Assessing the LPS-induced depressive-like model in mice of both sexes

Adult male and female C57BL/6 J mice were challenged with LPS or saline, weighed 24 h later, and then assessed for depressive-like behavior. Mice were then euthanized and their hypothalami collected. Forty mice (n = 10/group) were subjected to hypothalamus metabolomics and to statistical analysis of behavioral changes. The behavioral performance of LPS-related depression in SPT, TST, and FST, was evaluated. The time schedule for inducing depressive-like behavior is depicted in Fig. 1A. To validate the murine depression model, physiological sickness responses were assessed based on the changes in BW. Following LPS exposure, two-way ANOVA revealed a significant main effect of LPS treatment [F(1, 18) = 18.870, P < 0.001], and no significant interaction between the sex and LPS treatment [F(1, 18) = 0.400, P = 0.535] (Fig. 1B). Anhedonic-like behavior was evaluated using the SPT, which estimates an animal’s hedonic-seeking behavior by measuring the percentage of sucrose solution consumption. We detected a significant main effect of LPS challenge on the percentage of sucrose consumption [F(1, 18) = 38.568, P < 0.001] with a close significant interaction between sex and LPS treatment [F(1, 18) = 3.865, P = 0.065] (Fig. 1C). To assess the impact of inflammatory stress on depressive-like behaviors in both sexes, the TST and FST were employed. The two tests quantify an animal's “efforts” to fight in inescapable situation in the absence of apparent deficits in locomotor deficits. Regarding immobility time in the TST, two-way ANOVA revealed a significant main effect of LPS treatment [F(1, 18) = 64.402, P < 0.001], and no significant interaction between the sex and LPS treatment [F(1, 18) = 0.678, P = 0.421] (Fig. 1D). In the FST, a sex-dependent increase in immobility duration was noted in LPS-treated male mice compared to their female counterparts. The results of the two-way ANOVA indicated that the main effect of LPS treatment was significantly evident [F(1, 18) = 45.708, P < 0.001], as were the main effect of sex [F(1, 18) = 6.555, P = 0.020] and the interaction between sex and LPS treatment [F(1, 18) = 4.446, P = 0.049]. Subsequent examination using Bonferroni's post hoc test demonstrated a noteworthy increase in the duration of immobility in male and female mice subjected to LPS compared to their saline-administered counterparts. In contrast to their male counterparts, female mice exhibited a notably increased immobility in the FST (P < 0.05, Fig. 1E).

Fig. 1

Assessment of the LPS-induced depressive-like behaviors and hypothalamic inflammatory cytokines in female and male mice. A Experimental timeline and design. B–E Body weight changes (B) and sucrose preference test (C), tail suspension test (D), and forced swimming test (E) results in female and male mice after LPS exposure, n = 10 mice/group. F, G The mRNA (F) and protein (G) levels of hypothalamic cytokines (TNF-α, IL-1β, and IL-6), n = 5 (F) n = 6 (G) mice/group. All data are presented as means ± SEM and were analyzed by two-way ANOVA (sex × treatment) followed by Bonferroni's post hoc tests. *P < 0.05, **P < 0.01, ***P < 0.001. LPS lipopolysaccharide-treated group, CON saline-treated group, BW body weight, SPT sucrose preference test, TST tail suspension test, FST forced swimming test

Inflammatory cytokine expression in the hypothalamus.

The assessment of hypothalamic inflammatory responses induced by LPS stress involved the detection of the three inflammatory cytokines TNF-α, IL-1β, and IL-6 that play crucial roles in the pathogenesis of depression [5, 7, 11, 39]. The mRNA and protein levels of these cytokines were analyzed by qRT-PCR and ELISA, respectively. Regarding Tnf mRNA levels, a highly significant interaction between LPS treatment and sex was observed [F(1, 8) = 31.260, P < 0.001], and the main effect of LPS was found to be statistically significant [F(1, 8) = 120.863, P < 0.001], as was the main effect of sex [F(1, 8) = 41.985, P < 0.001]. A non-significant interaction between LPS treatment and sex was observed in the Il1b mRNA level [F(1, 8) = 1.735, P = 0.224]. Furthermore, the main effect of LPS was statistically significant [F(1, 8) = 45.219, P < 0.001], whereas the main effect of sex was not significant [F(1, 8) = 0.777, P = 0.404]. Additional statistical analyses revealed a significant interaction between LPS treatment and sex in relation to the Il6 mRNA level [F(1, 8) = 19.826, P = 0.002], as well as significant main effects for LPS [F(1, 8) = 20.729, P = 0.002] and sex [F(1, 8) = 17.091, P = 0.003] (Fig. 1F). Akin to transcriptional expression, the protein levels of TNF-α (LPS treatment × Sex: [F(1, 10) = 29.129, P < 0.001], LPS: [F(1, 10) = 90.964, P < 0.001], and Sex: [F(1, 10) = 26.067, P < 0.001]), IL-1β (LPS treatment × Sex: [F(1, 10) = 0.027, P = 0.874], and LPS: [F(1, 10) = 42.830, P < 0.001]), and IL-6 (LPS treatment × Sex: [F(1, 10) = 4.647, P = 0.057], and LPS: [F(1, 10) = 38.286, P < 0.001]) in the hypothalamus demonstrated a consistent response pattern to LPS intervention in male and female mice (Fig. 1G). These findings indicate that the inflammatory response in the hypothalamus was more pronounced in female mice than in male mice when exposed to LPS stress.

Sex-specific metabolic signatures in the hypothalamus

Next, we assessed whether substantial sex differences in the metabolic profiles of the hypothalamus occurred in mice in response to the LPS challenge. To obtain a preliminary understanding of the hypothalamic alterations in the metabolome, we performed a multivariate pattern recognition analysis (Fig. 2). The metabolites detected in the combined positive and negative ESI modes were used to perform OPLS–DA. Initially, the OPLS–DA score plots revealed distinct partitions between male and female control mice (R2X = 0.290, R2Y = 0.908; Fig. 2B), indicating divergence in hypothalamic metabolic patterns in the absence of inflammation. This finding is consistent with previous reports on sexual dimorphism in the hypothalamus [22, 25]. In addition, the OPLS–DA score plots demonstrated marked partitions between depressive-like mice of both sexes and their healthy counterparts, with minimal overlap (male, R2X = 0.463, R2Y = 0.801; female, R2X = 0.495, R2Y = 0.964; Fig. 2B). A notable dissimilarity in metabolic profiles was observed between depressive-like male and female mice (R2X = 0.702, R2Y = 0.989; Fig. 2B). Specifically, positive R2Y and Q2Y values indicated robust metabolic differences between groups. Furthermore, the validity of the constructed OPLS–DA models between different comparison groups was determined using 300-iteration permutation tests. The higher original Q2 and R2 values compared to their corresponding permutated values indicated that the models were not overfitted, thus corroborating the robustness of the results (Fig. 2C). Furthermore, to identify the metabolites responsible for distinguishing sex differences between groups based on scan acquisition modes, we performed independent OPLS–DA analyses of the metabolic data in both positive (ESI +) and negative (ESI–) ion modes. The OPLS–DA outcomes consistently demonstrated clear separations in both detection modes (see Additional file 3: Fig. S1) and indicated that certain DMs detected in the ESI + and ESI– modes contributed to the separations observed in the different groups. A metabolite level difference was deemed significant when the OPLS–DA model yielded a VIP value greater than 1.0 and the two-tailed Student's t test resulted in a P value of less than 0.05 between the two groups. The findings of this analysis indicated the detection of DMs that were accountable for distinguishing between two experimental groups, namely 121 DMs in CON-Female vs. CON-Male, 124 DMs in LPS-Male vs. CON-Male, 61 DMs in LPS-Female vs. CON-Female, and 37 DMs in LPS-Female vs. LPS-Male (see Additional file 3: Fig. S2, Additional file 4: Tables S3–S6).

Fig. 2

Metabolomics analysis of the hypothalamus from male and female mice with LPS-induced depressive-like behavior compared to their control counterparts. A 3D view of the OPLS-DA model showing four clearly separated groups, including saline-treated male controls (CON-Male, purple), LPS-treated depressive-like male mice (LPS-Male, green), saline-treated female controls (CON-Female, blue), and LPS-treated depressive-like female mice (LPS-Female, turquoise). B The OPLS–DA model shows a more noticeable separation between t each group pair. The colours correspond to those in (A). C The permutation test suggests the validity of the OPLS–DA model between each group pair, as the Q2 and R2 values yielded by the permutation test were higher than their original values

As shown in Fig. 2 and (see Additional file 3: Fig. S1), OPLS–DA indicated discernible differences in the metabolic features of the hypothalamus between female and male mice in the absence or presence of any inflammatory response. Consequently, we next focused on further characterizing these dissimilarities. The biochemical features of the DMs were annotated using HMDB, LIPID MAPS, KEGG, and PubChem database. A comprehensive biochemical classification of the DMs was clustered and visualized (Fig. 3). These analyses provided valuable insights into the biological functions and signaling pathways of the identified metabolites in depressive-like male and female mice. The majority (83.47%) of the 121 DMs differing between the CON-Female and CON-Male groups were lipids and lipid-like molecules [33], including 64 glycerophospholipids (GPs), 17 fatty acyls (FAs), 2 sterol lipids (STs), and 18 sphingolipids (SPs). The remaining DMs were carboxylic acids and their derivatives, organooxygen compounds, and imidazopyrimidines (Fig. 3A). These results indicated that the hypothalamic metabolic signatures that exhibited significant differences were related to lipid metabolism. To gain a more precise understanding of these variations, a more detailed categorisation of lipids and lipid-like molecules was performed (Fig. 3B–E). Based on annotations found in the LIPID MAPS database, 64 distinct GP molecules were categorized into 10 subclasses: lysophosphatidylcholine (LPC), lysophosphatidylethanolamine (LPE), lysophosphatidylinositol (LPI), lysophosphatidylserine (LyPS), phosphatidic acid (PA), phosphatidylcholine (PC), phosphatidylethanolamine (PE), phosphatidylglycerol (PG), phosphatidylinositol (PI), and phosphatidylserine (PS; Fig. 3B, purple column). PE constituted nearly 50% of GP molecules, exhibiting a uniform trend of elevated expression with positive z-scores in the hypothalamus of CON-Female mice compared to the CON-Male cohort. This trend is also evident in PI, PG, LyPS, LPI, LPE, and LPC. Conversely, only seven GP metabolites were detected at low levels in the hypothalamus of CON-Female mice (Fig. 3C). Furthermore, the CON-Female and CON-Male groups were distinguished by 17 FA metabolites (11 down- and 6 up-regulated, Fig. 3D), 18 SPs (all upregulated), and 3 STs (1 down- and 2 upregulated, Fig. 3E).

Fig. 3

Sex-specific hypothalamic metabolic signatures of male and female mice with LPS-evoked depressive-like behavior. A–E Hypothalamic metabolic signatures of saline-treated male and female controls. A Comprehensive biochemical classification of all identified differential metabolites (DMs) was clustered and visualized as a pie chart. B Lipid subgroups and lipid molecule counts according to the annotated lipid classifications of the LIPID MAPS and HMDB databases. C–E Levels of lipid metabolites belonging to glycerophospholipids (C), fatty acyls (D), sterol lipids, and sphingolipids (E) in CON-Male mice relative to CON-Female mice. F–J Hypothalamic metabolic signatures in saline- and LPS-treated male mice. F Pie chart depicting the biochemical classification of all DMs. G Lipid subgroups and lipid molecule counts based on LIPID MAPS and HMDB annotations. H–J Glycerophospholipids (H), fatty acyls (I), sterol lipids, and sphingolipids (J) levels in the LPS-Male group compared to the CON-Male group. K–N Hypothalamic metabolic signatures in saline- and LPS-treated female mice. K Biochemical classification of all DMs. L Lipid subgroups and lipid molecule counts. M, N Glycerophospholipids (M), fatty acyls, sterol lipids and sphingolipids (N) in the LPS-Female vs. CON-Female. Fold changes in the lipids are expressed as z-scores; a positive value indicates a higher level, whereas a negative value indicates a lower concentration in each group compared to the corresponding control group. LPA lysophosphatidic acid, PA phosphatidic acid, LPC lysophosphatidylcholine, PC phosphatidylcholine, LPE lysophosphatidylethanolamine, PE phosphatidylethanolamine, LPG lysophosphatidylglycerol, PG phosphatidylglycerol, LPI lysophosphatidylinositol, PI phosphatidylinositol, LyPS lysophosphatidylserine, PS phosphatidylserine, FA fatty acyls; sterol lipids, ST; and SP sphingolipids

Given sex-specific modifications in lipid metabolism in the plasma and brain during ageing and dietary variations [40, 41], it is reasonable to hypothesize that dissimilarities in hypothalamic lipid metabolism under physiological conditions (Fig. 3A–E) may serve as the underlying basis for sex-specific manner of mood alterations or psychiatric disorders in response to inflammatory stimuli. Consequently, additional analyses were conducted to disclose the lipid metabolism traits in the hypothalamus of male and female mice in the presence of inflammatory stress (Fig. 3F–N). Seventy metabolites differentially expressed in the LPS-Male group compared to the CON-Male group were classified as lipids and lipid-like molecules, comprising over 50% of all DMs (Fig. 3F). The profile of the lipid-like molecules consisted of 70 compounds, including 44 GPs, 19 FAs, 5 STs, and 2 SPs (Fig. 3G–J). Within the GP subclass, it was observed that 15 PE metabolites were the most prevalent subgroup and significantly increased in the LPS-Male group compared to the CON-Male group. The levels of three PC (PC 40:7, PC 40:6, and PC 33:0), three PG (PG 29:0, PG 27:0, and PG 31:0), one PS (PS 38:1), and two LPE (LPE 16:0 and LPE 18:0) metabolites were elevated, whereas those of three lysophosphatidic acid (LPA), two LPE, two LPI, and two PA metabolites were decreased (Fig. 3H). Similarly, the hypothalamus of LPS-treated male mice showed significant modifications in the FA subclass compared to their CON-Male counterparts, as evidenced by a reduction in the levels of 14 and an elevation in the levels of 5 FA species (Fig. 3I). Among the ST and SP subclasses, five ST metabolites had reduced concentrations, whereas two SP metabolites had elevated levels (Fig. 3J). Likewise, DMs detected in LPS-treated female depressive-like mice, as opposed to their sex-matched controls, were primarily composed of lipid species (> 70%), specifically 32 GPs, 5 FAs, 4 SPs, and 8 STs (Fig. 3K–N). Within the GP subclass, reduced levels were observed in 2 LPA, 3 LPC, 2 LPE, and 16 PE isoforms, whereas increases were noted in 3 PC and 5 PS isoforms (Fig. 3M). Notably, a comparison of female depressive-like mice and their sex-matched controls demonstrated a decrease in the quantity of DMs within the FA subclasses and an increase in the number of ST species (Fig. 3N), in contrast to the other two comparison groups (Fig. 3D, E, in CON-Female vs. CON-Male, and Fig. 3I, J, in LPS-Male vs. CON-Male). Interestingly, the fluctuation patterns of the disclosed PE isoforms, as depicted in Fig. 3C, H and M, led to a thorough categorization of their individual levels. Distinct expression patterns in PE isoforms were observed, with female mice exhibiting a significant increase, no difference, or a significant decrease in hypothalamic PE metabolites in the presence of neuroinflammation. In contrast, male mice generally showed higher levels of PE isoforms in the hypothalamus under inflammatory conditions (see Additional file 3: Fig. S3). Importantly, 37 DMs in LPS-treated female and male depressive-like mice revealed that the majority (> 75%, 29 lipids) were lipid species. Specifically, these DMs consisted of 6 GPs, 2 glycerolipids (GLs), 12 FAs, 3 SPs, and 6 STs (Fig. 4A, B). The results indicated that a decrease in the proportion of GP species in the DMs, while increases were observed in 12 FA and 6 ST metabolites (Fig. 4C–E), in comparison to the three comparisons illustrated in Fig. 3. It suggested that the difference in hypothalamic metabolism between female and male mice with depressive-like behaviors is not primarily attributable to GPs involved lipid metabolism signaling, but rather to metabolic fluctuations that may be dominated by FA or ST lipids (Fig. 4A–E). Collectively, these data imply that the dissimilarity in the hypothalamic lipid metabolism response to inflammatory stimuli may be involved in sex-specific susceptibility to depression in this mouse model.

Fig. 4

The hypothalamic steroid hormone biosynthesis is significantly affected by inflammation, particularly in female mice. A–E Hypothalamic metabolic signatures of LPS-treated female and male mice. A Biochemical classifications of all DMs. B Lipid subgroups and lipid molecule counts. C–E Glycerophospholipids (C), fatty acyls (D), sterols, and sphingolipids (E) in LPS-Female vs. LPS-Male. Relative levels of the lipids by z-scores, a positive value indicates a higher level, whereas a negative value indicates a lower concentration in the LPS-Female group. F, G Implications of DMs in biological processes (F) and metabolic pathways (G) in the LPS-Female vs. LPS-Male comparison based on MetaboAnalyst 5.0 analyses. H Venn plot of significantly affected metabolic pathways in LPS-treated female and male mice. Perturbations of steroid hormone biosynthesis are apparent in the hypothalamus of the female depression model. I Venn diagram of DMs from different comparisons. Red boxes represent increased levels, and green boxes represent decreased levels. FA fatty acyls, LPG lysophosphatidylglycerol, LPI lysophosphatidylinositol, MG monoacylglycerol, PC phosphatidylcholine, PG phosphatidylglycerol, PS phosphatidylserine, SP sphingolipid, ST sterol lipid, TG triacylglycerol

Perturbation of the steroidogenic pathway occurs in the hypothalamus of female mice with inflammation-evoked depression

To verify the BP and molecular pathways among the DMs, a comprehensive metabolic network was constructed using the MetaboAnalyst 5.0. Enrichment analyses were conducted on the enriched metabolite sets derived from various comparison groups to determine the presence of differential metabolite sets associated with specific BPs in the presence or absence of inflammation. Most DMs, as lipid molecules, have a strong association with the primary BPs and lipid metabolism. Specifically, the most prominent BPs were found to be linked to lipid molecules, such as glycerophospholipid metabolism, glycerolipid metabolism, phospholipid biosynthesis, sphingolipid metabolism, and steroid hormone biosynthesis. This suggests a strong association between hypothalamic lipid metabolism and sex disparities (Fig. 4F and see Additional file 3: Fig. S4A–S4C). Pathway analysis is an extension of the BP enrichment analysis of differential metabolite sets to obtain a better understanding of the pathways involved. We identified several pathways that may be aberrant in neuroinflammation-induced female and male depressive-like mice (Fig. 4G and see Additional file 3: Fig. S4D–S4F). In the absence of inflammation, the hypothalamus of healthy male and female mice exhibited distinct metabolic pathways for amino acids (alanine–aspartate–glutamate and purine metabolism) and lipids (glycerophospholipid and sphingolipid metabolism; see Additional file 3: Fig. S4D). In the context of neuroinflammation, metabolic perturbations in the hypothalamus of male depression model mice compared to control male littermates were primarily associated with glycerophospholipid and purine metabolism (see Additional file 3: Fig. S4E). The top three pathways identified in the hypothalamus of female depression model mice were glycerophospholipid metabolism, steroid hormone biosynthesis, and sphingolipid metabolism (see Additional file 3: Fig. S4F). The highest-ranked pathway was the same in female and male model groups compared to their sex-matched healthy controls, i.e., glycerophospholipid metabolism. Purine metabolism and steroid hormone biosynthesis ranked second in the male and female depression model groups, respectively. In our previous study, we verified that purine metabolism is a key hypothalamic pathway involved in male mice with LPS-induced depression [26]. Compared with male depressive-like mice, hypothalamic steroid biosynthesis was the most significantly affected pathway in female depressive-like mice (Fig. 4F–H). We next investigated common and discriminating metabolite expression patterns among the study groups comparing LPS-Male with CON-Male, LPS-Female with CON-Female, and LPS-Female with LPS-Male mice. As shown in Fig. 4I, none of the DMs were common in all three comparisons. The intersection of 10 DMs between LPS-Male vs. CON-Male and LPS-Female vs. CON-Female consisted of three non-lipid molecules and seven GP molecules: tetracosanoic acid, LPE 18:0, LPA 20:1, LPA 18:0, PE 44:10, PE 40:7, and PE O-18:0/19:1. The levels of tetracosanoic acid, LPE 18:0, PE 44:10, PE 40:7 and PE O-18:0/19:1 were increased in LPS-Male mice and decreased in LPS-Female mice relative to their sex-matched controls. In addition, 15 DMs were present in two other comparisons (7 in the intersection between LPS-Male vs. CON-Male and LPS-Female vs. LPS-Male; 8 in the intersection between LPS-Female vs. CON-Female and LPS-Female vs. LPS-Male). These 15 DMs, of which the five ST metabolites 5alpha-pregnane-3,20-dione (5α-DHP), tetrahydrocortisone, progesterone, allopregnanolone, and 5α-tetrahydrocortisol (Fig. 4I) are involved in steroid biosynthesis, are considered important neuroactive steroids [42, 43]. The results obtained from the DM pathway analyses suggest that the steroidogenic pathway of the hypothalamus exhibits sex specificity in the context of depression induced by inflammatory stress, leading to female proclivity.

The hypothalamus is not only a target area for neuroactive steroids, but also a steroidogenic region [44]. Interactions between neurosteroid production and action have been implicated in the emergence of various neuropsychiatric disorders, such as depression, anxiety, and schizophrenia [45]. However, the role of the hypothalamic steroidogenic changes in inflammation-associated depression remains unclear. To fully understand the alterations occurring within the hypothalamic steroidogenic pathways in response to inflammatory stress, the metabolites associated with these pathways were analyzed separately and in detail (Fig. 5A, B). Pregnenolone is considered a vital precursor of all steroid hormones and is related several different metabolic pathways [43, 44, 46], including the synthesis of corticosteroids (cortisol and cortisone), neuroactive steroids (particularly progesterone, another precursor of virtually all steroid hormones), and its metabolism into its reduced derivatives, such as 5α-DHP, allopregnanolone, and 11-deoxycorticosterone. The levels of pregnenolone and progesterone decreased, whereas 5α-DHP, allopregnanolone, and 11-deoxycorticosterone concentrations in LPS-Female mice increased compared to those in CON-Female mice (Fig. 5B). Pregnenolone, progesterone, and allopregnanolone levels in the hypothalamus were validated using ELISA (see Additional file 3: Fig. S5A). Subsequently, referring to KEGG annotations, the altered key metabolic molecules were designated within the steroidogenic pathway, and the essential enzymes implicated in this pathway are shown in Fig. 5C. Neurosteroids are steroid hormone derivatives locally synthesized from cholesterol in the brain. In this process, the steroidogenic acute regulatory protein (StAR, encoded by Star) mediates the trafficking of cholesterol to the mitochondria, which is the rate-limiting step in neurosteroid production, and it is subsequently metabolized by cytochrome P450 11A1 (CYP11A1, encoded by Cyp11a1) into pregnenolone [44]. Moreover, the enzymatic mechanisms implicated in the conversion of pregnenolone to neuroactive metabolites such as progesterone, 5α-DHP, and allopregnanolone, require the participation of several essential enzymes, namely cytochrome P450 species (CYP21A1, CYP11B1, and CYP11B2), as well as 5α-reductase type I, II, and III isozymes (SRD5A1, SRD5A2, and SRD5A3). The mRNA and protein levels of the associated key enzymes were verified using RT-PCR and WB, respectively. In the presence of inflammatory stress, two-way ANOVA revealed that hypothalamic mRNA [F(1, 20) = 41.787, P < 0.001] and protein [F(1, 12) = 11.996, P = 0.005] levels of SRD5A1 exhibited significant changes in both male and female mice, with a more pronounced increase observed in female mice. Bonferroni's post hoc analysis demonstrated a significant increase in the mRNA and protein levels of SRD5A1 in LPS-Female mice compared to LPS-Male mice (Fig. 5D, E). Collectively, these results indicate that the steroid biosynthesis pathway in the hypothalamus of female mice is perturbed by inflammatory stress.

Fig. 5

Verification of perturbations in steroidogenic pathways in the hypothalamus of LPS-induced depressive-like female mice. A, B Heatmap (A) and scatter plot (B) of the detected steroid metabolites in the hypothalamus of LPS-induced depressive-like mice and sex-matched controls. A Raw measurement mean was converted into z-scores, with downregulation denoted by blue, and upregulation indicated by red. B Relative levels of steroid metabolites were determined by comparing them to the raw measurement mean of the CON-Male group. C Metabolite–protein interaction network of the steroid biosynthesis pathway, according to the KEGG annotations. Differential steroids between LPS-Female and CON-Female mice are labelled with red (increased) and green (decreased) in the network, and essential proteins and unchanged steroid metabolites are designated. D, E mRNA transcriptional (D) and protein (E) levels of the key proteins involved in steroidogenic pathways, n = 6 (D) and n = 4 (E) mice/group. All data are presented as means ± SEM and were analyzed by two-way ANOVA followed by Bonferroni's post hoc tests. *P < 0.05, **P < 0.01, ***P < 0.001

Central pregnenolone delivery reverses hypothalamic inflammation and depressive-like behaviors in female mice

Inflammation plays a critical role in modulating the hypothalamic neurosteroid metabolic pathway, thereby increasing susceptibility to depression. Consequently, the regulation of inflammation could potentially yield comprehensive therapeutic advantages for individuals with depression [2, 5]. The neurosteroid pregnenolone, a crucial precursor of steroid hormones, suppresses inflammation [46] and exhibits potential antidepressant properties [45]. Indeed, the levels of pregnenolone were reduced in the LPS-Female group (Fig. 5B and see Additional file 3: Fig. S5A) and were more significantly correlated with depressive-like behaviors in female mice (see Additional file 3: Fig. S5B–S5G). To ascertain the potential effects of pregnenolone supplements on hypothalamic inflammation and their potential to ameliorate depressive-like behaviors, female mice received pregnenolone via intracerebroventricular (i.c.v.) administration (Fig. 6A). The administration of pregnenolone successfully counteracted the LPS-induced increases in the levels of the pro-inflammatory cytokine mRNAs in female mice (Tnf, [F = 21.267, P < 0.001]; Il1b, [F = 8.454, P < 0.001]; and Il6, [F(1, 8) = 9.194, P < 0.001]; Fig. 6B). As anticipated, the evaluation of the effects of pregnenolone on depressive-like behaviors revealed that pregnenolone treatment (i.c.v.) effectively reversed the diminished sucrose preference ([F = 5.323, P = 0.007], Fig. 6C) and extended immobility ([F = 5.983, P = 0.004], Fig. 6D) following LPS intervention. We also investigated the effect of pregnenolone on hypothalamic inflammation and depressive-like behaviors in male mice and found that pregnenolone reduced in the elevated levels of cytokine mRNAs, particularly Tnf ([F = 9.881, P < 0.001]) and Il1b ([F = 4.769, P = 0.012]) induced by LPS. However, it failed to mitigate the depressive-like behaviors induced by LPS. (see Additional file 3: Fig. S6A–C). Taken together, these results clearly indicate that maintaining homeostasis in the neurosteroid biosynthesis pathway is crucial for the hypothalamus of female mice. Pregnenolone supplementation may potentially support homeostasis, thereby mitigating excessive inflammation and improving depressive symptoms in female mice.

Fig. 6

Pregnenolone supplements reversed hypothalamic inflammation and depressive-like behaviors in female mice. A Schematic illustration of the central pregnenolone (Preg) delivery and subsequent behavioral tests in female mice. Blue arrows show the injection phases. Schematic representation of the injection site in the third ventricle stained with Evans blue (bottom right, Scale bar 2 mm). B Transcriptional levels of hypothalamic Tnf, Il1b, and Il6, after i.c.v. administration of artificial cerebrospinal fluid (ACSF) or pregnenolone (Preg), n = 6 mice/group. C, D Recorded parameters to assess sucrose preference in the SPT (C) and immobility in the FST (D) after i.c.v. administration of ACSF or pregnenolone, n = 6 mice/group. All data are presented as means ± SEM and were analyzed using one-way ANOVA with Bonferroni's post hoc tests. *P < 0.05, **P < 0.01, ***P < 0.001. AP anterior–posterior, ML medial–lateral, DV dorsal–ventral, SPT sucrose preference test, FST forced swimming test; i.c.v., intracerebroventricular

Pharmacological inhibition of 5α-reductase mimics the effects of central pregnenolone delivery in female mice

The neurosteroid biosynthesis pathway in the hypothalamus of female mice exhibited a noticeable disruption in response to inflammatory stress, characterized by decreased pregnenolone and progesterone levels, and increased 5α-DHP and allopregnanolone levels (Fig. 5B and see Additional file 3: Fig. S5A). Furthermore, SRD5A1 was upregulated in the hypothalamus of LPS-treated female mice (Fig. 5E). In the brain, the rate-limiting enzyme SRD5A1 is involved in neurosteroidogenesis, particularly the conversion from progesterone to 5α-DHP and allopregnanolone [38]. These findings suggest that rectifying the imbalance within the steroidogenic pathway may be a viable therapeutic strategy for neuroinflammation-associated depression. To test this hypothesis, dutasteride, a 5α-reductase inhibitor capable of inhibiting SRD5A1 [47], was stereotactically injected into the hypothalamus of female and male mice (Fig. 7A). As anticipated, inhibition of 5α-reductase effectively corrected the decrease in pregnenolone and proge