Association of anhedonia with brain-derived neurotrophic factor and oxidative stress in bipolar disorder Authors: Xueyan Zhu, Peiyun Zhang, Chao Liu, Yao Zhang, Xiaoxiao Wang, Xin Li, Daofeng Lu, Fang Guo, Yuan Yao, Jiayu Zhu, Chengjia Shen, Chongze Wang, Ping Zhang, Qi Yan, Zhenghui Yi, Qinyu Lv, Peijuan Wang Published: 2026 ABSTRACT AbstractBackgroundIndividuals with bipolar disorder (BD) undergo depressed episodes, frequently characterized by anhedonia. Neurotrophic factors and oxidative stress are linked to anhedonia. The precise association between these biomarkers and anhedonia in individuals with BD remains unclear. This study primarily examined the potential relationships among brain-derived neurotrophic factor (BDNF), superoxide dismutase (SOD), catalase (CAT), and anhedonia in individuals with BD.MethodsThis study had 85 people diagnosed with bipolar disorder, of whom 61 exhibited depression symptoms. Additionally, there were 88 healthy control volunteers. The study thoroughly evaluated anticipatory and consummatory pleasure, childhood adversity, oxidative stress indicators, and BDNF levels in depressed BD patients and healthy controls. Partial correlation analysis and stepwise multiple linear regression analysis were employed to explore the relationships among oxidative stress indicators, BDNF and anhedonia.ResultsIn comparison to the healthy control group, the depressed BD group demonstrated markedly diminished anticipatory and consummatory pleasure, elevated anxiety and depression ratings, and increased childhood adversity. There was a significant difference in CAT levels between the two groups. In depressed BD patients, BDNF and CAT were moderately and negatively correlated with consummatory pleasure. The regression analysis indicated BDNF as a significant predictor of consummatory pleasure.ConclusionThese data suggested that the anhedonia is affected by BDNF and oxidative stresses, so offering a critical insight for addressing anhedonia in BD.Clinical trial registration numberNot applicable.Keywords: Bipolar disorder, Anhedonia, Brain-derived neurotrophic factor, Catalase Introduction Bipolar disorder (BD), characterized by recurrent episodes of mania or hypomania and depression, affects over 1% of the global population and ranks as the 17th leading cause of disease burden worldwide [1]. Anhedonia is a fundamental symptom of severe depressive disorder, correlated with the intensity of depression episodes, suicidal inclinations, and unfavorable prognosis [2]. Anhedonia is primarily classified into two kinds. They are respectively anticipatory pleasure, which refers to a reduced ability to experience pleasure when anticipating a favorable event, and consummatory pleasure, which relates to a reduced ability to enjoy pleasure from activities [3]. Anhedonia represents a core and prominent symptom of the depressive phase of BD [4]. However, there is currently no defined treatment approach for anhedonia. Evidence from certain studies suggests that some commonly prescribed selective serotonin reuptake inhibitors (SSRIs) may exert limited therapeutic effects on hedonic deficits in patients with depression. Persistent residual anhedonia may subsequently compromise treatment adherence and elevate the risk of depressive relapse [5]. Consequently, investigating the determinants of anhedonia in individuals with BD may serve as a potential avenue for guiding their therapy. Certain researchers assert that oxidative stress(OS) is associated with the pathophysiology of depression and BD [6, 7]. Catalase (CAT) decomposes hydrogen peroxide into water and oxygen and serves as an enzyme that eliminates reactive oxygen species (ROS) in the body [8]. The relationship between CAT activity and depressive symptoms remains inconsistent across studies, with some reporting decreased CAT levels in depression [9], others finding elevated enzyme activity during acute depressive episodes [10], and a meta-analysis demonstrating no statistically significant difference between depressive and healthy cohorts [6]. The CAT levels of 43 individuals with BD undergoing therapy were significantly lower than those of 30 healthy controls (p = 0.021) [11]. Consequently, a primary objective of this study is to ascertain whether a disparity exists in oxidative stress levels, particularly the activity of the CAT enzyme, between depressed BD patients and healthy controls, and to evaluate the statistical significance of this difference. Brain-Derived Neurotrophic Factor (BDNF) is a neurotrophic protein expressed throughout the body, originating from brain neurons, vascular endothelial cells, and platelets [12]. A reduction in BDNF may diminish the plasticity of neural synapses, perhaps correlating with depression and treatment-resistant mood disorders, whereas antidepressant administration can elevate BDNF levels [13]. Serum BDNF levels were significantly correlated with the severity of depression [14]. Beyond mood disorders, BDNF dysregulation has been documented across a broader spectrum of psychiatric conditions, including schizophrenia, obsessive-compulsive disorder, post-traumatic stress disorder, and panic disorder, suggesting that it may reflect a transdiagnostic neurobiological vulnerability rather than a disorder-specific mechanism [15]. Notably, disrupted BDNF expression and an altered balance between neurotrophic and apoptotic proteins have been reported not only in patients with schizophrenia but also in their unaffected first-degree relatives, further supporting this transdiagnostic perspective [16]. Nevertheless, the specific role of BDNF in BD — particularly its association with anhedonia during depressive episodes — warrants dedicated investigation. Preliminary data indicated that BDNF levels in the blood of individuals with BD may be diminished compared to those in healthy individuals, and a prolonged duration of bipolar disorder in patients may correlate with elevated BDNF levels [17]. Recent research demonstrated a complicated interplay between neurotrophic factor dysregulation and oxidative stress damage, perhaps serving as one of the pathogenic causes of BD. Intracerebral administration of BDNF has been shown to protect the prefrontal cortex and hippocampus from ouabain-induced oxidative damage in a preclinical manic model [18]. Beyond its neuroprotective role, BDNF expression in reward-related neural circuits, which encompass dopamine neurons in the ventral tegmental area and their projections to the prefrontal cortex, has been correlated with anhedonia [19]. Moreover, BDNF deficiency has been shown to exacerbate hedonic deficits under inflammatory conditions [20]. In summary, BDNF may influence anhedonia in depressed BD patients. Consequently, the current investigation examined whether a comparable association exists between BDNF and anhedonia in depressed BD patients. Early life experiences significantly influence brain development, thus impacting human emotions. Early life stressors, such as childhood maltreatment and neglect, can elevate the chance of severe depression in later life [21]. There was evidence that early adversity has a major impact on the emotional elements connected to mental diseases in the field of psychopathology [22]. However, there were very few research of this type in the field of bipolar disorder. Therefore, this study explores the relationship between anhedonia in depressed BD patients and early adversity in individuals. The present study aimed to investigate the associations between anhedonia and three potential contributing factors — BDNF levels, oxidative stress, and early life adversity — in depressed BD patients, with the goal of advancing our understanding of the neurobiological and psychosocial underpinnings of anhedonia in this population. This research established the groundwork for the early identification and thorough comprehension of the mechanisms behind anhedonia in individuals with bipolar disorder. Methods Participants From September 2023 to December 2023, this research recruited 85 patients with BD from the outpatient and inpatient departments of Shanghai Mental Health Centre, as well as 88 healthy control patients through advertisements. At least two psychiatrists with senior positions independently diagnosed all patients. The patient group’s inclusion criteria were explicitly defined as follows: (1) The patient and their guardian voluntarily participated and signed the informed consent form. (2) The patient is of Han ethnicity, aged 18–50 years, and has a minimum of 6 years of education. They were diagnosed with DSM-V bipolar disorder (Type I or II) either currently or previously. The following were the exclusion criteria: (1) other comorbid mental disorders as defined by the DSM-V diagnostic criteria; (2) mental disorders caused by organic diseases; (3) mental disorders caused by psychoactive substances; (4) severe infections, tumors, cardiovascular diseases, and other major illnesses; and (5) uncooperative patients. Patients with BD were further classified as having current depressive symptoms if their Hamilton Depression Rating Scale (HAMD-17) score was ≥ 7 at the time of assessment. Of the 85 recruited patients, 61 met this criterion and were included in subsequent analyses. The remaining 24 patients (HAMD-17 < 7) were excluded as they did not meet the symptom criterion relevant to the present research question. A healthy control group was recruited, matched to the patient group in terms of age, sex, and years of education. The control group’s inclusion criteria were as follows: (1) a minimum of six years of education, Han ethnicity, and an age range of 18 to 50 years; (2) voluntary participation and the execution of voluntary informed consent. Exclusion criteria: (1) history of alcohol dependence or use of other psychoactive substances; (2) severe physical illness; (3) personal history of mental illness or first-degree relative with a history of mental illness. Clinical evaluation The following general information was obtained from each subject: age, gender, height, weight, years of education, marital status, and the frequency of smoking and alcohol consumption. The 17-Item Hamilton Rating Scale for Depression (HAMD-17) [23, 24], the Hamilton Anxiety Rating Scale (HAMA) [25], and Young Manic Rating Scale (YMRS) [26] were employed by patients with bipolar disorder to evaluate their manic, anxiety, and depressive moods. The patients were categorized as those with depressive symptoms based on a threshold score of 7 by HAMD-17, which encompassed those in the remission phase or mixed episode of BD [27]. Participants’ anhedonia was evaluated using the Temporal Experience of Pleasure Scale (TEPS), which comprises two components: consummatory and anticipatory pleasure [28]. The Social Anhedonia Scale of the Chinese version (RSAS-C) and Physical Anhedonia Scale of the Chinese version (RPAS-C) were well-received for their straightforward format and comprehensive coverage of information. A higher overall score indicates more severe symptoms of anhedonia. Additionally, they demonstrated strong robustness in clinical samples (Cronbach α: 0.83–0.93) [29, 30]. The Childhood Trauma Questionnaire (CTQ) assessed childhood adversity from five distinct dimensions, including sexual abuse, physical abuse, emotional abuse, physical neglect, and emotional neglect. This scale demonstrated strong internal consistency (Cronbach α: 0.89–0.95) [31]. Collection of biological markers Between 7:00 and 9:00 AM, a 5 ml peripheral venous blood sample was collected after fasting for a minimum of 12 h. The serum was stored at -80 °C until testing was performed after it was separated. BDNF, SOD, Glutathione peroxidase (GSH-Px), CAT, and malondialdehyde (MDA) levels were quantified using enzyme-linked immunosorbent assay (ELISA) (R&D Systems, USA). The researchers, who were uninformed of the clinical data of the samples, conducted the experiments in accordance with the manufacturer’s protocol. Duplicate testing was implemented for each sample. Samples were randomly selected for testing to confirm the detection’s repeatability. Statistical methods The normality of continuous variables was assessed using the Shapiro-Wilk test. Variables conforming to normal distribution were compared between groups using independent-samples t-tests, while variables that did not conform to normal distribution were analyzed using the Mann-Whitney U test. Categorical variables were analyzed using the chi-square test. A partial correlation analysis was performed controlling for potential confounders, including age, gender, educational background, alcohol consumption, smoking, and Body Mass Index (BMI). A stepwise multiple linear regression model was employed to identify factors influencing BDNF levels in individuals with BD presenting with depressive symptoms (HAMD ≥ 7), with the same confounders entered as candidate predictors alongside the primary variables of interest. All continuous data are reported as mean ± standard deviation, and categorical data are described as frequency and percentage. Statistical significance was set at p < 0.05, and all tests were two-tailed. Results Demographic comparison of the healthy control group and BD patients Our study comprised 61 individuals diagnosed with BD exhibiting depressive symptoms (HAMD-17 ≥ 7) and 88 healthy controls. In terms of age, BMI, years of education, marital status, drinking, and smoking, the intergroup analysis revealed substantial statistical differences between the two groups (Table 1). Table 1Comparison of demographic between depressed BD patients and healthy controlsHC(n = 88)BD_Dep (n = 61)X2/Z P Age(year)24.60 ± 2.5223.08 ± 6.84-2.520.012Sex2.250.163Male35(39.80%)17(27.90%)Female53(60.20%)44(72.10%)Marital status8.15 0.011 a Single85(96.60%)51 (83.60%)Married3(3.40%)8(13.10%)Divorced02(3.30%)Education year17.27 ± 2.2113.50 ± 3.95-6.10 < 0.001 DrinkingNever57(64.80%)40(65.60%)6.380.083 aSometimes31(35.20%)17(27.90%)Often02(3.30%)Everyday02(3.30%)Smoking15.16< 0.001aNever86(97.70%)48(78.70%)Sometimes2(2.30%)7(11.50%)Often02(3.30%)Everyday04(6.60%)BMI (kg/m2)22.28 ± 3.3024.50 ± 3.95-1.43 0.011 Abbreviation: BD_Dep, individuals diagnosed with depressed BD patientsData representation: Mean ± SD, count (N) and percentage (%), Median (IQR)a: Fisher’s Exact Test Comparison of demographic between depressed BD patients and healthy controls Abbreviation: BD_Dep, individuals diagnosed with depressed BD patients Data representation: Mean ± SD, count (N) and percentage (%), Median (IQR) a: Fisher’s Exact Test Comparison of clinical characteristics and biomarkers between BD patients and the healthy control group Figure 1A and B indicated that the RPAS-C and RSAS-C scores for the depressed BD group were markedly elevated compared to the healthy control group (p < 0.05). Figure 1C and D, and 1E demonstrated that the scores for the sub-dimensions of consummatory pleasure and anticipatory pleasure as well as the overall TEPS score, in the depressed BD group, were markedly lower than those in the healthy group (p < 0.001). Group differences in clinical characteristics between depressed bipolar disorder patients, and healthy controls. (A RPAS-C; B RSAS-C; C Anticipatory pleasure; D Consummatory pleasure; E TEPS Total Score) Depressed BD patients reported significantly higher scores on emotional abuse, physical abuse, and sexual abuse compared to healthy controls, whereas emotional neglect and physical neglect scores were significantly higher in healthy controls (Table 2). Table 2Comparison of clinical variables and biological markers between depressed BD patients and healthy controlsHC(n = 88)BD_Dep (n = 61)t/Z P HAMD-1719.72 ± 7.59HAMA-14b19.67 ± 9.25YMRS-1110.44 ± 9.41Total disease course(months)60.43 ± 53.74TEPSAnticipatory pleasure b39.17 ± 6.0132.59 ± 9.045.33 < 0.001 Consummatory pleasure b46.27 ± 7.2338.15 ± 9.405.96 < 0.001 Total score b90.24 ± 12.7874.80 ± 17.456.23 < 0.001 RSAS-C19.60 ± 3.8624.49 ± 3.56-6.97 < 0.001 RPAS-C28.97 ± 3.4331.75 ± 5.26-3.06 0.002 Emotional abuse6.85 ± 3.4111.36 ± 4.95-6.78 < 0.001 Physical abuse5.63 ± 1.867.28 ± 3.38-4.15 < 0.001 Sexual abuse5.40 ± 1.535.93 ± 2.40-2.09 0.037 Emotional neglect18.07 ± 6.3713.61 ± 4.99-4.72 < 0.001 Physical neglect11.10 ± 3.368.93 ± 3.00-4.03 < 0.001 BDNF (ng/ml)1.77 ± 1.301.61 ± 1.02-0.100.917CAT (ng/ml)4.67 ± 1.433.99 ± 1.54-2.86 0.004 GSH-PX (ng/ml)1.78 ± 1.442.07 ± 1.69-1.850.065SOD (ng/ml)7.56 ± 4.166.87 ± 4.10-1.630.103MDA (ng/ml)1.42 ± 0.971.62 ± 1.05-1.180.239b: Independent-samples t-test; all other continuous variables were analyzed using Mann-Whitney U test Comparison of clinical variables and biological markers between depressed BD patients and healthy controls b: Independent-samples t-test; all other continuous variables were analyzed using Mann-Whitney U test Patients with bipolar depression exhibited dysregulation across multiple oxidative stress and neurotrophic markers relative to healthy controls, with lower BDNF, SOD, and CAT and higher GSH-PX and MDA. Among these, only CAT differed significantly (p < 0.05), with healthy controls showing higher catalase activity (Table 2). Correlation examination of markers in depressed BD patients Upon controlling for age, gender, education, alcohol consumption, smoking status, and BMI, the correlation results are shown in Fig. 2. The study revealed a significant negative correlation between BDNF and consummatory pleasure (r = -0.359, p < 0.01) as well as the TEPS total score (r = -0.338, p < 0.05) in depressed BD patients. CAT had a demonstrated a moderate negative with consummatory pleasure (r = -0.279, p < 0.05). A substantial positive association existed between BDNF and CAT (r = 0.578, p < 0.001). Among CTQ subscales, emotional neglect was significantly correlated with the TEPS total score (r = -0.278, p < 0.05), while emotional abuse was significantly correlated with total disease duration (r = 0.319, p < 0.01). Partial correlations between clinical variables and biological markers in depressed bipolar disorder patients Regression analysis The subsequent stepwise multiple linear regression analysis indicated that both BDNF and emotional neglect were significant predictors of consummatory pleasure (R² = 0.167, F = 5.822, p = 0.005). Specifically, BDNF (β = −2.602, 95% CI: −4.817, − 0.388, p = 0.022) and emotional neglect (β = −0.511, 95% CI: −0.964, − 0.057, p = 0.028) were both negatively associated with consummatory pleasure (Fig. 3). Stepwise multiple linear regression analysis for consummatory pleasure Discussion This work is among the first to specifically examine the neuroimmunological underpinnings of anhedonia in depressed BD patients and its correlation with clinical manifestations. Our main findings are as follows: (1) Anhedonia and childhood trauma were prominent features distinguishing the depressed BD patients from healthy individuals; (2) The serum CAT levels in depressed BD patients were markedly lower than those in healthy controls; (3) In patients with BD experiencing anhedonia, a connection was observed between anhedonia and emotional neglect. (4) A notable correlation existed between CAT, BDNF and consummatory pleasure. (5) BDNF was associated with consummatory pleasure. Childhood trauma is a well-established risk factor for bipolar disorder, associated with earlier onset, emotional instability, and elevated suicidal risk [32, 33]. Recent studies have further demonstrated that BD patients with a history of childhood trauma exhibit more frequent depressive episodes, greater symptom severity, and poorer treatment response [34, 35]. Consistent with these findings, depressed BD patients in the present study reported significantly higher scores on emotional abuse, physical abuse, and sexual abuse than healthy controls, with emotional abuse further correlating with total disease duration. Unexpectedly, both neglect subscales were higher in healthy controls. This finding should be interpreted with caution, as the two groups differed significantly in years of education (17.27 vs. 13.50, p < 0.001), which may have contributed to this pattern. Chinese normative studies of the CTQ-SF have documented systematically elevated neglect subscale scores in community samples, attributable to culturally restrained emotional expression in parenting rather than pathological caregiving deficits [36, 37]. Within the depressed BD group, emotional neglect was significantly associated with anhedonia. Extending prior findings in major depressive disorder [38], this suggests that emotional neglect is a clinically relevant correlate of reward processing deficits in bipolar depression. Our investigation identified a significant disparity in CAT levels between the two groups. This aligned with prior research findings. The antioxidant activity of CAT was markedly diminished in the mice subjected to the depression stress paradigm [39]. This study initially posited that oxidative stress could be a pathogenic process in depressed BD patients, suggesting potential treatment implications. Consistent with the findings of Peng et al., CAT levels were considerably decreased in patients with severe depressive disorder compared to healthy controls. Nevertheless, we did not detect substantial variations among the groups regarding other antioxidant enzymes [40]. This finding is not unexpected, given the considerable inconsistency reported across the existing literature. For instance, certain studies indicated that CAT and SOD levels in individuals with major depression were markedly lower than those in healthy controls, while MDA levels were dramatically elevated [41]. However, the activities of SOD and CAT have been documented as raised, unaltered, or enhanced in individuals with depression in other researches [42–44]. While immunological variables are acknowledged as contributors to mood disorders, the expression of inflammatory responses differs among individuals [45]. The disparities in researches may be ascribed to the substantial variations in illness course, clinical symptoms, and etiology of mood disorders. Simultaneously, varying study cohorts will influence the homogeneity of the samples. Future investigations will provide a more precise examination of the correlation between oxidative stress markers in cerebrospinal fluid and mood disorders as they relate to brain tissue. In recent years, the investigation and study of biomarkers for BD has garnered significant attention. Among these, BDNF is the most extensively researched. A prior meta-analysis indicated that BDNF levels in patients with BD were modestly diminished during manic episodes and dramatically reduced during depressive episodes [46]. In our investigation, despite the absence of a significant difference between the two groups, a notable observation was that the BDNF levels in patients from the depressed BD group were somewhat lower. BDNF is crucial for the development, differentiation, and survival of neurons [47]. BDNF, prevalent in the cerebral cortex and hippocampus, can mitigate depressive symptoms through distinct mechanisms [48, 49]. For instance, BDNF levels were decreased in the hippocampus and prefrontal cortex of mice exhibiting depression-like behavior when exposed to lipopolysaccharide [50]. The role of BDNF in reward circuitry is not uniform across brain regions. While hippocampal and prefrontal BDNF exerts neuroprotective effects, BDNF signaling within the mesolimbic dopamine pathway—specifically from the ventral tegmental area to the nucleus accumbens (NAc)—has been shown to promote susceptibility to stress-induced aversive and anhedonic behaviors rather than confer resilience [51, 52]. Importantly, this mesolimbic BDNF elevation has also been observed in humans: postmortem analysis of nucleus accumbens tissue from individuals with major depressive disorder revealed a 40% increase in BDNF protein levels compared to unaffected controls [52], suggesting that heightened NAc BDNF signaling may represent a translatable correlate of depressive pathology rather than a neuroprotective response. Whether analogous mesolimbic BDNF dysregulation occurs in BD remains an important question warranting direct investigation, though shared mesolimbic vulnerabilities across depressive states lend plausibility to this extrapolation. Although global peripheral BDNF declines during depressive episodes in BD, this population-level pattern may obscure region-specific divergences in BDNF function. Partial correlation analysis in the present study revealed that BDNF levels were negatively associated with consummatory pleasure within the depressed BD group (r = − 0.359, p < 0.01), consistent with a mesolimbic-dominant model in which elevated ventral tegmental area–NAc BDNF signaling paradoxically impairs reward experience. In this context, the progressive, episode-dependent decline in BDNF signaling across BD stages may contribute a distinct neurobiological context within which the BDNF–consummatory pleasure relationship diverges from patterns observed in other psychiatric conditions [53]. This study’s strength resides in its examination of the pertinent aspects contributing to anhedonia in depressed BD patients. At present, there are no sanctioned pharmacological treatments for anhedonia. This study indicated that biomarkers like oxidative stress and BDNF may enhance anhedonia in patients with BD, offering novel insights for clinical diagnosis and treatment. The present study has several limitations that should be acknowledged. The research was constrained by the restricted sample size and the cross-sectional approach employed to investigate the factors affecting anhedonia in depressed BD patients. This constrained our initial comprehension of the dynamic interplay between anhedonia and biomarkers in patients, and also did not establish the causal relationship. To mitigate these constraints, subsequent research should implement longitudinal surveys, augment the sample size, and incorporate other factors to diminish the influence of confounding variables. Conclusion Depressed BD patients have anhedonia, and the dysregulation of BDNF levels alongside oxidative stress may influence the pathological course, suggesting a potential avenue for targeted therapy. Acknowledgements We are grateful to all the physicians and subjects that participated in this study. Author contributions Xueyan Zhu: Conceptualization, Data curation, Writing original draft. Peiyun Zhang and Chao Liu: Data curation, Writing original draft. Yao Zhang, Xiaoxiao Wang, Xin Li, Daofeng Lu, Fang Guo, Yuan Yao, Jiayu Zhu, Chengjia Shen, Chongze Wang, Ping Zhang: Investigation, Data curation. Qi Yan and Zhenghui Yi: Writing- Reviewing and Editing. Qinyu LV, Peijuan Wang: Funding acquisition, Methodology, Writing, review & editing. All authors read and approved the final manuscript. Funding This work was supported by multiple grants: STI 2030-Major Projects+2022ZD0208500; the Fundamental Research Funds for the Central Universities (2024GDND02); Health Commission of Nantong City of China (grant number MS2024074 to P.J.W); Shanghai Science and Technology Innovation Action Plan Natural Science Fund Project (21ZR1455400); Shanghai Jiaotong University “Jiaotong University Star” Program Medical-Industrial Crossover Research Fund Project (YG2019QNB07); Elderly Health Project of the Health Commission of Jiangsu Province, China (LKM2022062); Health Commission of Nantong City, China (grant number MSZ2023061 to Q.Y.); Science Foundation of Kangda College of Nanjing Medical University (grant number KD2024KYJJ306). Data availability The datasets used and analyzed during the current study are available from the corresponding author on reasonable request. Declarations Ethics approval and consent to participate The Ethics Committee of the Shanghai Mental Health Centre approved this investigation (application number: 2024-09R). The ethical principles delineated in the Declaration of Helsinki were rigorously followed throughout the course of the investigation. All participants provided written informed consent prior to their participation. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Xueyan Zhu, Peiyun Zhang and Chao Liu contributed equally to this work. Contributor Information Qinyu Lv, Email: lvqinyu_louis@163.com. Peijuan Wang, Email: wpjntsy@163.com. References Associated Data Data Availability Statement The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.