Abstract

   The microbiota-gut-brain axis plays a pivotal role in neuropsychiatric
   disorders, particularly in depression. Escitalopram (ESC) is a
   first-line antidepressant, however, its regulatory mechanisms on the
   microbiota-gut-brain axis in the treatment of depression remain
   unclear. The antidepressant effects of ESC were evaluated using the
   forced swim test in Wistar-Kyoto (WKY) rats, while damage in the gut
   and brain regions was assessed through H&E staining and
   immunohistochemistry. The therapeutic mechanisms in WKY rats with
   depression-like behavior were investigated through 16S rRNA sequencing
   of the gut microbiota, serum untargeted metabolomics, and hippocampal
   proteomics. Results indicated that ESC intervention improved
   depressive-like behaviors, as evidenced by increased swimming times in
   WKY rats, and also restored intestinal permeability and brain tissue
   integrity. Significant changes in the gut microbiota composition,
   particularly an increase in Bacteroides barnesiae, as well as increases
   in serum sphingolipid metabolites (Sphinganine 1-phosphate,
   Sphingosine, and Sphingosine-1-phosphate) and hippocampal proteins
   (Sptlc1, Enpp5, Enpp2), were strongly correlated. These robust
   correlations suggest that ESC may exert its antidepressant effects by
   modulating sphingolipid metabolism through the influence of gut
   microbiota. Accordingly, this research elucidates novel mechanisms
   underlying the antidepressant efficacy of ESC and highlights the
   pivotal importance of the microbiota-gut-brain axis in mediating these
   effects.

   Subject terms: Depression, Hippocampus

Introduction

   Depression is a common mental disorder, characterized by persistent low
   mood, loss of interest, and impaired cognitive function, affecting an
   estimated 300 million people of all ages globally [[38]1]. The impact
   of depression extends beyond psychological health, increasing the risk
   of chronic conditions such as cardiovascular diseases and diabetes, as
   well as being closely associated with a significantly higher suicide
   rate compared to the general population [[39]2]. The hippocampus plays
   a crucial role in the pathophysiology of depression, with alterations
   in its structure and function potentially being closely linked to the
   symptoms of the disorder [[40]3]. Studies have shown that the number of
   hippocampal neurons and glial cells may be reduced in individuals with
   depression, which is negatively correlated with the severity of the
   condition [[41]4]. Furthermore, chronic stress and depression are
   associated with synaptic inhibition and neuronal atrophy in the
   hippocampus [[42]5]. In addition to these neural changes, emerging
   evidence suggests that the microbiota-gut-brain (MGB) axis, which
   involves complex interactions between the gastrointestinal system and
   the central nervous system (CNS), may influence brain function through
   neural, endocrine, immune, metabolic, and other pathways [[43]6]. Over
   the past decade, mounting evidence has suggested a potential link
   between alterations in the fecal microbiome composition and the
   development of depression [[44]7, [45]8]. Depression phenotypes are
   accompanied by changes in the gut microbiota composition, which in turn
   influences depression-like behaviors and increases the risk of
   depression [[46]6].

   Currently, pharmacological interventions remain the cornerstone of
   depression treatment. The interaction between medications and the
   microbiome can alter bacterial metabolism as well as drug activity and
   efficacy [[47]9]. Antidepressants possess antimicrobial and
   anti-inflammatory properties, which are associated with the ecology of
   the gut microbiota [[48]10, [49]11]. The antidepressant effects of
   fluoxetine and ketamine are correlated with the gut microbiome. For
   example, Actinobacteria and genera producing short-chain fatty acids
   (SCFAs), such as Butyricimonas and Turicibacter, might be related to
   the antidepressant effects of ketamine [[50]12]. Mice treated with
   fluoxetine exhibit changes in gut microbiome diversity and composition,
   such as decreased abundance of Ruminococcus and Adlercreutzia [[51]13].
   Escitalopram (ESC), a selective serotonin reuptake inhibitor, is
   considered a first-line treatment option for patients with depression
   [[52]14]. In the depression CUMS model, significant differences in gut
   microbial communities and metabolites are observed between responders
   and non-responders to escitalopram treatment [[53]15]. Nowadays,
   targeting the gut microbiome and integrating MGB axis mechanisms in
   disease treatment has become a focal point of research. Therefore,
   studying the regulatory effects of escitalopram on the gut microbiome
   will aid in advancing our understanding of its antidepressant
   mechanisms.

   The Wistar-Kyoto rat (WKY rat) is a widely used spontaneous model of
   depression in research, exhibiting a variety of behavioral and
   physiological traits associated with depression, such as reduced social
   interaction, decreased exploratory behavior, and heightened sensitivity
   to stressful environments [[54]16]. These characteristics make the WKY
   rat an important model for understanding the pathophysiological
   mechanisms of depression and its treatment. The Forced Swim Test (FST)
   is one of the classic experimental methods for assessing antidepressant
   behavior in animals [[55]17]. In this test, animals are placed in a
   pool of water from which they cannot escape or touch the bottom,
   typically exhibiting a state referred to as “behavioral despair.” The
   effectiveness of antidepressants is generally assessed in the FST by
   reducing the time spent in passive floating and increasing the time
   struggling and swimming, thus reflecting the potential antidepressant
   activity of the compounds. This method is not only a standard tool for
   evaluating antidepressants but also key for understanding behavioral
   changes associated with depression [[56]18, [57]19]. However, despite
   the WKY rat and FST providing significant tools for depression
   research, the mechanisms by which the gut microbiota influences the
   antidepressant effects of escitalopram, a commonly used selective
   serotonin reuptake inhibitor, especially within the WKY rat model,
   remain unclear.

   In this study, we utilized the WKY rat model of depression to
   investigate changes in the gut microbiome, serum metabolomic profiles,
   and hippocampal proteomic patterns following administration of ESC.
   Furthermore, this research analyzed the interactions between microbes,
   metabolites, and proteins, exploring how these differential elements
   synergistically function within the MGB axis to mediate the
   antidepressant effects of ESC. This study aims to deepen our
   understanding of the gut-brain interactions in the treatment of
   depression, hoping to provide a theoretical basis and experimental
   support for novel antidepressant treatment strategies based on
   microbiome modulation.

Materials and methods

Animal and experimental design

   This study utilized 8 male Wistar rats and 16 WKY rats, a depression
   model, weighing 150 ± 20 g and aged 6–8 weeks, obtained from Vital
   River Laboratories (Beijing, China). No formal sample size estimation
   was performed. The chosen sample size was based on prior research
   experience and the feasibility of the experimental design. Animals were
   included in the study if they met the following criteria: healthy, male
   rats of the specified strains, with no signs of illness or physical
   abnormalities. Exclusion criteria included any animals that exhibited
   abnormal behavior, signs of injury, or illness during the acclimation
   period or prior to the start of the experiment. The animals were housed
   under conditions of 21–25 °C temperature, 35 ± 5% humidity, and a 12 h
   light/dark cycle. The study was approved by the First Affiliated
   Hospital of Henan University of Science and Technology Animal Ethics
   Committee (2023-D-H0006), and conducted in accordance with the
   Regulations for the Management of Laboratory Animals by the Ministry of
   Science and Technology of the People’s Republic of China.

   The compound escitalopram oxalate was acquired from MCE (HY-14258A,
   MCE, China). Escitalopram oxalate was dissolved in 0.85% saline to
   achieve a dose of 10 mg/kg. Age and weight-matched WKY rats were
   randomly divided into two groups using computer-based randomization
   software, each consisting of 8 animals: one group received the prepared
   ESC solution (ESC group), and the other received the vehicle solution
   (WKY group). Additionally, 8 Wistar rats served as the control group
   (CON group), receiving the vehicle solution without any treatment. Rats
   in the ESC group were administered 10 mg/kg/day of ESC solution, while
   those in the WKY and control groups received an equivalent volume of
   vehicle solution, all via intraperitoneal injection for 21 consecutive
   days. Post-administration, daily assessments and records of rat body
   weight were conducted. Furthermore, any physiological changes,
   including the occurrence of diarrhea and any lethal effects, were
   carefully monitored. Blinding was not implemented during the treatment
   phase, but all outcome assessments were conducted without prior
   knowledge of group assignment to minimize bias. The experimental
   procedures are shown in Fig. [58]1A.

Fig. 1. ESC improves behavioral changes and MGB axis damage in rats with
depression-like behavior.

   [59]Fig. 1
   [60]Open in a new tab

   A Schematic flow of the experiment. B The immobility (%) in the forced
   swim test between the three groups at DAY0 and DAY22 was analyzed using
   a two-way repeated-measures ANOVA, with post-hoc comparisons performed
   using Bonferroni correction. C Growth rates of body weight in different
   rat groups were analyzed using a two-way repeated-measures ANOVA, with
   post-hoc comparisons performed using Bonferroni correction. The
   significance between CON and WKY is indicated by #, and between CON and
   ESC by *. D HE staining of the cerebral cortex and hippocampus. Scale
   bar, 50 μm. E Immunohistochemical staining and H-score assessment of
   colon sections for claudin-1, occludin, and ZO-1. Scale bar, 50 μm.
   Data are expressed as mean ± SEM. ^*P < 0.05, ^**P < 0.01,
   ^***P < 0.001, ns, no significance.

Forced swim test

   A transparent cylindrical glass tank (20 × 60 cm) is filled with water
   to approximately two-thirds of its height, with the water temperature
   maintained at 24 ± 1 °C [[61]20, [62]21]. A backlight is placed behind
   the swim tank to prevent light reflection from affecting experimental
   outcomes. Rats are placed in the swim tank for a 5 min trial, with
   their activity continuously recorded by a camera. It should be noted
   that no pretest was performed before the initial 5 min trial. A pretest
   is typically used to acclimate animals to the testing environment,
   enhancing the consistency and reliability of behavioral data [[63]22].
   This absence should be considered when interpreting and extrapolating
   the results of this study. The primary evaluation metric is the
   immobility time, defined as minimal movement just sufficient to keep
   the rat afloat, with occasional gentle movements to prevent drowning.
   FST is conducted both before (Day0) and 21 days after drug
   administration (Day22) to assess changes in depressive-like behaviors
   in rats.

Hematoxylin-eosin staining

   The brain tissues were fixed in 4% paraformaldehyde solution for 48 h,
   followed by embedding in paraffin. Paraffin-embedded tissue sections
   were prepared, stained with hematoxylin and eosin (HE). Samples were
   immersed in alcohol, stained with hematoxylin for 5 min, immersed in
   alcohol for 1 min, stained with eosin for 15 s and re-immersed in
   alcohol and xylene. Then subsequently examined under a light microscope
   to observe the morphology and structure of the brain tissues and
   capture images.

Immunohistochemistry

   The colon tissues were fixed and embedded in paraffin, and
   immunohistochemistry was employed to analyze the expression and
   distribution of colonic tight junction proteins. Paraffin-embedded
   tissue sections were deparaffinized through a series of rehydration
   steps and subjected to antigen retrieval. To block endogenous
   peroxidases, each slide was treated with 3% hydrogen peroxide for
   25 min and subsequently blocked with 3% normal goat serum. Sections
   were incubated overnight at 4 °C with specific primary antibodies
   against claudin-1, occludin, and zonula occludens-1 (ZO-1). Following
   PBST washing, slides were incubated with secondary antibodies at room
   temperature for 1 h. Visualization was achieved using
   3,3′-diaminobenzidine (DAB) chromogen, followed by counterstaining with
   hematoxylin. Immunohistochemistry scoring (H-score) was conducted to
   evaluate staining intensity: H-score = (percentage of weak intensity
   ×1) + (percentage of moderate intensity ×2) + (percentage of strong
   intensity ×3). The percentage of stained area and intensity of stained
   cells were graded as 0, negative; 1+, weak; 2+, moderate; 3+, strong
   [[64]23].

16S rRNA gene sequencing analysis

   Fresh colonic content was collected into sterile tubes, rapidly frozen
   in liquid nitrogen, and stored at −80 °C pending further processing.
   Total genomic DNA extraction of microbial communities was performed,
   with PCR amplification and library construction targeting the V3-V4
   variable regions of the 16S rRNA gene. Sequencing was carried out on
   Illumina’s PE250 platform to generate 250 bp paired-end raw reads, and
   raw paired-end sequences were quality-controlled using fastp (version
   0.19.6) and assembled using FLASH (version 1.2.11). Post-quality
   control, optimized sequences were denoised using the DADA2 plugin
   within the Qiime2 pipeline with default parameters to obtain amplicon
   sequence variants (ASVs). ASVs underwent sequence rarefaction [[65]24]
   for bioinformatics analysis, where α-diversity indices (Chao, Ace,
   Simpson, and Shannon) were compared using Kruskal-Wallis tests. Beta
   diversity analysis was visualized using principal coordinates analysis
   (PCoA) to assess differences in microbial community structure between
   groups. Furthermore, we employed the Wilcoxon rank-sum test to analyze
   the gut microbiota health index (GMHI) and the microbiota dysbiosis
   index (MDI). The GMHI, a composite index used to assess the overall
   health status of the gut microbiota, is generally considered a marker
   of gut health and ecological dysbiosis [[66]25]. The MDI serves as an
   index to determine the extent of ecological imbalance, with higher
   values indicating greater degrees of microbial dysregulation [[67]26].
   Differential abundance of taxa between groups was determined using bar
   plot analysis and heatmap visualization. Taxonomic differences at the
   genus level between groups were identified using linear discriminant
   analysis (LDA) effect size (LEfSe) (LDA score >2, P < 0.05). Wilcoxon
   rank-sum tests and Kruskal-Wallis tests were used to respectively
   determine statistical differences in gut microbiota composition between
   two and three groups. Microbial functions were predicted using PICRUSt2
   [[68]27] in conjunction with KEGG pathways.

Metabolome

   Metabolites were extracted from serum samples. Briefly, each sample
   underwent metabolite precipitation with 400 μL of extraction solvent
   (methanol:water = 4:1). Solid samples were ground using a cryogenic
   tissue grinder for 6 min (−10 °C), followed by a 30 min freeze (5 °C)
   and a 30 min incubation at −20 °C. The supernatant was collected after
   centrifugation for 15 min (4 °C, 13,000 g) and subjected to LC-MS/MS
   analysis for metabolite identification. Chromatographic separation was
   performed using an ACQUITY HSS T3 column (100 mm * 2.1 mm, 1.8 μm,
   Waters Corporation, USA). Metabolites eluted from the column were
   detected using high-resolution tandem mass spectrometry on a
   UHPLC-Q-Exactive HF-X system (Thermo, USA) operating in both positive
   and negative ion modes. To assess the stability of LC-MS/MS throughout
   the collection process, a quality control sample (pool of all samples)
   was collected every 5–15 samples. Raw LC-MS/MS data were processed
   using Progenesis QI software (Waters Corporation, USA) to generate
   sample information, metabolite names, and mass spectrometry response
   intensity data. Metabolite annotation was conducted using the KEGG
   database (Kyoto Encyclopedia of Genes and Genomes,
   [69]https://www.kegg.jp/) and HMDB database ([70]http://www.hmdb.ca/).
   PCoA and Orthogonal Partial Least Squares Discriminant Analysis were
   performed. Significant differential metabolites were selected based on
   Variable Importance in Projection (VIP) values from the Orthogonal
   Partial Least Squares Discriminant Analysis model and student’s t-test
   (VIP > 1 and P < 0.05). Additionally, pathway enrichment analysis of
   differential metabolites was conducted using the KEGG database.

Proteome

   Rat bilateral hippocampal tissues were extracted, rapidly frozen in
   liquid nitrogen, and stored at −80 °C. For protein extraction,
   hippocampal tissues were homogenized in protein lysis buffer, incubated
   on ice for 30 min with vortex mixing (5–10 s every 5 min), followed by
   centrifugation at 12,000 g for 30 min at 4 °C to collect the
   supernatant. Protein concentration in the supernatant was determined
   using the BCA assay. Enzymatic digestion and peptide quantification of
   proteins were performed, and liquid chromatography separation was
   carried out on an Ionopticks UPLC C18 Column (1.6 µm, 250 mm × 75 μm,
   Ionopticks, USA). The resulting fractions were analyzed using the
   timsTOF Pro2 mass spectrometer (Bruker, Germany) for LC-MS/MS analysis
   of protein samples. Group differences in protein expression were
   calculated using the t.test function in R, with significance set at
   P < 0.05 and fold change criteria of >1.2 or <0.83 for differential
   expression proteins. Gene Ontology (GO) database
   ([71]http://geneontology.org/) was used for annotation analysis of the
   differentially expressed protein (DEPs) in terms of biological
   processes, cellular components, and molecular functions. Differential
   pathway analysis of DEPs was conducted using the KEGG pathway database.
   The rich factor was calculated as the ratio of the number of
   differential proteins in a specific pathway to the total number of
   proteins in that pathway.

Statistical analysis

   Statistical analyses were performed using GraphPad Prism version 9.0
   (GraphPad Software Inc, La Jolla, CA, USA) and SPSS version 23.0 (IBM
   Inc, Armonk, NY, USA). Data are presented as mean ± SEM. For
   independent sample experiments, we used the Shapiro-Wilk test for
   normality and Levene’s test for homogeneity of variances. For normally
   distributed and equal variance data, differences between two groups
   were assessed with a two-tailed Student’s t-test, and differences among
   three groups were assessed with one-way ANOVA followed by Bonferroni
   post-hoc tests. For non-normally distributed data, we used the Wilcoxon
   rank-sum test for two-group comparisons and the Kruskal-Wallis H test
   with Dunn’s post-hoc for three-group comparisons. For analysis of body
   weight gain and immobility time in the forced swim test, we applied
   two-way repeated measures ANOVA to examine group effects (CON, WKY,
   ESC), time effects, and group-time interactions, using Bonferroni
   post-hoc for multiple comparisons. A P-value of <0.05 was considered
   statistically significant. Notably, for comparisons between WKY and ESC
   groups in microbial and protein data, we did not apply additional
   P-value corrections, intending to identify differences of potential
   biological significance (Supplementary Table [72]1). Spearman
   correlation analysis was employed to explore co-occurrence
   relationships among gut microbiota, metabolites, and proteins, and
   visualized using Cytoscape (version 3.7.2), displaying only edges with
   correlations >0.5 and P < 0.05.

Results

ESC alleviates depressive-like behaviors and gut-brain damage in WKY rats

   To assess the antidepressant effects of ESC, we utilized the FST to
   evaluate depressive-like behaviors in rats. In the FST, using a two-way
   repeated measures ANOVA, we found a significant interaction between
   group and time (F(2,14) = 11.310, P = 0.001). At baseline, there was a
   significant difference in immobility time among the three groups (F(2,
   14) = 16.227, P < 0.001). After the intervention, the difference in
   immobility time between the CON and the ESC group was not statistically
   significant (P = 0.272). Only the ESC group showed a significant
   reduction in immobility time before and after treatment (F(1,
   7) = 36.902, P = 0.001). No significant changes were observed in the
   CON and WKY groups before and after treatment. The above results
   indicate that ESC treatment can alleviate depression-like behaviors,
   particularly behaviors associated with despair (Fig. [73]1B). We also
   employed a two-way repeated measures ANOVA to assess changes in body
   weight. The results indicated that the interaction between treatment
   and time was not statistically significant (F(4, 28) = 0.893,
   P = 0.481). However, significant differences in body weight between
   treatment groups were observed on day 7 (F(2, 14) = 27.449, P < 0.001),
   with significant differences between the CON group and the WKY group
   (P = 0.012), and between the CON group and the ESC group (P < 0.001),
   while the difference between the WKY and ESC groups was not significant
   (P = 0.099). Similar significant differences between groups were
   observed on days 14 and 21. In terms of the effect of time, all groups
   exhibited significant changes in weight at each of the three time
   points (F values of 320.284, 139.878, and 265.814, all P < 0.001), and
   there were significant differences in the rate of weight gain over
   time. Body weight measurements revealed that both the WKY and ESC
   groups exhibited slower weight gain compared to the control group (Fig.
   [74]1C). To further evaluate the impact of ESC on the MGB axis, we
   conducted histopathological examinations of brain and colon tissues in
   rats. HE staining of brain tissues showed that the control group had
   abundant, regularly shaped, and densely packed cortical and hippocampal
   neurons with no apparent abnormalities. In the WKY group, more neurons
   exhibited shrunken and deeply stained appearances in the cortex, with a
   few similarly affected neurons in the hippocampus. In the ESC group,
   fewer cortical neurons were shrunken and deeply stained, and
   hippocampal structure appeared normal (Fig. [75]1D). ESC also had a
   mitigating effect on intestinal barrier integrity; claudin-1 and ZO-1
   protein expressions were significantly reduced in the WKY rats compared
   to the control group (P < 0.05), but markedly improved after ESC
   treatment (Fig. [76]1E).

Intestinal homeostasis in WKY rats was affected by ESC

   We characterized changes in the gut microbiota induced by ESC treatment
   through 16S rRNA sequencing of rat colonic contents. The control group
   exhibited 1928 ASVs, whereas the WKY group showed a reduction in ASVs
   (1412). Additionally, the ESC group displayed increased ASV counts
   (1709), though still lower than the CON group (Fig. [77]2A). Alpha
   diversity indices Ace, Chao, Shannon, and Simpson were used to evaluate
   microbial diversity. Results indicated that the WKY group had lower Ace
   and Chao indices compared to CON, and higher Simpson index, suggesting
   reduced richness and diversity of microbial communities in rats with
   depression-like behavior. ESC intervention appeared to influence this
   trend, albeit not statistically significant (Fig. [78]S1A). PCoA
   analysis assessed differences in intestinal microbiota composition
   among groups, our results indicate significant differences in microbial
   β-diversity among the three groups (Fig. [79]2B). Specifically, the
   PCoA analysis based on the weighted UniFrac distance revealed
   significant differences along the PC2 axis (explaining 18.62% of the
   variation) between the CON and WKY groups, as well as between the ESC
   and WKY groups (Supplementary Table [80]1). However, no significant
   differences between the ESC and WKY groups were observed along the PC1
   axis (explaining 35.28% of the variation). Moreover, the PCoA analysis
   based on the Bray-Curtis distance showed that the CON group could be
   significantly separated from both the ESC and WKY groups, while no
   significant differences were observed between the ESC and WKY groups
   (Fig. [81]S1B). Thus, our findings suggest that ESC may partially alter
   the intestinal microbial diversity in the WKY rats. Subsequently, we
   analyzed GMHI and MDI. Both WKY and ESC groups showed significant
   decreases in GMHI compared to CON (P < 0.05, Fig. [82]2C, Fig.
   [83]S1C). Moreover, MDI was higher in both WKY and ESC groups than in
   CON (Fig. [84]S1D, P < 0.05), with ESC showing a slight decrease
   compared to WKY (Fig. [85]2C). We further evaluated microbial
   composition at different taxonomic levels. Based on the analysis of
   compositional ratios, at the phylum level, Firmicutes predominated in
   all groups, followed by Bacteroidota. Firmicutes were most abundant in
   CON (92.99%), followed by WKY (88.54%) and ESC (85.83%). Bacteroidota
   was highest in ESC (11.41%), followed by WKY (9.22%) and CON (3.83%)
   (Fig. [86]2D). At the genus level, Turicibacter and Blautia are found
   in higher proportions in the WKY and ESC groups compared to the CON
   group. Additionally, the proportion of Lactobacillus is highest in the
   CON group, followed by the ESC group, and lowest in the WKY group (Fig.
   [87]2D). Using LEfSe analysis, we identified consistent microbial
   changes between CON vs. WKY and CON vs. ESC, with 40 and 41
   differential microbes identified, respectively (Fig. [88]S2A). Analysis
   of these differential microbes revealed 31 common differential species.
   A heatmap depicted the expression abundance and LDA values of these 31
   differential microbes across the three groups (Fig. [89]S2B).
   Representatively, Bacillus and Lachnospiraceae_NK4A136 were predominant
   in CON (LDA > 4), while Turicibacter and Blautia were predominant in
   WKY and ESC groups (LDA > 4, Fig. [90]S2C). Furthermore, we conducted
   LEfSe analysis between WKY vs. ESC to identify important ESC-influenced
   microbes (Fig. [91]2E). Results revealed differential abundance at
   various taxonomic levels, including 1 at the phylum level
   (Patescibacteria), 1 at the class level (Saccharimonadia), 4 at the
   order level (Staphylococcales, Saccharimonadales, Corynebacteriales and
   Bifidobacteriales), 6 at the family level (Staphylococcaceae,
   Saccharimonadaceae, Rikenellaceae, Marinifilaceae, Corynebacteriaceae
   and Bifidobacteriaceae), 11 at the genus level
   (unclassified_f__Anaerovoracaceae, UCG-005, Staphylococcus,
   Odoribacter, norank_f__Christensenellaceae, Intestinimonas,
   Corynebacterium, CHKCI002, Candidatus_Saccharimonas, Bifidobacterium
   and Alistipes), and 17 at the species level (Only Bacteroides barnesiae
   has complete annotation information). Notably, Bacteroides barnesiae at
   the species level showed significantly higher expression in ESC than in
   WKY (Fig. [92]2F). The PICRUSt2-based functional predictions provide
   insights into the microbial community’s functional capabilities by
   assessing the composition and abundance of functional attributes. The
   KEGG pathways predicted by PICRUSt2 indicated significant differences
   between the WKY and ESC groups in the level 2 pathways, such as
   biosynthesis of other secondary metabolites and glycan biosynthesis and
   metabolism. Specifically, at the level 3 pathway within lipid
   metabolism, fatty acid degradation and glycerolipid metabolism were
   significantly enriched in the WKY group, whereas sphingolipid
   metabolism was significantly enriched in the ESC group (Fig. [93]2G,
   H).

Fig. 2. Characteristics of microbial communities in colon contents.

   [94]Fig. 2
   [95]Open in a new tab

   A Veen diagram of ASVs in three groups. B PCoA of the three groups. C
   Analysis of differences in MDI (left) and GMHI (right) between WKY and
   ESC groups. D Bacterial taxonomic analysis at the phylum level across
   the three groups (left) and bacterial taxonomic analysis at the genus
   level across the three groups (right). E Differential microbiota
   between the WKY and ESC groups based on LEfSe analysis (LDA > 2.0,
   P < 0.05). LDA effect size showing the most differentially significant
   abundant microbiota between WKY and ESC groups. F Bar graph showing the
   abundance of differential microbiota at the genus and species levels. G
   Microbiota function predicted based on the KEGG database (at level II)
   by using PICRUSt2. H Microbiota function predicted based on the KEGG
   database (at level III) by using PICRUSt2. ^*P < 0.05, ^**P < 0.01,
   ^***P < 0.001.

ESC affects the metabolic profile in rats with depression-like behavior

   To explore the potential association between ESC and its effects on
   serum metabolites relevant to communication within the MGB axis, we
   analyzed changes in serum metabolites. Initially, PCoA was used,
   revealing significant separation of metabolic profiles between the CON
   and WKY/ESC groups, while no significant separation was observed
   between the ESC and WKY groups (Fig. [96]3A). Subsequently,
   differential metabolites were screened based on a standard of P < 0.05
   and VIP values > 1. Compared with the CON group, 104 metabolites were
   significantly upregulated and 89 metabolites were significantly
   downregulated in the WKY group (Fig. [97]3B). Similarly, 125
   metabolites were significantly upregulated and 64 metabolites were
   significantly downregulated in the ESC group compared with the CON
   group (Fig. [98]3C). Additionally, there were 131 differential
   metabolites between the ESC and WKY group (Fig. [99]3D). Furthermore,
   compared to the CON group, there are 58 commonly upregulated and 105
   commonly downregulated differential metabolites in both the WKY and ESC
   groups, exhibiting the same trend of changes in these groups (Fig.
   [100]3E). KEGG pathway annotations of the metabolites revealed that the
   majority belonged to Phospholipids (Fig. [101]3F). Further analysis of
   131 metabolites affected by ESC intervention was conducted, there were
   83 significantly upregulated metabolites and 48 significantly
   downregulated metabolites in the ESC group compared with the WKY group
   (Fig. [102]3G). The top 5 differential metabolites based on fold change
   in the ESC group were Allylestrenol, Mipafox, P-Tolyl Sulfate,
   2-Hydroxypyridine and P-Cresol glucuronide. The top 5 differential
   metabolites based on fold change in the WKY group were
   CDP-DG(18:2(10E,12Z)+=O(9)/20:3(8Z,11Z,14Z)), CDP-DG(18:1(9Z)/PGJ2),
   PS(15:0/20:3(8Z,11Z,14Z)), PGP(20:4(6Z,8E,10E,14Z)-2OH(5S,12 R)/i-19:0)
   and Leukotriene F4. Metabolite enrichment analysis was performed on
   differential metabolites between the WKY and ESC groups. Results showed
   that the top two pathways by P value were Glycerophospholipid
   metabolism and Sphingolipid metabolism (Fig. [103]3H). Within the
   Glycerophospholipid metabolism pathway, 15 metabolites were enriched;
   compared to the WKY group, glycerylphosphorylcholine, LysoPC
   (18:0/0:0), PC (18:0/22:6 (4Z,7Z,10Z,13Z,16Z,19Z)), and PS (20:4
   (8Z,11Z,14Z,17Z)/22:2 (13Z,16Z)) were expressed at higher levels in the
   ESC group, while the remaining 11 differential metabolites were
   downregulated. In the Sphingolipid metabolism pathway, three
   metabolites, Sphinganine 1-phosphate, Sphingosine, and
   Sphingosine-1-phosphate, were highly expressed in the ESC group (Fig.
   [104]3I). These data indicate that ESC intervention promotes changes in
   serum metabolites in rats with depression-like behavior.

Fig. 3. Serum Metabolomic Features.

   [105]Fig. 3
   [106]Open in a new tab

   A Principal Coordinate Analysis (PCoA) plots for CON, WKY, and ESC
   groups. B Volcano plot for CON vs. WKY (red: increased in WKY group;
   blue: increased in CON group). C Volcano plot for CON vs. ESC (red:
   increased in ESC group; blue: increased in CON group). D Volcano plot
   for WKY vs. ESC (red: increased in ESC group; blue: increased in WKY
   group). E Venn diagram showing metabolites with the same trends between
   CON vs. WKY and CON vs. ESC. F KEGG pathway annotations for 142 common
   differential metabolites. G Bar chart displaying the Log2FC of
   differential metabolites (red: increased in ESC group; blue: increased
   in WKY group); and heatmap of differential metabolites between WKY and
   ESC groups. High expression is marked in red, and low expression is
   marked in blue. The “value” represents the relative expression value
   (Z-score) obtained after normalization. H KEGG topology analysis of
   differential metabolites between WKY and ESC groups. The impact value
   represents the importance of metabolites in the metabolic pathway’s
   topology, while the P-value reflects the statistical significance of
   the metabolic pathway. The horizontal coordinate and size of the bubble
   represent the impact factor of the pathway in the topology analysis,
   with larger bubbles indicating a larger impact factor. The vertical
   coordinate and color of the bubble represent the P-value of the
   enrichment analysis, with darker colors indicating lower P-values and
   greater enrichment significance. I Heatmap of differential metabolites
   in Glycerophospholipid metabolism and Sphingolipid metabolism, with
   high expression shown in red and low expression shown in blue. The
   “value” represents the relative expression value (Z-score) obtained
   after normalization. Higher positive values indicate higher expression,
   while larger negative values indicate lower expression.

ESC influences rat proteome in hippocampus

   To investigate changes in the hippocampal proteome of WKY rats
   following ESC intervention, we conducted a proteomic analysis.
   Principal component analysis showed no significant separation of
   protein profiles among the three groups (Fig. [107]4A). However,
   differential proteins between groups were identified based on a fold
   change criterion of >1.2 or <0.83 and a P-value < 0.05 for further
   analysis. The results indicated that there were 127 upregulated DEPs
   and 231 downregulated DEPs in the WKY group (Fig. [108]4B), and 94
   upregulated DEPs and 244 downregulated DEPs in the ESC group compared
   with the CON group (Fig. [109]4C). Meanwhile, 103 DEPs were upregulated
   and 160 DEPs were downregulated in the ESC group compared with the WKY
   group (Fig. [110]4D). Differential proteins that were upregulated or
   downregulated in both WKY and ESC groups, as compared to the control
   group, were considered potentially due to genetic factors, as they were
   unaffected by the treatment (Fig. [111]4E). GO annotation analysis of
   these proteins identified 44, 78, and 100 entries in molecular function
   (MF), cellular component (CC), and biological process (BP),
   respectively. Fig. [112]4F depicts the top five enriched terms. MF
   analysis showed that most identified proteins are involved in binding
   to small molecules, proteins, and compounds. CC category analysis
   indicated that most proteins are predicted to be localized in
   organelles. Many proteins in the BP category are related to metabolic
   processes. KEGG pathway enrichment analysis identified 26 significantly
   enriched signaling pathways (P value < 0.05), with the alpha-Linolenic
   acid metabolism pathway showing the highest Rich factor (Rich
   factor = 0.375, P value = 0.0017), enriched with one upregulated
   protein Acc1a and two downregulated proteins Fads1 and Cyb5a (Fig.
   [113]4G). Further analysis was conducted on differential proteins
   between the ESC and WKY groups. GO annotation results revealed that
   most identified proteins in the MF analysis are involved in binding to
   proteins and organic cyclic compounds, while CC analysis showed most
   proteins are predicted to be located in organelles and cell membranes.
   BP analysis indicated that most proteins are related to metabolic
   processes (Fig. [114]4H). Additionally, KEGG enrichment analysis of
   differential proteins, in conjunction with the above metabolomics
   results, focused on proteins enriched in the Glycerophospholipid
   metabolism and Sphingolipid metabolism pathways. Two proteins in the
   Glycerophospholipid metabolism pathway, Lysophosphatidylcholine
   acyltransferase 1 (Lpcat1) and Pnpla7, and four proteins in the
   Sphingolipid metabolism pathway, Bin2a, Ectonucleotide
   pyrophosphatase/phosphodiesterase 5 (Enpp5), Serine
   palmitoyltransferase 1 (Sptlc1), and Ectonucleotide
   pyrophosphatase/phosphodiesterase 2 (Enpp2), were analyzed. Compared to
   the WKY group, Lpcat1, Enpp5, Sptlc1, and Enpp2 were significantly
   upregulated (P value < 0.05) in the ESC group, while Pnpla7 and Bin2a
   were significantly downregulated (P value < 0.05, Fig. [115]4I, J).

Fig. 4. Proteomic Characteristics of the Hippocampus.

   [116]Fig. 4
   [117]Open in a new tab

   A Principal component analysis of hippocampal proteomics for CON, WKY,
   and ESC groups. B Volcano plot for CON vs. WKY (red: increased in WKY
   group; blue: increased in CON group). C Volcano plot for CON vs. ESC
   (red: increased in ESC group; blue: increased in CON group). D Volcano
   plot for WKY vs. ESC (red: increased in ESC group; blue: increased in
   WKY group). E Venn diagram showing proteins with the same expression
   trends between CON vs. WKY and CON vs. ESC. F GO annotation analysis
   for 163 differentially expressed proteins. G KEGG enrichment analysis
   for 163 differentially expressed proteins. The bubble color indicates
   the p-value and the size indicates the number of proteins. The
   horizontal axis indicates the Rich factor, which means the ration of
   the number of DEPs that enriched in one pathway and the number of all
   annotated proteins under this pathway term. H GO annotation analysis of
   differential proteins between WKY and ESC. I Bar graph of protein
   expression enriched in the Glycerophospholipid metabolism pathway. J
   Bar graph of protein expression enriched in the Sphingolipid metabolism
   pathway. ^*P < 0.05, ^**P < 0.01.

Co-occurrence analysis of intestinal bacteria, serum metabolites and
hippocampal proteins

   To investigate the complex interactions among the gut microbiome,
   metabolites, and proteins, spearman correlation analysis was performed.
   The analysis assessed the correlations among differential microbes,
   metabolites, proteins and the immobility time of FST between the WKY
   and ESC groups, including 9 differential microbes, 18 differential
   metabolites, and 6 differential proteins. Correlation pairs with a
   coefficient greater than 0.5 or less than −0.5 and a P value below 0.05
   were selected for presentation, revealing a network comprising eight
   interconnected clusters, including Bacteroidota, Actinobacteriota,
   Patescibacteria, Firmicutes, as well as metabolites and proteins
   enriched in the Glycerophospholipid metabolism and Sphingolipid
   metabolism pathways (Fig. [118]5A). Several key factors emerged within
   these clusters. For instance, Bacteroides barnesiae had the most
   connections (degree = 15), followed by Enpp2, Pnpla7, and Bin2a
   (degree = 10), Lpcat1 and PE-NMe2(15:0/20:2(11Z,14Z)) (degree = 8),
   Staphylococcus (degree = 7), and Sptlc1, Sphinganine 1-phosphate and
   PS(15:0/20:3(8Z,11Z,14Z)) (degree = 6) were involved (Supplementary
   Table [119]2). Bacteroides barnesiae showed a positive correlation with
   metabolites enriched in the Sphingolipid metabolism pathway
   (sphingosine-1-phosphate, sphingosine, and sphinganine 1-phosphate).
   Moreover, it negatively correlated with the protein Bin2a (r < −0.9,
   P < 0.05) in the Sphingolipid metabolism pathway, while positively
   correlating with Sptlc1 and Enpp2 (r > 0.8, P < 0.05). Enpp2 was
   positively correlated with sphingosine-1-phosphate, sphingosine, and
   sphinganine 1-phosphate (r > 0.8, P < 0.05), Enpp5 was positively
   correlated with sphingosine-1-phosphate, sphinganine 1-phosphate,
   LysoPC(18:0/0:0), PC(18:0/22:6(4Z,7Z,10Z,13Z,16Z,19Z)),
   PE-NMe2(15:0/18:1(11Z)) (r > 0.8, P < 0.05), whereas Bin2a showed
   negative correlations with these metabolites (r < −0.9, P < 0.05).
   Sptlc1 was positively correlated with sphingosine (r > 0.8, P < 0.05).
   Additionally, the immobility time of FST was positively correlated with
   1-Palmitoylphosphatidylcholine, PC(16:0/18:3(9Z,12Z,15Z)),
   PE(15:0/22:1(13Z)), PE(22:6(4Z,7Z,10Z,13Z,16Z,19Z)/22:1(13Z)),
   PE-NMe2(15:0/20:2(11Z,14Z)), PE-NMe2(18:0/18:2(9Z,12Z)),
   PE-NMe2(18:1(9Z)/16:0), PE-NMe2(18:2(9Z,12Z)/16:0) and
   PS(15:0/20:3(8Z,11Z,14Z)), while it was negatively correlated with
   Candidatus_Saccharimonas, Bacteroides_barnesiae, UCG-005, Sphinganine
   1-phosphate, Sphingosine-1-phosphate, Glycerylphosphorylcholine and
   LysoPC(18:0/0:0), PC(18:0/22:6(4Z,7Z,10Z,13Z,16Z,19Z)) (Supplementary
   Table [120]3). KEGG pathway diagrams indicated that Glycerophospholipid
   metabolism and Sphingolipid metabolism pathways were interconnected
   through the same metabolite (Phosphoethanolamine). Notably, Sptlc1
   could positively promote the feedback generation of sphinganine
   1-phosphate and sphingosine, and Enpp2/5 were involved in the catalysis
   of sphingosine, with sphingosine-1-phosphate being a downstream
   metabolite of sphingosine (Fig. [121]5B). These findings demonstrate
   that the gut microbe Bacteroides barnesiae, along with the phospholipid
   metabolism pathway and associated proteins Sptlc1 and Enpp2/5, may play
   a crucial role in the antidepressant efficacy of ESC.

Fig. 5. Correlation analysis of differential microbiota, metabolites,
proteins and FST.

   [122]Fig. 5
   [123]Open in a new tab

   A Co-occurrence network analysis. Purple nodes represent microbiota,
   yellow nodes represent metabolites, blue nodes represent FST and green
   nodes represent proteins; the size of the nodes indicates the degree,
   or the number of connections. Lines between nodes represent Spearman
   negative (gray) or positive (red) correlations. B Map of key components
   in glycerophospholipid and sphingolipid metabolism pathways. Red
   indicates high expression in the ESC group, while blue indicates low
   expression in the ESC group.

Discussion

   The interaction between the gut microbiome and the brain has
   increasingly become a focus of neuroscience research. It is well
   understood that the stability of the gut microbiota is important for
   maintaining host health, and specific changes in the gut microbiota,
   such as variations in alpha and beta diversity and alterations in the
   abundance of bacterial families like Christensenellaceae,
   Lachnospiraceae, and Ruminococcaceae, are associated with depression
   [[124]28]. Under stress and depression, the gut microbiota becomes
   disordered, adversely affecting the CNS and exacerbating the disease
   [[125]29]. Studies have confirmed that transplanting the fecal
   microbiota of patients with severe depression into germ-free mice
   induces depressive symptoms in the mice, whereas restoring the gut
   microbiota can alleviate depression [[126]30, [127]31]. In our study,
   we analyzed the changes in the gut microbiota response to ESC treatment
   in WKY rats and the associations between alterations in the gut
   microbiota and changes in the host metabolic profile and hippocampal
   proteome, to explore the potential mechanisms by which ESC exerts
   therapeutic effects on depressive-like phenotypes through the MGB axis.

   In this study, we employed a WKY rat depression model. The FST, a
   behavioral test widely used in preclinical studies, assesses
   depressive-like behaviors in animal models [[128]32]. ESC has been
   shown to ameliorate depressive-like behaviors in our study, as
   evidenced by the reduced immobility time in WKY rats during FST.
   Although the FST is a reliable measure of depressive-like behavior, the
   lack of general locomotor activity assessment may limit the
   interpretation of the results. Therefore, when testing antidepressants
   in WKY rats, incorporating behavioral paradigms not reliant on
   locomotor activity in addition to the FST may provide a more
   comprehensive evaluation of the depressive phenotype in WKY rats.
   Pathological findings suggest that ESC intervention may alleviate brain
   damage and reduce intestinal permeability in WKY rats, potentially
   decreasing the translocation of microbial metabolites or microbiota and
   the invasion of harmful substances into the CNS. The expression of
   tight junction proteins is a critical marker for assessing intestinal
   health and disease recovery. Occludin and claudins, the major
   components of intestinal tight junctions, regulate the barrier
   properties of the paracellular space by forming tight junction strands.
   ZO-1 plays a key role in assembling intestinal tight junctions by
   linking transmembrane proteins to the intracellular actin cytoskeleton
   [[129]33]. Our study found that ESC intervention alleviated the
   underexpression of these three tight junction proteins, indicating that
   ESC can influence intestinal barrier function. ESC is a first-line
   antidepressant and anxiolytic medication [[130]34]. Most research on
   ESC’s antidepressant effects has focused on the serotonin transporter,
   which is responsible for the uptake of serotonin into serotonergic
   neurons; ESC inhibits this uptake, subsequently increasing serotonin
   levels at the postsynaptic membrane in the serotonergic synapse
   [[131]35, [132]36]. Interestingly, the therapeutic effects of ESC
   cannot be solely attributed to the inhibition of the serotonin
   transporter and involve multiple mechanisms of action. Studies have
   reported that alterations in the gut microbiome, characterized by
   changes in microbial diversity and imbalances in key microbial taxa,
   are observed in patients with depression [[133]30]. This suggests that
   investigating how antidepressants interact with these microbiome
   changes could be a promising direction for future research. To date,
   few studies have focused on the relationship between ESC’s therapeutic
   effects and the gut microbiome, and currently, there is no research on
   changes in the gut microbiota following ESC treatment in WKY rats.
   Therefore, our study aims to provide evidence for the antidepressant
   mechanisms of ESC through the perspective of the MGB axis.

   Through analysis of 16S rRNA sequencing data, we observed that the gut
   microbiota diversity in WKY rats was lower compared to the CON group,
   but there was an increase in both richness and diversity of the gut
   microbiota in the depressive group following ESC treatment. These
   results indicate a potential link between ESC treatment and changes in
   gut microbiota composition in WKY rats. The gut microbiota primarily
   consisted of the phyla Firmicutes and Bacteroidota, which have been
   reported to be the most affected in depression [[134]37, [135]38].
   Additionally, our study found that the abundance of Bacteroidota was
   higher in the ESC group compared to the other two groups, while
   Firmicutes was less abundant. Antidepressant medications can influence
   the composition of the microbiome; higher abundances of Bacteroidetes
   and Proteobacteria and lower abundances of Firmicutes, Actinobacteria,
   and Fusobacteria have been observed during treatment with selective
   serotonin reuptake inhibitors or serotonin-norepinephrine reuptake
   inhibitors [[136]11]. The analysis of gut microbiota differences showed
   that 16 genera were significantly enriched in the WKY and ESC groups
   compared to the control group, and ESC intervention did not reverse the
   abundance of these bacteria, which may reflect potential genetic
   factors. Due to these genetic background differences, WKY rats
   inherently exhibit physiological and behavioral characteristics that
   differ fundamentally from those of the CON group, potentially directly
   affecting their microbiota structure. Future studies should consider
   using control and experimental groups with the same genetic backgrounds
   combined with depression modeling to validate the consistency and
   reproducibility of results. Our results identified 11 distinct microbes
   at the genus level and 17 at the species level between the WKY and ESC
   groups, of which 8 genera and 1 species had clear annotation
   information. Firmicutes included Intestinimonas, Staphylococcus, and
   UCG-005, while Bacteroidota included Alistipes, Odoribacter, and
   Bacteroides barnesiae. It was found that Intestinimonas, Alistipes, and
   Odoribacter, which produce SCFAs, correlate negatively with the
   severity of depression and increase in abundance following
   antidepressant treatment [[137]11, [138]39]. Consistently, in our
   study, the abundance of these three bacteria increased following ESC
   intervention. Furthermore, in the ESC group, the abundance of
   Staphylococcus and UCG-005 decreased. Staphylococcus can produce toxins
   and staphylococcal enterotoxins, which stimulate the vagus nerve to
   send signals to the brain, potentially inducing disease behavior
   [[139]40]. In this study, UCG-005 belongs to the Oscillospiraceae
   family, which has been reported to be a key pathogenic gut bacterium
   associated with the development of colitis [[140]41]. It is noteworthy
   that Bacteroides_barnesiae is the only bacterium at the species level
   that has been clearly annotated. It belongs to the Bacteroides, which
   is one of the most abundant Gram-negative genera in the human gut.
   Bacteroides acts as a commensal bacterium that can establish a stable
   relationship with the host and play a potentially probiotic role. It is
   considered to be a major synthesiser of vitamin K and also maintains
   host intestinal homeostasis by regulating the levels of SCFAs
   [[141]42]. Interestingly, we observed an increase in the abundance of
   Bacteroides barnesiae in the ESC group compared to the WKY group.
   Moreover, correlational analysis revealed that the abundance of
   Bacteroides barnesiae is negatively associated with immobility time in
   the FST. These results suggest a potential association that might
   contribute to understanding the complex interactions within the
   microbiota-gut-brain axis in depressive-like behaviors.

   Our study utilized a 3-week escitalopram regimen, based on prior rodent
   and human studies [[142]43–[143]48], which typically describe
   escitalopram treatment as chronic, ranging from 3 to 12 weeks [[144]13,
   [145]15, [146]49–[147]52]. This duration aligns with research
   indicating that chronic antidepressants like escitalopram and
   vortioxetine can reshape the gut microbiome in depression, impacting
   diversity and altering key microbial phyla [[148]15, [149]50, [150]52,
   [151]53]. While acute administration of escitalopram is less studied,
   evidence from other antidepressants suggests it also affects the gut
   microbiota. For instance, (S)-norketamine decreased harmful bacteria
   like Escherichia-Shigella and increased beneficial ones like
   Harryflintia 24 h after injection [[152]54]. Therefore, our findings
   highlight the influence of chronic antidepressant treatment on the gut
   microbiome, suggesting the need to also explore the impacts of acute
   dosing in future studies.

   The gut microbiome interacts with the CNS through microbial-derived
   metabolites and neurotransmitters such as gamma-aminobutyric acid and
   serotonin [[153]55]. In general, alterations in the composition of the
   gut microbiota may result in alterations in microbial metabolites,
   which may subsequently influence the entire metabolome of the host. In
   our research, we analyzed and compared the serum metabolomic profiles
   across groups to explore the mechanisms of communication along the MGB
   axis. Initial annotation analysis of differential metabolites between
   WKY and ESC group revealed that lipid metabolism abnormalities are one
   of the main manifestations. The CNS comprises three major classes of
   lipids: phospholipids, sphingolipids, and cholesterol. Lipids account
   for over 50% of the dry weight of the brain and their composition can
   influence perception and emotional behavior, potentially leading to
   conditions such as depression and anxiety [[154]56, [155]57].
   Specifically, phospholipids and sphingolipids are closely linked to the
   development of neuropsychiatric disorders such as depression and
   anxiety [[156]58]. Plasma metabolites involved in lipid metabolism are
   key intermediates in the MGB axis [[157]59]. Studies have shown that
   traditional herbal medicines like kudzu can reverse depressive-like
   behaviors by regulating disturbances in phospholipid metabolism
   [[158]60]. Through the enrichment analysis of serum differential
   metabolites with or without ESC intervention, we found significant
   changes in glycerophospholipid and sphingolipid metabolism in the ESC
   group. The focus was particularly on the sphingolipid metabolism
   pathway, with sphingosine identified as a key differential metabolite.
   Research has linked elevated levels of sphingosine with reductions in
   depressive-like behavior and improvements in cognitive function
   [[159]61]. Sphingosine, a crucial intermediate in the sphingolipid
   metabolic pathway, can be phosphorylated to form
   sphingosine-1-phosphate, which is then cleaved into phosphoethanolamine
   [[160]62]. Many studies report that sphingosine-1-phosphate is involved
   in various neurological and psychiatric disorders, including
   Alzheimer’s disease [[161]63], depression, and anxiety [[162]64].
   Sphingosine-1-phosphate has been identified as a pharmacological target
   in mental illness research, exemplified by FTY720 (fingolimod), an S1P
   receptor agonist that modulates neuroinflammation and is approved for
   treating multiple sclerosis [[163]65]. In our study, WKY rats exhibited
   extremely low levels of sphingosine, indicating an abnormality in
   sphingolipid metabolism. In the ESC group, we observed an increase in
   sphingosine levels, which may relate to the observed changes in
   depressive-like behaviors. Additionally, glycerophospholipid metabolism
   was another significant pathway enriched in differential metabolites
   with or without ESC intervention. Glycerophospholipids, critical
   components of neuronal membranes and brain phospholipids, play an
   essential role in regulating synaptic function and can be subdivided
   into phosphatidylserine (PS), phosphatidylethanolamine (PE), and
   phosphatidylcholine (PC). Our research observed changes in levels of
   PS, PE, and PC in the serum associated with ESC treatment, which may
   reflect alterations in glycerophospholipid metabolic pathways in the
   periphery. Although this study primarily focuses on the mechanisms by
   which ESC modulates sphingolipid metabolism through reshaping the gut
   microbiome, it is noteworthy that ESC might also affect sphingolipid
   metabolism through other pathways. Firstly, sphingolipid metabolism is
   closely related to immune responses and inflammatory states [[164]66],
   which are often exhibited in patients with depression [[165]67]. The
   anti-inflammatory properties of ESC may play a role by modulating the
   metabolism of sphingolipid signaling molecules associated with
   neuroinflammation [[166]68, [167]69]. Secondly, the role of
   antidepressants in reducing oxidative stress should not be overlooked
   [[168]70], as it may help maintain the lipid balance of neural cells,
   particularly the stability of membrane sphingolipids, which are
   essential for neuronal function and survival [[169]71]. Lastly, our
   study’s data also show that ESC has impact on the glycerophospholipid
   metabolism pathway (Fig. [170]3I), which may further affect the
   synthesis and degradation of sphingolipids, impacting neural cell
   function [[171]72]. The presence of these potential mechanisms suggests
   a multifaceted regulatory role of ESC on sphingolipid metabolism,
   providing new directions for future research.

   The hippocampus is a critical brain region for the generation and
   regulation of emotions and cognition, and depression is commonly
   characterized by memory loss and cognitive impairments, indicating
   hippocampal neuronal dysfunction as a pathophysiological feature of
   this disorder [[172]73, [173]74]. In our study, pathological analysis
   revealed hippocampal neuronal atrophy [[174]75]. Given that proteins
   are the primary executors of physiological functions in organisms, we
   conducted a proteomic analysis of the hippocampus to enhance our
   understanding of the antidepressant mechanisms of ESC. Based on our
   microbiota and metabolomics findings, we are particularly interested in
   proteins related to sphingolipid metabolism. The analysis revealed
   significant differences in sphingolipid metabolism-related proteins
   between the WKY and ESC groups. Specifically, Sptlc1, Enpp5, and Enpp2
   were significantly enriched in the ESC group, while Bin2a was
   significantly enriched in the WKY group. Notably, ENPP5 (Ectonucleotide
   Pyrophosphatase/ Phosphodiesterase 5) showed a significant negative
   correlation with immobility time in the FST. ENPP5 is an enzyme that
   plays a crucial role in extracellular signal transduction. Studies have
   confirmed that ENPP expression in immature oligodendrocytes can enhance
   myelin regulatory factors via the mTOR signaling pathway, providing
   oxidative damage protection to myelin membranes and thereby reducing
   the risk of psychiatric disorders such as depression [[175]76].
   Additionally, members of the ENPP family influence neuroprotection and
   damage responses by regulating autophagy and generating
   lysophosphatidic acid (LPA), which affects neuronal signaling in
   neurodegenerative diseases [[176]77]. This suggests that ENPP proteins
   play a critical role in maintaining neural homeostasis and mental
   health, highlighting the potential of this protein family as emerging
   targets for psychiatric disorders like depression.

   Finally, to further elucidate the potential role of the
   microbiota-gut-brain (MGB) axis in the antidepressant mechanism of
   escitalopram (ESC), we performed multi-omics correlation analysis. In
   our study, co-occurrence network analysis revealed a significant
   increase in the gut microbiota Bacteroides barnesiae in ESC group,
   which was strongly positively correlated with key sphingolipid
   metabolism-related metabolites (such as Sphingosine and
   Sphingosine-1-phosphate) in the serum, as well as with Sptlc1, a
   protein associated with sphingolipid metabolism in the hippocampus.
   Increased activity of Sptlc1 promotes the production of sphingosine and
   Sphinganine 1-phosphate, which essential for neuroprotection [[177]78].
   Correlation analysis revealed that UCG-005 was positively correlated
   with FST, while Candidatus_Saccharimonas and Bacteroides barnesiae were
   negatively correlated. Eighteen differential metabolites were
   identified, with positive correlations found with
   1-Palmitoylphosphatidylcholine, PC(16:0/18:3(9Z,12Z,15Z)),
   PE(15:0/22:1(13Z)), PE(22:6(4Z,7Z,10Z,13Z,16Z,19Z)/22:1(13Z)),
   PE-NMe2(15:0/20:2(11Z,14Z)), PE-NMe2(18:0/18:2(9Z,12Z)),
   PE-NMe2(18:1(9Z)/16:0), PE-NMe2(18:2(9Z,12Z)/16:0), and
   PS(15:0/20:3(8Z,11Z,14Z)); and negative correlations with Sphinganine
   1-phosphate, Sphingosine-1-phosphate, Glycerylphosphorylcholine,
   LysoPC(18:0/0:0), and PC(18:0/22:6(4Z,7Z,10Z,13Z,16Z,19Z)). Six
   differential proteins were identified, with Enpp5 showing a negative
   correlation with FST. The correlation results support the potential
   association of these biomarkers with the improvement of depressive
   symptoms. Previous studies have shown that Bacteroides can modify
   sphingolipids through its metabolic activities, influencing the
   biosynthesis and degradation pathways of sphingolipids [[178]79,
   [179]80]. Sphingolipid metabolites and related proteins play a crucial
   role in regulating neuronal function and maintaining neural health
   [[180]81]. These findings suggest that Bacteroides barnesiae may
   indirectly promote the antidepressant effects of ESC by modulating
   sphingolipid metabolism. Specifically, by increasing the levels of
   sphingosine and sphingosine-1-phosphate, Bacteroides barnesiae may
   enhance the function of Sptlc1 and Enpp5, key proteins in the
   sphingolipid metabolic pathway that are closely associated with
   neuroprotection and anti-inflammatory effects. In this way, Bacteroides
   barnesiae not only impacts the gut environment as a gut microbiota
   member but may also play an important role in the antidepressant
   effects of ESC by modulating key neuro-metabolic pathways.

   This study has several limitations: (1) The analysis of gut microbiota
   was conducted using 16S sequencing. Although this method is effective
   for taxonomic identification, it offers lower resolution and less
   comprehensive functional insights compared to metagenomic sequencing.
   Future studies employing metagenomics and other sequencing techniques
   could effectively address these limitations; (2) The conclusions of
   this study are largely based on correlational analyses, necessitating
   cautious interpretation. Future interventions using fecal microbiota
   transplantation (FMT) or targeted microbial interventions may serve as
   effective approaches to explore causal relationships; (3) We did not
   apply traditional multiple comparisons adjustments to the microbial and
   protein comparisons between WKY and ESC groups, aiming to uncover
   biologically significant differences. However, this may increase the
   risk of false positives. Future studies will involve more rigorous
   validation to confirm these findings; (4) In this study, the choice of
   intraperitoneal administration of escitalopram differs from the
   commonly used oral route in clinical settings. This route of
   administration may not fully replicate the effects of oral
   administration on the microbiota, and thus might limit the direct
   extrapolation of our results to clinical application. Future studies
   specifically testing oral administration of escitalopram are needed to
   more accurately assess its effects on the gut-brain axis; (5) This
   study included only male subjects, and did not investigate potential
   sex-related differences that might affect the results. Future research
   should include female subjects to ensure broader applicability and
   relevance of the findings.

Conclusion

   In summary, our study explored new mechanisms underlying the
   antidepressant efficacy of ESC. We utilized WKY rats as a model for
   depression and assessed the antidepressant effects of ESC using the
   forced swim test. Histological examination using Hematoxylin and Eosin
   (HE) staining and immunohistochemistry (IHC) assessed the impact of ESC
   on intestinal permeability and brain damage. Additionally, multi-omics
   analyses were conducted to elucidate the pivotal role of gut microbiota
   in this context. We observed that ESC significantly increased the
   abundance of certain bacteria, including Bacteroides barnesiae.
   Furthermore, in the ESC group compared to the WKY group, there were
   notable alterations in the serum metabolome and hippocampal proteome,
   which influenced key metabolites and proteins involved in sphingolipid
   metabolism, such as sphinganine 1-phosphate, sphingosine,
   sphingosine-1-phosphate, Sptlc1, and Enpp2. The strong correlations
   among these key substances are consistent with the hypothesis that ESC
   may influence the microbiota-gut-brain axis, potentially associated
   with depressive-like behaviors via the Bacteroides
   barnesiae/sphingolipid metabolism pathway. The discovery of this
   gut-brain interaction provides potential targets for antidepressant
   therapy, potentially optimizing treatment outcomes by regulating
   specific gut microbiota and sphingolipid metabolic pathways. Future
   research should further explore the specific roles of these
   differential microbiota, metabolites and proteins in the treatment of
   depression, providing a scientific basis for developing new therapeutic
   strategies.

Supplementary information

   [181]Supplementary information^ (14.1KB, docx)
   [182]Figure S1^ (7.3MB, tif)
   [183]Figure S2^ (11.6MB, tif)
   [184]Supplementary tables^ (21.9KB, docx)

Acknowledgements