Integrated multi-omics characterization identifies key microenvironmental biomarkers in major depressive disorder microenvironment

Integrated multi-omics characterization identifies key microenvironmental biomarkers in major depressive disorder microenvironment | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Integrated multi-omics characterization identifies key microenvironmental biomarkers in major depressive disorder microenvironment Hui Tang, Yunkai Wu, Weihao Deng, Jiaquan Liang, Chunguo Zhang, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8976053/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 7 You are reading this latest preprint version Abstract Characterizing the pathological microenvironment of major depressive disorder (MDD) is of clinical relevance yet remains poorly defined. Here, we systematically dissect microenvironment-associated regulatory mechanisms underlying MDD phenotypes through integrative analysis of ten transcriptomic datasets spanning human peripheral blood, human and macaque brain single-cell transcriptomes, and rat pharmacological models. A robust transcriptomic signature was identified that links systemic inflammation signals to central glial dysfunction. MDD-associated up-regulated genes were preferentially enriched in microglia, whereas down-regulated genes were predominantly associated with oligodendrocytes, indicating a coordinated shift toward a pro-inflammatory and demyelinating microenvironment. Multi-layer convergence prioritizes ANO10, HK2, and CCDC50 as core dysregulated genes. Among them, HK2 emerges as a central immunometabolic node, coupling glycolytic reprogramming to microglial activation. HK2 exhibits cell-type specificity, antidepressant-responsive dynamics, and independent clinical validation, supporting its role as a mechanistically anchored and translationally accessible marker. Together, these findings delineate a glial-centric immunometabolic framework for MDD and provide molecular evidence bridging peripheral signals with central microenvironmental pathology. Biological sciences/Biotechnology Health sciences/Biomarkers/Predictive markers Health sciences/Diseases/Psychiatric disorders/Depression Major Depressive Disorder Multi-omics Microenvironment Microglia Inflammation Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction Major Depressive Disorder (MDD) represents a pervasive and debilitating global health challenge, imposing a profound social and economic burden [1]. According to the Global Burden of Disease Study 2019, mental disorders remain among the top ten leading causes of disease burden worldwide, with MDD being a primary contributor to disability-adjusted life years [2]. As the prevalence of MDD continues to rise, the pathogenic mechanisms driven by the complex interplay between microenvironmental, environmental, and neurobiological factors remain poorly understood. At present, MDD diagnosis relies largely on subjective symptom assessment guided by DSM-5 (Diagnostic and Statistical Manual of Mental Disorders) or ICD-11 (International Classification of Diseases 11th Revision), most commonly using structured instruments such as the HAMD (Hamilton Depression Rating Scale) [1, 3]. However, this approach is inherently limited by inter-rater variability and the absence of objective biological criteria [4]. Moreover, an number of patients exhibit treatment resistance to current pharmacological therapies, often accompanied by poor tolerability and insufficient symptom improvement [5]. Moreover, the complex and heterogeneous nature of MDD, often together with its overlap with other psychiatric disorders and the difficulty of obtaining living brain tissue, has limited the development of reliable diagnostic biomarkers [6]. Therefore, there is a critical need to identify accessible microenvironment-associated phenotype biomarkers that can reliably reflect central pathological states and improve the diagnosis of MDD. Advances in multi-omics sequencing have facilitated cross-scale investigation of disease microenvironments and phenotypes, yet the neural microenvironment has been relatively under-integrated in biological studies of MDD [7]. New evidence suggests that the pathophysiology of depression is not confined to aberrant neurotransmission but involves a systemic dysregulation of the neurovascular unit, particularly the immune-glial interface [8, 9]. Postmortem studies link neuroinflammation, defective synaptic pruning, and impaired neurogenesis to the neuronal microenvironment [10]. In fact, MDD-associated genetic variants are enriched in regulatory regions, highlighting the potential role of transcriptional regulation and the neural microenvironment in disease pathophysiology [11]. These central observations are often constrained by postmortem intervals, medication confounds, and cause of death [12]. Fortunately, peripheral blood and single-cell sequencing provide a minimally invasive means to capture systemic physiological states [13, 14]. Although several studies have profiled peripheral transcriptomic changes, findings have been inconsistent, likely due to sample size limitations and the dilution of disease-associated signals in bulk tissues [15]. This raises a critical question: to what extent can molecular signals detected in the periphery serve as reliable proxies for central nervous system microenvironment alterations, particularly the inflammatory states of non-neuronal cells [16, 17]. Recent single-cell transcriptomic resolutions have underscored that non-neuronal cells-specifically microglia and oligodendrocytes-play pivotal roles in mediating neuroinflammation and white matter integrity compromise in MDD [18, 19]. Further, various neurobiological factors with diverse cell types have been linked to MDD, including monoaminergic or glutamatergic systems and abnormalities in astrocytes, oligodendrocytes or immune cells [20]. In spite of significant progress, the precise molecular and cellular mechanisms mediating the risk for MDD remain unknown. Traditional bulk RNA sequencing studies, based on clinically defined samples, provide rich clinical information but remain unable to resolve molecular changes at the level of individual cell types [21]. However, single-cell sequencing approaches often fail to capture the full spectrum of the disease due to the lack of systemic context and limited sample availability [22, 23]. A systematic, integrative multi-omics approach is needed to link molecular alterations across modalities and form closed-loop analyses and validations. Such strategies are essential to identify robust diagnostic biomarkers with biological relevance in the context of MDD [24, 25]. In this study, we established a comprehensive analytical framework to identify robust transcriptomic features of MDD present in both the blood and the brain, and to dissect their cell-type specificity. Firstly, we collected ten bulk RNA sequencing datasets with MDD and employed a dual strategy combining a gene intersection method and Robust Rank Aggregation (RRA) meta-analysis to rigorously screen for high-confidence differentially expressed genes (DEGs) across heterogeneous cohorts. Then, we integrated bulk-level findings with multiple single-cell RNA sequencing data using Bayesian deconvolution method and cell-cell communication analysis, enabling a cell-type-resolved investigation of the MDD microenvironment. Through multi-level integration, we identified three core genes- ANO10, HK2, and CCDC50 -as candidate microenvironmental biomarkers. Among them, HK2 (Hexokinase 2) showed the most consistent and pronounced dysregulation, specifically localized to microglia, indicating a potential immunometabolic reprogramming towards glycolysis. We further validated the robustness of these biomarkers in an independent clinical cohort using transcriptomic analysis and quantitative real-time PCR (qPCR). Collectively, this study systematically integrates existing multimodal sequencing data to provide a comprehensive view of the neuro-immune microenvironment in MDD and establishes HK2 as a robust candidate target for future diagnostic and therapeutic interventions. Materials and Methods Data collection and preprocessing To establish a comprehensive multi-omics framework, we collected ten transcriptomic datasets from the Gene Expression Omnibus (GEO), including six human peripheral blood bulk profiles (GSE98793 [26], GSE76826 [27], GSE290797 [28], GSE247998 [29], GSE99725 [30] and GSE185855 [31]), two cross-species single-cell RNA sequencing (scRNA-seq) datasets (macaque: GSE201687 [32]; human: GSE144136 [33]), and two rat bulk transcriptomes assessing antidepressant response (GSE222756 [34], GSE194289 [35]). Data processing and statistical analyses were executed within the R statistical environment (v4.4.2). To ensure data comparability, raw count datasets (GSE247998 and GSE185855) were subjected to log transformation followed by quantile normalization utilizing the limma package (v3.62.2). Datasets available as pre-processed normalized values were utilized without further modification to preserve original quality control standards. For single-cell data, processing pipelines were rigorously aligned with the protocols described in the primary study (GSE201687) [32]. Differential gene expression analysis To characterize transcriptomic alterations associated with MDD, differential expression analysis was performed across six independent peripheral blood RNA sequencing datasets using the limma package. Differentially expressed genes (DEGs) were identified based on a stringent threshold of Benjamini-Hochberg adjusted p -value log 2 (1.1). To systematically assess the consistency of DEGs across cohorts, gene overlaps were visualized using UpSetR (v1.4.0) and ggplot2 (v3.5.2). Identifying robust DEGs To address inter-study heterogeneity and derive a robust consensus transcriptional signature, we employed a dual-screening strategy that integrates gene intersection analysis and Robust Rank Aggregation (RRA) meta-analysis. First, an intersection analysis was conducted to identify DEGs consistently shared across datasets, with emphasis on genes exhibiting concordant directions of dysregulation. Second, high confidence genes were further screened using the RobustRankAggreg package (v1.2.1). This probabilistic framework integrates ranked gene lists-ordered by p -values-from individual datasets to evaluate the likelihood of each gene achieving consistently high ranks across heterogeneous cohorts [36]. Robust DEGs were defined as those meeting both an RRA score log 2 (1.1) across all including datasets. Pathway enrichment analysis To interpret the biological relevance of the identified genes, functional enrichment analysis was performed using the g:Profiler platform [37]. Enrichment testing was conducted against the "All known genes" statistical domain with a stringent user threshold of 0.01. Multiple functional annotation sources were queried, including Gene Ontology categories (GO: Biological Process, Cellular Component, and Molecular Function), as well as KEGG and Reactome pathway databases. Enrichment results were visualized using the tidyverse suite (v2.0.0). Single -cell data integration and gene module scoring To assess the cellular distribution of the robust DEGs signatures at single-cell resolution, gene module scores were computed using the AddModuleScore function within the Seurat package (v5.3.0). Consensus up-regulated and down-regulated gene sets were defined as separate modules, enabling quantification of their aggregated expression levels across individual cells in both scRNA-seq datasets. Module score distributions was visualized using ggplot2. Integration with single-cell RNA dataset To resolve the cellular heterogeneity in bulk transcriptomeic profiles, cell-type deconvolution was performed using the Bayesian inference framework BayesPrism (v2.2.2) [38]. A high-resolution human brain scRNA-seq dataset (GSE144136) served as the reference signature matrix to infer cellular proportions across all bulk transcriptomic cohorts. All analyses were conducted using default parameters. Cell-cell communication analysis Intercellular signaling networks were inferred using CellChat (v1.6.1) [39] to characterize signaling cell interactions within the neural microenvironment. Based on the CellChatDB.human ligand-receptor interaction database, the probability and strength of cell-cell communication events among distinct cell populations were systematically quantified in the single-cell datasets. Screen cell-type-specific DEGs for identification of core genes To delineate the cellular origin of the identified transcriptomic signatures, cell-type specific markers genes were first identified from the scRNA-seq datasets using the FindAllMarkers function in Seurat (parameters: only.pos = TRUE, min.pct = 0.1, logfc.threshold = 0.25). For each major cell population, the top 500 marker genes ranked by average log fold change were retained. These marker genes were then intersected with the robust up-regulated and down-regulated DEGs derived from the bulk transcriptomic meta-analysis. Overlap statistics were summarized and visualized using ggplot2. Screening of cell type specific and drug responsive candidates To assess the therapeutic relevance of the identified candidates, cross-species ortholog mapping from human to rat was performed using the gprofiler2 package (v0.2.3). The resulting homologous genes were subsequently projected onto two independent rat antidepressant treatment datasets. Gene expression distributions across treatment conditions visualized using violin and box plots. Core DEGs were stringently defined according to a therapeutic reversal criterion, whereby candidate genes exhibited aberrant expression in the depression-like state and demonstrated significant normalization or directional reveal following antidepressant administration. Protein-Protein interaction (PPI) network analysis To explore the functional connectivity among the identified core genes, protein–protein interaction (PPI) network was constructed using the STRING database (https://cn.string-db.org) [40]. The analysis was conducted using default parameters to visualize the potential interactions. Blood samples for the validation cohort To validate the in silico findings, an independent clinical cohort comprising six patients with MDD (aged 18-45 years) and five healthy controls (HC) was recruited from the Department of Clinical Psychology at The Third People’s Hospital of Foshan. The diagnosis of major depressive disorder (MDD) was made by two physicians, and its severity was assessed using the 24-item Hamilton Depression Rating Scale (HAMD-24). Strict exclusion criteria were applied to minimize confounding factors. Participants were excluded if they presented with: (1) comorbid psychiatric disorders (e.g., schizophrenia, bipolar disorder, or anxiety disorders); (2) a family history of mental illness; (3) acute or chronic somatic pathologies affecting the immune system (e.g., autoimmune diseases or cancer); (4) current pregnancy; or (5) use of psychotropic medications (antidepressants), immunomodulators, or hormonal drugs within the two weeks preceding sampling. Peripheral venous blood samples were collected into PAXgene Blood RNA Tubes to ensure transcript stability. Specimens were immediately processed and stored at −80°C until RNA extraction. The study protocol was approved by The Medical Ethics Committee of The Third People's Hospital of Foshan (Foshan Mental Health Center) and was conducted in strict accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants prior to inclusion. RNA extraction and transcriptome data construction Total RNA was isolated using Trizol Reagent (Invitrogen Life Technologies). RNA purity and concentration were assessed using a NanoDrop spectrophotometer (Thermo Scientific). Sequencing libraries were constructed using the NEBNext Ultra II RNA Library Prep Kit for Illumina (NEB), strictly adhering to the manufacturer’s instructions. Library quality and fragment size distribution were verified using the Agilent High Sensitivity DNA Assay on a Bioanalyzer 2100 system (Agilent). Subsequently, libraries were sequenced on the Illumina NovaSeq 6000 platform to generate paired-end reads. Raw sequencing data (FASTQ format) were subjected to quality control using Cutadapt (v1.15) to remove adaptor sequences and low-quality reads. The resulting clean reads were aligned to the reference genome using the splice-aware aligner HISAT2 (v2.0.5). Finally, gene expression levels were quantified by generating raw read counts using HTSeq (v0.9.1). Real-time quantitative PCR (RT-qPCR) validation To experimentally validate the identified core transcriptomic signatures, RT-qPCR was performed on whole blood samples derived from the independent validation cohort (n = 11). Total RNA was extracted using Trizol reagent, followed by first-strand cDNA synthesis utilizing the Advantage® RT-for-PCR Kit. Amplification reactions were conducted on a LightCycler 480 II instrument using AceQ® qPCR SYBR® Green Master Mix. Specific primer sequences are detailed in Table 1. The thermal cycling protocol encompassed an initial denaturation at 95℃ for 5 min, followed by 40 cycles of 95℃ for 15 s and 60℃ for 30 s. Melting curve analysis was subsequently performed to verify amplicon specificity. Relative mRNA expression levels were quantified using the comparative C T (2 −ΔΔCt ) method, normalized to the endogenous reference gene GAPDH . Table 1 PCR primers used in this study Primer Name Sequence (5' to 3') ANO10 ENSG00000160746-F TCAGAACCTTCAGCCAAT ANO10 ENSG00000160746-R ACAGTTAGTGACCACAGATAT HK2 ENSG00000159399-F AAGACATTAGAGCATCTG HK2 ENSG00000159399-R CTCCATTTCTACCTTCAT CCDC50 ENSG00000152492-F AAGGAGTTACAGGAAGAG CCDC50 ENSG00000152492-R TCGTGATGGAGTAGATAC gapdh ENSG00000111640(nei)-F CTCTGGTAAAGTGGATATTGT gapdh ENSG00000111640(nei)-R GGTGGAATCATATTGGAACA Data availability Raw RNA-sequencing data (count files) and code generated in this study have been deposited on GitHub and are publicly available at https://github.com/Reisen333/MDD-Datasets. Publicly available datasets analyzed in this study are available in the Gene Expression Omnibus (GEO) repository under the following accession codes: GSE98793, GSE76826, GSE290797, GSE247998, GSE99725, GSE185855 (human blood bulk transcriptomes); GSE201687 (macaque brain scRNA-seq); GSE144136 (human brain scRNA-seq); and GSE222756, GSE194289 (Rat pharmacological models). The human reference genome (GRCh38.p13) and gene annotation files used for alignment were downloaded from the Ensembl database (http://asia.ensembl.org/Homo_sapiens/Info/Index). Source data for the figures are provided with this paper. All other data supporting the findings of this study are available from the corresponding authors upon reasonable request. Results Robust transcriptomic genes of MDD Initially, we conducted a comprehensive differential expression analysis across six independent human peripheral blood transcriptomic cohorts. To ensure the inclusion of contemporary data, all selected datasets were published after 2016. These final datasets comprised 326 MDD patients and 170 control samples (Fig. 1A, Supplementary Table S1). We identified a reproducible transcriptomic signature ( p log 2 (1.1)). However, substantial variation in the number of identified differentially expressed genes (DEGs) was observed across six datasets (Fig. 1B-C, Supplementary Table S2-S7), with the limited overlap between individual cohorts. This pronounced inter-study inconsistency underscores the profound biological heterogeneity inherent in MDD (Supplementary Fig. 2-3). Such single-cohort analysis is inherently limited and may be insufficient to fully capture the underlying pathogenic mechanisms of the disorder. To circumvent potential batch effects inherent in merging raw transcriptomic data, we employ an integrative framework combining gene intersection analysis with Robust Rank Aggregation (RRA). The RRA approach designed to detect consistently high-ranking genes across, heterogeneous lists—identified a robust signature of 3161 dysregulated genes (Fig. 1D; Supplementary Table S8). Notably, this RRA-derived signature effectively captured the signals from the pairwise intersection while expanding the repertoire to include latent but reproducible markers. This integrative gain underscores the pronounced molecular heterogeneity of MDD, where true disease-associated signals may manifest as subtle but consistent patterns rather than large cohort-specific effects. Consequently, conventional single-cohort bulk transcriptomic analyses are often insufficient to identify robust and reproducible biomarkers in such complex neuropsychiatric disorders. Single-cell landscape and cross-species validation of the robust MDD signature To systematically evaluate the biological relevance of potential biomarkers and prepare for functional characterization, we defined three distinct gene sets, including the "Overlap method" set, comprising the 99 DEGs shared between the two datasets (GSE98793 and GSE76826) (Supplementary Table S9), the "RRA Method" set, comprising the 3161 robust genes identified by the meta-analysis; and the "ALL DEGs" set, comprising all 9577 DEGs identified across all cohorts. We first performed functional enrichment analysis, which revealed a convergent dysregulation of pathways governing immune/inflammatory responses, organelle integrity, and cellular stress adaptation in both intersection- and RRA-derived signatures (Fig. 2A, Supplementary Table S10-S12). These findings suggest that peripheral transcriptomic alterations in MDD are underpinned by a systemic immunometabolic disruption. To investigate whether these systemic signals mirror specific central cellular pathologies, we integrated these genes with scRNA-seq data from both human and macaque brains (Fig. 2B). Cell composition analysis demonstrated a significant expansion of the microglial compartment alongside a concurrent reduction of the oligodendrocyte cells in the depressed brain (Fig. 2C). Further, gene module scoring revealed a striking cellular dichotomy: the robust up-regulated genes were predominantly enriched in microglia, whereas the down-regulated genes mapped extensively to oligodendrocytes (Fig. 2D). Collectively, these results imply that the core MDD signature reflects a fundamental imbalance between microglial activation and oligodendroglial support within the neuro-microenvironment. To bridge the gap between peripheral signals and central pathology, we applied Bayesian deconvolution to the six peripheral blood transcriptomic datasets and two rat pharmacological cohorts. Remarkably, blood transcriptomes of MDD patients faithfully recapitulated the central neuro-microenvironmental landscape, exhibiting an estimated increase in microglial cells and decrease in oligodendrocyte cells (Fig. 2E, Supplementary Fig. 3A-B). Importantly, antidepressant treatment in rat models reversed these pathological cellular trends, leading to a significant reduction in estimated microglial and oligodendrocyte proportions relative to the depression model (Fig. 2F, Supplementary Fig. 3C). Drug-specific effects were also observed. For example, fluoxetine and desipramine treatment were associated with an expansion of astrocyte and interneuron populations and a relative reduction in excitatory neurons (Fig. 2F), suggesting a partial restoration of cellular homeostasis within the brain microenvironment. To further detect the functional connectivity within this altered microenvironment, cell-cell communication analysis was performed. The depressed brain exhibited a hyperconnected signaling network, characterized by increased ligand–receptor interactions across cell types in both human and macaque single-cell datasets. Notably, microglia emerged as a central signaling hub in this intensified communication network (Fig. 2G, Supplementary Fig. 4), indicating a dominant role in orchestrating aberrant intercellular crosstalk. Collectively, these findings indicate that peripheral molecular signatures faithfully reflect specific central microenvironment dysfunctions in MDD, including microglial hyperactivation, compromised oligodendrocyte integrity, and disrupted glial–neuronal communication. Single-cell and pharmacological validation of key microenvironmental markers Building on the delineation of MDD-associated cellular alterations, we further explored the specific molecular drivers underlying glial dysfunctions. To prioritize candidate genes possessing both cellular specificity and therapeutic relevance, an intersectional analysis was performed between the robust DEGs identified from bulk meta-analysis and the top 500 marker genes defined for each cell type in the scRNA-seq datasets. This analysis revealed that up-regulated consensus genes were significantly enriched for microglial markers, whereas down-regulated genes showed substantial overlap with oligodendrocyte markers (Fig. 3A, Supplementary Table S13-S14), reinforcing the cell-type specificity of the transcriptomic signatures. Importantly, these cell-type–resolved candidates were also highly ranked in the RRA meta-analysis, supporting their robustness across heterogeneous cohorts. Among the top-prioritized genes were ANO10, HK2, TFEB, and ZNF532 (Fig. 3B), highlighting their potential roles as key regulators of microenvironmental dysregulation in MDD. We then sought to validate the translational reliability of these candidates by examining their expression dynamics across independent human blood cohorts and rat pharmacological models. The results showed that the microglia-associated genes were robustly up-regulated, whereas oligodendrocyte-associated genes were down-regulated in MDD blood samples (Fig. 3C, Supplementary Fig. 5-6). Importantly, we found that antidepressant treatment significantly reversed these pathological expression patterns in rat treatment datasets. For instance, administration of fluoxetine significantly attenuated the aberrant up-regulation of ANO10 and HK2 , while concurrently restoring the suppressed levels of CCDC50 (Fig. 3C, Supplementary Fig. 7-8). Collectively, this multi-stage screening strategy identified a core set of microglia-specific genes and oligodendrocyte-specific genes that demonstrate dynamic responsiveness to therapeutic interventions, emphasizing their potential as promising targets for MDD treatment. Validation and functional characterization the core biomarkers Based on the integrated analysis, we prioritized ANO10 , HK2 , and CCDC50 as core candidate biomarkers. These genes were consistently identified by both intersectional and RRA strategies and ranked among the top-tier positions in the RRA significance rankings (Fig. 4A). Heatmap visualization demonstrated a highly consistent expression profile for these genes across all six independent human blood cohorts (Fig. 4B), underscoring their stability as systemic markers. To link these peripheral signals to the central neuro-microenvironment, we found ANO10 and HK2 were predominantly enriched in the microglial compartment, whereas CCDC50 mapped specifically to the oligodendrocyte lineage (Fig. 4C). These findings were corroborated in bulk datasets: the aberrant upregulation of microglial markers ( ANO10, HK2 ) and downregulation of the oligodendrocyte marker ( CCDC50 ) observed in MDD were effectively reversed by antidepressant administration in rat models (Fig. 4D), confirming their role as pharmacodynamically responsive targets. To clarify the molecular mechanisms underlying these core genes, we constructed a Protein-Protein Interaction (PPI) network. The results showed that ANO10 was linked to ion channel regulation, CCDC50 modulated EGFR and NOTCH pathways (Fig. 4E), and HK2 emerged as a central hub in glycolysis and glucose metabolism. Functional enrichment analysis further positioned these genes-particularly HK2 , at the intersection of response to stress, metabolic processes, and nervous system development (Fig. 4F). Together, these findings confirm ANO10, HK2, and CCDC50 as high-confidence peripheral biomarkers that reflect specific central glial dysfunctions: HK2 may drive microglial immunometabolic reprogramming, while CCDC50 dysregulation points to compromised oligodendrocyte signaling. Independent cohort recapitulates microglial signatures and establishes HK2 as a robust cross-platform To corroborate our in-silico findings, we performed transcriptomic profiling on an independent validation cohort comprising six MDD patients and five healthy controls (Fig. 5A). Cell-type enrichment analysis corroborated the central role of microglia, as the upregulated gene module was again significantly enriched in the microglial signature (Fig. 5B, Supplementary Table S15), substantiating a pro-inflammatory glial phenotype. Consistent with findings from the discovery cohorts, differential expression analysis revealed a transcriptional landscape strongly enriched for immune-related pathways, including Toll-like receptor (TLR), Th1/Th2 differentiation, and JAK–STAT signaling cascades (Fig. 5C, Supplementary Table S16). We then examined the expression of the core candidate genes- ANO10, HK2, and CCDC50 . RNA-sequencing data from the validation cohort confirmed significant up-regulation of ANO10 and HK2 , alongside down-regulation of CCDC50 , fully recapitulating the directionality observed in the discovery analyses (Fig. 5D). To further assess the technical robustness of these candidates, RT-qPCR validation was performed. Notably, HK2 exhibited consistent up-regulation, whereas ANO10 and CCDC50 failed to demonstrate reproducible dysregulation in the qPCR assays (Fig. 5E). This discrepancy may reflect differences in baseline expression levels, platform sensitivity, or increased inter-individual variability for these targets. Taken together, these results demonstrate that while a broader microglial-associated transcriptional signature is reproducibly observed in MDD, cross-platform validation effectively narrows the candidate pool, establishing HK2 as the most robust and reproducible peripheral biomarker reflecting central microenvironmental dysregulation in MDD. Discussion MDD is a neuropsychiatric disorder for which clinical diagnosis still relies predominantly on symptom-based interviews and psychometric questionnaires, limiting objective quantification of disease state and treatment response. Moreover, the substantial inter-individual heterogeneity inherent to MDD has constrained the utility of single-modality transcriptomic studies, where disease signals are often obscured by biological variability, yielding poor reproducibility of candidate biomarkers [9]. In addition, current antidepressant therapies frequently produce incomplete remission and intolerable side effects, underscoring the need for biologically grounded stratification and monitoring tools. In this study, a central finding is that peripheral blood transcriptomes, when interpreted through an integrative framework, can faithfully identify key features of the altered neuro-microenvironment in MDD, most notably microglial activation and oligodendrocyte dysfunction. The concordance observed across peripheral bulk transcriptomics, brain single-cell profiles, and pharmacological response datasets challenges the long-standing assumption that peripheral molecular signals are inherently too nonspecific to inform central pathology. Instead, these results support a model in which depression-related immune–metabolic dysregulation is structured at the microenvironmental level and can be consistently detected across biological compartments. By jointly leveraging large-scale peripheral transcriptomics, cell-type–resolved single-cell data, and drug-response signatures in animal models, this study establishes a closed-loop validation framework for biomarker discovery. Functional annotation of the identified robust signatures revealed a convergent enrichment in pathways governing immunomodulation, organelle integrity, and cellular stress adaptation. This prominent immune signature strongly corroborates the prevailing neuroinflammation hypothesis of MDD [41, 42]. Beyond immune activation, the dysregulation of pathways associated with cell membrane dynamics and cellular stress responses points to widespread subcellular compromise in MDD. Specifically, the implication of mechanisms such as mitochondrial dysfunction and oxidative stress suggests a cascade of bioenergetic failure. Such disruptions are known to impair glial-neuronal homeostasis, disrupting energy metabolism and ultimately compromising synaptic transmission [43]. The activation of these pathways indicated a maladaptive cellular stress response, wherein intrinsic homeostatic mechanisms are overwhelmed by chronic stressors, shifting the cellular phenotype from adaptive survival to pathological dysfunction. The consistent reversal of microenvironment-associated signatures following antidepressant treatment further highlights their functional relevance and disease specificity. Collectively, these findings underscore the necessity of multimodal integration to identify robust, microenvironment-informed diagnostic biomarkers in MDD—biomarkers that move beyond isolated molecular changes toward a systems-level representation of depressive pathology. Having delineated the cellular landscape of MDD, our ultimate objective was to pinpoint the specific molecular drivers orchestrating these glial dysfunctions. The identification of HK2 emerged as a convergent and biologically coherent finding rather than a dataset-specific observation. HK2 was consistently prioritized across peripheral transcriptomic cohorts, multiple single-cell brain datasets, and pharmacological response models, reflecting its central role at the intersection of microglial activation, metabolic regulation, and immune signaling. Given the tight coupling between microglial inflammatory states and glycolytic reprogramming, the microglia-specific upregulation of HK2 supports a model in which immunometabolic shifts contribute to microenvironmental dysfunction in MDD. Importantly, the translational relevance of HK2 was further supported by independent clinical validation, as both transcriptomic sequencing and RT–qPCR analyses of a hospital-derived cohort consistently confirmed its elevated expression in MDD. Despite the robustness of this integrative framework, several limitations merit consideration. First, while the RRA-based strategy effectively reduced inter-study heterogeneity, it may have obscured transcriptomic features linked to specific depressive subtypes or sex-specific effects. Second, the predominantly cross-sectional design limits causal inference, precluding definitive conclusions regarding whether glial dysregulation represents a primary driver or a downstream consequence of MDD. Finally, pharmacological validation relied mainly on rodent models, and although supported by non-human primate data, species-specific immune differences cannot be fully excluded. In summary, this study establishes a systematic multi-omics framework that bridges peripheral circulation and central pathology in MDD, revealing a convergent immunometabolic microenvironment characterized by microglial hyperactivation and oligodendrocyte dysfunction. Through cross-modal and cross-platform validation, HK2 emerged as a robust, cell-type–resolved peripheral biomarker with translational relevance. These findings provide molecular evidence narrowing the blood–brain gap and support a glial-centric paradigm for precision diagnosis and therapeutic monitoring in major depressive disorder. Declarations Acknowledgements This work was funded by the National Natural Science Foundation of China (12501670), Guangdong Basic and Applied Basic Research Foundation (No. 2025A1515011988). Competing of Interests The authors declare that they have no competing interests. References Han J, Feng Y, Li N, Feng L, Xiao L, Zhu X et al. Correlation Between Word Frequency and 17 Items of Hamilton Scale in Major Depressive Disorder. Frontiers in psychiatry 2022; 13: 902873. Chen Q, Huang S, Xu H, Peng J, Wang P, Li S et al. The burden of mental disorders in Asian countries, 1990-2019: an analysis for the global burden of disease study 2019. Translational psychiatry 2024; 14 (1) : 167. Fiorillo A, Albert U, Dell'Osso B, Pompili M, Sani G, Sampogna G. The clinical utility and relevance in clinical practice of DSM-5 specifiers for major depressive disorder: A Delphi expert consensus study. Comprehensive psychiatry 2024; 133: 152502. Cosci F, Fava GA. The clinical inadequacy of the DSM-5 classification of somatic symptom and related disorders: an alternative trans-diagnostic model. CNS spectrums 2016; 21 (4) : 310-317. Voineskos D, Daskalakis ZJ, Blumberger DM. Management of Treatment-Resistant Depression: Challenges and Strategies. Neuropsychiatric disease and treatment 2020; 16: 221-234. Anagha K, Shihabudheen P, Uvais NA. Side Effect Profiles of Selective Serotonin Reuptake Inhibitors: A Cross-Sectional Study in a Naturalistic Setting. Prim Care Companion CNS Disord 2021; 23 (4). Wohleb ES, Franklin T, Iwata M, Duman RS. Integrating neuroimmune systems in the neurobiology of depression. Nature reviews Neuroscience 2016; 17 (8) : 497-511. Collignon A, Dion-Albert L, Ménard C, Coelho-Santos V. Sex, hormones and cerebrovascular function: from development to disorder. Fluids and barriers of the CNS 2024; 21 (1) : 2. Carrier M, Dolhan K, Bobotis BC, Desjardins M, Tremblay M. The implication of a diversity of non-neuronal cells in disorders affecting brain networks. Frontiers in cellular neuroscience 2022; 16: 1015556. Sălcudean A, Bodo C-R, Popovici R-A, Cozma M-M, Păcurar M, Crăciun R-E et al. Neuroinflammation—A Crucial Factor in the Pathophysiology of Depression—A Comprehensive Review. Biomolecules 2025; 15 (4) : 502. Chawla A, Cakmakci D, Fiori LM, Zang W, Maitra M, Yang J et al. Single-nucleus chromatin accessibility profiling identifies cell types and functional variants contributing to major depression. Nat Genet 2025; 57 (8) : 1890-1904. Mariani N, Cattane N, Pariante C, Cattaneo A. Gene expression studies in Depression development and treatment: an overview of the underlying molecular mechanisms and biological processes to identify biomarkers. Transl Psychiatry 2021; 11 (1) : 354. Medeiros GC, Gould TD, Prueitt WL, Nanavati J, Grunebaum MF, Farber NB et al. Blood-based biomarkers of antidepressant response to ketamine and esketamine: A systematic review and meta-analysis. Molecular psychiatry 2022; 27 (9) : 3658-3669. Mokhtari A, Ibrahim EC, Gloaguen A, Barrot CC, Cohen D, Derouin M et al. Using multiomic integration to improve blood biomarkers of major depressive disorder: a case-control study. EBioMedicine 2025; 113: 105569. Ofengeim D, Giagtzoglou N, Huh D, Zou C, Yuan J. Single-Cell RNA Sequencing: Unraveling the Brain One Cell at a Time. Trends Mol Med 2017; 23 (6) : 563-576. Li JM, Jiang CL. Biological Diagnosis of Depression: A Biomarker Panel from Several Nonspecial Indicators Instead of the Specific Biomarker(s). Neuropsychiatr Dis Treat 2022; 18: 3067-3071. Yamamoto M, Sakai M, Yu Z, Nakanishi M, Yoshii H. Glial Markers of Suicidal Behavior in the Human Brain-A Systematic Review of Postmortem Studies. Int J Mol Sci 2024; 25 (11). Yirmiya R, Rimmerman N, Reshef R. Depression as a microglial disease. Trends in neurosciences 2015; 38 (10) : 637-658. Edgar N, Sibille E. A putative functional role for oligodendrocytes in mood regulation. Translational psychiatry 2012; 2 (5) : e109. Nagy C, Suderman M, Yang J, Szyf M, Mechawar N, Ernst C et al. Astrocytic abnormalities and global DNA methylation patterns in depression and suicide. Molecular psychiatry 2015; 20 (3) : 320-328. Buenrostro JD, Wu B, Litzenburger UM, Ruff D, Gonzales ML, Snyder MP et al. Single-cell chromatin accessibility reveals principles of regulatory variation. Nature 2015; 523 (7561) : 486-490. Denisenko E, Guo BB, Jones M, Hou R, de Kock L, Lassmann T et al. Systematic assessment of tissue dissociation and storage biases in single-cell and single-nucleus RNA-seq workflows. Genome biology 2020; 21 (1) : 130. Squair JW, Gautier M, Kathe C, Anderson MA, James ND, Hutson TH et al. Confronting false discoveries in single-cell differential expression. Nature communications 2021; 12 (1) : 5692. Strawbridge R, Young AH, Cleare AJ. Biomarkers for depression: recent insights, current challenges and future prospects. Neuropsychiatric disease and treatment 2017; 13: 1245-1262. Le-Niculescu H, Roseberry K, Gill SS, Levey DF, Phalen PL, Mullen J et al. Precision medicine for mood disorders: objective assessment, risk prediction, pharmacogenomics, and repurposed drugs. Molecular psychiatry 2021; 26 (7) : 2776-2804. Leday GGR, Vértes PE, Richardson S, Greene JR, Regan T, Khan S et al. Replicable and Coupled Changes in Innate and Adaptive Immune Gene Expression in Two Case-Control Studies of Blood Microarrays in Major Depressive Disorder. Biol Psychiatry 2018; 83 (1) : 70-80. Miyata S, Kurachi M, Okano Y, Sakurai N, Kobayashi A, Harada K et al. Blood Transcriptomic Markers in Patients with Late-Onset Major Depressive Disorder. PLoS One 2016; 11 (2) : e0150262. Di Benedetto MG, Priestley KMS, Cattane N, Genini P, Saleri S, Biaggi A et al. Blood transcriptomic signatures associated with depression, or the risk for depression, in pregnant women from the Psychiatry Research And Motherhood - Depression (PRAM-D) study. Transl Psychiatry 2025; 15 (1) : 110. Sun S, Liu Q, Wang Z, Huang YY, Sublette ME, Dwork AJ et al. Brain and blood transcriptome profiles delineate common genetic pathways across suicidal ideation and suicide. Mol Psychiatry 2024; 29 (5) : 1417-1426. Moisan MP, Foury A, Dexpert S, Cole SW, Beau C, Forestier D et al. Transcriptomic signaling pathways involved in a naturalistic model of inflammation-related depression and its remission. Transl Psychiatry 2021; 11 (1) : 203. Cathomas F, Bevilacqua L, Ramakrishnan A, Kronman H, Costi S, Schneider M et al. Whole blood transcriptional signatures associated with rapid antidepressant response to ketamine in patients with treatment resistant depression. Transl Psychiatry 2022; 12 (1) : 12. Wu J, Li Y, Huang Y, Liu L, Zhang H, Nagy C et al. Integrating spatial and single-nucleus transcriptomic data elucidates microglial-specific responses in female cynomolgus macaques with depressive-like behaviors. Nat Neurosci 2023; 26 (8) : 1352-1364. Nagy C, Maitra M, Tanti A, Suderman M, Théroux JF, Davoli MA et al. Single-nucleus transcriptomics of the prefrontal cortex in major depressive disorder implicates oligodendrocyte precursor cells and excitatory neurons. Nat Neurosci 2020; 23 (6) : 771-781. Rayan NA, Aow J, Lim MGL, Arcego DM, Ryan R, Nourbakhsh N et al. Shared and unique transcriptomic signatures of antidepressant and probiotics action in the mammalian brain. Mol Psychiatry 2024; 29 (11) : 3653-3668. Rayan NA, Kumar V, Aow J, Rastegar N, Lim MGL, O'Toole N et al. Integrative multi-omics landscape of fluoxetine action across 27 brain regions reveals global increase in energy metabolism and region-specific chromatin remodelling. Mol Psychiatry 2022; 27 (11) : 4510-4525. Kolde R, Laur S, Adler P, Vilo J. Robust rank aggregation for gene list integration and meta-analysis. Bioinformatics 2012; 28 (4) : 573-580. Raudvere U, Kolberg L, Kuzmin I, Arak T, Adler P, Peterson H et al. g:Profiler: a web server for functional enrichment analysis and conversions of gene lists (2019 update). Nucleic Acids Res 2019; 47 (W1) : W191-w198. Chu T, Wang Z, Pe'er D, Danko CG. Cell type and gene expression deconvolution with BayesPrism enables Bayesian integrative analysis across bulk and single-cell RNA sequencing in oncology. Nat Cancer 2022; 3 (4) : 505-517. Jin S, Plikus MV, Nie Q. CellChat for systematic analysis of cell-cell communication from single-cell transcriptomics. Nat Protoc 2025; 20 (1) : 180-219. Szklarczyk D, Nastou K, Koutrouli M, Kirsch R, Mehryary F, Hachilif R et al. The STRING database in 2025: protein networks with directionality of regulation. Nucleic Acids Res 2025; 53 (D1) : D730-d737. Ruiz NAL, Del Ángel DS, Brizuela NO, Peraza AV, Olguín HJ, Soto MP et al. Inflammatory Process and Immune System in Major Depressive Disorder. Int J Neuropsychopharmacol 2022; 25 (1) : 46-53. Cui L, Li S, Wang S, Wu X, Liu Y, Yu W et al. Major depressive disorder: hypothesis, mechanism, prevention and treatment. Signal Transduct Target Ther 2024; 9 (1) : 30. Yin Y, Ju T, Zeng D, Duan F, Zhu Y, Liu J et al. "Inflamed" depression: A review of the interactions between depression and inflammation and current anti-inflammatory strategies for depression. Pharmacol Res 2024; 207: 107322. Additional Declarations The authors have declared there is NO conflict of interest to disclose Supplementary Files SupplementaryTableS1Theinformationofthedatabase.xlsx Supplementary Table S1 finalSupplementaryinformation.docx Supplementary information SupplementaryTablesS6GSE99725DEGLists.xlsx Supplementary Table S6 SupplementaryTablesS8DEGsfoundbyRRAmethods.xlsx Supplementary Table S8 SupplementaryTablesS10ALLDEGpathwayenrichmentresult.xlsx Supplementary Table S10 SupplementaryTablesS13GSE144136CellTypeMarkers.xlsx Supplementary Table S13 SupplementaryTablesS9DEGsfoundbyGeneintersectionmethods.xlsx Supplementary Table S9 SupplementaryTableS15ValidationDEGs.xlsx Supplementary Table S15 SupplementaryTablesS12RRAmethodpathwayenrichmentresult.xlsx Supplementary Table S12 SF4HyperconnectedintercellularsignalinglandscapeintheMDDmicroenvironment.tif Supplementary Figure. 4 SupplementaryTablesS11Geneintersectionmethodpathwayenrichmentresult.xlsx Supplementary Table S11 SupplementaryTableS16ValidationPathwayenrichmentresult.xlsx Supplementary Table S16 SupplementaryTablesS3GSE76826DEGLists.xlsx Supplementary Table S3 SupplementaryTablesS5GSE247998DEGLists.xlsx Supplementary Table S5 SF7Pharmacologicalattenuationofglialmarkerdysregulationbyfluoxetineinratmodel.tif Supplementary Figure. 7 SupplementaryTablesS14GSE201687CellTypeMarkers.xlsx Supplementary Table S14 SF5RecapitulationoftheglialspecifictranscriptomicsignatureintheGSE98793cohort.tif Supplementary Figure. 5 SF6IndependentvalidationofcelltypespecificmarkerdysregulationintheGSE76826cohort.tif Supplementary Figure. 6 SF2Transcriptomicheterogeneityanddifferentialexpressionlandscapeacrosssixindependentperipheralbloodcohorts.tif Supplementary Figure. 2 SF3Deconvolutionvalidatescentralglialshiftsinperipheralbloodandtheirpharmacologicalreversal.tif Supplementary Figure. 3 SupplementaryTablesS4GSE290797DEGLists.xlsx Supplementary Table S4 SF1Visualizingtranscriptomicheterogeneityacrosscohorts.tif Supplementary Figure. 1 SF8BroadpharmacologicalreversaloftheMDDassociatedglialsignatureacrossmultipleantidepressanttreatment.tif Supplementary Figure. 8 SupplementaryTablesS7GSE185855DEGLists.xlsx Supplementary Table S7 SupplementaryTablesS2GSE98793DEGLists.xlsx Supplementary Table S2 Cite Share Download PDF Status: Under Review Version 1 posted Review # 1 received at journal 10 May, 2026 Reviewer # 1 agreed at journal 19 Apr, 2026 Reviewers invited by journal 18 Apr, 2026 Editor assigned by journal 27 Feb, 2026 Submission checks completed at journal 27 Feb, 2026 First submitted to journal 26 Feb, 2026 Unknown event 26 Feb, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8976053","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":625559455,"identity":"8e236a53-606d-4d1f-a5c6-40f236373dbc","order_by":0,"name":"Hui Tang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA00lEQVRIie3PsQrCMBCA4ZRCs5y43qK+ghIoQgdfJV3MEgUXcRDJlC4FZ5/GQsGpdXYVwUUHwV2Moo5t3ATzTxe4jySEuFw/mk8Itql6z3XBizDIviQkTq3JgJb5aaL7IqXJEcksihUts+pbYDyMVhpHKRQhkkLECsa85mEyZA1D1igD9HQeK4RuNWmen0QAiiN6NxuCkh0M4YA8RE/ZkN059GGLPfMX1ucbwTTIakKXkl1huugATfa7yzxqLWlRTUwBfkb+ONbtm/yLxZLL5XL9c3ex3jnVYGOZ8QAAAABJRU5ErkJggg==","orcid":"","institution":"Foshan University, Foshan, China","correspondingAuthor":true,"prefix":"","firstName":"Hui","middleName":"","lastName":"Tang","suffix":""},{"id":625559456,"identity":"8dae2d18-35e5-4c78-843b-821922fbb272","order_by":1,"name":"Yunkai Wu","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Yunkai","middleName":"","lastName":"Wu","suffix":""},{"id":625559457,"identity":"4ab2828c-5493-4edc-986f-e353a6d78421","order_by":2,"name":"Weihao Deng","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Weihao","middleName":"","lastName":"Deng","suffix":""},{"id":625559458,"identity":"bcd5529c-3471-44c2-84f8-11baebad7bfc","order_by":3,"name":"Jiaquan Liang","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Jiaquan","middleName":"","lastName":"Liang","suffix":""},{"id":625559459,"identity":"c16c3677-efb3-4a9d-b59c-deccb1c20289","order_by":4,"name":"Chunguo Zhang","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Chunguo","middleName":"","lastName":"Zhang","suffix":""},{"id":625559460,"identity":"4b6fb6c2-242d-4e12-807f-a86be42aacf8","order_by":5,"name":"Jinyan Sun","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Jinyan","middleName":"","lastName":"Sun","suffix":""}],"badges":[],"createdAt":"2026-02-26 09:36:19","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8976053/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8976053/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":107875538,"identity":"e02f2a75-f912-445c-9b48-abfa467fcdf0","added_by":"auto","created_at":"2026-04-27 08:10:05","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":19242526,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCharacterization of transcriptomic heterogeneity and integrative identification of consensus transcriptomic genes. A\u003c/strong\u003e Schematic of the multi-omics bioinformatics analysis pipeline. \u003cstrong\u003eB\u003c/strong\u003e Bar plots showing the number of up-regulated (red) and down-regulated (blue) differentially expressed genes (DEGs) across six independent peripheral blood datasets (adjusted \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.05 and |log\u003csub\u003e2\u003c/sub\u003e(FC)| \u0026gt; log\u003csub\u003e2\u003c/sub\u003e(1.1)). \u003cstrong\u003eC\u003c/strong\u003e Intersection analysis of DEGs. Left: UpSet plot quantifying the shared and unique DEGs across all datasets. Right: Venn diagram illustrating the overlap of up-regulated and down-regulated genes between two representative datasets (GSE98793 and GSE76826). \u003cstrong\u003eD\u003c/strong\u003e Volcano plot derived from Robust Rank Aggregation (RRA) meta-analysis integrating all six datasets. Red and blue points denote robustly up-regulated and down-regulated genes, respectively (RRA score \u0026lt; 0.05).\u003c/p\u003e","description":"","filename":"Figure1.png","url":"https://assets-eu.researchsquare.com/files/rs-8976053/v1/a90b4f89177dd17cdadd6c0d.png"},{"id":107875256,"identity":"feed016c-5747-401a-a82e-b2c178bb1b0d","added_by":"auto","created_at":"2026-04-27 08:09:23","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":25179961,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSystemic immunometabolic signatures mirror central microglial activation and oligodendrocyte dysfunction. A\u003c/strong\u003eDot plot displaying the top 20 enriched biological pathways (GO/KEGG/Reactome) for robust DEG sets identified by intersection and RRA methods. \u003cstrong\u003eB \u003c/strong\u003eUMAP visualization of major cell types integrated from human and macaque brain scRNA-seq datasets. \u003cstrong\u003eC \u003c/strong\u003eQuantification of cell-type proportions in primate brains, showing significant microglial expansion and oligodendrocyte depletion in the MDD group compared to controls. \u003cstrong\u003eD\u003c/strong\u003e Cell-type specific enrichment scores for robust molecular signatures, demonstrating the mapping of up-regulated genes to microglia and down-regulated genes to oligodendrocytes. \u003cstrong\u003eE \u003c/strong\u003eIntercellular communication networks inferred by CellChat. Comparison of interaction number and strength reveals altered glial-neuronal signaling in the depressive state. \u003cstrong\u003eF\u003c/strong\u003e Estimated cell-type proportions in human peripheral blood datasets, recapitulating the central \"microglia-high/oligodendrocyte-low\" phenotype. G Reversal of these pathological cellular proportions in rat brain tissues following antidepressant administration.\u003c/p\u003e","description":"","filename":"Figure2.png","url":"https://assets-eu.researchsquare.com/files/rs-8976053/v1/d61807a4f5e0a1376c437dcf.png"},{"id":107875261,"identity":"20b3897a-2bb3-4a45-9eba-5643fa53630b","added_by":"auto","created_at":"2026-04-27 08:09:23","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":24876857,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMulti-stage screening and cross-species validation identify cell-type-specific and pharmaco-responsive gene signatures.\u003c/strong\u003e \u003cstrong\u003eA\u003c/strong\u003e Heatmap quantifying the overlap between robust bulk DEGs (intersection method) and top 100 cell-type-specific markers derived from primate scRNA-seq data. Color intensity denotes the number of overlapping genes. \u003cstrong\u003eB\u003c/strong\u003e Ranking of microglia- (MIC) and oligodendrocyte- (OLI) specific candidates within the RRA meta-analysis results. Heatmap colors indicate log\u003csub\u003e2\u003c/sub\u003e(FC) values across six independent blood datasets.\u003cstrong\u003e C\u003c/strong\u003e Expression profiles of prioritized MIC- (blue background) and OLI-specific (red background) genes. Top: Validation in human peripheral blood datasets (GSE98793, GSE76826). Bottom: Reversal of expression patterns in rat antidepressant treatment models (GSE194289, GSE222756). Asterisks show statistical significance (p \u0026lt; 0.05, *p \u0026lt; 0.01, **p \u0026lt; 0.001, ***p \u0026lt; 0.0001; ns, not significant).\u003c/p\u003e","description":"","filename":"Figure3.png","url":"https://assets-eu.researchsquare.com/files/rs-8976053/v1/c873627d1da52ea556d9a7c3.png"},{"id":107875426,"identity":"b4cb56c1-8cd1-4bd0-b53d-1cc13f9613b0","added_by":"auto","created_at":"2026-04-27 08:09:55","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":41794829,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eValidation and functional characterization of core glial-specific biomarkers.\u003c/strong\u003e \u003cstrong\u003eA\u003c/strong\u003e Volcano plot highlighting the prioritization of core genes (\u003cem\u003eANO10, HK2, CCDC50\u003c/em\u003e) within the RRA meta-analysis, with ranks and expression statistics detailed. \u003cstrong\u003eB\u003c/strong\u003e Heatmap displaying the expression profile of the three core genes alongside the top 20 up- and down-regulated RRA signatures across six independent blood datasets.\u003cstrong\u003e C\u003c/strong\u003eUMAP plots visualizing the cell-type-specific expression of \u003cem\u003eHK2\u003c/em\u003e(enriched in microglia) and \u003cem\u003eCCDC50\u003c/em\u003e (enriched in oligodendrocytes) in primate brain scRNA-seq datasets. \u003cstrong\u003eD\u003c/strong\u003e Validation of core gene expression in human peripheral blood (top panels, GSE98793, GSE76826) and their pharmacological reversal in rat models (bottom panels, GSE194289, GSE222756). Asterisks show statistical significance (p \u0026lt; 0.05, *p \u0026lt; 0.01, **p \u0026lt; 0.001, ***p \u0026lt; 0.0001; ns, not significant). \u003cstrong\u003eE\u003c/strong\u003e Protein–protein interaction (PPI) network constructed using the STRING database, illustrating the known functional interactome of the core genes.\u003cstrong\u003e F\u003c/strong\u003e Pathway enrichment analysis demonstrating the involvement of the core gene signature in stress response, nervous system processes, and metabolic pathways.\u003c/p\u003e","description":"","filename":"Figure4.png","url":"https://assets-eu.researchsquare.com/files/rs-8976053/v1/6c57b3d076d6e60b623a0991.png"},{"id":107875428,"identity":"94e8e31d-118c-4ae7-bff1-8106bc984ca8","added_by":"auto","created_at":"2026-04-27 08:09:55","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":19476936,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eIndependent cohort recapitulates microglial signatures and establishes \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eHK2\u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003e as a robust microenvironmental biomarkers.\u003c/strong\u003e \u003cstrong\u003eA\u003c/strong\u003e Demographics and clinical characteristics of the validation cohort. \u003cstrong\u003eB\u003c/strong\u003e Cell-type enrichment analysis of transcriptomic data from the validation cohort. Up-regulated DEGs are preferentially enriched in microglia (MIC), while down-regulated DEGs are enriched in neurons. \u003cstrong\u003eC\u003c/strong\u003e Dot plot displaying the top 20 enriched biological pathways for DEGs identified in the validation cohort. \u003cstrong\u003eD\u003c/strong\u003e Comparative expression of the three core genes (\u003cem\u003eANO10, HK2, CCDC50\u003c/em\u003e) between MDD and HC groups as quantified by RNA-sequencing.\u0026nbsp; \u003cstrong\u003eE\u003c/strong\u003e Technical validation of core gene expression via RT-qPCR. Relative mRNA levels were quantified using the 2\u003csup\u003e−ΔΔCt\u003c/sup\u003e method, normalized to GAPDH.\u003c/p\u003e","description":"","filename":"Figure5.png","url":"https://assets-eu.researchsquare.com/files/rs-8976053/v1/673b2bccee6e9f2fe693d2c0.png"},{"id":108184598,"identity":"286f2c6a-593a-47e0-a6fd-4f530e955c27","added_by":"auto","created_at":"2026-04-30 09:04:31","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":89643762,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8976053/v1/90975652-2597-480a-8314-4c468457209d.pdf"},{"id":107875424,"identity":"ea68b598-b2b2-4e27-9837-3e11483a88e7","added_by":"auto","created_at":"2026-04-27 08:09:55","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":11651,"visible":true,"origin":"","legend":"Supplementary Table S1","description":"","filename":"SupplementaryTableS1Theinformationofthedatabase.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8976053/v1/5a776202f716d56d3a56709a.xlsx"},{"id":108180987,"identity":"341c6854-a2c0-4b39-b97e-95f89a66c329","added_by":"auto","created_at":"2026-04-30 08:55:57","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":29600,"visible":true,"origin":"","legend":"Supplementary information","description":"","filename":"finalSupplementaryinformation.docx","url":"https://assets-eu.researchsquare.com/files/rs-8976053/v1/457f5e4eff335647b8d5ac1b.docx"},{"id":107876122,"identity":"d5c69339-7745-4a5e-9a93-aedf341c6367","added_by":"auto","created_at":"2026-04-27 08:13:02","extension":"xlsx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":3097397,"visible":true,"origin":"","legend":"Supplementary Table S6","description":"","filename":"SupplementaryTablesS6GSE99725DEGLists.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8976053/v1/553f4a4cf9ac2ee31018cb8f.xlsx"},{"id":107875431,"identity":"a406a88f-38f1-41de-a8f4-e4be76be3608","added_by":"auto","created_at":"2026-04-27 08:09:56","extension":"xlsx","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":429281,"visible":true,"origin":"","legend":"Supplementary Table S8","description":"","filename":"SupplementaryTablesS8DEGsfoundbyRRAmethods.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8976053/v1/b3cd15bc339a27681ec69315.xlsx"},{"id":107875260,"identity":"bed4318c-0d62-41c0-b8c9-33ea7c3b12af","added_by":"auto","created_at":"2026-04-27 08:09:23","extension":"xlsx","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":3903844,"visible":true,"origin":"","legend":"Supplementary Table S10","description":"","filename":"SupplementaryTablesS10ALLDEGpathwayenrichmentresult.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8976053/v1/d1b2ae0de8b1a580b8657b55.xlsx"},{"id":108006794,"identity":"618c33ab-aaf1-4db9-a31d-189451b3fae1","added_by":"auto","created_at":"2026-04-28 12:57:16","extension":"xlsx","order_by":6,"title":"","display":"","copyAsset":false,"role":"supplement","size":787058,"visible":true,"origin":"","legend":"Supplementary Table S13","description":"","filename":"SupplementaryTablesS13GSE144136CellTypeMarkers.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8976053/v1/1f50b09c5a215efc96977485.xlsx"},{"id":107875255,"identity":"e2316931-1743-4328-ad4f-a52ef9c400fd","added_by":"auto","created_at":"2026-04-27 08:09:23","extension":"xlsx","order_by":7,"title":"","display":"","copyAsset":false,"role":"supplement","size":20060,"visible":true,"origin":"","legend":"Supplementary Table S9","description":"","filename":"SupplementaryTablesS9DEGsfoundbyGeneintersectionmethods.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8976053/v1/0f30a0785fbfc4d1e9bc8187.xlsx"},{"id":107875257,"identity":"2ae1124b-6315-41b0-a4ac-a7b2f35fdf64","added_by":"auto","created_at":"2026-04-27 08:09:23","extension":"xlsx","order_by":8,"title":"","display":"","copyAsset":false,"role":"supplement","size":1623805,"visible":true,"origin":"","legend":"Supplementary Table S15","description":"","filename":"SupplementaryTableS15ValidationDEGs.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8976053/v1/91224ed9fe7c728b2ff0006c.xlsx"},{"id":107875600,"identity":"164b5b0f-97f3-462a-b919-9973cae39d69","added_by":"auto","created_at":"2026-04-27 08:10:12","extension":"xlsx","order_by":9,"title":"","display":"","copyAsset":false,"role":"supplement","size":1083196,"visible":true,"origin":"","legend":"Supplementary Table S12","description":"","filename":"SupplementaryTablesS12RRAmethodpathwayenrichmentresult.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8976053/v1/ff57cb3a943f6722a6f65380.xlsx"},{"id":107875430,"identity":"4f3a4fb9-0828-4526-9369-df5a0db7b9db","added_by":"auto","created_at":"2026-04-27 08:09:56","extension":"tif","order_by":10,"title":"","display":"","copyAsset":false,"role":"supplement","size":15331220,"visible":true,"origin":"","legend":"Supplementary Figure. 4","description":"","filename":"SF4HyperconnectedintercellularsignalinglandscapeintheMDDmicroenvironment.tif","url":"https://assets-eu.researchsquare.com/files/rs-8976053/v1/a4ed7f91a46b86025a6ab3bf.tif"},{"id":107875159,"identity":"55e36f8a-c412-4fb9-b908-963791595bf5","added_by":"auto","created_at":"2026-04-27 08:09:09","extension":"xlsx","order_by":11,"title":"","display":"","copyAsset":false,"role":"supplement","size":48746,"visible":true,"origin":"","legend":"Supplementary Table S11","description":"","filename":"SupplementaryTablesS11Geneintersectionmethodpathwayenrichmentresult.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8976053/v1/a02032e0c871d983cc46106e.xlsx"},{"id":107875434,"identity":"f2c8f1ec-4b61-4691-a06b-d981cf45d28d","added_by":"auto","created_at":"2026-04-27 08:09:56","extension":"xlsx","order_by":12,"title":"","display":"","copyAsset":false,"role":"supplement","size":238519,"visible":true,"origin":"","legend":"Supplementary Table S16","description":"","filename":"SupplementaryTableS16ValidationPathwayenrichmentresult.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8976053/v1/fd0abab4a149be57ea4a271e.xlsx"},{"id":107875433,"identity":"1a00fd43-ef24-4e79-9fb9-127d3f915b52","added_by":"auto","created_at":"2026-04-27 08:09:56","extension":"xlsx","order_by":13,"title":"","display":"","copyAsset":false,"role":"supplement","size":3025567,"visible":true,"origin":"","legend":"Supplementary Table S3","description":"","filename":"SupplementaryTablesS3GSE76826DEGLists.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8976053/v1/6444fbbdc5f6b65d7b2d631d.xlsx"},{"id":107875427,"identity":"4785ec55-c186-445a-b949-3a7638b8b403","added_by":"auto","created_at":"2026-04-27 08:09:55","extension":"xlsx","order_by":14,"title":"","display":"","copyAsset":false,"role":"supplement","size":2763883,"visible":true,"origin":"","legend":"Supplementary Table S5","description":"","filename":"SupplementaryTablesS5GSE247998DEGLists.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8976053/v1/1cecbd6c0ba402a6604b9118.xlsx"},{"id":107875514,"identity":"947fdcc7-d3f7-431d-9f80-169c04d91f01","added_by":"auto","created_at":"2026-04-27 08:10:04","extension":"tif","order_by":15,"title":"","display":"","copyAsset":false,"role":"supplement","size":22853880,"visible":true,"origin":"","legend":"Supplementary Figure. 7","description":"","filename":"SF7Pharmacologicalattenuationofglialmarkerdysregulationbyfluoxetineinratmodel.tif","url":"https://assets-eu.researchsquare.com/files/rs-8976053/v1/e9546e6d6d6491297133a5a4.tif"},{"id":107875162,"identity":"092e28da-ca89-419d-802f-8ac280d768e5","added_by":"auto","created_at":"2026-04-27 08:09:10","extension":"xlsx","order_by":16,"title":"","display":"","copyAsset":false,"role":"supplement","size":1017707,"visible":true,"origin":"","legend":"Supplementary Table S14","description":"","filename":"SupplementaryTablesS14GSE201687CellTypeMarkers.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8976053/v1/6c81a9e0b12d004641e142bd.xlsx"},{"id":107875539,"identity":"189a1360-181e-4efa-9faa-f87d4ed6c0b2","added_by":"auto","created_at":"2026-04-27 08:10:05","extension":"tif","order_by":17,"title":"","display":"","copyAsset":false,"role":"supplement","size":32602092,"visible":true,"origin":"","legend":"Supplementary Figure. 5","description":"","filename":"SF5RecapitulationoftheglialspecifictranscriptomicsignatureintheGSE98793cohort.tif","url":"https://assets-eu.researchsquare.com/files/rs-8976053/v1/bd869a2ff8430ff94f10c7e0.tif"},{"id":107875163,"identity":"4521ca92-2c55-4faf-8e6a-c1fb41743744","added_by":"auto","created_at":"2026-04-27 08:09:10","extension":"tif","order_by":18,"title":"","display":"","copyAsset":false,"role":"supplement","size":33328936,"visible":true,"origin":"","legend":"Supplementary Figure. 6","description":"","filename":"SF6IndependentvalidationofcelltypespecificmarkerdysregulationintheGSE76826cohort.tif","url":"https://assets-eu.researchsquare.com/files/rs-8976053/v1/1d08688f01079b57714a8bf4.tif"},{"id":107875540,"identity":"4fce8503-327a-43a5-befc-1241c8e22412","added_by":"auto","created_at":"2026-04-27 08:10:05","extension":"tif","order_by":19,"title":"","display":"","copyAsset":false,"role":"supplement","size":32564944,"visible":true,"origin":"","legend":"Supplementary Figure. 2","description":"","filename":"SF2Transcriptomicheterogeneityanddifferentialexpressionlandscapeacrosssixindependentperipheralbloodcohorts.tif","url":"https://assets-eu.researchsquare.com/files/rs-8976053/v1/814b0c8153c8948284363104.tif"},{"id":107875432,"identity":"5d7d2129-edf1-4be8-89f7-4997cd50155e","added_by":"auto","created_at":"2026-04-27 08:09:56","extension":"tif","order_by":20,"title":"","display":"","copyAsset":false,"role":"supplement","size":35132464,"visible":true,"origin":"","legend":"Supplementary Figure. 3","description":"","filename":"SF3Deconvolutionvalidatescentralglialshiftsinperipheralbloodandtheirpharmacologicalreversal.tif","url":"https://assets-eu.researchsquare.com/files/rs-8976053/v1/ae8d75f30a8429b617733a7a.tif"},{"id":107875164,"identity":"652382e3-2795-43e1-85ed-8087680d1490","added_by":"auto","created_at":"2026-04-27 08:09:10","extension":"xlsx","order_by":21,"title":"","display":"","copyAsset":false,"role":"supplement","size":3050341,"visible":true,"origin":"","legend":"Supplementary Table S4","description":"","filename":"SupplementaryTablesS4GSE290797DEGLists.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8976053/v1/0bcb6ce48b39a5cae40ce804.xlsx"},{"id":107875262,"identity":"eb115b08-5958-4d17-ba5e-4d4201db4f14","added_by":"auto","created_at":"2026-04-27 08:09:23","extension":"tif","order_by":22,"title":"","display":"","copyAsset":false,"role":"supplement","size":20054792,"visible":true,"origin":"","legend":"Supplementary Figure. 1","description":"","filename":"SF1Visualizingtranscriptomicheterogeneityacrosscohorts.tif","url":"https://assets-eu.researchsquare.com/files/rs-8976053/v1/3fce1c98303ef619a106b257.tif"},{"id":107875254,"identity":"94b44c9f-4c30-4204-abdb-cc27a24e801a","added_by":"auto","created_at":"2026-04-27 08:09:23","extension":"tif","order_by":23,"title":"","display":"","copyAsset":false,"role":"supplement","size":60303512,"visible":true,"origin":"","legend":"Supplementary Figure. 8","description":"","filename":"SF8BroadpharmacologicalreversaloftheMDDassociatedglialsignatureacrossmultipleantidepressanttreatment.tif","url":"https://assets-eu.researchsquare.com/files/rs-8976053/v1/a7a1c87568be844860111a5e.tif"},{"id":107875436,"identity":"596f0234-e9d7-4564-9fd8-0bae4cc77d2b","added_by":"auto","created_at":"2026-04-27 08:09:56","extension":"xlsx","order_by":24,"title":"","display":"","copyAsset":false,"role":"supplement","size":2839679,"visible":true,"origin":"","legend":"Supplementary Table S7","description":"","filename":"SupplementaryTablesS7GSE185855DEGLists.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8976053/v1/d1e5e49e2e7b307fce0d758a.xlsx"},{"id":107875425,"identity":"436e6c73-a6f2-4c88-87cd-6877d3d6eff1","added_by":"auto","created_at":"2026-04-27 08:09:55","extension":"xlsx","order_by":25,"title":"","display":"","copyAsset":false,"role":"supplement","size":2070999,"visible":true,"origin":"","legend":"Supplementary Table S2","description":"","filename":"SupplementaryTablesS2GSE98793DEGLists.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8976053/v1/8c42e5703a21bf02001920e6.xlsx"}],"financialInterests":"The authors have declared there is \u003cb\u003eNO\u003c/b\u003e conflict of interest to disclose","formattedTitle":"Integrated multi-omics characterization identifies key microenvironmental biomarkers in major depressive disorder microenvironment","fulltext":[{"header":"Introduction","content":"\u003cp\u003eMajor Depressive Disorder (MDD) represents a pervasive and debilitating global health challenge, imposing a profound social and economic burden [1]. According to the Global Burden of Disease Study 2019, mental disorders remain among the top ten leading causes of disease burden worldwide, with MDD being a primary contributor to disability-adjusted life years [2].\u0026nbsp;As the prevalence of MDD continues to rise, the pathogenic mechanisms driven by the complex interplay between microenvironmental, environmental, and neurobiological factors remain poorly understood. At present, MDD diagnosis relies largely on subjective symptom assessment guided by DSM-5 (Diagnostic and Statistical Manual of Mental Disorders) or ICD-11 (International Classification of Diseases 11th Revision), most commonly using structured instruments such as the HAMD (Hamilton Depression Rating Scale) [1, 3]. However, this approach is inherently limited by inter-rater variability and the absence of objective biological criteria [4].\u0026nbsp;Moreover, an number of patients exhibit treatment resistance to current pharmacological therapies, often accompanied by poor tolerability and insufficient symptom improvement [5]. Moreover, the complex and heterogeneous nature of MDD, often together with its overlap with other psychiatric disorders and the difficulty of obtaining living brain tissue, has limited the development of reliable diagnostic biomarkers [6]. Therefore, there is a critical need to identify accessible microenvironment-associated phenotype biomarkers that can reliably reflect central pathological states and improve the diagnosis of MDD.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAdvances in multi-omics sequencing have facilitated cross-scale investigation of disease microenvironments and phenotypes, yet the neural microenvironment has been relatively under-integrated in biological studies of MDD [7]. New evidence suggests that the pathophysiology of depression is not confined to aberrant neurotransmission but involves a systemic dysregulation of the neurovascular unit, particularly the immune-glial interface [8, 9]. Postmortem studies link neuroinflammation, defective synaptic pruning, and impaired neurogenesis to the neuronal microenvironment [10]. In fact, MDD-associated genetic variants are enriched in regulatory regions, highlighting the potential role of transcriptional regulation and the neural microenvironment in disease pathophysiology [11]. These central observations are often constrained by postmortem intervals, medication confounds, and cause of death [12]. Fortunately, peripheral blood and single-cell sequencing provide a minimally invasive means to capture systemic physiological states [13, 14]. Although several studies have profiled peripheral transcriptomic changes, findings have been inconsistent, likely due to sample size limitations and the dilution of disease-associated signals in bulk tissues [15]. This raises a critical question: to what extent can molecular signals detected in the periphery serve as reliable proxies for central nervous system microenvironment alterations, particularly the inflammatory states of non-neuronal cells [16, 17].\u003c/p\u003e\n\u003cp\u003eRecent single-cell transcriptomic resolutions have underscored that non-neuronal cells-specifically microglia and oligodendrocytes-play pivotal roles in mediating neuroinflammation and white matter integrity compromise in MDD [18, 19]. Further, various neurobiological factors with diverse cell types have been linked to MDD, including monoaminergic or glutamatergic systems and abnormalities in astrocytes, oligodendrocytes or immune cells [20]. In spite of significant progress, the precise molecular and cellular mechanisms mediating the risk for MDD remain unknown. Traditional bulk RNA sequencing studies, based on clinically defined samples, provide rich clinical information but remain unable to resolve molecular changes at the level of individual cell types [21]. However, single-cell sequencing approaches often fail to capture the full spectrum of the disease due to the lack of systemic context and limited sample availability [22, 23].\u0026nbsp;A systematic, integrative multi-omics approach is needed to link molecular alterations across modalities and form closed-loop analyses and validations. Such strategies are essential to identify robust diagnostic biomarkers with biological relevance in the context of MDD [24, 25].\u003c/p\u003e\n\u003cp\u003eIn this study, we established a comprehensive analytical framework to identify robust transcriptomic features of MDD present in both the blood and the brain, and to dissect their cell-type specificity. Firstly, we collected ten bulk RNA sequencing datasets with MDD and employed a dual strategy combining a gene intersection method and Robust Rank Aggregation (RRA) meta-analysis to rigorously screen for high-confidence differentially expressed genes (DEGs) across heterogeneous cohorts. Then, we integrated bulk-level findings with multiple single-cell RNA sequencing data using Bayesian deconvolution method and cell-cell communication analysis, enabling a cell-type-resolved investigation of the MDD microenvironment. Through multi-level integration, we identified three core genes-\u003cem\u003eANO10, HK2,\u0026nbsp;\u003c/em\u003eand \u003cem\u003eCCDC50\u003c/em\u003e-as candidate microenvironmental biomarkers. Among them, \u003cem\u003eHK2\u003c/em\u003e (Hexokinase 2) showed the most consistent and pronounced dysregulation, specifically localized to microglia, indicating a potential immunometabolic reprogramming towards glycolysis. We further validated the robustness of these biomarkers in an independent clinical cohort using transcriptomic analysis and quantitative real-time PCR (qPCR). Collectively, this study systematically integrates existing multimodal sequencing data to provide a comprehensive view of the neuro-immune microenvironment in MDD and establishes \u003cem\u003eHK2\u003c/em\u003e as a robust candidate target for future diagnostic and therapeutic interventions.\u003c/p\u003e"},{"header":"Materials and Methods ","content":"\u003ch3\u003e\u003cstrong\u003eData collection and preprocessing\u003c/strong\u003e\u003c/h3\u003e\n\u003cp\u003eTo establish a comprehensive multi-omics framework, we collected ten transcriptomic datasets from the Gene Expression Omnibus (GEO), including six human peripheral blood bulk profiles (GSE98793 [26], GSE76826 [27], GSE290797 [28], GSE247998 [29], GSE99725 [30] and GSE185855 [31]), two cross-species single-cell RNA sequencing (scRNA-seq) datasets (macaque: GSE201687 [32]; human: GSE144136 [33]), and two rat bulk transcriptomes assessing antidepressant response (GSE222756 [34], GSE194289 [35]). Data processing and statistical analyses were executed within the R statistical environment (v4.4.2). To ensure data comparability, raw count datasets (GSE247998 and GSE185855) were subjected to log transformation followed by quantile normalization utilizing the limma package (v3.62.2). Datasets available as pre-processed normalized values were utilized without further modification to preserve original quality control standards. For single-cell data, processing pipelines were rigorously aligned with the protocols described in the primary study (GSE201687) \u0026nbsp;[32].\u003c/p\u003e\n\u003ch3\u003e\u003cstrong\u003eDifferential gene expression analysis\u003c/strong\u003e\u003c/h3\u003e\n\u003cp\u003eTo characterize transcriptomic alterations associated with MDD, differential expression analysis was performed across six independent peripheral blood RNA sequencing datasets using the limma package. Differentially expressed genes (DEGs) were identified based on a stringent threshold of Benjamini-Hochberg adjusted \u003cem\u003ep\u003c/em\u003e-value \u0026lt; 0.05 and a fold-change magnitude of |log\u003csub\u003e2\u003c/sub\u003eFC| \u0026gt; log\u003csub\u003e2\u003c/sub\u003e(1.1). To systematically assess the consistency of DEGs across cohorts, gene overlaps were visualized using UpSetR (v1.4.0) and ggplot2 (v3.5.2).\u003c/p\u003e\n\u003ch3\u003e\u003cstrong\u003eIdentifying robust DEGs\u003c/strong\u003e\u003c/h3\u003e\n\u003cp\u003eTo address inter-study heterogeneity and derive a robust consensus transcriptional signature, we employed a dual-screening strategy that integrates gene intersection analysis and Robust Rank Aggregation (RRA) meta-analysis. First, an intersection analysis was conducted to identify DEGs consistently shared across datasets, with emphasis on genes exhibiting concordant directions of dysregulation. Second, high confidence genes were further screened using the RobustRankAggreg package (v1.2.1). This probabilistic framework integrates ranked gene lists-ordered by \u003cem\u003ep\u003c/em\u003e-values-from individual datasets to evaluate the likelihood of each gene achieving consistently high ranks across heterogeneous cohorts [36]. Robust DEGs were defined as those meeting both an RRA score \u0026lt; 0.05 and a cumulative mean |log\u003csub\u003e2\u003c/sub\u003eFC| \u0026gt; log\u003csub\u003e2\u003c/sub\u003e(1.1) across all including datasets.\u003c/p\u003e\n\u003ch3\u003e\u003cstrong\u003ePathway enrichment analysis\u003c/strong\u003e\u003c/h3\u003e\n\u003cp\u003eTo interpret the biological relevance of the identified genes, functional enrichment analysis was performed using the g:Profiler platform [37]. Enrichment testing was conducted against the \"All known genes\" statistical domain with a stringent user threshold of 0.01. Multiple functional annotation sources were queried, including Gene Ontology categories (GO: Biological Process, Cellular Component, and Molecular Function), as well as KEGG and Reactome pathway databases. Enrichment results were visualized using the tidyverse suite (v2.0.0).\u003c/p\u003e\n\u003ch3\u003e\u003cstrong\u003eSingle\u003c/strong\u003e\u003cstrong\u003e-cell data integration and gene module scoring\u003c/strong\u003e\u003c/h3\u003e\n\u003cp\u003eTo assess the cellular distribution of the robust DEGs signatures at single-cell resolution, gene module scores were computed using the AddModuleScore function within the Seurat package (v5.3.0). Consensus up-regulated and down-regulated gene sets were defined as separate modules, enabling quantification of their aggregated expression levels across individual cells in both scRNA-seq datasets. Module score distributions was visualized using ggplot2.\u003c/p\u003e\n\u003ch4\u003e\u003cstrong\u003eIntegration with single-cell RNA dataset\u003c/strong\u003e\u003c/h4\u003e\n\u003cp\u003eTo resolve the cellular heterogeneity in bulk transcriptomeic profiles, cell-type deconvolution was performed using the Bayesian inference framework BayesPrism (v2.2.2) [38]. A high-resolution human brain scRNA-seq dataset (GSE144136) served as the reference signature matrix to infer cellular proportions across all bulk transcriptomic cohorts. All analyses were conducted using default parameters.\u003c/p\u003e\n\u003ch4\u003e\u003cstrong\u003eCell-cell communication analysis\u003c/strong\u003e\u003c/h4\u003e\n\u003cp\u003eIntercellular signaling networks were inferred using CellChat (v1.6.1) [39] to characterize signaling cell interactions within the neural microenvironment. Based on the CellChatDB.human ligand-receptor interaction database, the probability and strength of cell-cell communication events among distinct cell populations were systematically quantified in the single-cell datasets.\u003c/p\u003e\n\u003ch3\u003e\u003cstrong\u003eScreen cell-type-specific DEGs for identification of core genes\u003c/strong\u003e\u003c/h3\u003e\n\u003cp\u003eTo delineate the cellular origin of the identified transcriptomic signatures, cell-type specific markers genes were first identified from the scRNA-seq datasets using the FindAllMarkers function in Seurat (parameters: only.pos = TRUE, min.pct = 0.1, logfc.threshold = 0.25). For each major cell population, the top 500 marker genes ranked by average log fold change were retained. These marker genes were then intersected with the robust up-regulated and down-regulated DEGs derived from the bulk transcriptomic meta-analysis. Overlap statistics were summarized and visualized using ggplot2.\u0026nbsp;\u003c/p\u003e\n\u003ch4\u003e\u003cstrong\u003eScreening of cell type specific and drug responsive\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003ecandidates\u003c/strong\u003e\u003c/h4\u003e\n\u003cp\u003eTo assess the therapeutic relevance of the identified candidates, cross-species ortholog mapping from human to rat was performed using the gprofiler2 package (v0.2.3). The resulting homologous genes were subsequently projected onto two independent rat antidepressant treatment datasets. Gene expression distributions across treatment conditions visualized using violin and box plots. Core DEGs were stringently defined according to a therapeutic reversal criterion, whereby candidate genes exhibited aberrant expression in the depression-like state and demonstrated significant normalization or directional reveal following antidepressant administration.\u003c/p\u003e\n\u003ch3\u003e\u003cstrong\u003eProtein-Protein interaction (PPI) network analysis\u003c/strong\u003e\u003c/h3\u003e\n\u003cp\u003eTo explore the functional connectivity among the identified core genes, protein–protein interaction (PPI) network was constructed using the STRING database (https://cn.string-db.org) [40]. The analysis was conducted using default parameters to visualize the potential interactions.\u003c/p\u003e\n\u003ch3\u003e\u003cstrong\u003eBlood samples for the validation cohort\u003c/strong\u003e\u003c/h3\u003e\n\u003cp\u003eTo validate the in silico findings, an independent clinical cohort comprising six patients with MDD (aged 18-45 years) and five healthy controls (HC) was recruited from the Department of Clinical Psychology at The Third People’s Hospital of Foshan. The diagnosis of major depressive disorder (MDD) was made by two physicians, and its severity was assessed using the 24-item Hamilton Depression Rating Scale (HAMD-24).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eStrict exclusion criteria were applied to minimize confounding factors. Participants were excluded if they presented with: (1) comorbid psychiatric disorders (e.g., schizophrenia, bipolar disorder, or anxiety disorders); (2) a family history of mental illness; (3) acute or chronic somatic pathologies affecting the immune system (e.g., autoimmune diseases or cancer); (4) current pregnancy; or (5) use of psychotropic medications (antidepressants), immunomodulators, or hormonal drugs within the two weeks preceding sampling.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003ePeripheral venous blood samples were collected into PAXgene Blood RNA Tubes to ensure transcript stability. Specimens were immediately processed and stored at −80°C until RNA extraction. The study protocol was approved by The Medical Ethics Committee of The Third People's Hospital of Foshan (Foshan Mental Health Center) and was conducted in strict accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants prior to inclusion.\u0026nbsp;\u003c/p\u003e\n\u003ch3\u003e\u003cstrong\u003eRNA extraction and transcriptome data construction\u003c/strong\u003e\u003c/h3\u003e\n\u003cp\u003eTotal RNA was isolated using Trizol Reagent (Invitrogen Life Technologies). RNA purity and concentration were assessed using a NanoDrop spectrophotometer (Thermo Scientific). Sequencing libraries were constructed using the NEBNext Ultra II RNA Library Prep Kit for Illumina (NEB), strictly adhering to the manufacturer’s instructions. Library quality and fragment size distribution were verified using the Agilent High Sensitivity DNA Assay on a Bioanalyzer 2100 system (Agilent). Subsequently, libraries were sequenced on the Illumina NovaSeq 6000 platform to generate paired-end reads.\u003c/p\u003e\n\u003cp\u003eRaw sequencing data (FASTQ format) were subjected to quality control using Cutadapt (v1.15) to remove adaptor sequences and low-quality reads. The resulting clean reads were aligned to the reference genome using the splice-aware aligner HISAT2 (v2.0.5). Finally, gene expression levels were quantified by generating raw read counts using HTSeq (v0.9.1).\u003c/p\u003e\n\u003ch3\u003e\u003cstrong\u003eReal-time quantitative PCR (RT-qPCR) validation\u003c/strong\u003e\u003c/h3\u003e\n\u003cp\u003eTo experimentally validate the identified core transcriptomic signatures, RT-qPCR was performed on whole blood samples derived from the independent validation cohort (n = 11). Total RNA was extracted using Trizol reagent, followed by first-strand cDNA synthesis utilizing the Advantage® RT-for-PCR Kit. Amplification reactions were conducted on a LightCycler 480 II instrument using AceQ® qPCR SYBR® Green Master Mix. Specific primer sequences are detailed in Table 1. The thermal cycling protocol encompassed an initial denaturation at 95℃\u0026nbsp;for 5 min, followed by 40 cycles of 95℃\u0026nbsp;for 15 s and 60℃\u0026nbsp;for 30 s. Melting curve analysis was subsequently performed to verify amplicon specificity. Relative mRNA expression levels were quantified using the comparative C\u003csub\u003eT\u003c/sub\u003e(2\u003csup\u003e−ΔΔCt\u003c/sup\u003e) method, normalized to the endogenous reference gene \u003cem\u003eGAPDH\u003c/em\u003e.\u003c/p\u003e\n\u003cp\u003eTable\u0026nbsp;1\u0026nbsp;PCR primers used in this study\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePrimer Name\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSequence (5' to 3')\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003eANO10\u003c/em\u003e ENSG00000160746-F\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eTCAGAACCTTCAGCCAAT\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003eANO10\u003c/em\u003e ENSG00000160746-R\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eACAGTTAGTGACCACAGATAT\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003eHK2\u003c/em\u003e ENSG00000159399-F\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAAGACATTAGAGCATCTG\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003eHK2\u003c/em\u003e ENSG00000159399-R\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCTCCATTTCTACCTTCAT\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003eCCDC50\u003c/em\u003e ENSG00000152492-F\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAAGGAGTTACAGGAAGAG\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003eCCDC50\u003c/em\u003e ENSG00000152492-R\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eTCGTGATGGAGTAGATAC\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003egapdh\u003c/em\u003e ENSG00000111640(nei)-F\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCTCTGGTAAAGTGGATATTGT\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003egapdh\u003c/em\u003e ENSG00000111640(nei)-R\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eGGTGGAATCATATTGGAACA\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eRaw RNA-sequencing data (count files) and code generated in this study have been deposited on GitHub and are publicly available at https://github.com/Reisen333/MDD-Datasets. Publicly available datasets analyzed in this study are available in the Gene Expression Omnibus (GEO) repository under the following accession codes: GSE98793, GSE76826, GSE290797, GSE247998, GSE99725, GSE185855 (human blood bulk transcriptomes); GSE201687 (macaque brain scRNA-seq); GSE144136 (human brain scRNA-seq); and GSE222756, GSE194289 (Rat pharmacological models). The human reference genome (GRCh38.p13) and gene annotation files used for alignment were downloaded from the Ensembl database (http://asia.ensembl.org/Homo_sapiens/Info/Index). Source data for the figures are provided with this paper. All other data supporting the findings of this study are available from the corresponding authors upon reasonable request.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003eRobust transcriptomic genes of MDD\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eInitially, we conducted a comprehensive differential expression analysis across six independent human peripheral blood transcriptomic cohorts. To ensure the inclusion of contemporary data, all selected datasets were published after 2016. These final datasets comprised 326 MDD patients and 170 control samples (Fig. 1A, Supplementary Table S1). We identified a reproducible transcriptomic signature (\u003cem\u003ep\u003c/em\u003e \u0026lt; 0.05 and |log\u003csub\u003e2\u003c/sub\u003eFC| \u0026gt; log\u003csub\u003e2\u003c/sub\u003e(1.1)). However, substantial variation in the number of identified differentially expressed genes (DEGs) was observed across six datasets (Fig. 1B-C, Supplementary Table S2-S7), with the limited overlap between individual cohorts. This pronounced inter-study inconsistency underscores the profound biological heterogeneity inherent in MDD (Supplementary Fig. 2-3). Such single-cohort analysis is inherently limited and may be insufficient to fully capture the underlying pathogenic mechanisms of the disorder. To circumvent potential batch effects inherent in merging raw transcriptomic data, we employ an integrative framework combining gene intersection analysis with Robust Rank Aggregation (RRA). The RRA approach designed to detect consistently high-ranking genes across, heterogeneous lists—identified a robust signature of 3161 dysregulated genes (Fig. 1D; Supplementary Table S8). Notably, this RRA-derived signature effectively captured the signals from the pairwise intersection while expanding the repertoire to include latent but reproducible markers. This integrative gain underscores the pronounced molecular heterogeneity of MDD, where true disease-associated signals may manifest as subtle but consistent patterns rather than large cohort-specific effects. Consequently, conventional single-cohort bulk transcriptomic analyses are often insufficient to identify robust and reproducible biomarkers in such complex neuropsychiatric disorders.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSingle-cell landscape and cross-species validation of the robust MDD signature\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo systematically evaluate the biological relevance of potential biomarkers and prepare for functional characterization, we defined three distinct gene sets, including the \"Overlap method\" set, comprising the 99 DEGs shared between the two datasets (GSE98793 and GSE76826) (Supplementary Table S9), the \"RRA Method\" set, comprising the 3161 robust genes identified by the meta-analysis; and the \"ALL DEGs\" set, comprising all 9577 DEGs identified across all cohorts. We first performed functional enrichment analysis, which revealed a convergent dysregulation of pathways governing immune/inflammatory responses, organelle integrity, and cellular stress adaptation in both intersection- and RRA-derived signatures (Fig. 2A, Supplementary Table S10-S12). These findings suggest that peripheral transcriptomic alterations in MDD are underpinned by a systemic immunometabolic disruption. To investigate whether these systemic signals mirror specific central cellular pathologies, we integrated these genes with scRNA-seq data from both human and macaque brains (Fig. 2B). Cell composition analysis demonstrated a significant expansion of the microglial compartment alongside a concurrent reduction of the oligodendrocyte cells in the depressed brain (Fig. 2C). Further, gene module scoring revealed a striking cellular dichotomy: the robust up-regulated genes were predominantly enriched in microglia, whereas the down-regulated genes mapped extensively to oligodendrocytes (Fig. 2D). Collectively, these results imply that the core MDD signature reflects a fundamental imbalance between microglial activation and oligodendroglial support within the neuro-microenvironment.\u003c/p\u003e\n\u003cp\u003eTo bridge the gap between peripheral signals and central pathology, we applied Bayesian deconvolution to the six peripheral blood transcriptomic datasets and two rat pharmacological cohorts. Remarkably, blood transcriptomes of MDD patients faithfully recapitulated the central neuro-microenvironmental landscape, exhibiting an estimated increase in microglial cells and decrease in oligodendrocyte cells (Fig. 2E, Supplementary Fig. 3A-B). Importantly, antidepressant treatment in rat models reversed these pathological cellular trends, leading to a significant reduction in estimated microglial and oligodendrocyte proportions relative to the depression model (Fig. 2F, Supplementary Fig. 3C). Drug-specific effects were also observed. For example, fluoxetine and desipramine treatment were associated with an expansion of astrocyte and interneuron populations and a relative reduction in excitatory neurons (Fig. 2F), suggesting a partial restoration of cellular homeostasis within the brain microenvironment. To further detect the functional connectivity within this altered microenvironment, cell-cell communication analysis was performed. The depressed brain exhibited a hyperconnected signaling network, characterized by increased ligand–receptor interactions across cell types in both human and macaque single-cell datasets. Notably, microglia emerged as a central signaling hub in this intensified communication network (Fig. 2G, Supplementary Fig. 4), indicating a dominant role in orchestrating aberrant intercellular crosstalk. Collectively, these findings indicate that peripheral molecular signatures faithfully reflect specific central microenvironment dysfunctions in MDD, including microglial hyperactivation, compromised oligodendrocyte integrity, and disrupted glial–neuronal communication.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSingle-cell and pharmacological validation of key microenvironmental markers\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBuilding on the delineation of MDD-associated cellular alterations, we further explored the specific molecular drivers underlying glial dysfunctions. To prioritize candidate genes possessing both cellular specificity and therapeutic relevance, an intersectional analysis was performed between the robust DEGs identified from bulk meta-analysis and the top 500 marker genes defined for each cell type in the scRNA-seq datasets. This analysis revealed that up-regulated consensus genes were significantly enriched for microglial markers, whereas down-regulated genes showed substantial overlap with oligodendrocyte markers (Fig. 3A, Supplementary Table S13-S14), reinforcing the cell-type specificity of the transcriptomic signatures. Importantly, these cell-type–resolved candidates were also highly ranked in the RRA meta-analysis, supporting their robustness across heterogeneous cohorts. Among the top-prioritized genes were \u003cem\u003eANO10, HK2, TFEB,\u003c/em\u003e and \u003cem\u003eZNF532\u003c/em\u003e (Fig. 3B), highlighting their potential roles as key regulators of microenvironmental dysregulation in MDD.\u003c/p\u003e\n\u003cp\u003eWe then sought to validate the translational reliability of these candidates by examining their expression dynamics across independent human blood cohorts and rat pharmacological models. The results showed that the microglia-associated genes were robustly up-regulated, whereas oligodendrocyte-associated genes were down-regulated in MDD blood samples (Fig. 3C, Supplementary Fig. 5-6). Importantly, we found that antidepressant treatment significantly reversed these pathological expression patterns in rat treatment datasets. For instance, administration of fluoxetine significantly attenuated the aberrant up-regulation of \u003cem\u003eANO10\u003c/em\u003e and \u003cem\u003eHK2\u003c/em\u003e, while concurrently restoring the suppressed levels of \u003cem\u003eCCDC50\u003c/em\u003e (Fig. 3C,\u0026nbsp;Supplementary Fig. 7-8). Collectively, this multi-stage screening strategy identified a core set of microglia-specific genes and oligodendrocyte-specific genes that demonstrate dynamic responsiveness to therapeutic interventions, emphasizing their potential as promising targets for MDD treatment.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eValidation and functional characterization the core biomarkers\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBased on the integrated analysis, we prioritized \u003cem\u003eANO10\u003c/em\u003e, \u003cem\u003eHK2\u003c/em\u003e, and \u003cem\u003eCCDC50\u003c/em\u003e as core candidate biomarkers. These genes were consistently identified by both intersectional and RRA strategies and ranked among the top-tier positions in the RRA significance rankings (Fig. 4A). Heatmap visualization demonstrated a highly consistent expression profile for these genes across all six independent human blood cohorts (Fig. 4B), underscoring their stability as systemic markers. To link these peripheral signals to the central neuro-microenvironment, we found \u003cem\u003eANO10\u003c/em\u003e and \u003cem\u003eHK2\u003c/em\u003e were predominantly enriched in the microglial compartment, whereas \u003cem\u003eCCDC50\u003c/em\u003e mapped specifically to the oligodendrocyte lineage (Fig. 4C). These findings were corroborated in bulk datasets: the aberrant upregulation of microglial markers (\u003cem\u003eANO10, HK2\u003c/em\u003e) and downregulation of the oligodendrocyte marker (\u003cem\u003eCCDC50\u003c/em\u003e) observed in MDD were effectively reversed by antidepressant administration in rat models (Fig. 4D), confirming their role as pharmacodynamically responsive targets.\u003c/p\u003e\n\u003cp\u003eTo clarify the molecular mechanisms underlying these core genes, we constructed a Protein-Protein Interaction (PPI) network. The results showed that \u003cem\u003eANO10\u003c/em\u003e was linked to ion channel regulation, \u003cem\u003eCCDC50\u003c/em\u003e modulated EGFR and NOTCH pathways (Fig. 4E), and \u003cem\u003eHK2\u003c/em\u003e emerged as a central hub in glycolysis and glucose metabolism. Functional enrichment analysis further positioned these genes-particularly \u003cem\u003eHK2\u003c/em\u003e, at the intersection of response to stress, metabolic processes, and nervous system development (Fig. 4F). Together, these findings confirm \u003cem\u003eANO10, HK2,\u003c/em\u003e and \u003cem\u003eCCDC50\u003c/em\u003e as high-confidence peripheral biomarkers that reflect specific central glial dysfunctions: \u003cem\u003eHK2\u003c/em\u003e may drive microglial immunometabolic reprogramming, while \u003cem\u003eCCDC50\u003c/em\u003e dysregulation points to compromised oligodendrocyte signaling.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eIndependent cohort recapitulates microglial signatures and establishes HK2 as a robust cross-platform\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo corroborate our in-silico findings, we performed transcriptomic profiling on an independent validation cohort comprising six MDD patients and five healthy controls (Fig. 5A). Cell-type enrichment analysis corroborated the central role of microglia, as the upregulated gene module was again significantly enriched in the microglial signature (Fig. 5B, Supplementary Table S15), substantiating a pro-inflammatory glial phenotype. Consistent with findings from the discovery cohorts, differential expression analysis revealed a transcriptional landscape strongly enriched for immune-related pathways, including Toll-like receptor (TLR), Th1/Th2 differentiation, and JAK–STAT signaling cascades (Fig. 5C, Supplementary Table S16).\u003c/p\u003e\n\u003cp\u003eWe then examined the expression of the core candidate genes-\u003cem\u003eANO10, HK2,\u003c/em\u003e and \u003cem\u003eCCDC50\u003c/em\u003e. RNA-sequencing data from the validation cohort confirmed significant up-regulation of \u003cem\u003eANO10\u0026nbsp;\u003c/em\u003eand \u003cem\u003eHK2\u003c/em\u003e, alongside down-regulation of \u003cem\u003eCCDC50\u003c/em\u003e, fully recapitulating the directionality observed in the discovery analyses (Fig. 5D). To further assess the technical robustness of these candidates, RT-qPCR validation was performed. Notably, \u003cem\u003eHK2\u003c/em\u003e exhibited consistent up-regulation, whereas \u003cem\u003eANO10\u003c/em\u003e and\u003cem\u003e\u0026nbsp;CCDC50\u003c/em\u003e failed to demonstrate reproducible dysregulation in the qPCR assays (Fig. 5E). This discrepancy may reflect differences in baseline expression levels, platform sensitivity, or increased inter-individual variability for these targets. Taken together, these results demonstrate that while a broader microglial-associated transcriptional signature is reproducibly observed in MDD, cross-platform validation effectively narrows the candidate pool, establishing \u003cem\u003eHK2\u003c/em\u003e as the most robust and reproducible peripheral biomarker reflecting central microenvironmental dysregulation in MDD.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eMDD is a neuropsychiatric disorder for which clinical diagnosis still relies predominantly on symptom-based interviews and psychometric questionnaires, limiting objective quantification of disease state and treatment response. Moreover, the substantial inter-individual heterogeneity inherent to MDD has constrained the utility of single-modality transcriptomic studies, where disease signals are often obscured by biological variability, yielding poor reproducibility of candidate biomarkers [9]. In addition, current antidepressant therapies frequently produce incomplete remission and intolerable side effects, underscoring the need for biologically grounded stratification and monitoring tools. In this study, a central finding is that peripheral blood transcriptomes, when interpreted through an integrative framework, can faithfully identify key features of the altered neuro-microenvironment in MDD, most notably microglial activation and oligodendrocyte dysfunction. The concordance observed across peripheral bulk transcriptomics, brain single-cell profiles, and pharmacological response datasets challenges the long-standing assumption that peripheral molecular signals are inherently too nonspecific to inform central pathology. Instead, these results support a model in which depression-related immune\u0026ndash;metabolic dysregulation is structured at the microenvironmental level and can be consistently detected across biological compartments.\u003c/p\u003e\n\u003cp\u003eBy jointly leveraging large-scale peripheral transcriptomics, cell-type\u0026ndash;resolved single-cell data, and drug-response signatures in animal models, this study establishes a closed-loop validation framework for biomarker discovery. Functional annotation of the identified robust signatures revealed a convergent enrichment in pathways governing immunomodulation, organelle integrity, and cellular stress adaptation. This prominent immune signature strongly corroborates the prevailing neuroinflammation hypothesis of MDD [41, 42]. Beyond immune activation, the dysregulation of pathways associated with cell membrane dynamics and cellular stress responses points to widespread subcellular compromise in MDD. Specifically, the implication of mechanisms such as mitochondrial dysfunction and oxidative stress suggests a cascade of bioenergetic failure. Such disruptions are known to impair glial-neuronal homeostasis, disrupting energy metabolism and ultimately compromising synaptic transmission [43]. The activation of these pathways indicated a maladaptive cellular stress response, wherein intrinsic homeostatic mechanisms are overwhelmed by chronic stressors, shifting the cellular phenotype from adaptive survival to pathological dysfunction. The consistent reversal of microenvironment-associated signatures following antidepressant treatment further highlights their functional relevance and disease specificity. Collectively, these findings underscore the necessity of multimodal integration to identify robust, microenvironment-informed diagnostic biomarkers in MDD\u0026mdash;biomarkers that move beyond isolated molecular changes toward a systems-level representation of depressive pathology.\u003c/p\u003e\n\u003cp\u003eHaving delineated the cellular landscape of MDD, our ultimate objective was to pinpoint the specific molecular drivers orchestrating these glial dysfunctions. The identification of \u003cem\u003eHK2\u003c/em\u003e emerged as a convergent and biologically coherent finding rather than a dataset-specific observation. \u003cem\u003eHK2\u0026nbsp;\u003c/em\u003ewas consistently prioritized across peripheral transcriptomic cohorts, multiple single-cell brain datasets, and pharmacological response models, reflecting its central role at the intersection of microglial activation, metabolic regulation, and immune signaling. Given the tight coupling between microglial inflammatory states and glycolytic reprogramming, the microglia-specific upregulation of \u003cem\u003eHK2\u0026nbsp;\u003c/em\u003esupports a model in which immunometabolic shifts contribute to microenvironmental dysfunction in MDD. Importantly, the translational relevance of \u003cem\u003eHK2\u003c/em\u003e was further supported by independent clinical validation, as both transcriptomic sequencing and RT\u0026ndash;qPCR analyses of a hospital-derived cohort consistently confirmed its elevated expression in MDD.\u003c/p\u003e\n\u003cp\u003eDespite the robustness of this integrative framework, several limitations merit consideration. First, while the RRA-based strategy effectively reduced inter-study heterogeneity, it may have obscured transcriptomic features linked to specific depressive subtypes or sex-specific effects. Second, the predominantly cross-sectional design limits causal inference, precluding definitive conclusions regarding whether glial dysregulation represents a primary driver or a downstream consequence of MDD. Finally, pharmacological validation relied mainly on rodent models, and although supported by non-human primate data, species-specific immune differences cannot be fully excluded. In summary, this study establishes a systematic multi-omics framework that bridges peripheral circulation and central pathology in MDD, revealing a convergent immunometabolic microenvironment characterized by microglial hyperactivation and oligodendrocyte dysfunction. Through cross-modal and cross-platform validation, \u003cem\u003eHK2\u003c/em\u003e emerged as a robust, cell-type\u0026ndash;resolved peripheral biomarker with translational relevance. These findings provide molecular evidence narrowing the blood\u0026ndash;brain gap and support a glial-centric paradigm for precision diagnosis and therapeutic monitoring in major depressive disorder.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was funded by the National Natural Science Foundation of China (12501670), Guangdong Basic and Applied Basic Research Foundation (No. 2025A1515011988).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting of Interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing interests.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eHan J, Feng Y, Li N, Feng L, Xiao L, Zhu X\u003cem\u003e et al.\u003c/em\u003e Correlation Between Word Frequency and 17 Items of Hamilton Scale in Major Depressive Disorder. \u003cem\u003eFrontiers in psychiatry\u003c/em\u003e 2022; \u003cstrong\u003e13: \u003c/strong\u003e902873.\u003c/li\u003e\n\u003cli\u003eChen Q, Huang S, Xu H, Peng J, Wang P, Li S\u003cem\u003e et al.\u003c/em\u003e The burden of mental disorders in Asian countries, 1990-2019: an analysis for the global burden of disease study 2019. \u003cem\u003eTranslational psychiatry\u003c/em\u003e 2024; \u003cstrong\u003e14\u003c/strong\u003e(1)\u003cstrong\u003e: \u003c/strong\u003e167.\u003c/li\u003e\n\u003cli\u003eFiorillo A, Albert U, Dell\u0026apos;Osso B, Pompili M, Sani G, Sampogna G. The clinical utility and relevance in clinical practice of DSM-5 specifiers for major depressive disorder: A Delphi expert consensus study. \u003cem\u003eComprehensive psychiatry\u003c/em\u003e 2024; \u003cstrong\u003e133: \u003c/strong\u003e152502.\u003c/li\u003e\n\u003cli\u003eCosci F, Fava GA. The clinical inadequacy of the DSM-5 classification of somatic symptom and related disorders: an alternative trans-diagnostic model. \u003cem\u003eCNS spectrums\u003c/em\u003e 2016; \u003cstrong\u003e21\u003c/strong\u003e(4)\u003cstrong\u003e: \u003c/strong\u003e310-317.\u003c/li\u003e\n\u003cli\u003eVoineskos D, Daskalakis ZJ, Blumberger DM. Management of Treatment-Resistant Depression: Challenges and Strategies. \u003cem\u003eNeuropsychiatric disease and treatment\u003c/em\u003e 2020; \u003cstrong\u003e16: \u003c/strong\u003e221-234.\u003c/li\u003e\n\u003cli\u003eAnagha K, Shihabudheen P, Uvais NA. Side Effect Profiles of Selective Serotonin Reuptake Inhibitors: A Cross-Sectional Study in a Naturalistic Setting. \u003cem\u003ePrim Care Companion CNS Disord\u003c/em\u003e 2021; \u003cstrong\u003e23\u003c/strong\u003e(4).\u003c/li\u003e\n\u003cli\u003eWohleb ES, Franklin T, Iwata M, Duman RS. Integrating neuroimmune systems in the neurobiology of depression. \u003cem\u003eNature reviews Neuroscience\u003c/em\u003e 2016; \u003cstrong\u003e17\u003c/strong\u003e(8)\u003cstrong\u003e: \u003c/strong\u003e497-511.\u003c/li\u003e\n\u003cli\u003eCollignon A, Dion-Albert L, M\u0026eacute;nard C, Coelho-Santos V. Sex, hormones and cerebrovascular function: from development to disorder. \u003cem\u003eFluids and barriers of the CNS\u003c/em\u003e 2024; \u003cstrong\u003e21\u003c/strong\u003e(1)\u003cstrong\u003e: \u003c/strong\u003e2.\u003c/li\u003e\n\u003cli\u003eCarrier M, Dolhan K, Bobotis BC, Desjardins M, Tremblay M. The implication of a diversity of non-neuronal cells in disorders affecting brain networks. \u003cem\u003eFrontiers in cellular neuroscience\u003c/em\u003e 2022; \u003cstrong\u003e16: \u003c/strong\u003e1015556.\u003c/li\u003e\n\u003cli\u003eSălcudean A, Bodo C-R, Popovici R-A, Cozma M-M, Păcurar M, Crăciun R-E\u003cem\u003e et al.\u003c/em\u003e Neuroinflammation\u0026mdash;A Crucial Factor in the Pathophysiology of Depression\u0026mdash;A Comprehensive Review. \u003cem\u003eBiomolecules\u003c/em\u003e 2025; \u003cstrong\u003e15\u003c/strong\u003e(4)\u003cstrong\u003e: \u003c/strong\u003e502.\u003c/li\u003e\n\u003cli\u003eChawla A, Cakmakci D, Fiori LM, Zang W, Maitra M, Yang J\u003cem\u003e et al.\u003c/em\u003e Single-nucleus chromatin accessibility profiling identifies cell types and functional variants contributing to major depression. \u003cem\u003eNat Genet\u003c/em\u003e 2025; \u003cstrong\u003e57\u003c/strong\u003e(8)\u003cstrong\u003e: \u003c/strong\u003e1890-1904.\u003c/li\u003e\n\u003cli\u003eMariani N, Cattane N, Pariante C, Cattaneo A. Gene expression studies in Depression development and treatment: an overview of the underlying molecular mechanisms and biological processes to identify biomarkers. \u003cem\u003eTransl Psychiatry\u003c/em\u003e 2021; \u003cstrong\u003e11\u003c/strong\u003e(1)\u003cstrong\u003e: \u003c/strong\u003e354.\u003c/li\u003e\n\u003cli\u003eMedeiros GC, Gould TD, Prueitt WL, Nanavati J, Grunebaum MF, Farber NB\u003cem\u003e et al.\u003c/em\u003e Blood-based biomarkers of antidepressant response to ketamine and esketamine: A systematic review and meta-analysis. \u003cem\u003eMolecular psychiatry\u003c/em\u003e 2022; \u003cstrong\u003e27\u003c/strong\u003e(9)\u003cstrong\u003e: \u003c/strong\u003e3658-3669.\u003c/li\u003e\n\u003cli\u003eMokhtari A, Ibrahim EC, Gloaguen A, Barrot CC, Cohen D, Derouin M\u003cem\u003e et al.\u003c/em\u003e Using multiomic integration to improve blood biomarkers of major depressive disorder: a case-control study. \u003cem\u003eEBioMedicine\u003c/em\u003e 2025; \u003cstrong\u003e113: \u003c/strong\u003e105569.\u003c/li\u003e\n\u003cli\u003eOfengeim D, Giagtzoglou N, Huh D, Zou C, Yuan J. Single-Cell RNA Sequencing: Unraveling the Brain One Cell at a Time. \u003cem\u003eTrends Mol Med\u003c/em\u003e 2017; \u003cstrong\u003e23\u003c/strong\u003e(6)\u003cstrong\u003e: \u003c/strong\u003e563-576.\u003c/li\u003e\n\u003cli\u003eLi JM, Jiang CL. Biological Diagnosis of Depression: A Biomarker Panel from Several Nonspecial Indicators Instead of the Specific Biomarker(s). \u003cem\u003eNeuropsychiatr Dis Treat\u003c/em\u003e 2022; \u003cstrong\u003e18: \u003c/strong\u003e3067-3071.\u003c/li\u003e\n\u003cli\u003eYamamoto M, Sakai M, Yu Z, Nakanishi M, Yoshii H. Glial Markers of Suicidal Behavior in the Human Brain-A Systematic Review of Postmortem Studies. \u003cem\u003eInt J Mol Sci\u003c/em\u003e 2024; \u003cstrong\u003e25\u003c/strong\u003e(11).\u003c/li\u003e\n\u003cli\u003eYirmiya R, Rimmerman N, Reshef R. Depression as a microglial disease. \u003cem\u003eTrends in neurosciences\u003c/em\u003e 2015; \u003cstrong\u003e38\u003c/strong\u003e(10)\u003cstrong\u003e: \u003c/strong\u003e637-658.\u003c/li\u003e\n\u003cli\u003eEdgar N, Sibille E. A putative functional role for oligodendrocytes in mood regulation. \u003cem\u003eTranslational psychiatry\u003c/em\u003e 2012; \u003cstrong\u003e2\u003c/strong\u003e(5)\u003cstrong\u003e: \u003c/strong\u003ee109.\u003c/li\u003e\n\u003cli\u003eNagy C, Suderman M, Yang J, Szyf M, Mechawar N, Ernst C\u003cem\u003e et al.\u003c/em\u003e Astrocytic abnormalities and global DNA methylation patterns in depression and suicide. \u003cem\u003eMolecular psychiatry\u003c/em\u003e 2015; \u003cstrong\u003e20\u003c/strong\u003e(3)\u003cstrong\u003e: \u003c/strong\u003e320-328.\u003c/li\u003e\n\u003cli\u003eBuenrostro JD, Wu B, Litzenburger UM, Ruff D, Gonzales ML, Snyder MP\u003cem\u003e et al.\u003c/em\u003e Single-cell chromatin accessibility reveals principles of regulatory variation. \u003cem\u003eNature\u003c/em\u003e 2015; \u003cstrong\u003e523\u003c/strong\u003e(7561)\u003cstrong\u003e: \u003c/strong\u003e486-490.\u003c/li\u003e\n\u003cli\u003eDenisenko E, Guo BB, Jones M, Hou R, de Kock L, Lassmann T\u003cem\u003e et al.\u003c/em\u003e Systematic assessment of tissue dissociation and storage biases in single-cell and single-nucleus RNA-seq workflows. \u003cem\u003eGenome biology\u003c/em\u003e 2020; \u003cstrong\u003e21\u003c/strong\u003e(1)\u003cstrong\u003e: \u003c/strong\u003e130.\u003c/li\u003e\n\u003cli\u003eSquair JW, Gautier M, Kathe C, Anderson MA, James ND, Hutson TH\u003cem\u003e et al.\u003c/em\u003e Confronting false discoveries in single-cell differential expression. \u003cem\u003eNature communications\u003c/em\u003e 2021; \u003cstrong\u003e12\u003c/strong\u003e(1)\u003cstrong\u003e: \u003c/strong\u003e5692.\u003c/li\u003e\n\u003cli\u003eStrawbridge R, Young AH, Cleare AJ. Biomarkers for depression: recent insights, current challenges and future prospects. \u003cem\u003eNeuropsychiatric disease and treatment\u003c/em\u003e 2017; \u003cstrong\u003e13: \u003c/strong\u003e1245-1262.\u003c/li\u003e\n\u003cli\u003eLe-Niculescu H, Roseberry K, Gill SS, Levey DF, Phalen PL, Mullen J\u003cem\u003e et al.\u003c/em\u003e Precision medicine for mood disorders: objective assessment, risk prediction, pharmacogenomics, and repurposed drugs. \u003cem\u003eMolecular psychiatry\u003c/em\u003e 2021; \u003cstrong\u003e26\u003c/strong\u003e(7)\u003cstrong\u003e: \u003c/strong\u003e2776-2804.\u003c/li\u003e\n\u003cli\u003eLeday GGR, V\u0026eacute;rtes PE, Richardson S, Greene JR, Regan T, Khan S\u003cem\u003e et al.\u003c/em\u003e Replicable and Coupled Changes in Innate and Adaptive Immune Gene Expression in Two Case-Control Studies of Blood Microarrays in Major Depressive Disorder. \u003cem\u003eBiol Psychiatry\u003c/em\u003e 2018; \u003cstrong\u003e83\u003c/strong\u003e(1)\u003cstrong\u003e: \u003c/strong\u003e70-80.\u003c/li\u003e\n\u003cli\u003eMiyata S, Kurachi M, Okano Y, Sakurai N, Kobayashi A, Harada K\u003cem\u003e et al.\u003c/em\u003e Blood Transcriptomic Markers in Patients with Late-Onset Major Depressive Disorder. \u003cem\u003ePLoS One\u003c/em\u003e 2016; \u003cstrong\u003e11\u003c/strong\u003e(2)\u003cstrong\u003e: \u003c/strong\u003ee0150262.\u003c/li\u003e\n\u003cli\u003eDi Benedetto MG, Priestley KMS, Cattane N, Genini P, Saleri S, Biaggi A\u003cem\u003e et al.\u003c/em\u003e Blood transcriptomic signatures associated with depression, or the risk for depression, in pregnant women from the Psychiatry Research And Motherhood - Depression (PRAM-D) study. \u003cem\u003eTransl Psychiatry\u003c/em\u003e 2025; \u003cstrong\u003e15\u003c/strong\u003e(1)\u003cstrong\u003e: \u003c/strong\u003e110.\u003c/li\u003e\n\u003cli\u003eSun S, Liu Q, Wang Z, Huang YY, Sublette ME, Dwork AJ\u003cem\u003e et al.\u003c/em\u003e Brain and blood transcriptome profiles delineate common genetic pathways across suicidal ideation and suicide. \u003cem\u003eMol Psychiatry\u003c/em\u003e 2024; \u003cstrong\u003e29\u003c/strong\u003e(5)\u003cstrong\u003e: \u003c/strong\u003e1417-1426.\u003c/li\u003e\n\u003cli\u003eMoisan MP, Foury A, Dexpert S, Cole SW, Beau C, Forestier D\u003cem\u003e et al.\u003c/em\u003e Transcriptomic signaling pathways involved in a naturalistic model of inflammation-related depression and its remission. \u003cem\u003eTransl Psychiatry\u003c/em\u003e 2021; \u003cstrong\u003e11\u003c/strong\u003e(1)\u003cstrong\u003e: \u003c/strong\u003e203.\u003c/li\u003e\n\u003cli\u003eCathomas F, Bevilacqua L, Ramakrishnan A, Kronman H, Costi S, Schneider M\u003cem\u003e et al.\u003c/em\u003e Whole blood transcriptional signatures associated with rapid antidepressant response to ketamine in patients with treatment resistant depression. \u003cem\u003eTransl Psychiatry\u003c/em\u003e 2022; \u003cstrong\u003e12\u003c/strong\u003e(1)\u003cstrong\u003e: \u003c/strong\u003e12.\u003c/li\u003e\n\u003cli\u003eWu J, Li Y, Huang Y, Liu L, Zhang H, Nagy C\u003cem\u003e et al.\u003c/em\u003e Integrating spatial and single-nucleus transcriptomic data elucidates microglial-specific responses in female cynomolgus macaques with depressive-like behaviors. \u003cem\u003eNat Neurosci\u003c/em\u003e 2023; \u003cstrong\u003e26\u003c/strong\u003e(8)\u003cstrong\u003e: \u003c/strong\u003e1352-1364.\u003c/li\u003e\n\u003cli\u003eNagy C, Maitra M, Tanti A, Suderman M, Th\u0026eacute;roux JF, Davoli MA\u003cem\u003e et al.\u003c/em\u003e Single-nucleus transcriptomics of the prefrontal cortex in major depressive disorder implicates oligodendrocyte precursor cells and excitatory neurons. \u003cem\u003eNat Neurosci\u003c/em\u003e 2020; \u003cstrong\u003e23\u003c/strong\u003e(6)\u003cstrong\u003e: \u003c/strong\u003e771-781.\u003c/li\u003e\n\u003cli\u003eRayan NA, Aow J, Lim MGL, Arcego DM, Ryan R, Nourbakhsh N\u003cem\u003e et al.\u003c/em\u003e Shared and unique transcriptomic signatures of antidepressant and probiotics action in the mammalian brain. \u003cem\u003eMol Psychiatry\u003c/em\u003e 2024; \u003cstrong\u003e29\u003c/strong\u003e(11)\u003cstrong\u003e: \u003c/strong\u003e3653-3668.\u003c/li\u003e\n\u003cli\u003eRayan NA, Kumar V, Aow J, Rastegar N, Lim MGL, O\u0026apos;Toole N\u003cem\u003e et al.\u003c/em\u003e Integrative multi-omics landscape of fluoxetine action across 27 brain regions reveals global increase in energy metabolism and region-specific chromatin remodelling. \u003cem\u003eMol Psychiatry\u003c/em\u003e 2022; \u003cstrong\u003e27\u003c/strong\u003e(11)\u003cstrong\u003e: \u003c/strong\u003e4510-4525.\u003c/li\u003e\n\u003cli\u003eKolde R, Laur S, Adler P, Vilo J. Robust rank aggregation for gene list integration and meta-analysis. \u003cem\u003eBioinformatics\u003c/em\u003e 2012; \u003cstrong\u003e28\u003c/strong\u003e(4)\u003cstrong\u003e: \u003c/strong\u003e573-580.\u003c/li\u003e\n\u003cli\u003eRaudvere U, Kolberg L, Kuzmin I, Arak T, Adler P, Peterson H\u003cem\u003e et al.\u003c/em\u003e g:Profiler: a web server for functional enrichment analysis and conversions of gene lists (2019 update). \u003cem\u003eNucleic Acids Res\u003c/em\u003e 2019; \u003cstrong\u003e47\u003c/strong\u003e(W1)\u003cstrong\u003e: \u003c/strong\u003eW191-w198.\u003c/li\u003e\n\u003cli\u003eChu T, Wang Z, Pe\u0026apos;er D, Danko CG. Cell type and gene expression deconvolution with BayesPrism enables Bayesian integrative analysis across bulk and single-cell RNA sequencing in oncology. \u003cem\u003eNat Cancer\u003c/em\u003e 2022; \u003cstrong\u003e3\u003c/strong\u003e(4)\u003cstrong\u003e: \u003c/strong\u003e505-517.\u003c/li\u003e\n\u003cli\u003eJin S, Plikus MV, Nie Q. CellChat for systematic analysis of cell-cell communication from single-cell transcriptomics. \u003cem\u003eNat Protoc\u003c/em\u003e 2025; \u003cstrong\u003e20\u003c/strong\u003e(1)\u003cstrong\u003e: \u003c/strong\u003e180-219.\u003c/li\u003e\n\u003cli\u003eSzklarczyk D, Nastou K, Koutrouli M, Kirsch R, Mehryary F, Hachilif R\u003cem\u003e et al.\u003c/em\u003e The STRING database in 2025: protein networks with directionality of regulation. \u003cem\u003eNucleic Acids Res\u003c/em\u003e 2025; \u003cstrong\u003e53\u003c/strong\u003e(D1)\u003cstrong\u003e: \u003c/strong\u003eD730-d737.\u003c/li\u003e\n\u003cli\u003eRuiz NAL, Del \u0026Aacute;ngel DS, Brizuela NO, Peraza AV, Olgu\u0026iacute;n HJ, Soto MP\u003cem\u003e et al.\u003c/em\u003e Inflammatory Process and Immune System in Major Depressive Disorder. \u003cem\u003eInt J Neuropsychopharmacol\u003c/em\u003e 2022; \u003cstrong\u003e25\u003c/strong\u003e(1)\u003cstrong\u003e: \u003c/strong\u003e46-53.\u003c/li\u003e\n\u003cli\u003eCui L, Li S, Wang S, Wu X, Liu Y, Yu W\u003cem\u003e et al.\u003c/em\u003e Major depressive disorder: hypothesis, mechanism, prevention and treatment. \u003cem\u003eSignal Transduct Target Ther\u003c/em\u003e 2024; \u003cstrong\u003e9\u003c/strong\u003e(1)\u003cstrong\u003e: \u003c/strong\u003e30.\u003c/li\u003e\n\u003cli\u003eYin Y, Ju T, Zeng D, Duan F, Zhu Y, Liu J\u003cem\u003e et al.\u003c/em\u003e \u0026quot;Inflamed\u0026quot; depression: A review of the interactions between depression and inflammation and current anti-inflammatory strategies for depression. \u003cem\u003ePharmacol Res\u003c/em\u003e 2024; \u003cstrong\u003e207: \u003c/strong\u003e107322.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"molecular-psychiatry","isNatureJournal":false,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"mp","sideBox":"Learn more about [Molecular Psychiatry](http://www.nature.com/mp/)","snPcode":"41380","submissionUrl":"https://mts-mp.nature.com/cgi-bin/main.plex","title":"Molecular Psychiatry","twitterHandle":"@molpsychiatry","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Major Depressive Disorder, Multi-omics, Microenvironment, Microglia, Inflammation","lastPublishedDoi":"10.21203/rs.3.rs-8976053/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8976053/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Characterizing the pathological microenvironment of major depressive disorder (MDD) is of clinical relevance yet remains poorly defined. Here, we systematically dissect microenvironment-associated regulatory mechanisms underlying MDD phenotypes through integrative analysis of ten transcriptomic datasets spanning human peripheral blood, human and macaque brain single-cell transcriptomes, and rat pharmacological models. A robust transcriptomic signature was identified that links systemic inflammation signals to central glial dysfunction. MDD-associated up-regulated genes were preferentially enriched in microglia, whereas down-regulated genes were predominantly associated with oligodendrocytes, indicating a coordinated shift toward a pro-inflammatory and demyelinating microenvironment. Multi-layer convergence prioritizes ANO10, HK2, and CCDC50 as core dysregulated genes. Among them, HK2 emerges as a central immunometabolic node, coupling glycolytic reprogramming to microglial activation. HK2 exhibits cell-type specificity, antidepressant-responsive dynamics, and independent clinical validation, supporting its role as a mechanistically anchored and translationally accessible marker. Together, these findings delineate a glial-centric immunometabolic framework for MDD and provide molecular evidence bridging peripheral signals with central microenvironmental pathology.","manuscriptTitle":"Integrated multi-omics characterization identifies key microenvironmental biomarkers in major depressive disorder microenvironment","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-27 07:56:46","doi":"10.21203/rs.3.rs-8976053/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorInvitedReview","content":"This content is not available.","date":"2026-05-11T02:01:33+00:00","index":1,"fulltext":"This content is not available."},{"type":"reviewerAgreed","content":"This content is not available.","date":"2026-04-19T17:59:17+00:00","index":1,"fulltext":"This content is not available."},{"type":"reviewersInvited","content":"","date":"2026-04-18T22:43:35+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-02-27T15:07:59+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-02-27T14:47:16+00:00","index":"","fulltext":""},{"type":"submitted","content":"Molecular Psychiatry","date":"2026-02-26T13:01:10+00:00","index":"","fulltext":""},{"type":"checksFailed","content":"","date":"2026-02-26T12:01:22+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"molecular-psychiatry","isNatureJournal":false,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"mp","sideBox":"Learn more about [Molecular Psychiatry](http://www.nature.com/mp/)","snPcode":"41380","submissionUrl":"https://mts-mp.nature.com/cgi-bin/main.plex","title":"Molecular Psychiatry","twitterHandle":"@molpsychiatry","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"4156bf70-08e2-4431-89d8-987e9880bb65","owner":[],"postedDate":"April 27th, 2026","published":true,"recentEditorialEvents":[{"type":"editorInvitedReview","content":"This content is not available.","date":"2026-05-11T02:01:33+00:00","index":1,"fulltext":"This content is not available."}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":66584833,"name":"Biological sciences/Biotechnology"},{"id":66584834,"name":"Health sciences/Biomarkers/Predictive markers"},{"id":66584835,"name":"Health sciences/Diseases/Psychiatric disorders/Depression"}],"tags":[],"updatedAt":"2026-04-27T07:56:47+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-27 07:56:46","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8976053","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8976053","identity":"rs-8976053","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}