Sex‐specific plasma metabolome signatures in major depressive disorder

Major depressive disorder (MDD) is a highly prevalent mental disorder with obvious sex differences.1 Numerous metabolomics studies have investigated the differential expressed metabolic signatures in MDD; however, few studies have investigated the sex differences at the large-scale molecular level other than transcriptome.2, 3 Here, we investigated the sex differences of plasma metabolome signatures in antidepressant-naive patients with MDD. In this study, the plasma metabolomes of antidepressant-naive patients with MDD and healthy controls (HC) from our previous study (Appendix S1) were reanalyzed.4 Briefly, 60 patients with MDD and 59 HC were enrolled. All patients were single episode and diagnosed using the DSM-IV. In addition, patients with MDD had a Hamilton Depression Rating Scale score ≥ 18. No participants had any severe systemic disease or mental disorder other than MDD. For patients with MDD, blood samples were taken during a clinic visit and the median fasting length before blood sampling was 4 h. For HC, fasting overnight blood samples were collected in the morning. The metabolomics platform consisted of an Acquity UPLC System (Waters, Milford, MA, USA) and a TripleTOF 5600 System (AB SCIEX, Framingham, MA, USA). We identified 3274 and 3424 features in the positive and negative modes, respectively. One male patient was excluded in the present study because of missing metabolome data. Therefore, 118 participants were included in the analysis (MDD: male, n = 29, female, n = 30; HC: male, n = 30, female, n = 29; age range, 17–66 years). There were no significant differences in sample size, age, or body mass index between groups (Table S1). The study was approved by the ethics committee of Chongqing Medical University. Written informed consent was obtained from all participants. All procedures were performed according to the principles expressed in the Declaration of Helsinki. The details of bioinformatics analyses are provided in Appendix S2. Briefly, principal component analysis and partial least square discriminant analysis were performed to reveal the intrinsic metabolome patterns in different groups using SIMCA-P 13.0 (Umetrics, Umea, Sweden). Weighted gene co-expression network analysis, a method to identify co-expressed molecules, was used to identify sex-specific and shared metabolic modules associated with MDD in R (version 3.5.1, R Project; https://www.r-project.org).5 The overlapping metabolite trends of the shared modules were evaluated using the rank–rank hypergeometric overlap test.6 Individual metabolites in the modules significantly associated with depression were uploaded to MetaboAnalyst 4.0 (https://www.metaboanalyst.ca) for mass spectrometry (MS) peaks to pathways analysis.7 Principal component analysis score plots showed that quality control samples were clustered tightly in both positive (R2X = 0.648; Fig. S1a) and negative (R2X = 0.652; Fig. S1b) modes, indicating satisfactory reproducibility. Partial least square discriminant analysis score plots showed that samples from these four groups were visibly separated in both positive (R2Y = 0.797, Q2 = 0.353; Fig. S1c) and negative modes (R2Y = 0.688, Q2 = 0.411; Fig. S1d), suggesting distinct metabolic differences between male and female patients with MDD and HC. Obvious outliers were removed before weighted gene co-expression network analysis network constructions (Fig. S2). Based on network topology, the soft-thresholding power was set at 5 and 9 for the male and female datasets, respectively (Fig. S3). A total of 37 metabolic modules’ metabolite peaks in the male dataset, and 30 modules’ metabolite peaks in the female dataset were identified (Fig. S4). Six and nine modules in the male and female datasets, respectively, were significantly correlated to MDD (uncorrected P < 0.05; Fig. S5). Two thirds of these modules (four and six modules for males and females, respectively) were sex-specific (Fig. S6a), while the metabolites of the shared modules (457 metabolites) showed similar overlapping trends (Fig. S6b). Moreover, sex-specific modules were mainly involved in bile acid biosynthesis and the carnitine shuttle for males and in glycosphingolipid metabolism and tryptophan metabolism for females; additionally, the porphyrin metabolism was a common pathway in both sexes (Table S2). Further analysis of all the MDD-related modules in the male and female datasets was performed, which validated these results (Table S3). The results of this study should be interpreted cautiously because of certain limitations. A single liquid chromatography mass spectrometry (LC–MS) platform cannot cover all plasma metabolomes; therefore, additional studies that combine LC–MS with other platforms are required. Additionally, we used a relatively small number of participants in each group, which may be insufficient to generalize our findings. Moreover, data from the MS peaks to pathways analysis need further validation on the proteome or transcriptome level. This study is supported by the National Key Research and Development Program of China (Grant No. 2017YFA0505700) and the Foundation Research and Frontier Exploration Program in Chongqing (Grant No. cstc2018jcyjAX0341). The authors declare no conflicts of interest. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.