qi-blood deficiency patterns, and the most significantly enriched term in the analysis of healthy controls versus patients with CSF and liver depression-spleen deficiency patterns. The intracellular signal transduction enables cells to convert extracellular signals, including hormones, growth factors, and neurotransmitters, into specific intracellular responses. These responses orchestrate critical cellular events, including alterations in gene expression, metabolic adaptation, or the initiation of proliferation. Notably, this process is critically implicated in the immune function and neuroendocrine pathway dysregulation, contributing to associated diseases. Intracellular signal transduction is a basic BP through which cells convert extracellular stimuli, including hormones, growth factors, and neurotransmitters, into specific intracellular responses, thereby regulating gene expression, metabolic activity, and cellular proliferation. Dysregulation of this process is critically implicated in the pathogenesis of immune and neuroendocrine abnormalities [11]. As early as 2005, Kaminska B. et al. revealed that impaired p38 mitogen-activated protein kinase (MAPK) phosphorylation is mechanistically associated with CFS pathogenesis and other chronic inflammatory disorders [12]. Reduced MAPK phosphorylation may impair immune function by disrupting cytokine translation, cytolytic granule polarization, and lytic protein release, which are key processes mediated by MAPK signaling in fundamental immune responses. Furthermore, Huth T.K. et al. (2016) demonstrated that reduced MAPK kinase phosphorylation in natural killer (NK) cells from patients with CFS significantly impairs intracellular MAPK signaling and consequently reduces the synergistic cytotoxic capacity of NK cells against target cells [13]. In neuroendocrine signaling pathways, substantial evidence indicates that patients with CFS frequently exhibit hypothalamic-pituitary-adrenal (HPA) axis hypofunction, resulting in aberrant cortisol levels [14]. Pro-inflammatory cytokines activate cell surface receptors to initiate intracellular signaling pathways, including nuclear factor kappa B (NF-κB) and p38 MAPK. These pathways directly stimulate the HPA axis. However, persistent stimulation disrupts the regulatory feedback mechanisms, leading to HPA axis hypofunction. The dysregulation of intracellular signal transduction is critically implicated in metabolic disorder pathogenesis. Specifically, impaired sensitivity of the insulin signaling pathway and its downstream PI3K/AKT/mTOR cascade disrupt glucose uptake into the muscle and adipose tissue. This disruption in cellular energy metabolism manifests systemic fatigue and weakness, which are hallmark features of CFS [15, 16]. Consequently, intracellular signal transduction represents a critically dysregulated BP in CFS pathogenesis. The subsequently identified enriched functional term was GO:0032432, describing the CC “actin filament bundle.” To date, only the mitochondrial structure has been specifically demonstrated to be associated with CFS pathogenesis [17]. Actin filament bundles, which are essential structural components for maintaining cellular morphology, motility, and intracellular transport, are particularly crucial for muscular function, forming a core structural foundation for striated muscle (including skeletal and cardiac muscle) contraction. Specifically, patients with CFS exhibit disrupted organization and dynamic interactions of actin filament bundles, which directly impair proper muscle contraction-relaxation cycles [18]. In immunological research on CFS, Günther et al. revealed that cytotoxic lymphocytes (for example, NK cells) from patients with CFS fail to polarize actin toward the immunological synapse upon target cell contact, resulting in impaired immune synapse formation and consequent dysfunction in cytotoxic activity [19, 20]. In metabolic diseases, a vital step in insulin-stimulated glucose uptake by muscle and adipose cells is glucose transporter glucose transporter type 4(GLUT4) translocation from intracellular vesicles to the plasma membrane. GLUT4-containing vesicles are transported to the subplasmalemmal region through a process dependent on actin cytoskeleton reorganization, where they subsequently dock and fuse with the plasma membrane. Actin cytoskeletal organization disruption impairs GLUT4 vesicle trafficking, leading to defective glucose uptake and subsequently contributing to the manifestation of systemic fatigue [21, 22]. Additionally, insulin secretion from pancreatic β-cells relies on vesicular trafficking and exocytosis, which require dynamic reorganization of the actin cytoskeleton. Abnormal actin dynamics disrupt the pulsatile insulin secretion pattern, compromising its physiological regulation [23]. The identified GO term, GO:0017124, corresponds to the generic functional annotation “SH3 domain binding.” This GO term includes the molecular recognition and binding of proline-rich peptide motifs in target proteins, a process critical for signal transduction pathways regulating diverse functions, including protein metabolism, glycolytic modulation, and mitochondrial energetics. Dysfunctional SH3 domain-mediated interactions may impair cellular energy supply efficiency [24]. The SH3 domain is functionally associated with immune cell dysfunction in CFS through its critical involvement in the PI3K/AKT signaling pathway. This pathway is a central intracellular hub that integrates inputs from diverse receptors, including those for growth factors, cytokines, and insulin, to regulate glucose metabolism and mitochondrial function. Furthermore, PI3K/AKT activation modulates immunometabolism by enhancing T and NK cell activity. Notably, PI3K/AKT pathway activation is highly dependent on protein interactions mediated by the SH3 domain [25]. Therefore, SH3 domain-mediated interactions may represent a critical pathogenic mechanism underlying CFS development. The cortical cytoskeleton (GO:0030863) was identified as the primary enriched CC in the comparison between healthy controls and patients with CFS and the liver depression-spleen deficiency pattern. The cortical cytoskeleton is a specialized cytoskeletal network composed of actin filaments, microtubules, and associated binding proteins (including myosin), which collectively maintain cell shape and mediate mechanical responses. Consequently, aberrant reorganization of the cortical cytoskeleton also disrupts immunological synapse formation, forming a link with the classical immunophenotype of chronic fatigue syndrome with reduced NK cytotoxicity. Studies demonstrate that during sustained muscle contraction, the cortical cytoskeleton buffers mechanical tension to safeguard sarcolemmal integrity when muscle fibers undergo stretching. The structural abnormalities in the cortical cytoskeleton may increase sarcolemmal permeability, leading to the aberrant release of muscle contraction signals and the exacerbation of contractile dysfunction [26]. Although no direct studies have established an association between the cortical cytoskeleton and neuroendocrine dysregulation in CFS, a plausible hypothesis can be advanced based on established biology: The precise regulation of the HPA axis relies on intricate neural circuits within the hypothalamus and neuron-glia interactions. These processes, including dendritic spine morphology, synaptic plasticity, and ensheathment of blood vessels by astrocytic end-feet, are all dynamically driven and supported by the actin cytoskeleton [27]. In the context of metabolic dysfunction, the cortical cytoskeleton, whose core architecture is composed of actin filament bundles, shares an analogous mechanistic basis. Dysregulated reorganization of this cytoskeletal network impairs the GLUT4-containing vesicle fusion with the plasma membrane, preventing their proper incorporation [21, 22]. Furthermore, cortical cytoskeleton integrity facilitates the downstream insulin signaling molecule organization into functional complexes. Conversely, disruption of this architecture disperses these molecules, attenuates insulin signaling, and exacerbates systemic insulin resistance [28]. The subsequent GO term identified was the general functional term GO:0005522, describing “profilin binding. “Profilin regulates cytoskeletal dynamics and immune cell function, critically governing cell motility and activation processes. In immunological studies, dysfunctional profilin activity impairs immune cell migration and attenuates immune response efficacy, leading to pro-inflammatory cytokine dysregulation and subsequent exacerbation of fatigue symptoms in patients with CFS [29]. Profilin is a critical regulatory factor that promotes actin polymerization. Profilin orchestrates actin filament assembly and synaptic maintenance through dynamic interactions with actin monomers and signaling molecules. Its functional impairment directly compromises synaptic stability and neurotransmission, thereby exacerbating neurocognitive fatigue, which is a core symptom of CFS, potentially positioning it as a critical node in CFS pathogenesis [30]. In pancreatic β-cells, glucose stimulation induces a reorganization of the cortical actin network, which allows insulin vesicles to dock and fuse with the plasma membrane for insulin release. This process is primarily regulated by Rho family GTPases. Profilin, a direct downstream effector of these GTPases, regulates insulin secretion [31]. The first significantly enriched biological function term identified was GO:0099171 (presynaptic regulation of chemical synaptic transmission) for Qi-blood deficiency and liver-Qi stagnation with spleen deficiency subtypes of CFS. The presynaptic modulation of synaptic transmission involves facilitation and inhibition, mediated through ion channel regulation, second messenger systems, and control of the readily releasable vesicle pool. These mechanisms are fundamentally associated with physical and cognitive fatigue. Presynaptic modulation directly influences central nervous system control over somatic functions by regulating neurotransmitter release at presynaptic terminals. Prolonged neural activity may trigger presynaptic modulation mechanisms that alter neurotransmitter release from presynaptic terminals, leading to reduced motor coordination, diminished force output, and exercise-induced fatigue [32]. The cytoskeleton is vital for the polarized cytotoxic granule transport to the secretory domain of the immune synapse. Dysregulation of this regulatory mechanism directly compromises the cytotoxic function in patients [19, 20]. In the framework of presynaptic modulation in chemical synaptic transmission, seminal research has implicated serotonin (5-HT) in fatigue mechanisms. A study by Florence Cotel and colleagues demonstrated that following repetitive muscle contraction, serotonergic neurons in the spinal cord exhibit sustained firing. This leads to a massive release and subsequent spillover of serotonin into the synaptic cleft. The serotonin binding to extrasynaptic receptors opens potassium channels, resulting in the inhibition of action potential initiation and reduced neuronal excitability. Consequently, upon receiving motor commands from the brain, impaired signal transmission in motor neurons fails to effectively relay excitation to muscle fibers, leading to diminished contraction force and systemic fatigue [33]. Reduced glutamate release from presynaptically inhibited glutamatergic synapses in cortical circuits decreases neuronal excitability, resulting in cognitive fatigue manifestations, including impaired attention and slowed processing speed [34]. Consequently, presynaptic modulation of chemical synapses is a critical BP implicated in CFS pathophysiology. The cortical cytoskeleton (GO:0030863) was identified as the primary enriched CC in the comparison between healthy controls and patients with CSF and liver depression-spleen deficiency pattern. The cortical cytoskeleton contributes to fatigue pathogenesis through the dynamic coordination of microfilament systems, regulation of neural signal transmission and metabolism, and energy-dependent maintenance of cytoskeletal homeostasis [35]. A previous study by Günther’s team demonstrated that actin cytoskeleton polarization failure at the cell-cell contact site impairs the efficient assembly of the immunological synapse, thereby compromising cytotoxic immune function in patients [19, 20]. Although no direct evidence currently associates cytoskeletal abnormalities with neuroendocrine dysregulation in CFS, existing studies demonstrate that corticotropin-releasing hormone (CRH) neurons in the hypothalamus and endocrine cells in the pituitary gland rely on vesicular secretion for hormone release, a process precisely orchestrated by the cortical actin cytoskeleton. Consequently, cytoskeleton dysfunction in hypothalamic or pituitary cells could disrupt pulsatile hormone secretion patterns or lead to insufficient hormone release. This mechanism provides a plausible cellular explanation for the HPA axis dysregulation observed in patients with CFS [36]. In contrast to the roles of actin filament bundles and the cortical cytoskeleton, cytoskeletal stability disruption directly compromises the contractile capacity of muscle fibers, leading to diminished force generation and premature fatigue [37]. These mechanisms establish an essential association between cytoskeletal abnormalities and physical fatigue development. The subsequently identified GO term was GO:0004697, which describes “diacylglycerol-dependent serine/threonine kinase activity. “This GO term describes protein kinase C (PKC) as a signal transducer that senses intracellular diacylglycerol (DAG) concentration changes and regulates cellular processes, including proliferation, differentiation, and metabolism. A key hub of its function involves regulating phosphorylated actin-binding proteins to orchestrate actin filament reorganization, thereby facilitating cell migration and morphological changes. This GO term defines kinases that monitor intracellular diacylglycerol concentrations to participate in cellular signaling cascades, thereby critically regulating proliferation, differentiation, and metabolic processes. These enzymes function as central hubs that control the phosphorylation state of actin-binding proteins to direct actin filament reorganization, thereby facilitating cellular migration and morphological plasticity [38]. Furthermore, various neurotransmitters and hormones, including vasopressin and catecholamines, activate the phospholipase C/PKC pathway in the hypothalamus and pituitary gland to stimulate CRH release. Aguilera G et al. have identified PKC signaling as a key intracellular pathway regulating HPA axis activity, suggesting that HPA axis dysfunction in CFS may originate from altered sensitivity or aberrant output of this pathway within critical neuroendocrine cells. In type 2 diabetes, significantly elevated intracellular DAG levels lead to persistent PKC activation, which disrupts insulin signaling cascades and promotes insulin resistance. Furthermore, DAG-PKC pathway activation has been implicated in various vascular abnormalities [39, 40]. Consequently, diacylglycerol-dependent serine/threonine kinase activity may be a critical pathogenic node in CFS development. We identified several significantly enriched KEGG pathways, including hsa04330 (notch signaling pathway), hsa05017 (spinocerebellar ataxia), and hsa04726 (serotonergic synapse). The Notch signaling pathway is an evolutionarily conserved transduction mechanism vital for cellular proliferation, differentiation, and apoptosis in multiple cell types [41]. A study has demonstrated that Notch signaling pathway dysregulation can disrupt muscle satellite cell function, leading to impaired muscle repair, premature satellite cell depletion through excessive differentiation, and reduced muscular endurance with exacerbated post-exertional fatigue [42]. In neuroscience, Notch signaling is a key regulator of the hypothalamic neuroendocrine cell pool. Subtle alterations in Notch signal strength may affect the development and maintenance of core CRH neurons governing the HPA axis, thereby potentially modulating stress response [43, 44]. In the context of metabolic regulation, Notch signaling is aberrantly activated in adipose tissue under obese conditions. This hyperactivation promotes adipose tissue inflammation, suppresses normal adipocyte differentiation, and compromises adipocyte function, thereby driving systemic insulin resistance [45]. These findings provide direct experimental validation of our results (Fig. 7). Notch signaling pathway The spinocerebellar ataxia pathway (hsa05017) is clinically characterized by gait instability, limb ataxia, and dysmetria, where cerebellar dysfunction leads to impaired precision in muscle contraction regulation (Fig. 8), resulting in pathological muscle overactivation and significantly increased energy expenditure [46]. The relationship between spinocerebellar ataxia (SCA) and immune function remains unclear. In the neuroendocrine domain, multiple studies have confirmed the high prevalence of autonomic dysfunction in patients with SCA, manifesting as impaired cardiovascular reflexes, orthostatic hypotension, gastrointestinal dysmotility, and sweating abnormalities [47]. The autonomic nervous system directly regulates insulin secretion, hepatic glycogenolysis, and lipid metabolism. Consequently, autonomic dysfunction can lead to dysregulated blood glucose control and reduced energy utilization efficiency, predisposing individuals to metabolic disorders [48, 49]. Spinocerebellar ataxia The serotonergic synapse pathway (hsa04726) is a specialized synaptic structure centered on the neurotransmitter serotonin (5-HT), which regulates diverse physiological functions, including mood, sleep, appetite, nociception, and cognition (Fig. 9). Our findings experimentally justify Newsholme’s central fatigue hypothesis, which first proposed that 5-HT is a key mediator of central fatigue. Gerard P. Ahern et al. demonstrated that various immune cells express multiple serotonergic receptors, enabling serotonin to directly modulate immune cell proliferation, cytokine secretion, and chemotaxis [50]. Notably, research teams have established that most serotonin in the human body is produced in the gut. Gut microbiota disruption can trigger systemic immune activation and alter the synthesis and release of intestinal serotonin [51]. In neuroendocrine studies, Anthony J. Cleare’s study identified distinct neuroendocrine abnormalities in patients with CFS, with serotonergic system hyperactivity being one of the central hypotheses. During prolonged physical exertion, increased central serotonergic activity stimulates CRH release from the paraventricular nucleus of the hypothalamus, activating the HPA axis. Chronic dysregulation of this mechanism can lead to HPA axis hypofunction, manifesting as a blunted response to stress [52, 53]. In metabolic disease, elevated serotonin levels suppress adipose tissue thermogenesis, reduce skeletal muscle oxidative capacity, and inhibit β-cell proliferation. Consequently, enhanced serotonergic activity reduces whole-body energy expenditure, promoting obesity and insulin resistance [54]. Serotonergic synapse We conducted separate analyses of the most significantly enriched genes associated with BP, CC, and MFs across the comparative groups. For BP, the MAST3 gene, associated with the Qi and blood deficiency syndrome, encodes a protein that directly facilitates the NF-κB signaling pathway activation. Elevated MAST3 protein phosphorylation levels were observed in patients with CFS, correlating with enhanced inflammatory responses [55]. Concurrently, MAST3 is closely associated with the intracellular microtubule network. Dysregulation of its role in stabilizing microtubules may lead to cellular morphology and intracellular transport abnormalities [56]. The ARFGEF2 gene encodes a protein that is critically involved in neurodevelopment [57]; however, its downstream effector, ARF1, is vital for inflammasome activation. Studies demonstrate that ARF1 deficiency leads to NLRP3 inflammasome activation, contributing to the low-grade chronic inflammation observed in patients [58]. Genes associated with liver depression and spleen deficiency syndromes include AKAP11 and MAP4K1. AKAP11 interacts with glycogen synthase kinase-3β (GSK3β) through specific binding domains. GSK3β is vital for oxidative stress and inflammatory networks by suppressing the Nrf2 antioxidant function; however, it promotes the pro-inflammatory activity of NF-κB [59]. Another significant gene functions in T cells and broader immune cell populations. In T cells, deficiency of the MAP4K1-encoded protein leads to aberrant T cell overactivation. Within the broader immune landscape, MAP4K1 loss amplifies inflammatory responses, elevating pro-inflammatory cytokines, including interleukin (IL)-12, tumor necrosis factor-α, and IL-6 [60, 61]. The genes associated with distinct CFS syndrome types include GIT1, PRKCB, and ADRA2A. GIT1 dysfunction may disrupt key cellular processes, including migration, morphological maintenance, and intracellular communication. GIT1 promotes dendritic spine formation, growth, and motility in cultured neurons. GIT1-knockout mice exhibit impaired learning task performance and significantly reduced adaptability to novel environments [62]. PRKCB encodes the protein kinase PKCβ. A study by Thomas T. Su’s team has demonstrated that PKCβ is strictly required for B cell development, activation, and antibody production. PKCβ-deficient B cells fail to activate the NF-κB pathway, resulting in severely impaired proliferation and defective antibody production [63]. Furthermore, pharmacological inhibition of PKCβ activity mitigates pathological endothelial damage in diabetic and ischemic injury models [64]. In glucose homeostasis regulation, the ADRA2A receptor transmits sympathetic signals and suppresses postprandial GLP-1 release, thereby indirectly elevating the blood glucose levels. Given the established role of GLP-1 as a neuroprotective hormone, its deficiency may contribute to impairments in learning and memory formation [65, 66]. In cognitive function, ADRA2A activation in the prefrontal cortex ameliorates symptoms of attention deficit and hyperactivity. Consequently, ADRA2A agonists are commonly used to treat attention-deficit hyperactivity disorder [67]. ADRA2A activation also exerts analgesic effects, particularly by modulating musculoskeletal pain under stress conditions. Therefore, ADRA2A pathway dysregulation may contribute to reduced pain thresholds and sustained chronic pain states in patients with CFS [68]. Regarding CC, the most prominent composition-related genes associated with the CFS Qi and blood deficiency syndrome include RFLNB, FLNA, and LCP1. RFLNB encodes filamin B (FLNB), a critical cross-linking protein in cytoskeletal architecture. FLNB transmits extracellular matrix signals to the cytoskeleton, and its mutations cause various skeletal dysplasias. Functional FLNB impairment may necessitate increased energy expenditure for cytoskeletal remodeling [69]. FLNA is essential for T cell receptor downstream signaling and immune synapse stabilization. Its functional impairment disrupts effective T cell activation, leading to diminished immune responses. This provides a direct molecular mechanism for T cell dysfunction in CFS [70]. Furthermore, FLNA knockdown in endothelial cells inhibits vascular endothelial growth factor-induced proliferation and migration, disrupts vascular-like structure formation, and compromises blood-brain barrier integrity [71]. FLNA serves a dual function as a mechanosensor and signaling hub. A study by T.P. Stossel et al. revealed that FLNA undergoes conformational changes to convert mechanical forces into biochemical signals, thereby regulating cellular behavior. FLNA abnormalities may lead to broad dysregulation across multiple signaling pathways, including inflammatory responses, stress adaptation, and energy metabolism [72]. LCP1 facilitates cytoskeletal remodeling and drives immune cell activation. LCP1 knockdown reduces neuroinflammation and ameliorates lymphopenia while concurrently impairing T cell activation [73, 74]. Consequently, LCP1 dysfunction in patients with CFS may prevent T cells from forming effective immune synapses, leading to attenuated activation signals and impaired viral clearance or immunological memory formation. The most prominent CC-related genes associated with CFS liver depression and spleen deficiency syndrome include CAPZB and MPP1. CAPZB is a convergence point for multiple signaling pathways, modulating immune response intensity and energy expenditure through actin dynamics regulation. Furthermore, it facilitates immune synapse formation by restricting excessive actin polymerization. CAPZB knockdown disrupts actin architecture at the immune synapse, impairing complete T cell activation [75, 76]. MPP1 encodes membrane palmitoylated protein 1 (p55), a key cytoskeletal organizer essential for maintaining plasma membrane mechanical stability, regulating signal transduction, and ensuring cell polarity. A study by Magdalena Trybus and colleagues demonstrated that MPP1 gene mutations cause defective membrane-cytoskeleton attachment in erythrocytes, leading to spherocyte formation. These spherical erythrocytes undergo splenic sequestration and destruction, resulting in chronic hemolytic anemia and potential long-term tissue hypoxia [77]. The most significant CC-related genes shared between CFS Qi and blood deficiency, liver depression, and spleen deficiency syndromes include FRMD4B and ADD1. FRMD4B encodes the scaffold protein FERM domain-containing protein 4B. FRMD4B knockdown reduces axonal elongation and impairs learning and memory processes. This provides a molecular basis for the core CFS symptoms, including poor concentration and memory deficits [78]. ADD1 encodes adducin 1. Its deficiency or dysfunction destabilizes the erythrocyte membrane skeleton, resulting in abnormal cell morphology and increased erythrocyte fragility [79]. This mechanism parallels that of MPP1, a gene associated with liver depression and spleen deficiency syndromes. Both contribute to a state of chronic mild hypoxia and inadequate energy supply to highly metabolic tissues, including muscles and the brain, providing a hemorheological explanation for persistent fatigue in CFS. Of particular significance, a study by Cai Qi’s team revealed that ADD1 modulates synaptic plasticity through glutamate receptor regulation, establishing a direct link to human memory performance. ADD1-knockout mice exhibit pleiotropic hydrocephalus and axonal degeneration, which disrupt neural information transmission efficiency [80], manifesting as poor concentration, slowed cognitive processing, and memory impairment. Regarding MF, the most significantly enriched genes associated with the CFS Qi and blood deficiency syndrome include PTPN12 and HCLS1. PTPN12 encodes the protein tyrosine phosphatase non-receptor type 12, commonly termed PTP-PEST. It is a critical negative regulator of cell migration and is vital for endothelial cell migration and adhesion [81]. PTPN12 dysfunction impairs immune cell migration and stable adhesion, preventing their recruitment to inflammatory sites and contributing to delayed immune responses in patients with CFS. Furthermore, PTPN12 is identified as a potential therapeutic target for various severe diseases, including cancer, diabetes, metabolic disorders, and autoimmune conditions [82]. HCLS1 is predominantly expressed in hematopoietic cells. As an actin-binding protein, it drives actin cytoskeleton reorganization upon immune receptor activation to initiate immune cell activation and proliferation [83]. For MF, the most significantly enriched genes in the CFS liver depression and spleen deficiency syndrome include VASP and EVL. The VASP-encoded protein is an actin regulatory factor involved in actin-based cellular migration and T-cell polarization. In neurons, VASP localizes to dendritic spines, where it modulates actin cytoskeleton dynamics to regulate spine size and stability, which are the structural foundations of synaptic plasticity. Consequently, VASP dysfunction impairs dendritic spine maturation, thereby adversely affecting learning and memory. This manifests as cognitive fatigue, poor concentration, and memory deficits. The EVL-encoded protein exhibits functional similarities to VASP in this regard [84]. Impaired VASP function disrupts effective immune cell migration, hindering their recruitment to inflammation or infection sites. This provides a key mechanism for the potential chemotactic deficiencies observed in patients with CFS [85, 86]. EVL is similarly crucial for T cell migration. Notably, EVL and VASP are selectively required for activated T cell trafficking, whereas they are dispensable for normal T cell development or homing to lymph nodes and the spleen. Genetic deletion of EVL and VASP does not substantially impair T cell development [87]. At the MF level, the most significantly enriched genes across different CFS syndrome types include PKN3, PRKCB, and PKN1. PKN3 encodes a protein kinase N3. A study by Kristin Möpert’s team demonstrated that PKN3 is vital for maintaining endothelial cell morphology, cell-cell junctions, and motility. Moreover, PKN3 function loss leads to aberrant actin rearrangement, thereby attenuating the pro-inflammatory endothelial cell activation [88]. Concurrently, PKN3 expression dysregulation is closely associated with tumor metastasis. Studies have confirmed that PKN3 is significantly upregulated in multiple malignancies and is vital for promoting metastatic progression [89, 90]. PKN3 additionally contributes to angiogenesis regulation. PKN3 knockout suppresses angiogenesis, specifically impairing microvascular sprout formation and in vivo neovascularization [90]. PKN1 encodes protein kinase N1, which belongs to the AGC kinase family along with PKN3. Studies reveal that while PKN1-deficient mice exhibit normal lymphatic system development, their peripheral blood lymphocyte counts are reduced by half, indicating significantly impaired lymphocyte migration upon PKN1 knockout [91]. This reveals a novel mechanism underlying immune deficiency in CFS patients. Furthermore, PKN1 upregulation may serve as a prognostic biomarker in endometrial cancer. Studies indicate elevated PKN1 expression at mRNA and protein levels in patients with an unfavorable prognosis, with significantly higher PKN1 mRNA levels observed in those with metastatic disease [92]. Increasing evidence indicates that cytoskeletal dysregulation represents a convergent pathological mechanism underlying the neuroendocrine, immune, and metabolic abnormalities observed in CFS. Syndrome-specific patterns emerge when comparing the Qi and blood deficiency and liver depression and spleen deficiency subtypes. The Qi and blood deficiency syndrome is characterized by alterations in MAST3-mediated NF-κB hyperactivation and FLNB-related skeletal signaling defects, contributing to inflammatory tone and energy expenditure imbalances. Conversely, the liver depression and spleen deficiency syndrome exhibits prominent disruptions in CAPZB-controlled immune synapse formation and MPP1-mediated membrane stability, leading to impaired T-cell activation and erythrocyte fragility. Despite these distinct molecular profiles, both syndromes converge on PKN1/3-regulated endothelial dysfunction and share overlapping deficits in VASP/EVL-guided neuronal plasticity, providing a unified explanation for the core symptoms of fatigue, cognitive dysfunction, and immune impairment in CFS. This syndromic stratification clarifies the heterogeneity of CFS manifestations and offers potential pathways for personalized therapeutic strategies. Our enrichment analysis results (GO and KEGG pathways) identified key functional categories that can be used for CFS-related gene screening and functional prediction. These findings will be further validated through integrated clinical and statistical analyses in subsequent CFS studies. Conclusion This study aimed to identify the key GO and KEGG pathways associated with CFS-related proteins. We identified 397 critical features and established 25 quantitative rules using machine learning approaches. These findings have been substantiated by a recent study in the field. This study contributes to a deeper understanding of the pathological mechanisms underlying CFS and innovative tool development for analyzing the functional characteristics of the disease. Core Objective: To elucidate CFS mechanisms through functional annotation of disease-associated proteins (GO and KEGG pathways) at the molecular level. Key Findings: Machine learning identified 397 critical features and 25 quantitative rules, which were validated against existing research, demonstrating robust reliability. Significance: This study deepens the multidimensional understanding of CFS pathology (immune regulation, microenvironment, and others) and establishes a methodological framework for analyzing the functional characteristics of complex diseases, thereby advancing precise “gene-function-disease” correlation studies. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Acknowledgements We thank Home for Researchers editorial team (www.home-for-researchers.com) for language editing service. Abbreviations Biological Processes Cellular Components Chronic Fatigue Syndrome Corticotropin-releasing Hormone Diacylglycerol Data-dependent Acquisition Data-independent Acquisition Gene Ontology Hypothalamic-pituitary-adrenal Kyoto Encyclopedia of Genes and Genomes Mitogen-activated Protein Kinase Molecular Functions Natural Killer cells Protein Kinase C Spinocerebellar Ataxia Traditional Chinese Medicine Author contributions Tingting Xu and Yonghui Yin contributed to the study conception and design. Material preparation, data collection and analysis were performed by Tingting Xu, Shuo Gao, Xinxin Cheng, Wenxuan Man, and Yanjie Wang. The first draft of the manuscript was written by Tingting Xu; all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript. Funding This research was supported by the Shandong Provincial Key Research and Development Program (Nos. 2021CXGC010510). Data availability The datasets generated and analyzed during the current study are publicly available. Declarations Ethics approval and consent to participate All participants provided written informed consent. The study was conducted in accordance with the Declaration of Helsinki and was approved by the Ethics Committee of the Affiliated Hospital of Shandong University of Traditional Chinese Medicine (Approval No: (2023) Lun Shen No. (053) - KY). Consent for publication All authors approved the final manuscript and the submission to this journal. Conflict of interest There are no conflicts of interest in this study. All authors have read and approved this version of the article, and due care has been taken to ensure the integrity of the work. Neither the entire paper nor any part of its content has been published or has been accepted elsewhere. It is not being submitted to any other journal. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1 Data Availability Statement The datasets generated and analyzed during the current study are publicly available.