Three-week-old male C57 mice from the Experimental Animal Centre of Army Medical University were housed in a specific pathogen-free (SPF) environment at 18°C–26°C and 50%–60% humidity, with a 12-h-on/12-h-off light cycle. The mice had free access to feed and distilled water.

After 2 weeks of co-feeding (keep in the same cage), the mice were randomly divided into two groups with12 mice each for the alcohol treatment experiment: the control group (control; free access to feed and distilled water for 10 weeks), and the chronic drinking group (alcohol; free access to feed and 10% ethanol [99%, vol/vol] for 10 weeks). The average weekly alcohol intake of per mouse in the alcohol group is shown in Supplementary Table 1. All mice were sacrificed after anesthesia induction with 1% pentobarbital sodium solution by intraperitoneal injection.

Fecal microbiota transplantation (FMT) experiment: Three-week-old male C57 mice were co-fed (in the same cage) and then randomly divided into two groups with five mice each 2 weeks later: the alcohol-fecal microbiota transplantation (FMT) group (alcohol-FMT; fecal microbiota from the alcohol group was transplanted into normal mice; free access to feed and distilled water for 10 weeks), and the control-FMT group (fecal microbiota from the control group was transplanted into normal mice; free access to feed and distilled water for 10 weeks). All mice were sacrificed after anesthesia induction with 1% pentobarbital sodium solution by intraperitoneal injection.

Five hundred milligrams of fresh donor stool samples were continuously collected from the alcohol (five mice) and control (five mice) groups immediately upon defecation, resuspended in 4 ml of saline, vortexed for 5 min, and filtered with a 360-mesh sterile strainer. Transplantation into the recipient mice was achieved by gavage with 200 μl of the supernatant from the fecal sample every Monday, Wednesday, and Friday for 10 weeks (Figure 1A). Fecal samples from donors were collected during the whole fecal microbiota transplantation process.

The cauda epididymis was minced in 100 ml of phosphate-buffered saline (PBS, pH 7.2) and incubated for 20 min at 37°C in a 5% CO₂ incubator to release sperm. Sperm were observed under a Nikon N200 microscope and measured with a Suiplus semen analysis (SSA) system.

Mouse testes and small intestines were harvested. The entire testis was fixed with a special testis solution (Servicebio Cat. No.: G1121) for 24 h. The small intestine was cut into small sections of approximately 1 cm, which were then cleaned and fixed with paraformaldehyde (Servicebio Cat. No.: G1101) for 24 h. Hematoxylin and eosin (H&E) staining was performed after dehydration in graded ethanol solutions, paraffin embedding, and sectioning (2.5 μm). Three random fields of sections on each slide were observed under an Olympus upright microscope (×200).

The deparaffinized and rehydrated intestine and testis sections were blocked with 5% bovine serum albumin (BSA) containing 0.1% Triton X-100 for 1 h and incubated with a primary antibody against F4/80 (1:500, ab111101, Abcam, Cambridge, MA, United States) at 4°C overnight. Then, the sections were incubated with the secondary antibody (BA1039, Boster, Wuhan, Hubei, China) at room temperature for 1 h. After color development, the sections were observed under an Olympus microscope, and the proportions of positively stained cells were analyzed. Protein expression was quantified using ImageJ software (National Institutes of Health; http://www.imagej.softonic.de).

For immunofluorescence staining, paraffin sections of the intestines (2.5 μm thick) were assayed with 4′,6-diamidino-2-phenylindole (DAPI) as a counterstain. The primary antibody used in this experiment was rabbit anti-mouse CD3e (1:500; ab231775, Abcam, Cambridge, MA, United States). The secondary antibody (1:500; A11034, Invitrogen, United States) was an Alexa Fluor 488-conjugated donkey anti-rabbit antibody. Sections were examined under a ZEISS 880 confocal microscope. Protein expression was quantified using ImageJ software (National Institutes of Health; http://www.imagej.softonic.de).

DNA fragmentation indicative of apoptosis was examined using the terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) method. The TUNEL assay was performed using the In Situ Cell Death Detection Kit (Cat. No. 11684817910, Roche Molecular Biochemicals, Germany) according to the manufacturer’s instructions. Sections (4 μm thick) were transferred to polylysine-coated slides and incubated overnight at 60°C. Then, they were deparaffinized in xylene. Sections were rehydrated in a graded series of 75%, 80%, 90%, and 100% ethanol and washed with PBS. Slides immersed in 0.1 M citrate buffer were microwaved at 350 W for 5 min for antigen retrieval. PBS was used for washing. Sections were incubated with the TUNEL reaction mixture (labeling solution + terminal deoxynucleotidyl transferase (TdT) enzyme solution) at 37°C for 1 h in a humidified chamber and washed twice with PBS. After this step, Converter-POD was added, and the slides were incubated for 30 min at 37°C and then rinsed twice with PBS. Next, the sections were treated with 3,3-diaminobenzidine at room temperature and rinsed twice with PBS. Sections were counterstained with Harris’ hematoxylin for 10 min. The sections were dehydrated with a series of graded alcohol concentrations. Xylene was used as a clearing agent, and then, the sections were coverslipped with Entellan for light microscopy examination.

Mouse small intestine was harvested, cut into sections of 1 mm × 1 mm × 2 mm, and placed in a 1.5-ml Eppendorf tube (Servicebio Cat. No. G1101) containing 2.5% glutaraldehyde. The samples were then embedded, sectioned (2.5 μm), stained, and observed and imaged by transmission electron microscopy (HITACHI HT7800/HT7700). Image-Pro Plus 6.0 (Media Cybernetics, Inc., Rockville, MD, United States) software was used to measure the intercellular space (nm) at five equal positions with an intact intercellular space on each slide [reference: (El-Sokkary, 2001) 0.0 k, 500 nm ruler].

Mouse serum endotoxin levels were measured using ELISA kits (Quanzhou Ruixin Biological Technology Co. Ltd., China) according to the manufacturer’s protocols.

RayBio Mouse Cytokine Antibody Arrays (QAM-TH17-1-1, RayBiotech, Norcross, GA) were employed to evaluate chrysin-mediated inhibition of signaling in the serum and testis according to the manufacturer’s specification. Eighteen different cytokines were evaluated: interleukin 1 beta (IL-1β), IL-2, IL-4, IL-5, IL-6, IL-10, IL-12p70, IL-13, IL-17, IL-17F, IL-21, IL-22, IL-23, IL-28, interferon gamma (IFN-γ), macrophage inflammatory protein 3 alpha (MIP-3α), transforming growth factor beta 1 (TGF β1), and tumor necrosis factor alpha (TNF-α). Each sample was prepared in triplicate. An Axon scanner 4000B with GenePix software was used to record fluorescence intensities.

DNA from intestinal contents was prepared from animals in the alcohol, control, alcohol-FMT, and control-FMT groups using a fecal DNA extraction kit (DP712, Tiangen Company, Beijing, China); three biological replicates per group were analyzed. DNA quality was monitored on 1% agarose gels. The hypervariable V3–V4 region of the 16S-rDNA gene was amplified by PCR using the primers 341F: CCTACGGGNGGCWGCAG and 806R: GGACTACHVGGGTATCTAAT, where the barcode was an eight-base sequence unique to each sample. PCR was performed in 30-μL reactions composed of 15 μl of Phusion® High-Fidelity PCR Master Mix (New England Biolabs), 0.2 μM forward and reverse primers, and 10 ng of template DNA. Thermal cycling consisted of an initial denaturation step at 98°C for 1 min, followed by 30 cycles of denaturation at 98°C for 10 s, annealing at 50°C for 30 s, and elongation at 72°C for 30 s. Finally, the reactions were incubated at 72°C for 5 min. PCR products were detected on 2% agarose gels by electrophoresis and then purified using the GeneJET Gel Extraction Kit (Thermo Scientific). Sequencing libraries were generated using the Illumina TruSeq DNA PCR-Free Library Preparation Kit (Illumina, United States) according to the manufacturer’s recommendations, and index codes were added. Sequencing was performed using the Illumina NovaSeq 6,000 platform (Genecloud Co. Ltd., Chongqing, China). Sequences were analyzed using QIIME software package 2 (version 2020.2). The effective tags were clustered into operational taxonomic units (OTUs) of ≥97% similarity using the MOTHUR pipeline. The tag sequence with the highest abundance was selected as the representative sequence within each cluster. Bacterial alpha diversity was determined by performing a sampling-based OTU analysis and presented as the observed OTUs, Chao index, Shannon index, and Simpson index, which were calculated using the R program package “vegan.” Principal coordinate analysis (pCoA) was conducted using the R package¹ to display and compare the beta diversity of the microbiome between samples. The Bray–Curtis metric distances, unweighted UniFrac distances, and weighted UniFrac distances were calculated with the phyloseq package. Bacterial taxonomic analyses and comparisons, including bacterial phyla, classes, orders, families, and genera, were conducted between two groups using the Wilcoxon rank-sum test. The predominance of bacterial communities between groups was analyzed using the linear discriminant analysis (LDA) effect size (LEfSe) method.² Based on the normalized relative abundance matrix, features with significantly different abundance levels between assigned taxa were determined using LEfSe with the Kruskal–Wallis rank-sum test (p < 0.05), and LDA was used to assess the effect size of each feature [LDA score (log10) = 3.5 as the cutoff value]. Correlation analysis was performed by calculating Pearson’s correlation coefficient using the R language mixOmics package to calculate the correlation coefficient r² and the p value of differentially intestinal flora and differentially metabolites. All 16S rRNA datasets are publicly available (NCBI SRA accession numbers: SRP789987 for mouse intestinal content samples).

Tissues (100 mg) were individually ground in liquid nitrogen, and the homogenate was resuspended in prechilled 80% methanol by thoroughly vortexing the mixture. The samples were incubated on ice for 5 min and then centrifuged at 15,000 × g for 20 min at 4°C. Some of the supernatant was diluted with liquid chromatography–mass spectrometry (LC–MS)-grade water to a final concentration of 53% methanol. The samples were subsequently transferred to a fresh Eppendorf tube and then centrifuged at 15,000 × g for 20 min at 4°C. Finally, the supernatant was injected into a liquid chromatography tandem mass spectrometry (LC–MS/MS) system. Ultrahigh-performance liquid chromatography (UHPLC)-MS/MS analyses were performed using a Vanquish UHPLC system (Thermo Fisher, Germany) coupled with an Orbitrap Q Exactive™ HF-X mass spectrometer (Thermo Fisher, Germany) at Novogene Co., Ltd. (Beijing, China). The raw data files generated by the UHPLC–MS/MS system were processed using the Compound Discoverer 3.1 platform (CD3.1, Thermo Fisher) to perform peak alignment, peak picking, and quantitation of each metabolite. The normalized data were used to predict the molecular formula based on additive ions, molecular ion peaks, and fragment ions. Then, peaks were matched with the mzCloud,³ mzVault, and MassList databases to obtain accurate qualitative and relative quantitative results. Statistical analyses were performed using the statistical software R (R version R-3.4.3), Python (Python version 2.7.6), and CentOS (CentOS release 6.6). When data were not normally distributed, normal transformations were attempted using the area normalization method.

These metabolites were annotated using the Kyoto Encyclopedia of Genes and Genomes (KEGG) database,⁴ Human Metabolome Database (HMDB),⁵ and LIPID MAPS database.⁶ Principal component analysis (PCA) and partial least squares discriminant analysis (PLS-DA) were performed using metaX (a flexible and comprehensive software for processing metabolomics data). We employed a univariate analysis (t test) to calculate statistical significance (p-value). Metabolites with a variable importance in projection (VIP) value > 1, p-value < 0.05, and fold change (FC) ≥ 2 or ≤ 0.5 were considered differentially altered metabolites. Volcano plots were constructed to filter the metabolites of interest based on the log2 (FC) and log10 (p-value) of metabolites. The functions of these metabolites and metabolic pathways were analyzed using the KEGG database. Next, we evaluated the metabolic pathway enrichment of the differentially altered metabolites. When the ratios were satisfied by x/n > y/N, the metabolic pathway was considered enriched, while when the p value of the metabolic pathway was <0.05, the enrichment of the metabolic pathway was considered statistically significant. Correlation analysis was performed by calculating Pearson’s correlation coefficient using the R language mixOmics package to calculate the correlation coefficient r² and the p-value of differentially expressed genes and differentially expressed metabolites. All differentially expressed genes and differentially altered metabolites obtained in the alcohol-FMT group compared with the control-FMT group were mapped to the KEGG database to obtain their common pathway information and for statistical analysis.

Total RNA was extracted from the testicular tissues of animals in the alcohol-FMT and control-FMT groups, with three replicates in each group. RNA integrity was assessed using an RNA Nano 6000 Assay Kit and a Bioanalyzer 2100 system (Agilent Technologies, Santa Clara, CA, United States). The cDNA libraries were prepared using an Illumina TruSeq RNA Sample Prep Kit. The average size of the library cDNAs was 150 bp (excluding adapters). The integrity and quality of cDNA libraries were assessed using an Agilent 2100 Bioanalyzer and an ABI StepOne Plus real-time PCR system. RNA sequencing (RNA-seq) was performed by Novogene Company (Tianjin, China). RNA-seq reads from the FASTQ files were mapped to the mouse reference genome using TopHat2. The accession number for the RNA-seq data is GEO GSE192522.

The original data were filtered to obtain clean data and ensure the quality and reliability of the analysis. Trinity software was used for transcript assembly. Corset hierarchical clustering was performed for splicing. BUSCO software was used to evaluate the splicing quality of Trinity.fasta, unigene.fa, and cluster.fasta files. The gene function was annotated using the following databases: Nr [National Center for Biotechnology Information (NCBI nonredundant protein sequences), Nt (NCBI nonredundant nucleotide sequences), Pfam (protein family), KOG/COG (clusters of orthologous groups of proteins), Swiss-Prot (a manually annotated and reviewed protein sequence database), KO (KEGG Orthology database), and GO (Gene Ontology)]. The differential expression analysis of the two conditions/groups was performed using the DESeq2 R package (1.20.0). DESeq2 provides statistical routines for determining differential expression in digital gene expression data using a model based on a negative binomial distribution. The resulting p values were adjusted using the Benjamini–Hochberg approach to control the false discovery rate. Genes with an adjusted p-value < 0.05 identified by DESeq2 were considered differentially expressed. The GO enrichment analysis of differentially expressed genes was performed using the clusterProfiler R package, and gene length bias was corrected. We used the clusterProfiler R package to test the statistical enrichment of differentially expressed genes in KEGG pathways.

The data are expressed as the mean ± standard error of the mean (SEM). Statistical comparisons between two measurements were analyzed by independent-samples t tests using SPSS 26.0 software (SPSS, Inc., Chicago, IL, United States). When analyzing gut microbiota sequencing data, we performed a two-tailed Wilcoxon rank-sum test by R Project; p < 0.05 was considered statistically significant.

Through its interaction with the host, the gut microbiota plays an important role in human diseases. We established a mouse model of chronic drinking (10% ethanol; alcohol) to investigate the effect of chronic drinking-induced intestinal dysbiosis on sperm quality. No significant difference was observed in body weight between the alcohol group and the control group (p = 0.479, Supplementary Table 2). We harvested the intestinal contents from each group for 16S rDNA sequencing and found that chronic drinking induced significant changes in the gut microbiota. We measured alpha diversity using the Shannon index, Chao1 index, ACE index, Simpson index, observed features, ENSPIE, and Fisher’s alpha (Figure 2A; Supplementary Figure S1A). Most of these indices manifested similar tendencies, as the gut microbiota of mice in the alcohol group exhibited a higher alpha diversity than that of control mice (p = 0.0104, p = 0.0373, p = 0.0373, p = 0.0249, p = 0.0373, p = 0.0249, and p = 0.0373, respectively). The beta diversity based on weighted PCoA clearly showed a difference in the microbiota composition between the control group and the alcohol group, with a significant difference along the PCO1 axis (accounting for 54.83% of the overall variation, Figure 2B). Next, we compared the species composition of the gut microbiota between the alcohol group and the control group at different levels. Chronic drinking induced significant differences in the gut microbiota at different levels (Supplementary Figure S1B). The relative abundance of the dominant phyla (> 0.5%) Bacteroides and Tenericutes decreased significantly, while that of the phylum Firmicutes increased significantly. The relative abundance levels of the dominant classes (> 0.5%) Bacteroidia, Mollicutes, and Erysipelotrichia decreased significantly, while those of the classes Clostridia and Verrucomicrobia increased significantly. The relative abundance of the dominant orders (> 0.5%) Bacteroidales, Mycoplasmatales, and Erysipelotrichales decreased significantly, while that of the orders Clostridiales and Verrucomicrobiales increased significantly. At the family level, the relative abundance of the dominant taxa (> 0.5%) Erysipelotrichaceae, Muribaculaceae, and Mycoplasmataceae decreased significantly, while that of Akkermansiaceae, Lachnospiraceae, Rikenellaceae, and Ruminococcaceae increased significantly. At the genus level, the relative abundance of the dominant taxa (˃0.5%) Allobaculum, Alloprevotella, Adlercreutzia, Gordonibacter, and Mycoplasma decreased significantly, while that of Alistipes, Eubacterium_nodatum_group, Lachnospiraceae_NK4A136_group, Family_XIII_UCG_001, Angelakisella, Ruminiclostridium_5, Ruminiclostridium_6, UBA1819, Moorella, Marvinbryantia, Ruminococcus__gnavus _ group, and Ruminococcaceae_UGG_014 increased significantly (Figure 2C). The heatmap shows the most differentially abundant features selected at the genus level between mice in the alcohol and control groups (Figure 2D). We performed high-dimensional class comparisons using LEfSe (LDA score > 3.5) to identify bacterial biomarkers of the two groups (Figure 2E). Collectively, the gut microbiota differed between mice in the alcohol and control groups.

We also investigated the effect of chronic drinking on male sperm quality. Sperm count, sperm concentration, sperm hyperactivation, and forward progression were significantly reduced in the alcohol group (Figure 2F), along with significantly decreased numbers of spermatocytes (red arrow), round spermatids (green arrow), and sperm (black arrow) in the seminiferous tubules (Figure 2G), and the infiltration of a largenumber of macrophages in the testicular interstitium (Figure 2H). These findings are consistent with previous reports (Anderson et al., 1982; Talebi et al., 2011; Ezzatabadipour et al., 2012; Chen and Li, 2014).

Furthermore, we observed intestinal mucosal epithelial cells and cell junction complexes with transmission electron microscopy. The small intestinal mucosa was significantly more permeable in the alcohol group than in the control group (Figure 2I; Supplementary Table 3), which manifested as a widened gap junction. Intestinal dysbiosis increases the permeability of the intestinal mucosa, which aggravates intestinal dysbiosis and causes gut metabolites to enter the blood circulation through the intestinal barrier. In summary, chronic alcohol consumption induces intestinal dysbiosis and significantly affects sperm quality in mice.

We investigated the effect of chronic drinking-induced intestinal dysbiosis on sperm quality by transplanting the fecal microbiota from the alcohol group into the intestine of normal mice by gavage and then analyzed the sperm from the alcohol-FMT group and control-FMT group. At the end of FMT, no significant difference was observed in the mean body weight between the alcohol-FMT group and the control-FMT group (21.58 ± 3.87 g vs. 23.42 ± 3.49 g, p > 0.05; Supplementary Table 4). Sperm counts, sperm concentrations and sperm hyperactivation were significantly lower in the alcohol-FMT group than in the control-FMT group. The straightness (straight-line (rectilinear) velocity/average path velocity) of sperm declined but not significantly (Figure 1B), along with significantly decreased numbers of spermatocytes and sperm in the seminiferous tubules (Figure 1C). The cell apoptosis analysis (TUNEL) results further showed increased apoptosis of spermatocytes in the seminiferous tubules of the alcohol-FMT group compared with the control-FMT group, although the difference did not reach statistical significance (Figures 1D,E). Moreover, serum endotoxin levels were significantly higher in the alcohol-FMT group than in the control-FMT group (Figure 1F). Based on these findings, FMT from the alcohol group into normal mice increases serum endotoxin levels and impairs sperm quality in otherwise normal mice.

The intestinal contents were harvested from the alcohol-FMT group and control-FMT group for high-throughput 16S rDNA sequencing. We measured alpha diversity using the Shannon index, Chao1 index, ACE index, Simpson index, observed features, ENSPIE, and Fisher’s alpha (Figure 3A; Supplementary Figure S3A). Most of the indices exhibited similar tendencies, although the gut microbiota of the alcohol-FMT group displayed higher alpha diversity than that of the control-FMT group, but the difference did not reach statistical significance (p = 0.1745, p = 0.2505, p = 0.2505, p = 0.4647, p = 0.2505, p = 0.4647, and p = 0.2505, respectively), which was consistent with the results described above. A distinct clustering of the microbiota composition for the two groups was observed using Bray–Curtis-based pCoA (Figure 3B). Next, we analyzed the microbiota composition and observed intergroup differences in species composition at different levels (Supplementary Figure S3B). At the phylum level, the relative abundance levels of the dominant taxa (˃0.5%) Bacteroides and Epsilonbacteraeota decreased, while those of Patescibacteria and Cyanobacteria increased. At the class level, the relative abundance of the dominant taxa (> 0.5%) Bacteroidia, Campylobacter, and Erysipelotrichia decreased, while that of Anaerolineae, Oxyphotobacteria, and Saccharimonadia increased. The relative abundance of the dominant orders (> 0.5%) Bacteroidales, Campylobacterales, and Erysipelotrichales decreased, while that of the orders Chloroplast, Clostridiales, and Saccharimonadales increased. At the family level, the relative abundance of the dominant taxa (> 0.5%) Bacteroidaceae, Erysipelotrichaceae, Helicobacteraceae, Muribaculaceae, and Prevotellaceae decreased, while that of Lachnospiraceae and Saccharimonadaceae increased. At the genus level, differential clustering and features were observed between the alcohol-FMT group and the control-FMT group. The relative abundance of the dominant genera (> 0.5%) Helicobacter, Bacteroides, Oscillibacter, Lachnospiraceae_UCG_001, Blautia, and Alloprevotella decreased, while that of the genera Lachnospiraceae_NK4A136_group, Ruminococcaceae_UCG_014, Ruminiclostridium_6, Ruminococcus_torques_group, and GCA_900066575 increased (Figure 3C). The heatmap shows the most differentially abundant features selected at the genus level between the alcohol-FMT and control-FMT mice (Figure 3D). Some of the changes described above were also observed between the alcohol group and the control group.

As described above, serum endotoxin levels were significantly elevated in the alcohol-FMT group. Pearson’s correlation analysis showed that serum endotoxin levels were correlated with the relative abundance of Lachnospiraceae_NK4A136_group, Helicobacter, Bacteroides, Candidatus_Saccharimonas, and Akkermansia, although the intergroup difference did not reach statistical significance (Pearson’s correlation coefficient, p ≥ 0.05). Pearson’s correlation analysis of sperm quality and the relative abundance of gut microbiota showed a significant negative correlation between sperm motility and the abundance of Lachnospiraceae_NK4A136_group (p = 0.025) (Figure 3E). These findings indicate that transplantation of the fecal microbiota from the alcohol group to normal mice also altered the gut microbiota and led to impaired sperm quality in otherwise normal mice. We suggest that Lachnospiraceae_NK4A136_group may be one of the most important species causing impaired sperm quality, although further research is needed to investigate the mechanism.

As described above, the difference in the gut microbiota may be an important factor contributing to impaired sperm quality in the alcohol-FMT group. We adopted KEGG pathway analysis to detect the relative abundance levels of metabolic pathways related to the gut microbiota to investigate how chronic drinking-induced intestinal dysbiosis affects spermatogenesis. Fifteen KEGG pathways differed between alcohol-FMT and control-FMT mice (Figure 4A). In particular, the primary and secondary bile acid biosynthesis pathways were significantly less enriched in the alcohol-FMT group, although the intergroup difference did not reach statistical significance (p = 0.081 and p = 0.73, respectively). The correlation analysis result of differentially enriched KEGG pathways with the species at the genus level showed that differentially abundant microbiota species were enriched in a series of metabolic pathways (Figure 4B), such as steroid biosynthesis, Wnt signaling pathway, primary bile acid biosynthesis, secondary bile acid biosynthesis, folate biosynthesis, and oxidative phosphorylation, all of which have been shown to be critical for spermatogenesis (Ding et al., 2020; Zhang et al., 2021). Next, we performed UHPLC–MS/MS to analyze the nontargeted metabolomics of the intestinal contents from the alcohol-FMT group and the control-FMT group (Supplementary Table 5). PLS-DA showed apparent metabolite clustering in both groups (Figure 4C). Transplantation of the fecal microbiota from the alcohol group to normal mice caused extensive changes in the gut metabolites in otherwise normal mice (Figure 4D). A total of 105 differentially altered metabolites were identified, including 22 upregulated metabolites and 83 downregulated metabolites in the alcohol-FMT group compared with the control-FMT group (Figure 4E). These metabolites were annotated to metabolic pathways related to amino acids, lipids and lipid molecules, fatty acids and conjugates, steroids, and flavonoids using the KEGG database, HMDB, and LIPID MAPS database (Figure 4F). The differentially altered metabolites were enriched in 20 KEGG pathways, including pentose and glucuronate interconversions (p = 0.0150), riboflavin metabolism (p = 0.0287), nicotinate and nicotinamide metabolism (p = 0.1638), morphine addiction (p = 0.0739), lysine degradation (p = 0.1638), glyoxylate and dicarboxylate metabolism (p = 0.0456), cortisol synthesis and secretion (p = 0.1115), citrate cycle (tricarboxylic acid (TCA) cycle) (p = 0.0287), caffeine metabolism (p = 0.0456), alanine, aspartate, and glutamate metabolism (p = 0.1370), and oxidative phosphorylation (p = 0.2065; Figure 4G), which was consistent with the KEGG analysis results of the 16S sequencing data. The significantly differentially altered metabolites enriched in these metabolic pathways included alpha-ketoglutaric acid, oxaloacetic acid, ribitol, riboflavin-5-phosphate, 1-methylxanthine, and 1-methyluric acid (Supplementary Table 6). Analyses of the 16S rDNA sequencing data and nontargeted metabolomics data identified a positive or negative correlation between 49 differentially abundant microbiota species and 50 differentially altered metabolites (Figure 4H), indicating that changes in the gut microbiota caused changes in the gut metabolites. Interestingly, the differentially altered metabolites were mainly lipids and amino acids, with a significantly less enriched bile acid metabolic pathway in the alcohol-FMT group. In summary, the differences in the gut microbiota between the alcohol-FMT group and the control-FMT group caused significant differences in gut metabolites, which is one possible explanation for the difference in sperm quality, although further research is needed to validate the results.

As described above, serum endotoxin levels were elevated in the alcohol-FMT group, which was correlated with certain microbiota, especially the dominant phyla Firmicutes and Bacteroides. We first analyzed intestinal inflammation in the recipient mice. H&E staining showed pronounced lymphocyte infiltration in the lamina propria of the small intestine in the alcohol-FMT group compared with the control-FMT group (Figure 5A). Immunohistochemistry and immunofluorescence staining also showed that a large amount of macrophages (Figures 5B,C) and T cells (Figure 5D) had infiltrated the lamina propria of the small intestine in alcohol-FMT recipient mice, suggesting that chronic drinking induced changes in the gut microbiota to promote inflammation. The testes and epididymis are where sperm are produced, mature, and are mobilized (Cornwall, 2009; Mäkelä et al., 2019), and they are the foundation of male fertility and reproductive health. Elevated endotoxin levels cause testicular and epididymal inflammation and affect spermatogenesis and sperm motility (Cao et al., 2010; Wang et al., 2019). Testicular immunohistochemistry also revealed that a large number of macrophages infiltrated the testicular interstitium in the alcohol-FMT group compared with the control-FMT group (Figures 5E,F), indicating the presence of inflammation. The results described above confirmed that FMT from the alcohol group to normal mice increased serum endotoxin levels and caused chronic inflammation in otherwise normal mice. We used antibody array technology to analyze the serum levels of 18 specific inflammatory cytokines in the alcohol-FMT and control-FMT groups. PCA showed differential clustering, with a significant difference along the PCO2 axis (responsible for 52.2% of the overall change rate, Figure 5G). All 18 inflammatory cytokines were present at higher levels in the alcohol-FMT group than in the control-FMT group, although the difference did not reach statistical significance in most cases (Figure 5H). In particular, serum IL-12p70 (p = 0.0148) and IL-22 (p = 0.0246) levels were significantly higher in the alcohol-FMT group, indicating significantly increased inflammation in this group compared with the control-FMT group. We also analyzed inflammatory cytokine levels in the testes and detected higher levels of the proinflammatory cytokines IL-2 and IL-Ib and lower levels of the anti-inflammatory cytokines TGF β1, IL-12p70, and IL-22 in the alcohol-FMT group than in the control-FMT group, although the differences did not reach statistical significance (p > 0.05).

Studies have demonstrated that chronic inflammation causes sustained activation of endothelial cells, which promotes angiogenesis, increased blood flow, and macrophage aggregation (Cao et al., 2010). We observed the testicular blood flow in each group using the Vevo 2,100 imaging system (Visual sonic Inc., Toronto ON, Canada) with a 30-MHz phased array transducer (MS400) at a frame rate of 235/s. The results indicated increased testicular blood flow in the alcohol-FMT group compared with the control-FMT group (Figure 5I), which, along with ultrasound images, confirmed that FMT from the alcohol group to normal mice increased testicular blood flow and the level of testicular inflammation in otherwise normal mice. In summary, FMT from the alcohol group to normal mice induced changes in the gut microbiota, which caused intestinal and testicular inflammation and ultimately impaired sperm quality in the alcohol-FMT group.

As described above, FMT from the alcohol group impaired sperm quality. We next investigated the mechanism of impaired sperm quality in the alcohol-FMT group by collecting testis samples from the alcohol-FMT group and the control-FMT group for transcriptome sequencing, with fragments per kilobase per million (FPKM) ≥ 1 as the threshold for gene expression. Among the eligible genes, 1,025 genes showed significant differences in expression (p ≤ 0.05, or |log2 FC| > 0), including 618 upregulated genes and 407 downregulated genes. We then used the H-cluster method with log2(FPKM + 1) to analyze differential expression and centralization correction to observe clustering (Figures 6A,B). Moreover, KEGG pathway analysis revealed a series of enriched pathways, such as ribosome (p = 3.14E-12), oxidative phosphorylation (p = 1.16E-11), Wnt signaling pathway (p = 0.0077), synthesis and degradation of ketone bodies (p = 0.0091), proteasome (p = 0.0403), thermogenesis (p = 1.52E-07), and retrograde endocannabinoid signaling (p = 0.0003) (Figure 6C), all of which play important roles in spermatogenesis and sperm quality. In particular, the expression levels of the mitochondrial genes MT-ND1 (mitochondrially encoded NADH:ubiquinone oxidoreductase core subunit 1), MT-ND2, MT-ND3, and MT-ND4, which encode different subunits of the mitochondrial dehydrogenase complex; oncogenes Myb and Myb12, which regulate cell growth and differentiation; E2F transcription factor 7 (E2F7) and matrix metallopeptidase 14 (MMP14), which regulate cell cycle progression; chromosomal structure maintenance protein genes structural maintenance of chromosomes 1A (Smc1a) and Smc3, which are involved in the regulation of meiosis; and the spermatocyte-specific regulatory factor TATA-binding protein (Tbp) were significantly decreased (Figure 6B).

Next, we performed differential KEGG pathway coenrichment analysis using metabolomics and transcriptomics data with KEGG pathways as the entry and the R language ggplot2 package to plot a KEGG coenrichment bubble chart of the differentially altered metabolites and the differentially expressed genes. Oxidative phosphorylation, the key pathway required for spermatogenesis, was significantly enriched in both transcriptomics and metabolomics data (p = 1.16E-11) (Figure 6D). Oxidative phosphorylation was also a differentially enriched metabolic pathway in the KEGG pathway analysis of the gut microbiota data. Next, we calculated the correlation coefficient r² and p-value to identify the correlations between the differentially expressed genes and the differentially altered metabolites with the R language mixOmics package and Pearson’s correlation analysis. The results indicated a significant correlation between 200 differentially expressed genes and 50 differentially altered metabolites (Figure 6E). In summary, FMT from the alcohol group to normal mice affected sperm quality by inducing intestinal dysbiosis and abnormal expression of testis-related genes in otherwise normal mice.