Adolescent exposure to polystyrene nanoplastics induces male reproductive damage via the microbiome-gut-testis axis

Abstract

Polystyrene nanoplastics (PS-NPs), are increasingly associated with reduced male fertility, yet the underlying mechanisms remain poorly defined. Here, we systematically unraveled a novel microbiome-gut-testis axis mediating PS-NPs–induced reproductive toxicity. Adolescent rats exposed to PS-NPs for 5 weeks induced dose-dependent testicular injury, characterized by disrupted spermatogenesis, and compromised blood-testis barrier. Single-cell atlases revealed spermatogenic arrest, abnormal immune microenvironment, and perturbed testicular cell communication upon exposure to PS-NPs. Furthermore, multi-omics analysis highlighted the activation of NF-κB/IL-17/HIF-1 and inhibition of PPAR-γ signaling, contributing to increased DNA damage and apoptosis, suppressed autophagy, and dysregulated energy–lipid metabolism. Additionally, PS-NPs exposure initiated gut microbial dysbiosis, significantly increasing pro-inflammatory bacteria, while reducing beneficial commensals. This microbial disruption compromised intestinal barrier integrity, leading to elevated circulating LPS levels. Subsequent activation of the TLR4/MyD88/NF-κB signaling pathway propagated inflammatory responses to testes. Crucially, FMT from PS-NPs–exposed donors reproduced the damage in healthy recipients, thus suggesting gut microbiota as a causal mediator. Therapeutically, DI intervention effectively mitigated the reproductive toxicity by restoring gut barrier integrity, rebalancing microbial communities, and suppressing inflammation. Our findings unveil a gut microbiome-centric mechanism for nanoplastic-induced male reproductive toxicity, and identify DI as a promising therapeutic candidate, accordingly providing critical insights for environmental risk assessment.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s12951-026-04069-y.

Introduction

Polystyrene nanoparticles (PS-NPs) are nanoplastics widely used as model particles in materials science and increasingly detected in environmental matrices and food webs because of the fragmentation of larger plastics and pervasive human activities [1]. Their small size, high specific surface area, and hydrophobicity confer prolonged environmental persistence and a strong tendency to adsorb co-contaminants and interact with biological interfaces, contributing to bio-accumulation across trophic levels [2–4]. Growing evidence has suggested that PS-NPs exposure may pose serious health risks, with effects on reproductive health being a particular cause for concern. Epidemiological studies have highlighted a global crisis in male reproductive health. A systematic review reported a 50–60% decline in sperm counts among men globally between 1973 and 2011, with recent updates indicating an accelerated decline of approximately 2.64% per year since 2000 [5]. Recent studies link PS-NPs exposure with disrupted endocrine function, reduced fertility, and poor developmental and reproductive outcomes in males and females [6]. PS-NPs can act as endocrine disruptors, interfering with hormone synthesis, transport, or receptor signaling along the hypothalamic-pituitary-gonadal axis [7–9]. Disturbances noted to date include disrupted estrous or menstrual cycling, defective spermatogenesis, diminished sperm quality, delayed sexual maturation, and compromised gonadal structure and function. Animal and in vitro studies further demonstrate that PS-NPs can penetrate biological barriers, including the intestinal epithelium, blood-testis barrier (BTB), and placenta, thereby eliciting developmental and reproductive toxicity during critical windows of exposure [10]. These effects are proposed to be mediated through oxidative stress, inflammatory responses, mitochondrial dysfunction, and dysregulation of steroidogenic pathways, collectively with potential lifelong impacts on fertility and reproduction. However, the mechanisms whereby PS-NPs disrupt spermatogenesis remain incompletely defined, and the spatiotemporal or cell-type-specific effects in the testis are virtually uncharacterized.

The gut microbiome is a complex community of microorganisms inhabiting the gastrointestinal tract and has emerged as a key regulator of systemic physiology, including reproduction, through a bidirectional “microbiome-gut-reproduction axis” [11]. Gut microbes modulate circulating sex hormones via enterohepatic cycling, produce bioactive metabolites, and shape mucosal and systemic immune tone, hence impacting reproductive outcomes [12]. Perturbations in microbial composition and function have been associated with hormonal imbalance, subfertility, and reproductive disorders [13, 14]. Notably, PS-NPs exposure itself has been shown to perturb the gut microbiome [15]. Experimental data indicate that PS-NPs can alter microbial diversity and community structure, decrease short-chain fatty acid–producing taxa, and impair epithelial tight junctions, leading to increased intestinal permeability, low-grade inflammation, and endocrine dysregulation [16]. Such shifts may disrupt immune and metabolic pathways crucial for maintaining reproductive homeostasis. For example, gestational exposure to PS-NPs triggers maternal gut dysbiosis, causing placental NAD⁺ deficiency, ferroptosis, and oxidative stress, ultimately leading to adverse pregnancy outcomes [17].

Itaconate is an IRG1-derived macrophage immunometabolite characterized by multimodal anti-inflammatory activity through coordinated immunometabolic modulation [18–21]. Cell-permeable itaconate esters such as dimethyl itaconate (DI) pharmacologically augment this pathway and have shown protection in preclinical models of endotoxemia, colitis, and sterile inflammation by restoring epithelial barrier integrity and curbing cytokine cascades [22, 23].

While prior research has described the environmental behavior of PS-NPs and their organ toxicities, the specific mechanistic pathways connecting PS-NPs–induced gut dysbiosis with male reproductive outcomes remain poorly defined. In particular, whether and how PS-NPs–driven alterations in gut microbial communities influence testicular physiology, hormone biosynthesis, and spermatogenesis is largely unexplored. Here, we test the hypothesis that PS-NPs exposure perturbs the gut microbiota and compromises the intestinal barrier, thereby increasing endotoxemia and activating TLR4/MyD88–centered signaling in both intestine and testis to drive BTB disruption, energetic dysfunction, and impaired spermatogenesis. We integrate 16 S rRNA profiling with single-cell and bulk RNA-seq of testes and untargeted metabolomics to delineate microbiota-metabolite-testis signaling axes. Causality is probed by fecal microbiota transplantation (FMT) from PS-NPs–exposed donors into microbiota-depleted recipients. Finally, we evaluate DI as an experimental intervention to remodel the gut microbiota, blunt LPS/TLR4/MyD88 and NF-κB signaling, restore gut and testicular barrier integrity, and rescue reproductive endpoints. This multi-layered approach bridges critical knowledge gaps at the interface of nanoplastic toxicology, reproduction, and the microbiome, and provides evidence for a mechanistic pathway with translational implications for nanoplastic risk assessment and therapy.

Materials and methods

PS-NPs Preparation

PS-NPs, both unmodified (Cat#: 6-1-0008) and green fluorescent (Cat#: 7-3-0008), with a nominal diameter of 80 nm, were procured from the Tianjin Baseline Chromatography Technology Development Center (Keyuan Science and Trade Building, Tianjin, China). The stock suspensions were stored at 4 °C. Prior to experimentation, the PS-NPs were vortexed to ensure homogeneity. Working suspensions at desired concentrations were then prepared by diluting the stock in either ddH₂O or saline (Beyotime, ST341-500 mL).

Characterization of PS-NPs

PS-NPs were ultrasonicated for 30 min to ensure dispersion and subsequently characterized by transmission electron microscopy (TEM; FEI Talos F200X G2). Hydrodynamic size and ζ-potential were determined by dynamic light scattering using a Zetasizer NanoZS90 instrument (Malvern, UK). The chemical properties of PS-NPs were analyzed by Fourier transform infrared spectroscopy (FTIR; Thermo, Germany) and Raman spectroscopy (HORIBA LabRAM Soleil, France) to identify characteristic functional groups. The endotoxin levels associated with PS-NPs were quantified using a Chromogenic Limulus Amebocyte Lysate (LAL) Assay Kit (Beyotime, C0276S), following the manufacturer’s protocol.

PS-NPs exposure and dimethyl Itaconate (DI) treatment

Male Sprague-Dawley rats (3 weeks old) were obtained from the Weitong Lihua Experimental Animal Center (Beijing, China). The rats were maintained in an environment free of specific pathogens, with unrestricted access to a standard commercial chow diet (formulated with corn, soybean meal, and wheat bran) and water. Following a one-week adjustment period, the rats were randomly assigned to one of three categories: control, low-dose PS-NPs (1.5 mg/kg/day), and high-dose PS-NPs (70 mg/kg/day), with 18 rats per group. Dosages of PS-NPs were determined based on estimated human environmental exposure levels (the basis for dose selection was provided in Supplementary methods). We recorded daily measurements of body weight, food intake, and behavioral status to monitor the health of the rats. Rats were weighed and euthanized via cervical dislocation under anesthesia, twenty-four hours following the final dose. Notably, rats exposed to non-fluorescently labeled PS-NPs were assigned to scRNA-seq, bulk RNA-seq, metabolomic profiling, 16 S rRNA sequencing, and downstream validation assays, whereas those given green-fluorescent PS-NPs were used exclusively for localization imaging on cryosections. Blood, testis, epididymis, fecal, and intestinal tissues were collected. The right testis was cryopreserved in liquid nitrogen and kept at − 80 °C for future analysis, whereas the left testis was preserved in 4% paraformaldehyde for histology. The Institutional Animal Care and Use Committee at Nanjing Medical University approved this research (IACUC-2510084).

To evaluate the protective efficacy of DI (Sigma-Aldrich, Cat. 617527) against PS-NPs–induced testicular injury, rats were administered daily intraperitoneal injections of DI at a validated dosage of 50 mg/kg (dissolved in sesame oil) for 30 days [24]. This dosage was selected based on previous pharmacodynamic studies demonstrating that 50 mg/kg of DI effectively suppressed systemic inflammation in rodent models of colitis and metabolic disorders. Following this intervention, blood, semen, intestinal segments, and testicular tissues were systematically collected for downstream analysis.

Single-cell dissociation, library generation, and sequencing

Single-cell suspensions were prepared following the published protocols [25]. Testes were harvested from two randomly selected rats in both the control and high-dose groups. Seminiferous tubules were enzymatically dissociated, and the resulting cell suspensions were filtered, washed, and re-suspended for droplet-based scRNA-seq. 3’-tag libraries were constructed and sequenced to a depth of ≥ 20,000 reads per cell utilizing 150-bp paired-end sequencing.

After aligning the raw reads to the rn6 genome and quantifying gene counts with Cell Ranger v5.0.1 (10x Genomics), the resulting expression matrix was analyzed with the Seurat package [26]. Genes detected in fewer than three cells were removed, as were cells with > 10% mitochondrial reads, < 500&>8000 expressed genes, or doublet scores flagged by the scds package [27]. Highly variable genes were selected for unsupervised clustering and cell-type annotation. PCA was performed, and the first 30 PCs (resolution = 0.5) were supplied to Seurat’s FindClusters function to define cell clusters. Two-dimensional visualization was generated with UMAP.

Bulk RNA sequencing analysis

Total RNA was extracted with Trizol Reagent, and quality-checked by NanoDrop. For RNA sample preparation, 3 µg of RNA was used to construct sequencing libraries with TruSeq RNA sample preparation kits. The mRNA was purified, fragmented, and reverse transcribed to cDNA. The cDNA pieces were adenylated, linked with Illumina PE adapters, purified, and amplified through PCR. The end result was measured with a Bioanalyzer 2100 device. Sequencing was performed on the Nova Seq 6000 platform, producing raw data in FASTQ format. Low-quality reads and adaptors were trimmed using fastp software to obtain clean reads. Trinity software was used for de novo transcript assembly, generating transcript sequences in FASTA format. Gene function annotation was conducted using databases such as GO, and KEGG. DESeq was employed to pinpoint differentially expressed genes (DEGs), applying thresholds of |log₂FoldChange| greater than 0.5 and a P-value below 0.05. The heatmap package in R was utilized to create heat maps and conduct cluster analyses.

Metabolomics analysis

Frozen lyophilized testis samples were ground with tungsten beads and were treated with the methanol, acetonitrile, water solution (2:2:1). The testis samples were subjected to an ultrasonic bath and afterwards, the samples were spun down. The supernatant was collected, dried under vacuum, and analyzed by UPLC-ESI-Q-orbitrap-MS. For the mass spectrometry, separation was done by ACQUITY UPLC HSS T3 and the mobile phase was changed with time by putting 0.1% formic acid in the water and 100% acetonitrile. The mass spectrometer was operated in positive and negative mode at spray voltages of 3.8 kV and 3.2 kV, respectively with the capillary at 320 °C. Full MS scans were collected over an m/z range of 70-1050 at 70,000 resolution and MS/MS scans at 17,500 resolution. QC samples were injected to normalize the data, with blanks and QC samples injected every four samples. Data handling was carried out using MS-DIAL for peak alignment and correction of the retention time, along with extracting peak area. Candidate metabolites were identified by mass and MS/MS data, utilizing both an in-house metabolite standard library and publicly available resources. The analysis only considered variables in which any given set contained more than 50% non-zero value entries. R software was then used to carry out multivariate data analyses. Overfitting of the models was then assessed using permutation testing. Lastly, to determine biologically relevant pathways from the differential metabolites, we performed KEGG analysis which considered only pathways statistically significant at P < 0.05.

Supplementary methods

Detailed descriptions of the additional methods are provided in the Supplementary methods.

Results

Characterization of PS-NPs and analysis of their dose-dependent intestinal uptake and testicular deposition

TEM revealed near-spherical PS-NPs with smooth surfaces and minimal aggregation (Fig. 1A), and single-particle sizing yielded a unimodal, close-to-Gaussian diameter distribution in the 80 nm range (Fig. 1B). FTIR spectra revealed the characteristic signature of an aqueous dispersion: broad O–H stretching (3258 cm⁻¹) and bending (1635 cm⁻¹) bands, together with the collective vibration of the hydrogen-bond network near 485 cm⁻¹. In addition, a distinct peak at 1045 cm⁻¹, assigned to the C–O stretching mode of the polymer backbone, confirmed the presence of nano-sized polystyrene fragments in the sample (Fig. 1C). Moreover, Raman spectroscopy was employed to verify the chemical identity of the PS-NPs (Fig. 1D). Zeta-potential measurements revealed a surface charge of − 21.4 mV (Fig. 1E), and dynamic light scattering gave a Z-average hydrodynamic diameter of 63.05 nm in distilled water (Fig. 1F). The endotoxin contamination of the PS-NPs was assessed and was found to be negligible (Fig. S1A).

Fig. 1.

Physicochemical characterization of PS-NPs and their dose-dependent intestinal uptake and testicular deposition in vivo. (A) TEM of PS-NPs shows near-spherical particles. Right, magnified view of the boxed area. (B) Size distribution of PS-NPs calculated by DLS analysis. (C) FTIR spectroscopy revealed the characteristic transmittance profile of PS-NPs across the infrared wavenumber range. (D) Raman spectrum for PS-NPs. (E) Zeta potential distribution of PS-NPs. (F) Hydrodynamic size distribution profile of PS-NPs. (G, H) Direct visualization by TEM identified the dose-dependent retention of undissolved PS-NPs in colonic tissues (G) and testicular tissues (H) (red arrows). (I, J) Green fluorescent PS-NPs showed dose-dependent accumulation in colonic tissues (I) and testicular tissues (J) (white arrows). PS-NPs, polystyrene nanoplastics; TEM, transmission electron microscope; DLS, dynamic light scattering; FTIR, Fourier transform infrared

Since the internalization and tissue distribution of PS-NPs governed their biotoxicity, we characterized unmodified particles via TEM (Fig. 1G, H), and subsequently performed localization of green-fluorescent PS-NPs in tissue sections using DAPI counterstaining and high-resolution fluorescence microscopy (Fig. 1I, J). The results demonstrated that the burden of internalized PS-NPs progressively increased in the colonic and testicular tissues with higher exposure doses.

PS-NPs exposure impaired spermatogenesis, lowered testosterone, and disrupted the blood–testis barrier and Sertoli-cell support

After 5 weeks gavage, the body weight and food intake were decreased after PS-NPs exposure. Daily qualitative health monitoring demonstrated no significant changes in behavior (Fig. S1B, C). Organ size and weight were measured, and tissue morphology was assessed. While no significant alterations in testicular or epididymal size and weight were observed in the low-dose group, the high-dose group exhibited markedly smaller organs with reduced weight (Fig. 2A-D). Sperm quality parameters, including quantity, vitality, normal morphology, and motility, were significantly compromised in the PS-NPs exposure groups (Fig. S2A-D). Histological examination corroborated the functional deficits, with H&E staining revealing marked disorganization of seminiferous tubules, evident loss of spermatogenic cells, prominent vacuolization, and a pronounced thinning of the epithelium following PS-NPs exposure (Fig. 2E). Masson staining demonstrated that the thickening of the interstitial and basement membranes in the testes was more pronounced in the PS-NPs groups compared to the controls (Fig. S3). Furthermore, serum testosterone decreased in a dose-dependent manner (Fig. 2F). Besides, immunofluorescence of germ cell lineage markers demonstrated that PS-NPs treatment reduced c-KIT-positive spermatogonia (Fig. 2G), SCP3-positive spermatocytes (Fig. 2H), and MVH-positive germ cells (Fig. 2J) in a dose-dependent manner. PNA staining, which specifically labels the sperm acrosome, was significantly reduced in PS-NPs groups relative to controls (Fig. 2I). CYP11A1 expression, critical for steroidogenesis, was also reduced, suggesting impaired steroidogenic function (Fig. S4A). However, SOX9-positive cells, marking Sertoli cells, did not exhibit significant differences (Fig. S4B).

Fig. 2.

PS-NPs exposure impaired spermatogenesis, lowered testosterone, disrupted the BTB and Sertoli-cell support. (A, C) Representative gross images of testes (A) and epididymides (C). (B, D) Bar graph showing the average testis weight (B) and epididymis weight (D) across the three groups. Data are presented as mean ± SEM (n = 8). (E) H&E of testicular tissue. Scale bars, 200 μm. (F) Serum testosterone concentration decreased with PS-NPs dose. Data are presented as mean ± SEM (n = 8). (G-J) IF images showing c-KIT (spermatogonia marker) expression (G), SCP3 (spermatocyte marker) expression (H), PNA in acrosome structures (I), and MVH (germ cell marker) expression (J) across the three groups. Right: quantification. Scale bars, 100 μm. Data are presented as mean ± SEM (n = 4). (K-P) IF images showing the expression of BTB-related proteins, including the tight-junction proteins ZO-1 (K) and Occludin (M), the adherens-junction proteins N-cadherin (L) and E-cadherin (N), the gap-junction protein Cx43 (O), and the intermediate filament Vimentin (P). Right: quantification. Scale bars, 100 μm. Data are presented as mean ± SEM (n = 4). (Q, R) The characteristic bands and quantitative analysis of BTB-related proteins in PS-NPs exposed groups. Data are presented as mean ± SEM (n = 3). (S, T) The characteristic bands and quantitative analysis of SCF, GDNF, and BMP4 proteins. Data are presented as mean ± SEM (n = 3). One-way ANOVA followed by Tukey’s post hoc test was used. Ns means no significance. *P < 0.05, **P < 0.01. ***P < 0.001. PS-NPs, polystyrene nanoplastics; BTB, blood–testis barrier; H&E, Hematoxylin and eosin; IF, Immunofluorescence

Furthermore, key components of the BTB and cell-cell communication were compromised by PS-NPs. The tight-junction proteins ZO-1 and Occludin, the adherens-junction proteins N-cadherin and E-cadherin, the gap-junction protein Cx43, and the intermediate filament Vimentin all showed visibly weaker immunofluorescence in PS-NPs–treated groups, with corresponding quantifications showing significant reductions (Fig. 2K-P). Consistently, WB validated the downregulation of these proteins in the rat primary Sertoli cells (Fig. 2Q, R). Further analysis revealed that PS-NPs exposure also downregulated critical growth factors, including SCF, GDNF, and BMP4. These proteins showed a clear, dose-dependent decrease in the PS-NPs–exposed groups compared to the controls (Fig. 2S, T).

Construction of a single-cell transcriptomic landscape of testis after PS-NPs exposure

To closely investigate how PS-NPs affect the testis at the cellular level, we conducted scRNA-seq. After applying stringent quality control measures and removing cell doublets, we categorized the cells into 21 distinct clusters (Fig. 3A). A total of 19,653 cells and 12,931 cells were yielded from the control and PS-NPs groups, respectively. Through joint UMAP dimensional reduction, we identified nine transcriptionally unique cell populations that were consistently present across both datasets (Fig. 3B). These clusters were annotated as major testicular cell types based on the expression of canonical markers: Dazl (spermatogonia), Sycp3 (spermatocytes), Odf3 (round spermatids), Prm2 (elongating spermatids), Clu (Sertoli cells), Cyp17a1 (Leydig cells), Myh11 (myoid cells), Apoe (macrophages), and Pecam1 (endothelial cells) (Fig. 3C). A dot-plot confirmed the specificity and detection frequency of these markers within each cluster, suggesting a reliable cellular reference framework for further comparison between the control and exposed groups (Fig. 3D). Next, we examined differential gene expression in each cell type using volcano plots (Fig. 3E). GO analysis revealed PS-NPs exposure led to a coordinated downregulation of genes involved in mitochondrial energy production and spermatogenesis. KEGG analysis extended on these findings, highlighting oxidative phosphorylation and thermogenesis, alongside ribosome and protein processing in endoplasmic reticulum pathways, reflecting concomitant disturbance in mitochondrial function, protein homeostasis, and cellular energy metabolism (Fig. S5A). Moreover, scRNA-seq revealed a dramatic reduction in post-meiotic germ cells—especially elongating and round spermatids—in the high-dose group. Notably, we observed a concomitant increase in the population of testicular macrophages (Fig. S5B, C).

Fig. 3.

Single-cell atlas of the testis revealed cell-type shifts and pathway reprogramming after PS-NPs exposure. (A) Experimental workflow. Rats received saline or PS-NPs. Testes were collected, dissociated, and single cells were 10× barcoded, followed by sequencing and downstream bioinformatic analysis. (B) UMAP embeddings of the integrated dataset (left, combined; right up, Control; right down, PS-NPs). Each dot was a cell, colored by annotated identity. (C) Feature plots showing canonical marker expression used for annotation: Dazl (spermatogonia), Sycp3 (spermatocytes), Odf3 (round spermatids), Prm2 (elongating spermatids), Clu (Sertoli cells), Cyp17a1 (Leydig cells), Myh11 (myoid cells), Apoe (macrophages), and Pecam1 (endothelial cells). (D) Dot plot of marker genes across different clusters. Dot size indicated detection rate; color scale reflected average scaled expression, confirming specificity of the chosen markers for each cell type. (E) Volcano plots of differentially expressed genes (DEGs) between Control and PS-NPs within each annotated cell type. Red and blue points denoted significantly up- and down-regulated genes, respectively

PS-NPs exposure weakened and rewired testicular cell–cell communication networks

To delineate intercellular communication networks among testicular cell types, the CellChat algorithm was employed to infer putative ligand–receptor interactions. The ligand–receptor inference showed a contraction of the intercellular connectome after PS-NPs exposure, with fewer predicted edges (Fig. 4A, B) and reduced aggregate interaction strength (Fig. 4C, D).

Fig. 4.

CellChat-based mapping showed contraction and rewiring of testicular cell–cell communication after PS-NPs exposure. (A) Difference network for the number of inferred interactions (PS-NPs vs. Control). Edge thickness reflected the change in ligand–receptor edges among nine cell types. (B) Bar plot of the total number of interactions per group: Control: 106; PS-NPs: 76. (C) Difference network for aggregate interaction strength (sum of edge weights). (D) Bar plot of total interaction strength: Control: 3,423; PS-NPs: 2,565. (E, F) Analysis of differential information flow highlighted key signaling pathways altered between control and PS-NPs conditions. Pathways exhibiting stronger signaling in control and PS-NPs groups were color-coded in red and cyan, respectively. (G, H) Heatmap of overall signaling patterns in control (G) and PS-NPs (H) groups. (I, J) Heatmap of incoming signaling in control (I) and PS-NPs (J) groups. (K, L) Scatter plots showed the prominent cellular senders and receivers in the control (K) and PS-NPs (L) groups, with communication strength quantified by signals sent (x-axis) and received (y-axis)

Ranking pathways by relative information flow revealed a group-specific portfolio: the control group preferentially engaged endothelial/adhesion and ECM programs, such as CD39, VCAM, CDH, THBS, JAM, COLLAGEN, ESAM, GRN, and PTN, whereas the PS-NPs–exposed group showed relatively higher contribution in CLDN, PTPRM, NRG and BMP signaling (Fig. 4E, F). Furthermore, heatmaps of overall signaling patterns showed attenuated outgoing communications from somatic compartments toward germ cells in the PS-NPs group (Fig. 4G, H). Consistently, analysis of incoming signaling revealed a concerted attenuation across germ cell populations, affecting multiple ligand-receptor families (Fig. 4I, J). Across all cells, the macrophages emerged as the principal signaling node, with the highest incoming and outgoing strengths regardless of PS-NPs exposure status (Fig. 4K, L). However, the outgoing interactions and strengths were both reduced after PS-NPs exposure (Fig. S6A-D). The specific receptor-ligand interactions involving macrophages were shown in Fig. S6E. Taken together, these changes point to a breakdown in the testicular communication network.

PS-NPs exposure caused DNA damage, increased apoptosis, and suppressed autophagy in testes

Single-cell gene-set scoring revealed a global increase in DNA Damage signaling after PS-NPs exposure. The DNA Damage score shifted to higher values in the PS-NPs group (Fig. 5A), and this increase was consistently observed across spermatogonia, spermatocytes, and elongating spermatids (Fig. 5B). Besides, the Apoptosis program was markedly enhanced. The Apoptosis gene set score increased in the PS-NPs–treated group (Fig. 5C), and was elevated in each germ-cell compartment (Fig. 5D). Furthermore, Autophagy was attenuated rather than activated. The Autophagy gene set score was significantly decreased in the PS-NPs group (Fig. 5E), with concordant reductions across spermatocytes, round spermatids, and elongating spermatids (Fig. 5F). Gene set score also revealed a global down-shift of the Spermatid Development program in PS-NPs–exposed testes, with lower median scores across all cells (Fig. S7). Furthermore, the oxidative phosphorylation gene set score was suppressed in the PS-NPs group, concomitant with a reduction in Sertoli cell score, potentially compromising the energy supply for spermatogenesis (Fig. S8). Similarly, suppression of the bile acid metabolism gene set coincided with a lowered Leydig cell score, likely affecting testosterone synthesis (Fig. S9).

Fig. 5.

PS-NPs exposure elevated DNA damage and apoptosis while suppressing autophagy in the testis. (A) Density plot showing elevated gene set scores for DNA Damage-related pathways. (B) Violin plots displaying DNA Damage-related gene set scores across different spermatogenic cell types. (C) Density plot showing elevated gene set scores for Apoptosis-related pathways. (D) Violin plots displaying Apoptosis-related gene set scores across different spermatogenic cell types. (E) Density plot showing decreased gene set scores for Autophagy-related pathways. (F) Violin plots displaying Autophagy-related gene set scores across different spermatogenic cell types. (G) IF for γ-H2AX in testicular tissues, with quantification of γ-H2AX–positive cells per tubule. Scale bars, 100 μm. Data are presented as mean ± SEM (n = 4). (H) TUNEL assay showing DNA fragmentation due to apoptosis in testicular tissues, with quantification of TUNEL-positive nuclei per tubule. Scale bars, 100 μm. Data are presented as mean ± SEM (n = 4). (I, J) IHC for Caspase 3 (I) and Caspase 9 (J), with quantitative analysis of Caspase 3 and Caspase 9 staining intensity. Scale bars, 200 μm. Data are presented as mean ± SEM (n = 4). (K-M) IF for autophagy markers LC3B (K), P62 (L), and Beclin1 (M) in testicular tissues, with quantitative analysis of LC3B, P62, and Beclin1 fluorescence intensity. Scale bars, 100 μm. Data are presented as mean ± SEM (n = 4). One-way ANOVA followed by Tukey’s post hoc test was used. Ns means no significance. *P < 0.05, **P < 0.01, ***P < 0.001. PS-NPs, polystyrene nanoplastics; IF, Immunofluorescence; IHC, Immunohistochemistry

At the tissue level, these transcriptomic changes were validated by immunostaining. Immunofluorescence results showed a clear increase in γ-H2AX-positive nuclei in PS-NPs groups, signaling elevated levels of DNA double-strand breaks (Fig. 5G). TUNEL staining revealed a significant rise in germ-cell apoptosis following exposure (Fig. 5H). IHC demonstrated higher expression of Caspase 3 and Caspase 9, further confirming activation of apoptotic pathways (Fig. 5I, J). Additionally, the expression profile of autophagy markers indicated suppressed autophagic initiation and impaired flux. Levels of LC3B and Beclin1 were decreased in PS-NPs–treated groups, whereas p62 accumulated significantly (Fig. 5K-M).

PS-NPs exposure activated immune and inflammatory responses and reprogrammed key pathways in testes

To comprehensively characterize the transcriptional landscape, we combined bulk RNA-seq with single-cell RNA sequencing. Bulk RNA-seq analysis uncovered significant gene expression changes, with 128 genes upregulated and 96 genes downregulated compared to controls (Fig. 6A). Hierarchical clustering of these DEGs produced a clear separation between exposed and control groups (Fig. 6B). GO enrichment highlighted biological processes such as DNA packaging complex, and positive regulation of cell killing (Fig. 6C). KEGG analysis indicated that HIF-1, NF-κB, IL-17, and PPAR signaling pathways were prominently affected in the PS-NPs group (Fig. 6D). Besides, single-cell gene-set scoring revealed elevated immune and inflammatory activation, especially in macrophages, which are key mediators of testicular immunity. Both the Immune and Inflammatory Response scores were significantly higher in the PS-NPs–exposed group (Fig. 6E-H).

Fig. 6.

PS-NPs exposure triggered testicular immune and inflammatory responses by dysregulating the HIF-1/NF-κB/IL-17 and PPAR-γ signaling. (A) Volcano plot illustrating DEGs in the control (Con) and PS-NPs–treated groups via bulk RNA-seq. (B) Heatmap representing hierarchical clustering of the DEGs. (C) GO enrichment highlighting significantly altered biological processes, such as DNA packaging complex and positive regulation of cell killing. (D) KEGG analysis indicating disrupted significant signaling pathways, such as HIF-1, NF-κB, IL-17, and PPAR signaling. (E, G) Density plot showing elevated gene set scores for Immune (E) and Inflammatory Response-related (G) pathways in PS-NPs–treated groups. (F, H) Violin plots displaying Immune (F) and Inflammatory Response-related (H) gene set scores in macrophages. (I-N) IF images for F4/80 (I), CD3 (J), TNF-α (K), IL-17 A (L), IL-6 (M) and NLRP3 (N) expression across the three groups. Right: quantification. Scale bars, 100 μm. Data are presented as mean ± SEM (n = 4). (O-T) IHC staining for Cox-2 (O), CCL2 (P) IL-18 (Q) NF-κB (R), PPAR-γ (S), and HIF-1α (T) expression across the three groups. Right: quantification. Scale bars, 200 μm. Data are presented as mean ± SEM (n = 4). (U) Western blot and quantitative analysis validating the upregulation of IL-17 A, CCL2, NF-κB, and NLRP3 in PS-NPs–exposed groups. Data are presented as mean ± SEM (n = 3). (V) Western blot and densitometry showing reduced PPAR-γ protein in PS-NPs–exposed groups. Data are presented as mean ± SEM (n = 3). (W) Western blot and densitometry showing elevated HIF-1α protein in PS-NPs–exposed groups. Data are presented as mean ± SEM (n = 3). One-way ANOVA followed by Tukey’s post hoc test was used. Ns means no significance. *P < 0.05, **P < 0.01, ***P < 0.001. PS-NPs, polystyrene nanoplastics; DEGs, differentially expressed genes; IF, Immunofluorescence; IHC, Immunohistochemistry

Furthermore, tissue validation confirmed immune cell infiltration and cytokine upregulation. There was a marked increase in F4/80-positive macrophages and CD3-positive T cells in PS-NPs–treated testes compared to controls (Fig. 6I, J). Pro-inflammatory mediators, including TNF-α, IL-17 A, IL-6, and the inflammasome component NLRP3, displayed stronger signals after PS-NPs exposure (Fig. 6K-N). Moreover, COX-2, CCL2, IL-18, and NF-κB were also elevated (Fig. 6O-R). WB analysis validated these findings, confirming the significant upregulation of IL-17 A, CCL2, NLRP3, and NF-κB in PS-NPs–exposed groups compared to controls (Fig. 6U). PS-NPs exposure led to significant disruptions in key signaling pathways in testicular tissue. IHC and WB revealed that PPAR-γ was reduced in PS-NPs groups (Fig. 6S and V). Conversely, HIF-1α, was markedly elevated in groups treated with PS-NPs, as evidenced by stronger staining and higher protein levels (Fig. 6T and W).

PS-NPs exposure reprogrammed cell-type pathways and remodeled the testicular metabolome toward energy and lipid dysregulation

To better understand how PS-NPs disrupt testicular function at the cellular level, we performed a detailed cell-type-specific pathway enrichment analysis, which revealed widespread metabolic disturbances across both somatic and germ cells (Fig. 7A-D and Fig. S10A-E). To validate and expand these transcriptomic findings, metabolomics was performed on testicular tissue. The consistent overlap observed in the total ion and basepeak chromatograms of QC samples confirmed the good reproducibility and robustness of the metabolomics platform (Fig. S11 and S12). OPLS-DA showed a clear separation between two groups (Fig. 7E), with robust performance confirmed by permutation testing (Fig. 7F). Chemical classification of the altered metabolites revealed major shifts in carboxylic acids, glycerophospholipids, fatty acyls, organooxygen compounds, and steroid derivatives (Fig. 7G). The heatmap provided a closer look at the individual metabolic changes, revealing widespread shifts across several pathways (Fig. 7H). KEGG analysis emphasized broad disturbances in energy metabolism, amino acid metabolism, lipid metabolism, and carbohydrate metabolism (Fig. 7I). Mapping onto the TCA cycle suggested an overall attenuation of oxidative metabolism (Fig. 7J). Quantitatively, acetyl-CoA and NAD were reduced in PS-NPs–treated testes (Fig. 7K, L), accompanied by a decline of the high-energy buffer phosphocreatine (Fig. 7M). Metabolomics also uncovered changes in polyamine and lipid transport. Spermidine—a polyamine involved in cell growth and sperm maturation—was reduced. Several lipid-related molecules essential for membrane structure and sperm motility also declined, including linoleic acid, L-acetylcarnitine, butyrylcarnitine, and LysoPA (16:0/0:0) (Fig. 7N-R).

Fig. 7.

Cell-type pathway reprogramming and metabolomic remodeling of the testis after PS-NPs exposure. (A–D) Cell-type–specific pathway enrichment of differentially expressed genes (DEGs) from scRNA-seq. Bubble plots summarized enriched GO terms and KEGG pathways for Sertoli cells (A), Leydig cells (B), peritubular myoid cells (C), and elongating spermatids (D). (E) The OPLS-DA model demonstrated distinct separations between control (Con) and PS-NPs–treated groups. (F) Permutation testing validated the robustness of the OPLS-DA model. (G) Chemical class distribution of differential features. (H) Heatmap detailing significant alterations in metabolite profiles between control and PS-NPs–treated groups. (I) KEGG pathway enrichment of altered metabolites. (J) TCA cycle map with directional changes. Schematic indicates attenuation of oxidative metabolism with reductions in acetyl-CoA and NAD (red arrows). (K–R) quantification of key metabolites associated with spermiogenesis. Bar graphs showing significant reductions in acetyl-CoA (K), NAD (L), phosphocreatine (M), spermidine (N), linoleic acid (O), L-acetylcarnitine (P), butyrylcarnitine (Q), and LysoPA (16:0/0:0) (R) in PS-NPs–exposed testes compared with controls. Data are presented as mean ± SEM (n = 5). T test was used

PS-NPs exposure reshaped gut microbiota and impaired intestinal barrier integrity

To explore whether PS-NPs influence gut microbial balance, we conducted 16 S rRNA sequencing on fecal samples. The results revealed substantial shifts in microbial composition. A Venn diagram highlighted large but distinct microbial communities between the groups, with 2,561 shared OTUs, but 415 and 527 unique to control and PS-NPs groups, respectively (Fig. 8A). PS-NPs exposure specifically increased microbial richness, as indicated by elevated Chao-1 and observed species indices. However, the unchanged Shannon diversity suggests that this increased richness did not alter community evenness (Fig. 8B-D). Beta-diversity analyses (PCoA, NMDS, and UPGMA clustering) showed a clear separation in microbial community structure, underscoring a strong impact of PS-NPs on gut microbiota composition (Fig. 8E and Fig. S13). Additionally, at the phylum level, Firmicutes and Bacteroidota dominated both groups (Fig. 8F). At the genus level, significant disruptions were observed (Fig. 8G). The PS-NPs–exposed group exhibited increased abundance of inflammation-associated bacteria such as Bacteroides and Escherichia-Shigella, while beneficial microbes like Lachnospiraceae_UCG-008, Lachnospira, and Lactobacillus were significantly reduced (Fig. 8H and Fig. S14). The LEfSe identified key microbial clades driving these group differences (Fig. 8I), and a cladogram illustrated the altered taxonomic tree from phylum to genus (Fig. 8J). Network analysis revealed a rewiring of bacterial interactions in the PS-NPs group, with pro-inflammatory taxa gaining more connections, suggesting a shift in microbial cooperation or dominance (Fig. 8K). Moreover, the functional potential of the microbial communities was inferred using PICRUSt2 and subsequently mapped to KEGG. This analysis identified functions associated with various metabolic processes, such as biotin, sphingolipid, and histidine metabolism (Fig. S15).

Fig. 8.

PS-NPs exposure reshaped the gut microbiota and impaired intestinal barrier integrity. (A) Venn diagram illustrating the unique and shared OTUs between the control (Con) and PS-NPs–treated groups. (B-D) Alpha diversity indices, including Chao-1 (B), Observed species (C), and Shannon index (D). Data are presented as mean ± SEM (n = 8). Wilcoxon rank-sum test was used. (E) Beta-diversity analysis visualized by PCoA based on Bray-Curtis dissimilarity. (F, G) Gut bacteria distribution at the phylum (F) and genus (G) level. (H) Heatmap of bacterial genera, illustrating the differential abundance of specific taxa between the control and PS-NPs–treated groups. (I) LEfSe results, identifying bacterial taxa that contributed to the differences between the control and PS-NPs groups, with taxa in green enriched in PS-NPs and taxa in red enriched in control. (J) Cladogram visualizing the phylogenetic distribution of significantly different taxa. (K) Network analysis depicting co-occurrence patterns among microbial genera. (L) H&E staining showing histological alterations in the intestines. (M) Masson’s trichrome staining indicating tissue fibrosis in PS-NPs–treated groups. (N-P) Immunofluorescence staining for ZO-1 (N), Occludin (O), and E-cadherin (P), showing decreased expression in PS-NPs–treated groups. (Q-S) Quantitative analysis of ZO-1 (Q), Occludin (R), and E-cadherin (S) fluorescence intensity. Data are presented as mean ± SEM (n = 4). One-way ANOVA followed by Tukey’s post hoc test was used. Ns means no significance. *P < 0.05, **P < 0.01, ***P < 0.001. PS-NPs, polystyrene nanoplastics; OTUs, operational taxonomic units; PCoA, principal coordinate analysis

Furthermore, histology and barrier markers demonstrated intestinal injury. H&E staining revealed inflammatory cell infiltration in the PS-NPs groups (Fig. 8L), while Masson’s trichrome showed increased collagen deposition and fibrosis (Fig. 8M). Immunofluorescence showed reduced tight-junction proteins ZO-1 and Occludin, and the adhesion molecule E-cadherin in PS-NPs–treated groups (Fig. 8O-S).

PS-NPs exposure induced intestinal inflammation through the LPS/TLR4/MyD88 axis and linked dysbiosis to reproductive outcomes

Immunofluorescence staining for F4/80 and CD3 showed an increased presence of macrophages and T cells in both low and high PS-NPs groups compared to the controls, suggesting immune disruption (Fig. 9A, B), and quantification confirmed significant elevations (Fig. 9C, D). Besides, circulating endotoxin burden was elevated. Serum LPS levels were higher in PS-NPs–treated groups (Fig. 9E). The bacterial product LPS has been demonstrated to regulate intestinal inflammation via the activation of TLR4 and MyD88-dependent pathway [28]. Both intestinal and testicular samples were analyzed by WB to verify the hypothesis (Fig. 9F, G and Fig. S16).

Fig. 9.

PS-NPs exposure induced intestinal LPS/TLR4/MyD88–driven inflammation and linked dysbiosis to reproductive outcomes. (A, B) Immunofluorescence staining for F4/80 (A) and CD3 (B). (C, D) Quantitative analysis of F4/80 (C) and CD3 (D) fluorescence intensity. Data are presented as mean ± SEM (n = 4). (E) Serum levels of LPS in control and PS-NPs–treated groups. Data are presented as mean ± SEM (n = 8). (F) The characteristic bands of TLR4 and MyD88 proteins. (G) Quantitative analysis of protein levels for TLR4 and MyD88. Data are presented as mean ± SEM (n = 3). (H-M) Immunohistochemistry staining for inflammatory markers IL-6 (H), COX-2 (I), IL-18 (J), IL-17 A (K), TNF-α (L), and NLRP3 (M) in control (Con), low-dose (Low), and high-dose (High) PS-NPs groups. (N) Quantitative analysis of staining intensity for IL-6, IL-18, TNF-α, COX-2, IL-17 A, and NLRP3. Data are presented as mean ± SEM (n = 4). (O-S) Serum levels of TNF-α (O), IL-6 (P), IL-18 (Q), COX-2 (R), and IL-17 A (S) in control and PS-NPs–treated groups. Data are presented as mean ± SEM (n = 8). (T) Correlation heatmap linking inflammatory cytokines with specific gut microbiota. (U) Correlation heatmap linking reproductive parameters with specific gut microbiota. One-way ANOVA followed by Tukey’s post hoc test was used. Ns means no significance. *P < 0.05, **P < 0.01, ***P < 0.001. PS-NPs, polystyrene nanoplastics

Furthermore, pro-inflammatory mediators were consistently increased in the gut. IHC demonstrated stronger staining of IL-6, COX-2, IL-18, IL-17 A, TNF-α, and NLRP3 in PS-NPs groups (Fig. 9H-N). ELISA revealed significantly increased concentrations of inflammatory cytokines in PS-NPs–treated groups, indicating systemic inflammation as a result of exposure (Fig. 9O-S). To explore potential associations between the gut microbiome and host phenotype, we performed Spearman correlation analyses. Bacteroides and Escherichia–Shigella exhibited significant positive correlations with serum LPS and multiple cytokines, whereas beneficial genera, such as Lachnospira, Lactobacillus, and Lachnospiraceae_UCG-008, showed inverse associations (Fig. 9T and Fig. S17). Consistent with these inflammatory patterns, the pro-inflammatory genera were negatively correlated with testosterone and sperm parameters, while protective genera showed positive associations with these reproductive parameters (Fig. 9U and Fig. S18). Moreover, to delineate cross-omics interactions, we performed integrative analyses. The heatmap of correlation analyses were conducted to connect transcriptomic and metabolomic datasets, as well as to link metabolomic profiles with microbial composition (Fig. S19A, B). Besides, a comprehensive network was constructed, highlighting alterations in key metabolites NAD, spermidine, FMA, and L-glutamate, and critical genes including Nmnat1, Acp2, Srm, and Gss. This combined analysis further indicated notable impacts on various pathways, including the biosynthesis of cofactors, glutathione metabolism, alanine, aspartate, and glutamate metabolism and riboflavin metabolism, thereby mapping a multi-omics landscape of PS-NPs-induced disruption (Fig. S19C, D).

Gut microbiota from PS-NPs–exposed rats resulted in testicular and intestinal injuries in recipient rats

To explore a mechanistic link whether the gut microbiome is a causal mediator of PS-NPs–induced testicular and intestinal damage, we performed the FMT experiment. After a 7-d antibiotic pretreatment to deplete the resident microbiota, recipient rats received daily gavage of fecal suspensions collected from either saline or PS-NPs–treated donors. PS-NPs-FMT recipients demonstrated reduced epididymal and testis weights (Fig. S20A, B), with H&E revealing vacuolization and disorganization of the seminiferous epithelium (Fig. S20C). Serum testosterone levels were significantly reduced in the PS-NPs-FMT group, paralleling declines in sperm quality (Fig. S20D-H). Immunofluorescence staining revealed decline in c-KIT⁺ spermatogonia, SCP3⁺ spermatocytes, and PNA-labeled acrosomal structures (Fig. S20I-K). Additionally, the structural integrity of BTB was disrupted in the PS-NPs-FMT group (Fig. S20L, M). Furthermore, the FMT-PS-NPs group exhibited enhanced immune infiltration, indicated by an increased abundance of F4/80⁺ macrophages and CD3⁺ T cells (Fig. S21A, B). This was further supported by increased protein expression of TLR4 and MyD88, indicating activation of the TLR4/MyD88 signaling axis (Fig. S21C).

Next, we also evaluated how the transferred gut microbiota contributed to systemic inflammation and intestinal injury. H&E staining showed mucosal inflammation in the colon of rats that received FMT from PS-NPs–exposed donors compared to controls (Fig. S22A), and immunofluorescence revealed a breakdown of intestinal tight junctions (Fig. S22B, C). This was accompanied by enhanced immune cell infiltration in the PS-NPs-FMT group, demonstrated by increased densities of F4/80⁺ macrophages and CD3⁺ T cells (Fig. S22D, E). At the signaling level, we observed a corresponding upregulation of TLR4 and MyD88 proteins (Fig. S22F). Additionally, transplantation of gut microbiota from PS-NPs–exposed donors recapitulated the elevation of systemic inflammatory mediators (Fig. S22G-L). Moreover, FMT from PS-NPs-exposed donors successfully transferred a dysbiotic gut microbiota state to recipient rats. Compared to the Con-FMT group, recipients of PS-NPs-FMT exhibited increased alpha diversity, a complete separation in community structure (beta-diversity), and a characteristic shift in microbial composition (Fig. S23A-F). Specifically, pro-inflammatory genera (e.g., Bacteroides, Ruminococcus) were enriched, while beneficial taxa (e.g., Lactobacillus, Lachnospiraceae_UCG-006) were depleted (Fig. S23G). A cladogram illustrated the phylogenetic distribution of these differentially abundant taxa, and network analysis revealed altered topological properties in the microbial interaction network of the PS-NPs-FMT group (Fig. S23H, I).

DI alleviated PS-NPs–induced testicular injury by dampening TLR4/MyD88/NF-κB–driven inflammation

Given the anti-inflammatory potential of the itaconate, we evaluated its derivative DI to explore whether it could be utilized as the therapeutic agent against PS-NPs–induced toxicity. Histological analysis revealed a clear structural improvement in the testes following DI co-treatment (Fig. 10A). This morphological rescue coincided with a significant restoration of serum testosterone levels (Fig. 10B).

Fig. 10.

DI mitigated PS-NPs–induced testicular injury by dampening TLR4/MyD88/NF-κB–driven inflammation and supporting germline recovery. (A) H&E staining of testes. Scale bars, 200 μm. (B) Serum testosterone. DI partially rescued testosterone levels relative to PS-NPs alone. Data are presented as mean ± SEM (n = 8). (C) IF images showing PNA staining in acrosome structures between the PS-NPs and PS-NPs + DI groups. Right: quantification. Scale bars, 100 μm. Data are presented as mean ± SEM (n = 4). (D) The volcano plot of Bulk RNA-seq showing DI-responsive DEGs. (E) Heatmap of representative DI-responsive DEGs demonstrating clear separation between PS-NPs and PS-NPs + DI groups. (F) GO enrichment indicating inflammatory response, defense response and extracellular matrix. (G) KEGG enrichment highlighting immune pathways, such as NF-κB signaling, cytokine–cytokine receptor interaction, phagosome, and Th1 and Th2 cell differentiation. (H-L) IF images of F4/80 (H) CD3 (I) IL-17 A (J) NLRP3 (K), and IL-6 (L) between PS-NPs and PS-NPs + DI groups. Right, quantification. Scale bars, 100 μm. Data are presented as mean ± SEM (n = 4). (M) Western blots of testis tissue for TLR4, MyD88, and NF-κB. DI lowered protein abundance of all three components of the LPS/TLR4/MyD88/NF-κB axis relative to PS-NPs alone. (N) Densitometry for confirming significant suppression of TLR4, MyD88, and NF-κB by DI. Data are presented as mean ± SEM (n = 3). T test was used. Ns means no significance. *P < 0.05, **P < 0.01, ***P < 0.001. DI, dimethyl itaconate; PS-NPs, polystyrene nanoplastics; H&E, Hematoxylin and eosin; IF, immunofluorescence; DEGs, differentially expressed genes

Furthermore, immunofluorescence staining demonstrated a stronger acrosomal PNA signal in co-treated rats, suggesting partial restoration of spermiogenesis (Fig. 10C). Furthermore, DI significantly improved the sperm quality in PS-NPs–exposed rats (Fig. S24A). This recovery extended to structural elements of the testis. The expression of ZO-1, N-cadherin and Vimentin were also upregulated (Fig. S24B-D). At the transcriptional level, bulk RNA-seq revealed clear molecular shifts. The volcano plot identified 106 up-regulated and 52 down-regulated DEGs (Fig. 10D), with hierarchical clustering showing distinct transcriptomic profiles between the two groups (Fig. 10E). GO enrichment pointed to inflammatory responses and extracellular matrix organization (Fig. 10F). This was corroborated by KEGG analysis, which identified NF-κB signaling alongside key immune processes including phagosome, cytokine-cytokine receptor interaction, and Th1/Th2 cell differentiation as top-ranked categories (Fig. 10G). Immunostaining confirmed these anti-inflammatory effects of DI. There was a marked reduction in testicular immune infiltration (Fig. 10H, I), accompanied by a concomitant decrease in the key pro-inflammatory effectors (Fig. 10J-L). WB analysis confirmed the suppression of the upstream LPS/TLR4/MyD88/NF-κB signaling axis in DI-treated testes, with reduced protein levels of TLR4, MyD88, and NF-κB (Fig. 10M, N).

DI restored gut barrier integrity, suppressed intestinal TLR4/MyD88 signaling, rebalanced dysbiosis, and lowered systemic inflammation

Administration of DI to PS-NPs–exposed rats significantly attenuated inflammation of colonic tissue, as evidenced by H&E staining (Fig. 11A). DI treatment improved structural integrity by reestablishing tight junctions—both ZO-1 and Occludin were restored (Fig. 11B, C). Besides, immunofluorescence showed a marked decline in F4/80⁺ macrophages and CD3⁺ T cells (Fig. 11D, E). At the molecular level, DI suppressed the LPS/TLR4/MyD88 inflammatory pathway. WB data showed reduced expression of TLR4 and MyD88 in the colon (Fig. 11F), accompanied by a significant decrease in LPS levels (Fig. 11G). Systemic inflammation was notably relieved, as reflected by decreased circulating levels of TNF-α, COX-2, IL-6, IL-17 A, and IL-18 (Fig. 11H-L). Co-treatment with DI also reshaped the community structure. Despite the absence of significant changes in alpha diversity, beta-diversity analyses demonstrated a significant restructuring of the gut microbiota composition after DI treatment (Fig. S25A-F). The compositional changes of the corresponding samples at the phylum and genus levels were shown in the Fig. 11M and N, respectively. Quantitative analysis revealed that DI administration significantly reduced the abundance of LPS-producing genera (Bacteroides, Escherichia–Shigella, and Paraprevotella), while concurrently increasing beneficial commensals, such as Faecalibaculum, Lachnospira, and Akkermansia, suggesting a microbial mechanism for its anti-inflammatory effects (Fig. 11O).

Fig. 11.

DI restored gut barrier integrity, suppressed intestinal TLR4/MyD88 signaling, rebalanced dysbiosis, and lowered systemic inflammation. (A) H&E staining of the Colon. Scale bars, 200 μm. (B-E) IF imaging for ZO-1 (B), Occludin (C), F4/80 (D), and CD3 (E). Right, normalized fluorescence quantification. Scale bars, 100 μm. Data are presented as mean ± SEM (n = 4). T test was employed. (F) Western blots of colon tissue for TLR4 and MyD88. DI lowered both proteins relative to PS-NPs alone. Right, normalized quantification. Data are presented as mean ± SEM (n = 3). T test was employed. (G-L) Serum measurements of circulating LPS (G), TNF-α (H), COX-2 (I), IL-6 (J), IL-18 (K), and IL-17 A (L). Data are presented as mean ± SEM (n = 8). T test was employed. (M, N) Phylum (M) and genus (N) level composition via the 16 S rRNA sequencing (stacked bars per sample). (O) Differential genera between PS-NPs and PS-NPs + DI groups. DI decreased pro-inflammatory taxa, such as Paraprevotella, Bacteroides and Escherichia–Shigella, and increased beneficial taxa, such as Faecalibaculum, Lachnospira, and Akkermansia. Data are presented as mean ± SEM (n = 8). Wilcoxon rank-sum test was used. Ns means no significance. *P < 0.05, **P < 0.01, ***P < 0.001. DI, dimethyl itaconate; PS-NPs, polystyrene nanoplastics; H&E, Hematoxylin and eosin; IF, immunofluorescence

Discussion

The study provides mechanistic insights into a gut-centered route through which PS-NPs compromise male reproductive health. By combining particle characterization, in vivo distribution, single-cell and bulk transcriptomics, untargeted metabolomics, 16 S rRNA profiling, FMT, and a targeted immunometabolic intervention with DI, we establish a systems-level perspective that PS-NPs exposure does not act on the testis in isolation. Furthermore, it perturbs the intestinal ecosystem and barrier, increases systemic endotoxin load, and then propagates inflammation and metabolic signals to the testis through the LPS/TLR4/MyD88 axis, ultimately destabilizing the spermatogenic microenvironment, and impairing spermatogenesis.

The PS-NPs used in our study were small, monodisperse, and colloidally stable. Importantly, TEM and confocal imaging demonstrated dose-dependent internalization in colonic epithelium and deposition in testicular tissue, suggesting an exposure-distribution axis that earlier toxicology study had to infer indirectly [29]. Histological analysis of testicular tissues further revealed structural disruptions, which were crucial for spermatogenesis. The reduction in c-KIT, SCP3, and PNA positive cells underscored the detrimental impact of PS-NPs on spermatogonia, spermatocytes, and spermatids, respectively. The decrease in CYP11A1 expression demonstrated a disruption in steroidogenesis, consistent with the observed reduction in testosterone levels. Notably, SOX9⁺ Sertoli cells were numerically preserved, but their functional outputs were not: SCF, GDNF, and BMP4 were all downregulated, which is exactly the set of trophic factors required to maintain and differentiate spermatogonia [30]. Going beyond phenotypic toxicity, our study uncovered the molecular drivers of testicular failure. We observed a coordinated loss of junctional proteins and an upregulation of the HIF-1α/NF-κB/IL-17 A inflammatory axis. Crucially, our data suggested that this was not merely a local response to physical particle stress, but a consequence of a compromised immune microenvironment. The breakdown of the BTB, consistent with recent findings on environmental toxicants [31, 32], exposed the germline to systemic inflammatory mediators, thereby locking the testis into a chronic inflammatory state that suppressed PPAR-γ and disrupted steroidogenesis [31–33].

By integrating the control and PS-NPs–exposed testes into a joint single-cell RNA-seq atlas, we identified nine stable testicular cell types. Differential expression and GO/KEGG enrichment consistently pointed to mitochondrial energetics, flagellum-dependent motility, microtubule-based movement, and ER/ribosome homeostasis as disrupted modules. These are exactly the pathways that support spermiogenesis and axoneme formation, and their loss provides a molecular explanation for the poor sperm quality we observed. The CellChat analysis added another layer: the global intercellular connectome contracted after PS-NPs. Total numbers and strengths of inferred ligand-receptor interactions dropped and germ-cell nodes shifted toward lower outgoing strengths. The macrophages served as the principal signaling hub, but their outgoing strength was lower after exposure, suggesting that macrophages were also in a dysregulated state. This macrophage-centered network is important, because it provides a resident immune population ready to respond to extra-testicular inflammatory cues [34, 35]. Although scRNA-seq provided high-resolution insights, the limited sample size is an important caveat. Nonetheless, our key findings were substantiated by bulk RNA-seq analyses and experiment validations.

Untargeted metabolomics, interpreted together with cell-type-specific DEGs, showed that PS-NPs caused a concerted metabolic reprogramming. At the whole-tissue level, OPLS-DA clearly separated the exposed group from controls, and key metabolites in energy, amino acid, carbohydrate, and lipid metabolism were perturbed. Because spermatogenesis is an energetically demanding process and relies on precise lipid remodeling of germ-cell membranes, this energy-lipid deficit provides a plausible upstream cause of the widespread apoptosis, DNA double-strand breaks, and suppressed autophagy that we observed. A similar pattern—oxidative phosphorylation down, autophagy flux blocked, apoptosis up—has been described in testes exposed to other nanomaterials and is frequently interpreted as ROS-mediated mitochondrial failure [36–38]. NF-κB activation, along with increased TNF-α, IL-6, IL-18, CCL2, COX-2, IL-17 A, and NLRP3, effectively locks the tissue into a pro-inflammatory state and further interferes with steroidogenesis and BTB maintenance [39–41]. An important mechanistic question regards the crosstalk between inflammatory and autophagic flux. While autophagy and NF-κB often engage in reciprocal regulation, the bioinformatic analyses and experimental evidence collectively support the hypothesis that TLR4/NF-κB activation acts as the upstream driver that subsequently suppresses autophagy in the PS-NPs exposure model. Previous researches have indicated that NF-κB activation can upregulate mTORC1 activity, a master inhibitor of autophagy, thereby blocking autophagosome formation [42]. TLR4 activation by LPS can inhibit autophagy in tissues through the PI3K/Akt/mTOR pathway or by inhibiting FOXO3-mediated transcription of autophagy genes [43, 44]. In this study, the inflammation-dependent block in autophagy likely exacerbates cellular stress and apoptosis, forming a vicious cycle that contributes to spermatogenic failure.

Our study was among the first to explore the impact of PS-NPs on the gut microbiome in the male reproduction. Venn diagram and alpha diversity metrics demonstrated that PS-NPs caused notable alterations in gut microbiota, with increases in Bacteroides, and Escherichia/Shigella and decrease in Lachnospiraceae_UCG-008, Lachnospira, and Lactobacillis. These findings are significant, as the gut microbiota play central roles in maintaining systemic immune balance and hormonal balance. Previous studies have demonstrated that disruptions in gut microbial composition can lead to hormonal imbalances and heightened immune activity, both of which negatively affect reproductive health [45, 46]. The immunofluorescence and histological analyses of intestinal tissues demonstrated that PS-NPs exposure compromised the intestinal barrier, as evidenced by reduced decreased ZO-1 and Occludin expression. This disturbance likely drives widespread inflammation, evidenced by an increase in F4/80 and CD3 positive cells, along with elevated levels of pro-inflammatory cytokines. Correlation heatmaps further highlighted the strong statistical link between PS-NPs–induced changes in gut microbiota and markers of reproductive dysfunction. Although these correlations were associative in nature, they aligned with the FMT results to support the concept of a microbiome–gut–reproduction axis. A key component of this axis appears to be the elevated production of LPS, an endotoxin derived from Gram-negative bacteria. Notably, PS-NPs exposure led to a significant increase in the abundance of Escherichia–Shigella and Bacteroides, two LPS-producing genera [47, 48]. LPS is known to activate the NF-κB pathway by binding to TLR4, triggering inflammatory cascades via both MyD88-dependent and -independent signaling mechanisms [49, 50]. In our study, WB confirmed that TLR4 and MyD88 protein levels were significantly upregulated in both the intestinal and testicular tissues following PS-NPs exposure. This activation coincided with higher levels of pro-inflammatory markers in PS-NPs–exposed groups, indicating a robust inflammatory response. The dysbiosis caused by PS-NPs contributed to heightened LPS production. LPS likely entered the bloodstream due to compromised gut barrier integrity, triggering systemic inflammation. Besides, the elevated LPS load promoted cytokine release, thereby amplifying immune responses and contributing to widespread inflammation, including in reproductive tissues. Systemic inflammation induced by PS-NPs and LPS in testicular tissue led to immune imbalance and the activation of multiple signaling pathways, such as HIF-1, PPAR, and NF-κB. These disruptions led to lipid metabolism disorders, autophagy dysregulation, and increased apoptosis. The cumulative effect of these processes included the destruction of the BTB, lowered testosterone levels due to impaired steroidogenesis, and disturbed spermatogenesis, ultimately leading to decreased sperm quality and fertility. Critically, FMT from PS-NPs–exposed donors into antibiotic–pretreated recipients recapitulated key aspects of the injury phenotype—including testicular dysfunction, BTB disruption, intestinal inflammation, TLR4/MyD88 pathway activation, and systemic inflammation—in the absence of the original nanoparticles. To our knowledge, FMT-based causality for plastic-induced reproductive injury, has been rarely shown [51]. Prompted by metabolomic signatures pointing to TCA cycle rewiring and by recent work on the itaconate/IRG1 axis as an endogenous brake on TLR-driven inflammation [52], we tested DI. DI is a cell-permeable itaconate derivative widely used to model the electrophilic, and NF-κB–suppressive actions of endogenous itaconate, although it may not be converted stoichiometrically to itaconate in cells [20, 53, 54]. In our experimental model, DI administration produced a dual-site rescue. Within the testicular tissues, DI restored seminiferous tubule morphology, improved sperm parameters, decreased inflammation and immune infiltration, and suppressed the TLR4/MyD88/NF-κB signaling axis. Mechanistically, DI targeted the intestinal compartment by enhancing gut barrier function and reconfiguring microbial community networks, thereby reducing endotoxemia and inflammatory flux to the distal gonadal tissues. These data suggest a microbiota-based protective role for DI in counteracting PS-NPs–induced reproductive toxicity in males.

This study has limitations that warrant further investigation. First, regarding the nanoplastic exposure model, specific limitations regarding the translational relevance must be acknowledged. Although the high dose (70 mg/kg/day) was derived from theoretical upper limits of human ingestion to identify potential hazards, it likely exceeds typical daily exposure levels for the general population. Furthermore, we utilized uniform, spherical, and pristine PS-NPs to ensure experimental consistency. In contrast, real-world human exposure involves a complex mixture of weathered, irregular microplastics with varying chemical compositions and aged surfaces, which may exhibit different bioavailability and toxicity profiles compared to laboratory-grade particles. Future epidemiological studies are needed to validate whether the specific gut-testis inflammatory axis identified in this rodent model is conserved in human populations. Second, while the 50 mg/kg dosage of DI effectively mitigated reproductive toxicity in this study, we did not perform a comprehensive dose-response analysis to determine the minimum effective dose or the maximum tolerated dose. Furthermore, although DI is a cell-permeable derivative of itaconate, its long-term pharmacokinetics and safety profile in treating chronic environmental toxicity remain to be fully characterized. Future studies should focus on optimizing the therapeutic regimen and evaluating potential off-target effects to facilitate clinical translation.

Conclusion

PS-NPs impair male reproduction through a gut-centered inflammatory cascade. Exposure reshapes the microbiota, weakens intestinal junctions, and elevates circulating LPS, which activates the TLR4/MyD88 axis in gut and testis, amplifying NF-κB/IL-17/HIF-1 signaling and suppressing PPAR-γ. In the testis, this compromises BTB integrity, diminishes Sertoli cell trophic factor secretion, and reduces Leydig cell endocrine input and steroidogenesis, thereby disrupting spermatogenesis and impairing fertility. Multi-omics reveal convergent energetic/lipid deficits with suppressed autophagy and heightened DNA damage and apoptosis, explaining defective spermatogenesis and poor sperm quality. Causality is supported by FMT, which reproduces the phenotype in recipients. Therapeutically, DI restores gut and testicular barrier integrity, dampens TLR4/MyD88/NF-κB signaling, reduces cytokinemia, and improves spermatogenesis. Together, these findings suggest a microbiome–gut–testis axis linking endotoxin-driven inflammation to metabolic failure and reproductive toxicity, and highlight immunometabolic interventions and barrier-repair strategies to mitigate nanoplastic risk (Fig. 12).

Fig. 12.

Diagram illustrating the microbiome–gut–testis mechanism of PS-NPs–induced male reproductive toxicity and the treatment of DI. DI, dimethyl itaconate; PS-NPs, polystyrene nanoplastics

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Jiaochen Luan: Investigation, Experimental verification, Software & data curation, Writing, review & editing; Xu Zhang: Investigation, Software & data curation, Writing-Original draft preparation; Tong Chen: Resources, Data curation, Funding acquisition; Shihang Pu: Software & data curation, Writing-Original draft preparation; Zhiyi Shen: Investigation, Writing-Original draft preparation; Chunlu Xu: Methodology, Software & data curation; Zhijun Chen: Investigation, Methodology, Resources; Jiayi Zhang: Methodology, Supervision, Funding acquisition, Writing, Reviewing & Editing; Danni Chen: Conceptualization, Project administration, Supervision, Writing, Reviewing & Editing.

Funding

This study was funded by the National Natural Science Foundation of China (82103580), Jiangsu Province Capability Improvement Project through Science, Technology and Education (ZDXK202219), Jiangsu Funding Program for Excellent Postdoctoral Talent (2022ZB730), and Young Scholars Fostering Fund of the First Affiliated Hospital of Nanjing Medical University (PY2023002, PY2025017).

Data availability

Datasets in this work are available from the corresponding authors on reasonable request.

Declarations

Ethics approval and consent to participate

All animal procedures were carried out according to the guidelines of Institutional Animal Care and Use Committee of Nanjing Medical University (Approval number: IACUC-2510084), and the animals were housed and maintained at the Experimental Animal Center of Nanjing Medical University.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.