Long-term polystyrene nanoplastics exposure aggravates retinal inflammation and photoreceptor degeneration through microglial SPP1 signaling and neutrophil extracellular traps formation Authors: Wenchuan Zhou, Xuanyi Chen, Jiagen Li, Jincan He Published: 2026 ABSTRACT BackgroundMicro/nanoplastics (MNPs), as emerging environmental contaminants, present a growing concern for human health. This study aims to investigate the effects of polystyrene nanoplastics (PS-NPs) exposure on retinal pathology and underlying mechanisms.MethodsRetinal detachment (RD) model was established on adult mice following PS-NPs exposure (10 and 50 mg/L) through drinking water for two months. In vitro, oxygen glucose deprivation (OGD) model was established on BV2 microglia-661W photoreceptor co-culture system following PS-NPs exposure (100 mg/L) for 24 h. SPP1 neutralizing antibody and recombinant protein were administrated by subretinal injection. DNase I and Cl-amidine were utilized to achieve neutrophil extracellular traps (NETs) inhibition. Electroretinogram was used to assess retinal function. Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL), immunofluorescent staining, western blot analysis and enzyme activity assays were used to analyze photoreceptor apoptosis, microglial responses and oxidative stress. Microglia were purified with CD11b MicroBeads. Transcriptomic profiles of PS-NPs-exposed microglia and human retinas of proliferative vitreoretinopathy (PVR) were analyzed.ResultsPS-NPs were able to breach the blood-retina barrier, disrupt phototransduction, aggravate oxidative stress and apoptosis in RD-induced photoreceptor degeneration model dose-dependently. Mechanistically, PS-NPs exposure triggered retinal inflammation, microglial activation and microglial SPP1-mediated peripheral neutrophil recruitment. SPP1 neutralization mitigated PS-NPs-aggravated chemokine secretion, neutrophil infiltration and NETs formation. Recombinant SPP1 protein treatment heightened neutrophil-driven retinal damage, while this could be partially reversed by chemokine receptor inhibition. NETs inhibition alleviated PS-NPs-exacerbated microglial proinflammatory activation and photoreceptor degeneration. Furthermore, transcriptomic profiling showed parallels between PS-NPs-exposed microglia and human PVR specimens in SPP1 signaling and stress/stimulus response pathways.ConclusionsOur findings demonstrated that PS-NPs exposure aggravated retinal inflammation and photoreceptor degeneration by microglial SPP1 signaling activation and NETs formation, underscoring new insights into the effects and potential targets of MNPs exposure on retinal disorders.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12967-026-08405-6. Background In recent years, plastic pollution has become a pressing global concern [1]. Degraded from plastic wastes, microplastics (MPs) and nanoplastics (NPs) were found to enter living organisms through dermal absorption, gastrointestinal ingestion and respiratory inhalation, breaching endothelial and epithelial barriers to exert multi-organ toxicity [2–5]. Emerging evidence suggested that micro/nanoplastics (MNPs) deposited in the central nervous system (CNS) exhibit significant neurotoxicity dose-dependently [6]. MNPs could disrupt the permeability and integrity of the blood-brain barrier (BBB), induce microglial activation and neuron dysfunction. Mechanistically, MNPs provoke oxidative stress, immune-inflammatory responses, organelle dysfunction and metabolic dysregulation, collectively reducing neuron activity and synaptic plasticity, and ultimately impairing neurodevelopment and exacerbating neurodegenerative pathologies [7–9]. As a specialized extension of the CNS, the retina plays a pivotal role in visual perception and phototransduction. Photoreceptors, the light-sensitive primary sensory neurons, are relatively susceptible to environmental stress and genetic abnormalities [10]. Photoreceptor degeneration, characterized by progressive neuronal death, represents a leading cause of irreversible vision loss [11, 12]. During photoreceptor degeneration, retinal microglia undergo rapid activation and initiate retinal inflammation through cytokine secretion and peripheral immune cell recruitment, thereby critically modulating photoreceptor survival [13–15]. MPs have been detected in human ocular tissues including the retina, vitreous humor, tear fluid and meibum, with potential association to ophthalmologic disorders [16–18]. Zhang et al. collected human retinal tissues and identified several types of MPs, in which polystyrene (PS), polyethylene (PE), polypropylene (PP) and polyvinyl chloride (PVC) were detected in all samples [16]. Although technical approaches could not precisely detect NPs deposition in human ocular tissues, research has indicated the underestimated abundance, penetrability and toxicity of NPs compared to MPs [19, 20]. Our previous research established that maternal polystyrene nanoplastics (PS-NPs) exposure disrupted retinal development and metabolic profiles in murine progenies [21]. In adult mice, long-term PS-NPs exposure impaired retinal function but preserved retinal structural integrity [22]. Under light challenge, PS-NPs exposure potentiated microglial activation, retinal inflammation and photoreceptor apoptosis, while the underlying mechanisms remained unclear. In this study, we exposed adult mice with two doses of PS-NPs (10 mg/L and 50 mg/L) via drinking water for two months, followed by retinal detachment (RD)-induced mild photoreceptor degeneration. The effects of PS-NPs exposure on phototransduction, oxidative stress and photoreceptor apoptosis were assessed. Next, we identified microglial proinflammatory polarization and peripheral immune cell recruitment as key contributors to PS-NPs-aggravated retinal injury. Combined with transcriptome analysis and wet experiment validation, we demonstrated activated SPP1 and chemokine-associated signaling in PS-NPs-exposed microglia. Furthermore, neutrophil extracellular traps (NETs) were revealed as a potential amplifier for microglia-neutrophil crosstalk following PS-NPs exposure and RD challenge. Finally, we identified shared signaling pathways between PS-NPs-induced cytotoxicity and human proliferative vitreoretinopathy (PVR) features, providing novel perspectives of MNPs exposure as a potential environmental risk factor for retinal disease pathogenesis. Methods PS-NPs characterization and preparation Non-fluorescent and red fluorescent protein (RFP) conjugated PS-NPs were purchased from Zhichuan Technology (Nantong, China), with the average diameter of 100 nm. The HITACHI (Tokyo, Japan) SU8010 scanning electron microscope (SEM) was used for morphological observation. Fourier-transform infrared spectroscopy (FTIR) was employed to detect functional groups of PS-NPs, with the scanning range of 400–4000 cm− 1. Surface charge and stability of particles were assessed by Zeta potential determination using dynamic light scattering. The nanoparticle tracking analysis (NTA) was conducted using Zetaview-PMX120-Z (Particle Metrix, Meerbusch, Germany). PS‑NPs suspensions were prepared under sterile conditions using endotoxin‑free consumables and reagents. To avoid aggregations, the PS-NPs suspensions were sufficiently sonicated before use. Limulus amebocyte lysate (LAL) endotoxin assay Endotoxin levels in PS-NPs suspensions were assessed using a chromogenic LAL assay kit (#L00350C, GenScript, Nanjing, China) according to the manufacturer’s protocols. Tubes, tips and reagent water utilized in this experiment were endotoxin-free. The endotoxin concentrations for all tested PS-NPs suspensions were undetectable (< 0.01 EU/mL). Animal husbandry, PS-NPs exposure and RD induction Male C57BL/6J mice (8 weeks old) were purchased from Vital River Laboratory Animal Technology (Beijing, China). Animals were housed in a specific-pathogen-free facility under controlled conditions, including regulated temperature, humidity and a 12-hour light/dark cycle, with ad libitum access to food and water. Mice were allocated into three groups: control, PS-NPs-exposed (10 mg/L) and PS-NPs-exposed (50 mg/L) group. The exposure doses translated to 1.6 mg/kg body weight/day and 8.0 mg/kg body weight/day, assuming the average water intake as 4 mL and the average body weight as 25 g. The control group received autoclaved water for 2 months. Mice in PS-NPs-exposed group were given PS-NPs via drinking water for 2 months at two doses (10 mg/L and 50 mg/L). To confirm the penetration of blood-retina barrier (BRB), we exposed a subset of mice with RFP-PS-NPs. RD was established following the two-month PS-NPs exposure. Briefly, subretinal injection of 1% sodium hyaluronate (2 µL) was conducted to achieve the detachment of neural retina from the underlying retinal pigment epithelium (RPE). Surgeries on all groups were performed by the same operator within a single batch to minimize inter-batch variability. For control mice, a small gauge needle was inserted through the sclera into the subretinal space without sodium hyaluronate injection. Eyes with subretinal hemorrhage or unsuccessful detachment were excluded. On Day 1 post-RD challenge, when mild photoreceptor degeneration was induced, mice were sacrificed for further analysis. Mice were randomly assigned to experimental groups using a computer-generated random number sequence at treatment initiation. Outcome assessments were performed by investigators blinded to group allocation, using coded sample labels. SPP1 neutralizing antibody (#AF808, R&D Systems, MN, USA) and corresponding IgG antibody (#16-4714-82, ThermoFisher Scientific, CA, USA) as control were administrated by subretinal injection of 2 µL. Recombinant mouse SPP1 protein (#441-OP, R&D Systems) and phosphate-buffered saline (PBS) vehicle were administrated by subretinal injection of 2 µL. TAK-779 (#HY-13406, MedChemExpress, NJ, USA), a CCR5/CXCR3 antagonist, was intraperitoneally injected at the dose of 6 mg/kg body weight during RD induction. NETs inhibition was achieved by subretinal injection of DNase I (#EN0521, ThermoFisher Scientific) at 2 µL and intraperitoneal injection of Cl-amidine (#HY-100574, MedChemExpress) at 10 mg/kg body weight. Electroretinogram (ERG) ERG is commonly used for assessing retinal function. ERG responses were recorded using the Ganzfeld stimulator, Espion e3 recording and Color Dome (Diagnosys LLC, Lowell, USA). We employed white-flash stimuli at 0.01, 0.1, 1.0, 3.0 and 10.0 log cd∙s/m2 after dark adaption (scotopic) and 10.0 log cd∙s/m2 after light adaption (photopic). Five responses were recorded following each stimulus. Tissue processing, immunohistochemistry and terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) of retinal sections After eye enucleation, the cornea, iris, ciliary body, and lens were excised. The eyecup was then PBS-rinsed and fixed in 4% paraformaldehyde (PFA). For hematoxylin & eosin (H&E) staining, we embedded the fixed eyecup with paraffin and sectioned at 10 μm. For immunofluorescent staining, we embedded it with optimal cutting temperature (OCT) compound and sectioned at 10 μm. TUNEL staining was conducted according to the manufacturer’s instructions (#11684795910, Roche, Basel, Switzerland). Immunohistochemistry was performed as previously described [22]. Slides were treated with an autofluorescence quencher (#G1221, Servicebio, Wuhan, China) before observation. Imaging was performed under a fluorescent microscopy (Olympus BX51, Tokyo, Japan), using a standardized fluorescence excitation protocol to ensure consistent exposure time across all samples. For each retina, three non-adjacent sections were analyzed as technical replicates. The primary and secondary antibodies used in this study were presented in Supplementary table S1. IBA1 immunofluorescent staining on retinal wholemounts and skeleton analysis Following fixation with PFA, the neural retinal tissues were carefully removed, flattened and stained with IBA1. The skeleton analysis of IBA1+ microglia was conducted using MotiQ plugin in ImageJ as described in our previous work [23]. Oxidative stress-related analysis Oxidative stress level of retinal tissue was evaluated through reactive oxygen species (ROS) quantification, lipid peroxidation measurement and antioxidant enzyme activity analysis. For ROS detection, we stained retinal sections with 10 µM Dihydroethidium (DHE) (#104821–25 − 2, MedChemExpress) for 30 min at 37 ℃. For lipid peroxidation assessment, malondialdehyde (MDA) concentration was calculated according to the manufacturer’s specifications (#ab118979, Abcam, California, US). Activities of antioxidant enzymes including superoxide dismutase (SOD), catalase (CAT) and glutathione peroxidase (GPx) were analyzed using commercial kits (#S0101S, #S0051, #S0056, Beyotime Biotech, Shanghai, China). Western blot analysis Pre-cooled radioimmunoprecipitation assay (RIPA) buffer (#P0013B, Beyotime Biotech) with protease and phosphatase inhibitors (#P1045, Beyotime Biotech) was used for tissue lysis. We then centrifuged the lysis mix and quantified protein concentration of the supernatant using the BCA assay kit (#P0010, Beyotime Biotech). Ten micrograms of protein were separated by SDS-PAGE and transferred to PVDF membranes (#1620177, Bio-Rad, Shanghai, China). Following blocking and antibodies incubation, we visualized protein bands using ECL reagent under ChemiDoc imaging system (#1705061, Bio-Rad). ImageJ was used for analyzing band intensity normalized to β-actin as the loading control. Primary and secondary antibodies used in this study were presented in Supplementary table S1. Caspase-3 activity assessment Caspase-3 pathway activation is a hallmark of apoptosis. To assess photoreceptor apoptosis after PS-NPs exposure and RD induction, a caspase-3 activity assay kit was utilized according to the manufacturers’ instructions (#C1115, Beyotime Biotech). Retinal tissue dissociation and microglia purification The whole retina was isolated, dissociated and sorted as described previously [24]. Briefly, the neuroretina was isolated and digested at 37 °C for 25 min using Neural Tissue Dissociation Kit (#130-094-802, Miltenyi Biotec, MD, USA). After terminating the digestion process by adding Ames’ buffer with 0.5% BSA (#A1420, Merck, Darmstadt, Germany), we centrifuged the mix to obtain the single-cell suspension. For microglia enrichment, CD11b MicroBeads (#130-093-634, Miltenyi Biotec) with separation columns (#130-042-401, Miltenyi Biotec) were utilized. BV2 microglia-661W photoreceptor co-culture system, PS-NPs exposure and oxygen glucose deprivation (OGD) model establishment Murine microglial cell line BV2 and murine photoreceptor cell line 661W were purchased from Fuheng Biology Inc. (Fuheng, Shanghai, China) and cultured in Dulbecco’s modified Eagle’s medium (DMEM) containing 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin (P/S). OGD represents the in vitro model of RD [25, 26], characterized by limited oxygen and nutrients supply. In this study, an indirect co-culture system was built using a 6.5 mm Transwell® with 0.4 μm pore (#14312, LabSelect, Beijing, China). BV2 microglia and 661W photoreceptors were passaged to the plates and inserts respectively. Based on MNPs deposition in human retina [16], we determined the appropriate dose (100 mg/L) and duration (24 h) of PS-NPs in the cellular co-culture system, which have been validated in previous research [22, 27]. Following PS-NPs exposure, we changed the DMEM medium to Earle’s balanced salt solution (#E7510, Merck) in 1% oxygen hypoxia conditions for 4 h. After OGD induction, BV2 microglia were collected for subsequent analysis. RNA extraction, reverse transcription and real-time quantitative real-time PCR (RT-qPCR) Total RNA was extracted using the FastPure Cell/Tissue Total RNA Isolation Kit (#RC101-01, Vazyme Biotech, Nanjing, China). Reverse transcription and RT-qPCR were performed with the HiScript III All-in-One RT SuperMix Kit (#R333-01, Vazyme Biotech) and the ChamQ Universal SYBR qPCR Master Mix (#Q711-02, Vazyme Biotech) respectively. Relative gene expression levels were calculated using the 2−ΔΔCt method with Actb as the internal control. The primer sequences used in this study were provided in Supplementary table S2. Enzyme-linked immunosorbent assay (ELISA) For retinal tissues, isolated retinas were lysed in pre-cooled RIPA buffer containing protease and phosphatase inhibitors. For BV2 microglia, we collected and centrifuged the cell culture media to obtain the clear supernatant. After quantification of protein content, we utilized commercial ELISA kits to determine the concentration of TNF-α (#PT512, Beyotime Biotech), IL-1β (#PI301, Beyotime Biotech), IL-6 (#PI326, Beyotime Biotech), SPP1 (#MOST00, R&D Systems, MN, USA), CCL4 (MMB00, R&D Systems), CXCL9 (MCX900, R&D Systems), CXCL10 (#PC206, Beyotime Biotech), MPO-DNA (#U96-1089E, YOBIBIO Biotech, Shanghai, China) and CitH3 (#U96-1160E, YOBIBIO Biotech) according to the manufacturer’s protocols. Each sample was measured in duplicate technical replicates. Transcriptome analysis Two public RNA-Sequencing (RNA-Seq) datasets were downloaded from GEO database (GSE243612, control and PS-NPs-exposed primary rat microglia samples; GSE41019, healthy and PVR human retinal samples) and analyzed. Differentially expressed genes (DEGs) were identified using DESeq2 at the criteria of |Log2Fold Change (FC)| > 1 and padj < 0.05. Functional enrichment analysis was performed on Metascape (http://metascape.org). Gene set enrichment analysis (GSEA) was conducted with clusterProfiler and enrichplot package at the criteria of |NES| > 1, p value < 0.05 and FDR q < 0.25. Search Tool for the Retrieval of Interacting Genes (STRING) online database (https://string-db.org/) was utilized for protein-protein interaction (PPI) network construction. In STRING, each protein-protein interaction was annotated with scores which ranged from 0 to 1 as indicator of confidence [28]. Statistical analysis All values were expressed as mean ± standard error of mean (SEM). Normality of the data was assessed using the Shapiro–Wilk test. For data that passed normality (p > 0.05), parametric tests, including unpaired Students’ t test and one-way analysis of variance (ANOVA) with Tukey’s multiple comparisons post-test, were used. For data that violated normality, non‑parametric tests (Mann–Whitney U or Kruskal–Wallis) were used. Visualization of the statistical results were performed using GraphPad Prism 8.0.2 (GraphPad Software Inc., San Diego, CA, USA). Statistical significance was set at p < 0.05. Results PS-NPs characterization and localization Spherical PS-NPs with the diameter of 100 nm were employed for long-term oral exposure on adult mice in this study. SEM confirmed uniformly spherical shape (Fig. 1A) and NTA revealed the size distribution with the average diameter of 118.47 ± 0.43 nm (Fig. 1B). FTIR exhibited absorption peaks at 3026, 1601, 1452 and 755 cm− 1 (Fig. 1C). Zeta potential of PS-NPs was − 16.06 ± 1.25 mV in water, indicating a negative surface charge (Fig. 1D). At the working dose of 10 mg/L and 50 mg/L, the particle concentrations were 2.6 * 109 particles/mL and 1.2 * 1010 particles/mL. Fig. 1PS-NPs characterization and localization. (A) Shape and size of PS-NPs under SEM. (B) NTA showing size distribution. (C) FTIR analysis displaying functional groups, as indicated by arrows. (D) Zeta potential of PS-NPs used in this study. (E) Localization of RFP-PS-NPs in murine retinas following long-term exposure via drinking water at two doses (10 mg/L, 50 mg/L). Data points and error bars indicate mean ± SEM. GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; OPL, outer plexiform layer; ONL, outer nuclear layer; RPE, retinal pigment epithelium PS-NPs characterization and localization. (A) Shape and size of PS-NPs under SEM. (B) NTA showing size distribution. (C) FTIR analysis displaying functional groups, as indicated by arrows. (D) Zeta potential of PS-NPs used in this study. (E) Localization of RFP-PS-NPs in murine retinas following long-term exposure via drinking water at two doses (10 mg/L, 50 mg/L). Data points and error bars indicate mean ± SEM. GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; OPL, outer plexiform layer; ONL, outer nuclear layer; RPE, retinal pigment epithelium Following a two-month exposure, RFP-PS-NPs were detected across retinal layers, indicating BRB penetration (Fig. 1E). The relative fluorescence intensity of RFP/DAPI was quantified (Supplementary Fig. S1). However, further studies using more direct and comprehensive approaches are warranted to investigate the transport process of PS-NPs across BRB. For subsequent experiments, we used PS-NPs without RFP conjugation to eliminate potential fluorescence interference. PS-NPs exposure impairs phototransduction dose-dependently We have previously investigated the damage of PS-NPs exposure on retinal function in adult mice [22]. In this study, we employed PS-NPs exposure at two doses (10 mg/L and 50 mg/L) and established RD model on mice after long-term exposure. We recorded and compared ERG responses across 6 groups: Control, PS-NPs-exposed (10 mg/L), PS-NPs-exposed (50 mg/L), RD, PS-NPs-exposed (10 mg/L) + RD and PS-NPs-exposed (10 mg/L) + RD (Fig. 2A). Under scotopic conditions, both doses of PS-NP exposure significantly lowered a-wave and b-wave amplitudes compared to controls, with greater reduction at 50 mg/L. This dose-dependent impairment persisted in RD-induced retinas (Fig. 2B, C). Photopic ERG responses showed similar trend without statistical significance (Fig. 2D, E). Collectively, we confirmed the PS-NPs-induced phototransduction deficits and demonstrated the dose-dependent impairment in murine retina with or without RD challenge. Fig. 2PS-NPs exposure impairs phototransduction dose-dependently. (A) Representative ERG recordings of Control, PS-NPs-exposed (10 mg/L), PS-NPs-exposed (50 mg/L), RD, PS-NPs-exposed (10 mg/L) + RD, PS-NPs-exposed (50 mg/L) + RD mice under flash stimuli at 0.01, 0.1, 1.0, 3.0 and 10.0 log cd∙s/m2 after dark adaptation (scotopic), and 10.0 log cd∙s/m2 after light adaptation (photopic). (B, C) Quantification of scotopic a-wave (B) and b-wave (C) amplitudes among six groups. n = 5, one-way ANOVA with Tukey’s test. (D, E) Quantification of photopic a-wave (D) and b-wave (E) amplitudes. n = 5, one‐way ANOVA with Tukey’s test. Data points and error bars indicate mean ± SEM. ns, not significant. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001 PS-NPs exposure impairs phototransduction dose-dependently. (A) Representative ERG recordings of Control, PS-NPs-exposed (10 mg/L), PS-NPs-exposed (50 mg/L), RD, PS-NPs-exposed (10 mg/L) + RD, PS-NPs-exposed (50 mg/L) + RD mice under flash stimuli at 0.01, 0.1, 1.0, 3.0 and 10.0 log cd∙s/m2 after dark adaptation (scotopic), and 10.0 log cd∙s/m2 after light adaptation (photopic). (B, C) Quantification of scotopic a-wave (B) and b-wave (C) amplitudes among six groups. n = 5, one-way ANOVA with Tukey’s test. (D, E) Quantification of photopic a-wave (D) and b-wave (E) amplitudes. n = 5, one‐way ANOVA with Tukey’s test. Data points and error bars indicate mean ± SEM. ns, not significant. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001 PS-NPs exposure aggravates oxidative stress and photoreceptor degeneration in RD dose-dependently Previous studies have demonstrated that PS-NPs exposure increases oxidative stress level in various tissues and cell types [29, 30]. In the present study, we stained retinal sections with DHE, a probe dye detecting ROS, and observed significantly elevated ROS level in PS-NPs-exposed mice compared to controls (Fig. 3A, B). Biochemical analyses showed that PS-NPs exposure reduced the activities of key antioxidant enzymes including SOD, CAT and GPx in murine retinas (Fig. 3C-E). In addition, increased MDA concentration confirmed enhanced lipid peroxidation in PS-NPs-exposed group (Supplementary Fig. S2). These effects exhibited dose-dependent responses, with 50 mg/L PS-NPs group showing more pronounced effects than the 10 mg/L group, consistent across both RD and non-RD conditions. Fig. 3PS-NPs exposure aggravates oxidative stress and photoreceptor degeneration in RD dose-dependently. (A) Representative images of DHE staining in murine retinas showing ROS accumulation. (B) Quantification of DHE/DAPI fluorescence intensity ratio. n = 5, one-way ANOVA with Tukey’s test. (C-E) SOD (C), CAT (D) and GPx (E) activities of murine retinas. n = 5, one‐way ANOVA with Tukey’s test. (F) Representative images of TUNEL staining. (G, H) Assessing photoreceptor degeneration by quantification of TUNEL+ nuclei proportion in ONL (G) and ONL thickness (H). n = 5, one‐way ANOVA with Tukey’s test. (I) Caspase-3 activity analysis of retinas. n = 5, one‐way ANOVA with Tukey’s test. (J, K) Western blot analysis of cleaved caspase-3 confirmed PS-NPs-aggravated photoreceptor degeneration in RD retinas. n = 3, one‐way ANOVA with Tukey’s test. Data were curated and presented as described in Fig. 2 PS-NPs exposure aggravates oxidative stress and photoreceptor degeneration in RD dose-dependently. (A) Representative images of DHE staining in murine retinas showing ROS accumulation. (B) Quantification of DHE/DAPI fluorescence intensity ratio. n = 5, one-way ANOVA with Tukey’s test. (C-E) SOD (C), CAT (D) and GPx (E) activities of murine retinas. n = 5, one‐way ANOVA with Tukey’s test. (F) Representative images of TUNEL staining. (G, H) Assessing photoreceptor degeneration by quantification of TUNEL+ nuclei proportion in ONL (G) and ONL thickness (H). n = 5, one‐way ANOVA with Tukey’s test. (I) Caspase-3 activity analysis of retinas. n = 5, one‐way ANOVA with Tukey’s test. (J, K) Western blot analysis of cleaved caspase-3 confirmed PS-NPs-aggravated photoreceptor degeneration in RD retinas. n = 3, one‐way ANOVA with Tukey’s test. Data were curated and presented as described in Fig. 2 Next, we performed TUNEL staining and quantified ONL thickness and proportion of apoptotic cells (TUNEL+) in ONL (Fig. 3F-H). Neither dose of PS-NPs exposure alone could induce significant photoreceptor apoptosis or ONL thinning, consistent with our previous report [22]. However, PS-NPs exposure markedly exacerbated RD-induced photoreceptor degeneration, particularly at the higher dose (50 mg/L). These findings were further corroborated by increased caspase-3 activity and elevated expression of cleaved caspase-3 using western blot analysis (Fig. 3I-K). PS-NPs exposure promotes microglial M1 polarization and peripheral immune cell infiltration in RD To elucidate the mechanisms by which PS-NPs exacerbated photoreceptor degeneration in RD, we selected the 10 mg/L exposure dose for subsequent experiments. Given the essential role of neuroinflammation in modulating photoreceptor survival, we quantified the level of proinflammatory cytokine (TNF-α, IL-1β, IL-6) secretion and their corresponding gene expression (Fig. 4A-C, Supplementary Fig. S3). While PS-NPs exposure alone failed to induce significant retinal inflammation, it markedly amplified inflammatory responses in RD-challenged retinas. Fig. 4PS-NPs exposure promotes microglial M1 polarization and peripheral immune cell infiltration in RD. (A-C) ELISA analysis showed PS-NPs exposure promoted secretion of proinflammatory cytokines including TNF-α (A), IL-1β (B) and IL-6 (C) in RD retinas. n = 5, one-way ANOVA with Tukey’s test. (D) Representative images of IBA1 immunofluorescent staining on retinal wholemounts and corresponding skeleton analysis images. (E-G) Quantification of soma area (E), number of endpoints/cell (F) and process length (G). n = 5, one‐way ANOVA with Tukey’s test. (H) Flow charts illustrating microglial purification by CD11b magnetic sorting. (I, J) Relative mRNA expression of microglial M1 (I) and M2 (J) markers. n = 3, one‐way ANOVA with Tukey’s test. (K) Representative images of H&E staining showed exacerbated peripheral immune cell infiltration in PS-NPs-exposed RD retinas. (L) Quantification of density of subretinal immune cells. n = 5, one‐way ANOVA with Tukey’s test. Data were curated and presented as described in Fig. 2 PS-NPs exposure promotes microglial M1 polarization and peripheral immune cell infiltration in RD. (A-C) ELISA analysis showed PS-NPs exposure promoted secretion of proinflammatory cytokines including TNF-α (A), IL-1β (B) and IL-6 (C) in RD retinas. n = 5, one-way ANOVA with Tukey’s test. (D) Representative images of IBA1 immunofluorescent staining on retinal wholemounts and corresponding skeleton analysis images. (E-G) Quantification of soma area (E), number of endpoints/cell (F) and process length (G). n = 5, one‐way ANOVA with Tukey’s test. (H) Flow charts illustrating microglial purification by CD11b magnetic sorting. (I, J) Relative mRNA expression of microglial M1 (I) and M2 (J) markers. n = 3, one‐way ANOVA with Tukey’s test. (K) Representative images of H&E staining showed exacerbated peripheral immune cell infiltration in PS-NPs-exposed RD retinas. (L) Quantification of density of subretinal immune cells. n = 5, one‐way ANOVA with Tukey’s test. Data were curated and presented as described in Fig. 2 Microglia are primary initiators of retinal inflammatory responses. We then performed IBA1 immunofluorescent staining on retinal sections and wholemounts to assess microglial activation (Fig. 4D-G, Supplementary Fig. S4). Similarly, PS-NPs exposure alone did not induce microglial activation in unchallenged retinas. However, following RD induction, PS-NP-exposed retinas demonstrated enhanced microglial migration and morphological activation, characterized by enlarged soma and reduced branching. We also purified retinal microglia using CD11b magnetic sorting strategy and found PS-NPs exposure promoted M1-polarization (proinflammatory state) exclusively in RD conditions (Fig. 4H-J). The proinflammatory microglial shift correlated with the increase in peripheral immune cell infiltration, as confirmed by H&E staining (Fig. 4K, L). Collectively, these findings suggested that PS-NPs exposure promoted microglial proinflammatory polarization and peripheral immune cell invasion in RD. PS-NPs exposure upregulates SPP1 and chemokine-associated signaling in microglia To investigate molecular pathways mediating PS-NPs-induced microglial activation, we analyzed a public transcriptomic dataset (GSE243612) of primary rat microglia exposed to PS-NPs for 24 h. Differential gene expression analysis (Fig. 5A) revealed coordinated upregulation of inflammation-associated mediators (Spp1, Tspo, Nfkbiz, Gas6) and lipid metabolism-associated regulators (Cd36, Fabp4), alongside downregulation of microglial homeostatic markers (Tmem119, P2ry12, P2ry13). Functional enrichment of upregulated DEGs (|Log2FC| > 1, padj < 0.05) demonstrated chemokine signaling pathway, cytokine-receptor interaction, TNF signaling pathway, neutrophil chemotaxis and positive regulation of inflammatory response (Fig. 5B, C). Furthermore, GSEA (|NES| > 1, p value < 0.05, FDR q < 0.25) confirmed leukocyte chemotaxis and chemokine signaling pathway activation, with key genes including Spp1, Ccl4, Cxcl9 and Cxcl10 (Fig. 5D, E). Concurrent activation of peroxisome and lipid oxidation pathways also aligned with our oxidative stress analyses above (Supplementary Fig. S5). Fig. 5PS-NPs exposure upregulates SPP1 and chemokine-associated signaling in microglia. (A) Volcano plot showing DEGs in PS-NPs-exposed microglia compared to controls. (B, C) KEGG (B) and GO (C) enrichment analysis of upregulated DEGs. (D, E) GSEA showing activated leukocyte chemotaxis (D) and chemokine signaling pathway (E) in PS-NPs-exposed microglia. (F) Flow charts illustrating OGD establishment in BV2-661W co-culture system. (G-J) ELISA analysis quantifying SPP1 (G), CCL4 (H), CXCL9 (I) and CXCL10 (J) secretion of BV2 microglia. n = 5, one-way ANOVA with Tukey’s test. (K-N) Relative mRNA expression of Spp1 (K), Ccl4 (L), Cxcl9 (M) and Cxcl10 (N) in BV2 microglia. n = 3, one‐way ANOVA with Tukey’s test. Data were curated and presented as described in Fig. 2 PS-NPs exposure upregulates SPP1 and chemokine-associated signaling in microglia. (A) Volcano plot showing DEGs in PS-NPs-exposed microglia compared to controls. (B, C) KEGG (B) and GO (C) enrichment analysis of upregulated DEGs. (D, E) GSEA showing activated leukocyte chemotaxis (D) and chemokine signaling pathway (E) in PS-NPs-exposed microglia. (F) Flow charts illustrating OGD establishment in BV2-661W co-culture system. (G-J) ELISA analysis quantifying SPP1 (G), CCL4 (H), CXCL9 (I) and CXCL10 (J) secretion of BV2 microglia. n = 5, one-way ANOVA with Tukey’s test. (K-N) Relative mRNA expression of Spp1 (K), Ccl4 (L), Cxcl9 (M) and Cxcl10 (N) in BV2 microglia. n = 3, one‐way ANOVA with Tukey’s test. Data were curated and presented as described in Fig. 2 We then conducted the co-culture of BV2 microglia and 661W photoreceptors with 100 mg/L PS-NPs for 24 h (Fig. 5F), with exposure dose and duration optimized from previous publications [22, 27]. After PS-NPs exposure, OGD was established on the co-culture system to mimic RD model in vitro. ELISA and RT-qPCR analyses demonstrated increased secretion and gene expression of SPP1, CCL4, CXCL9 and CXCL10 (Fig. 5G-N). These results suggested the potential involvement of microglial SPP1 and chemokine-associated signaling in PS-NPs-aggravated neutrophil infiltration during photoreceptor degeneration. SPP1 promotes neutrophil infiltration and NETs formation in PS-NPs-exposed RD retinas To functionally validate SPP1’s role in microglia, we first performed SPP1 and IBA1 immunofluorescent co-localization analysis (Supplementary Fig. S6). The average SPP1/IBA1 fluorescence intensity ratio in IBA1+ cells of Control, PS-NPs-exposed and PS-NPs-exposed + RD group was 0, 96.7% and 129.8% (ω2 = 0.95), confirming the activation of microglial SPP1 signaling in RD, with PS-NPs exposure further amplifying this response. Next, subretinal delivery of SPP1 neutralizing antibody was conducted to achieve SPP1 signaling blockade. Ly6G (a marker for neutrophil) and TUNEL co-staining revealed that SPP1 neutralization significantly attenuated PS-NPs-aggravated neutrophil (Ly6G+) infiltration and photoreceptor apoptosis (Fig. 6A-C). We also identified that SPP1 neutralization decreased the secretion of chemokines including CCL4, CXCL9 and CXCL10 (Fig. 6D-F). In neurodegenerative diseases, infiltrated neutrophils could exacerbate tissue damage through NETs formation and subsequent release of inflammatory factors [31, 32]. Quantitative analysis of MPO-DNA and CitH3 (markers for NETs) secretion demonstrated that PS-NPs exposure enhanced NETs formation in RD, which was partially reversed by SPP1 neutralization (Supplementary Fig. S7). Ly6G and CitH3 co-localization further confirmed the protective role of SPP1 neutralization against PS-NPs-induced NETs formation (Fig. 6G, H). To test whether chemokines mediated SPP1’s effects, we conducted recombinant SPP1 protein administration and chemokine receptor inhibition experiments. The results suggested that recombinant SPP1 protein aggravated neutrophil infiltration, NETs formation and photoreceptor apoptosis, while pharmacological inhibition of CCR5 (receptor for CCL4) and CXCR3 (receptor for CXCL9 and CXCL10) by TAK-779 alleviated these pathological phenotypes (Fig. 6I-M). Collectively, these findings supported a causal role of SPP1 in promoting neutrophil-driven retinal damage in part through CCL4-CCR5 and CXCL9/10-CXCR3 axes. Fig. 6SPP1 promotes neutrophil infiltration and NETs formation in PS-NPs-exposed RD retinas. (A) Representative images of Ly6G and TUNEL co-staining showing decreased neutrophil (Ly6G+) infiltration and photoreceptor apoptosis by SPP1 neutralization. The white arrows indicated infiltrated neutrophils in subretinal space. (B, C) Quantification of Ly6G/DAPI fluorescence intensity ratio (B) and TUNEL+ nuclei proportion in ONL (C). n = 5, one-way ANOVA with Tukey’s test. (D-F) Quantification of CCL4 (D), CXCL9 (E) and CXCL10 (F) secretion. n = 5, one‐way ANOVA with Tukey’s test. (G) Representative images of Ly6G and CitH3 co-staining. (H) Quantification of CitH3/Ly6G fluorescence intensity ratio in Ly6G+ cells. n = 5, one‐way ANOVA with Tukey’s test. (I) Representative images of Ly6G and TUNEL co-staining on PS-NPs-exposed + RD, PS-NPs-exposed + RD + recombinant SPP1 protein, PS-NPs-exposed + RD + recombinant SPP1 protein + TAK-779 retinas. TAK-779 is a CCR5/CXCR3 antagonist. (J, K) Quantification of Ly6G/DAPI fluorescence intensity ratio (J) and TUNEL+ nuclei proportion in ONL (K). n = 5, one‐way ANOVA with Tukey’s test. (L, M) ELISA analysis of NETs markers including MPO-DNA (L) and CitH3 (M). n = 5, one‐way ANOVA with Tukey’s test. Data were curated and presented as described in Fig. 2 SPP1 promotes neutrophil infiltration and NETs formation in PS-NPs-exposed RD retinas. (A) Representative images of Ly6G and TUNEL co-staining showing decreased neutrophil (Ly6G+) infiltration and photoreceptor apoptosis by SPP1 neutralization. The white arrows indicated infiltrated neutrophils in subretinal space. (B, C) Quantification of Ly6G/DAPI fluorescence intensity ratio (B) and TUNEL+ nuclei proportion in ONL (C). n = 5, one-way ANOVA with Tukey’s test. (D-F) Quantification of CCL4 (D), CXCL9 (E) and CXCL10 (F) secretion. n = 5, one‐way ANOVA with Tukey’s test. (G) Representative images of Ly6G and CitH3 co-staining. (H) Quantification of CitH3/Ly6G fluorescence intensity ratio in Ly6G+ cells. n = 5, one‐way ANOVA with Tukey’s test. (I) Representative images of Ly6G and TUNEL co-staining on PS-NPs-exposed + RD, PS-NPs-exposed + RD + recombinant SPP1 protein, PS-NPs-exposed + RD + recombinant SPP1 protein + TAK-779 retinas. TAK-779 is a CCR5/CXCR3 antagonist. (J, K) Quantification of Ly6G/DAPI fluorescence intensity ratio (J) and TUNEL+ nuclei proportion in ONL (K). n = 5, one‐way ANOVA with Tukey’s test. (L, M) ELISA analysis of NETs markers including MPO-DNA (L) and CitH3 (M). n = 5, one‐way ANOVA with Tukey’s test. Data were curated and presented as described in Fig. 2 NETs inhibition attenuates PS-NPs-exacerbated microglial proinflammatory activation and photoreceptor apoptosis in RD To analyze the effects of PS-NPs-potentiated NETs formation on retinal inflammation and photoreceptor degeneration, we administrated DNase I via subretinal injection. Decreased NETs formation was confirmed by Ly6G-CitH3 co-staining and ELISA analysis of MPO-DNA and CitH3 (Fig. 7A-D). We found that DNase I treatment significantly attenuated retinal inflammation, microglial migration and photoreceptor apoptosis (Fig. 7E-G, Supplementary Fig. S8). To validate these findings, Cl-amidine (a NETs inhibitor) was administrated via intraperitoneal injection. ELISA and RT-qPCR analyses indicated that NETs inhibition demonstrated a mitigating effect against retinal inflammation (Supplementary Fig. S9). Using CD11b magnetic sorting strategy, we identified a shift toward M2 polarization (anti-inflammatory) in purified microglia following Cl-amidine administration (Fig. 7H, I). IBA1 and TUNEL co-staining analysis indicated that the average density of microglia in ONL of PS-NPs-exposed + RD and PS-NPs-exposed + RD + Cl-amidine group was 17.2 and 7.8 (Hedge’s g = 2.42). Reduced photoreceptor apoptosis was also observed in NETs inhibition group (Fig. 7J-L). Caspase-3 activity assays and cleaved caspase-3 immunoblotting confirmed the neuroprotective effect of NETs inhibition in PS-NPs-exposed + RD retinas (Fig. 7M, Supplementary Fig. S10). Collectively, these results suggested NETs as potential amplifiers of the PS-NPs-promoted neuroinflammatory cascade, wherein “microglial recruitment of neutrophils - NETs formation - microglial proinflammatory activation” forms a self-perpetuating loop exacerbating photoreceptor degeneration. Fig. 7NETs inhibition attenuates PS-NPs-exacerbated microglial proinflammatory activation and photoreceptor apoptosis in RD. (A) Representative images of Ly6G and CitH3 co-staining indicated the reduction of NETs formation following DNase I treatment. (B) Quantification of CitH3/Ly6G fluorescence intensity ratio in Ly6G+ cells. n = 5, unpaired Students’ t test. (C, D) ELISA analysis of NETs markers including MPO-DNA (C) and CitH3 (D). n = 5, unpaired Students’ t test. (E-G) ELISA analysis of TNF-α (E), IL-1β (F) and IL-6 (G) in retinas. n = 5, unpaired Students’ t test. (H, I) Relative mRNA expression of M1 (H) and M2 (I) markers in purified retinal microglia following Cl-amidine administration. n = 3, unpaired Students’ t test. (J) Representative images of IBA1 and TUNEL co-staining showing alleviated microglial activation and photoreceptor apoptosis by NETs inhibition. (K, L) Quantification of density of microglia (K) and TUNEL+ nuclei proportion (L) in ONL. n = 5, unpaired Students’ t test. (M) Quantification of caspase-3 activity. n = 5, unpaired Students’ t test. Data were curated and presented as described in Fig. 2 NETs inhibition attenuates PS-NPs-exacerbated microglial proinflammatory activation and photoreceptor apoptosis in RD. (A) Representative images of Ly6G and CitH3 co-staining indicated the reduction of NETs formation following DNase I treatment. (B) Quantification of CitH3/Ly6G fluorescence intensity ratio in Ly6G+ cells. n = 5, unpaired Students’ t test. (C, D) ELISA analysis of NETs markers including MPO-DNA (C) and CitH3 (D). n = 5, unpaired Students’ t test. (E-G) ELISA analysis of TNF-α (E), IL-1β (F) and IL-6 (G) in retinas. n = 5, unpaired Students’ t test. (H, I) Relative mRNA expression of M1 (H) and M2 (I) markers in purified retinal microglia following Cl-amidine administration. n = 3, unpaired Students’ t test. (J) Representative images of IBA1 and TUNEL co-staining showing alleviated microglial activation and photoreceptor apoptosis by NETs inhibition. (K, L) Quantification of density of microglia (K) and TUNEL+ nuclei proportion (L) in ONL. n = 5, unpaired Students’ t test. (M) Quantification of caspase-3 activity. n = 5, unpaired Students’ t test. Data were curated and presented as described in Fig. 2 PS-NPs-induced cytotoxicity shares transcriptomic features with human tractional retinal detachment As described above, PS-NPs exposure acted as an environmental hazard exacerbating retinal inflammatory responses and photoreceptor degeneration in the mouse model of RD. To investigate whether PS-NPs exposure could constitute a risk factor for human retinal injury, we obtained a transcriptomic dataset from GEO database (GSE41019) containing human samples with PVR, which is characterized by tractional RD and fibrous membrane formation. A total of 255 upregulated DEGs (|Log2FC| > 1, padj < 0.05) were identified in fibrous membrane tissues compared to healthy retinal tissues. We also extracted the upregulated DEGs from PS-NPs-exposed primary microglia in Fig. 5, and performed an overlap analysis between these two datasets. This revealed 38 common upregulated DEGs, including SPP1, FAS, PLAU and ANGPT2, which were functionally associated with stress and stimulus response pathways (Fig. 8A). Fig. 8PS-NPs-induced cytotoxicity shares transcriptomic features with human tractional retinal detachment. (A) The Venn diagram showed 38 overlapped upregulated DEGs between PS-NPs-exposed microglia dataset and human PVR dataset. (B) The PPI network showed SPP1 (highlighted by the red square) as a hub gene among overlapped upregulated DEGs associated with stress and stimulus response pathways. (C) The schematic diagram of this study PS-NPs-induced cytotoxicity shares transcriptomic features with human tractional retinal detachment. (A) The Venn diagram showed 38 overlapped upregulated DEGs between PS-NPs-exposed microglia dataset and human PVR dataset. (B) The PPI network showed SPP1 (highlighted by the red square) as a hub gene among overlapped upregulated DEGs associated with stress and stimulus response pathways. (C) The schematic diagram of this study Furthermore, a PPI network was constructed using the STRING online database based on these overlapped DEGs (Fig. 8B). The network consisted of 38 nodes and 33 edges, with SPP1 emerging as a hub gene exhibiting high connectivity. Previous studies have also reported upregulated SPP1 signaling closely related to tissue remodeling processes in PVR [33]. Collectively, these findings suggested that PS-NPs exposure induced molecular alterations similar to those observed in PVR patients, which indicated PS-NPs as a potential modifying environmental factor. The schematic diagram of this study was presented as Fig. 8C. Discussion As emerging environmental pollutants, MNPs have been detected in various human tissues and excreta, inducing adverse health impacts through multiple toxicological mechanisms. Photoreceptor degeneration, a leading cause of irreversible blindness, is modulated by both genetic and environmental factors [34]. In this study, we demonstrated the dose-dependent retinal damage induced by long-term PS-NPs exposure and elucidated the underlying molecular mechanisms. We showed that PS-NPs exposure impaired phototransduction, exacerbated oxidative stress, and promoted photoreceptor apoptosis dose-dependently in RD retinas. We revealed microglial proinflammatory activation and recruitment of peripheral immune cells as pivotal mechanisms underlying PS-NPs-aggravated photoreceptor degeneration. Combined with transcriptome analysis, cellular co-culture system and OGD establishment, we identified significant activation of SPP1 and chemokine-associated signaling in PS-NPs-exposed microglia. SPP1 neutralization, recombinant SPP1 protein administration, chemokine receptor inhibition and NETs inhibition experiments further delineated a pathogenic axis linking microglial SPP1 and chemokine-associated signaling to neutrophil infiltration and NETs formation. Finally, we observed shared transcriptomic features between PS-NPs-exposed microglia and human PVR patients, particularly in the upregulation of SPP1 signaling, further implicating MNPs as a potential modifying environmental factor. This study demonstrated that PS-NPs exposure elicited minimal retinal morphological alterations but increased retinal susceptibility to injury, as evidenced by exacerbated retinal thinning and apoptosis following RD challenge. However, previous studies reported that PS-NPs exposure alone caused significant retinal morphological disruption, oxidative stress exacerbation and cell death [27, 35]. These different outcomes could be associated with exposure parameters, including exposure doses, particle size, administration route and exposure duration. Therefore, future studies are warranted to systematically investigate the critical values of these parameters to induce direct retinal damage without additional challenge, which may correspond to the exposure thresholds capable of causing overt human retinal injury in real world. SPP1 is a pleiotropic SIBLING family glycoprotein involved in cell adhesion, migration, and immune modulation. Our gain- and loss-of-function experiments supported a causal role of SPP1 in promoting neutrophil infiltration, NETs formation and photoreceptor degeneration under PS-NPs exposure. Mechanistically, SPP1 upregulated CCL4, CXCL9, and CXCL10 secretion in the retina, and pharmacological inhibition of their receptors CCR5/CXCR3 with TAK-779 partially reversed these effects, indicating that SPP1 acts in part via the CCL4-CCR5 and CXCL9/10-CXCR3 axes. However, SPP1 could directly engage cell surface receptors such as CD44 or integrins. As reported previously, SPP1 binding CD44 and αvβ3 activated breast fibroblast inflammation and promoted tumor growth [36]. Similarly, in acute kidney injury, SPP1 mediated the recruitment of CD44⁺ neutrophils [37]. Therefore, future studies are warranted to clarify whether SPP1 directly recruits neutrophils via these receptors or indirectly modulates the retinal microenvironment through microglia. NETs are fibrous structures composed of DNA scaffolds, histones, and granular proteins. Released by activated neutrophils, NETs play essential roles in host defense [38]. Our study revealed that PS-NPs exposure enhanced microglial recruitment of peripheral neutrophils, while excessive NETs further sustained microglial activation. Based on these observations, we inferred that NETs might serve as potential amplifiers in PS-NPs-driven positive feedback inflammatory loop and orchestrators of microglia-neutrophil crosstalk. This aligns with prior evidence showing that NETs participated in self-sustaining macrophage-neutrophil interactions during HCC progression [39] and mediated neutrophil-Th17 cell crosstalk via the NETs-CD44-IL-17 A axis in autoimmune uveitis [40]. Nevertheless, further studies employing fluorescence-based tracking approaches are needed to validate the PS-NPs-promoted neuroinflammation loop and elucidate the role of NETs in microglia-neutrophil crosstalk in RD. The exposure doses were selected based on our previous research [21, 22]. Senathirajah et al. identified that human intake of MPs ranged from 0.1 to 5 g/week/person, which transformed to 0.23–11.9 mg/kg body weight/day assuming the average human body weight as 60 kg [41]. The Environmental Working Group then reported human MPs intake as 18.33–71.69 g/year/person, translating to 0.84–3.27 mg/kg body weight/day. Recently, a modeling study analyzing MPs consumption from 1990 to 2018 across 109 countries demonstrated human MPs intake as 0.85–15 g/month/person, translating to 0.47–8.33 mg/kg body weight/day [42]. Based on the correction factor (human = 37, mouse = 3), our in vivo doses (1.6 mg/kg body weight/day and 8.0 mg/kg body weight/day for mice) transformed to 0.13 mg/kg body weight/day and 0.65 mg/kg body weight/day for humans, which fell at or below exposure levels in real world. However, Mohamed Nor et al. proposed that previous mass-based estimation of human MPs intake might not accurately reflect real-world exposure levels, as the diameter, shape and density of particles were diverse [43]. Moreover, Qian et al. detected 2.4 * 105 MNPs fragments per liter of bottled water (NPs accounting for 90%), which could be orders of magnitude higher than previous MPs estimates [20]. These findings highlighted the need for more precise detection techniques of MPs and NPs, and more standardized particle-mass dose transformations to provide environmentally relevant dose ranges for toxicological research. Therefore, the translational limitations of toxicological results in this study should be noted, and more real-world-relevant exposure patterns are needed in future studies. The reported MNPs deposition in human retina was 8.93–91.05 µg/g [16]. Considering the wet weight (3.3 mg) and total volume (1.56 mm3) of murine retina, adjusted by the correction factor, the in vitro dose for murine BV2 microglia and 661W photoreceptors was translated to 3.83 µg/g in human tissues, which was conservatively below the deposition concentration. However, the deposition-based translation might underestimate the effects of polymer type, particle size and exposure duration on deposition fraction and kinetics. The acute in vitro exposure-induced cytotoxicity could also be different from deposition-induced toxicity in tissues. Besides, the tissue volume-culture medium volume translation should be noted, as the uniform exposure in culture medium might not fully reflect the uneven distribution of MNPs in different retinal layers. Therefore, further studies on MNPs content detection in specific regions and more appropriate dose translation strategies are warranted. A limitation of this study is the homogeneity of exposure materials and exposure paradigms. Future toxicological studies should employ MNPs with varying particle sizes, shapes and surface chemistry. Another pivotal limitation is that our gain- and loss-of-function experiments of SPP1 were not microglia-specific, and we were unable to distinguish between retinal microglia and infiltrating immune cells (e.g., monocytes/macrophages) in the subretinal space. To address these limitations, future work could explore lineage-tracing techniques, microglia-specific depletion strategies and more direct neutrophil chemotaxis assays. Furthermore, the use of the murine microglial cell line BV2 might not fully recapitulate the complex biology of primary retinal microglia, including baseline immune activation and potentially their interaction with photoreceptors. Therefore, findings from BV2 cells should be validated in further investigations using primary retinal microglia. Conclusions In summary, this study demonstrated the dose-dependent toxicity of long-term PS-NPs exposure on retinal function and photoreceptor degeneration. Mechanistically, a PS-NPs-promoted neuroinflammatory loop associated with SPP1 signaling and NETs formation was identified in retinal degenerative diseases. Our study provided new insights into the impacts and potential targets of MNPs exposure on retinal health. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1 Supplementary Material 2 Supplementary Material 2 Supplementary Material 3 Supplementary Material 3