The protective potential of intestinal Bifidobacterium-derived extracellular vesicles with unique lipid composition against blue light-induced damage in retinal cells Authors: Wei-Hsuan Hsu, Pin-Hsuan Su, Hao-Yuan Chien, Tzu-Ming Pan, Bao-Hong Lee Published: 2026 ABSTRACT AbstractBifidobacterium longum is a probiotic that resides in the intestine, and extracellular vesicles (EVs) released by microorganisms have garnered attention in recent years for their role in regulating host health. This study found that triacylglycerol, cholesteryl esters, and free fatty acid (18:0 and 16:0) are primarily composed of lipids in B. longum-derived EVs by lipidomics assay. Such lipid composition facilitates the entry of EVs into ARPE-19 cells and protects the cells from oxidative death induced by blue light exposure. Transcriptomic analysis revealed that treatment with B. longum-derived EVs can regulate phosphoinositide 3-kinase and mTOR, thereby enhancing the survival capacity of blue light-induced ARPE-19 cells. Intestinal microbiota is strongly associated with host disease, and this study provides a strategy suggesting that intestinal probiotic EVs might potentially be used in the future to protect eyes from damage caused by blue light sensitivity through the gut-eye axis.Keywords: Bifidobacterium longum, Blue light, Extracellular vesicles, Gut-eye axis, Lipidomics, Oxidative death 1. Introduction Extracellular vesicles (EVs) not only function as signaling mediators across different species [1], but recent studies have also demonstrated that EVs derived from plants and probiotics possess the ability to modulate gut microbiota and inhibit pathogenic and opportunistic bacteria [2–4]. In addition to the interactions between microorganisms, EVs derived from lactic acid bacteria have also been found to exert protective effects on mammalian cells through cross-species mechanisms [5]. The bioactivities of probiotic EVs have been widely explored in various studies. For instance, EVs derived from Lactobacillus plantarum exhibit antibacterial activity [6], while those from Lactobacillus rhamnosus GG have shown anti-cancer effects [7]. Additionally, EVs from Bifidobacterium breve and Lactobacillus rhamnosus can be recognized by intestinal epithelial cells and dendritic cells, thereby modulating gut immune functions [8]. Blue light-induced damage to the retina is a slow, progressive dynamic process [9]. Many compounds are considered to be associated with retinal protection due to their antioxidant properties [10,11]. Current research increasingly highlights the importance of gut microbiota in host health, demonstrating protective effects on various organs through their metabolites [12,13]. Beyond metabolites, EVs derived from gut microbiota have also been shown to provide protective benefits to host organs [1,14]. The gut-eye axis theory has recently been proposed, suggesting that gut microbiota plays a crucial role in eye protection [15]. Based on this body of evidence, this study aims to screen different strains of Bifidobacterium longum and evaluate the protective potential of EVs secreted by these B. longum strains against blue light-induced damage in retinal cells. 2. Materials and methods 2.1. Culture of lactic acid bacteria The 8 strains of Bifidobacterium longum were isolated from human feces, including BLBI7, BLBI16, BLBI17, BLBI20, BLBI24, BLBI332, BLBI262, and BLMR22. These lactic acid bacteria were cultured in De Man Rogosa and Sharp (MRS) broth (BD Biosciences, CA, USA) containing 0.5% cysteine in anaerobic condition at 37 °C. 2.2. Isolation and characterization of EVs These lactic acid bacteria were cultured and grown until reaching the stationary phase. The cultured-medium were first centrifuged at 3000×g for 15 min to collect the supernatant. The supernatant was then centrifuged at 35,000×g for 60 min to remove cell debris, organelles, and aggregated proteins, followed by filtration through a 0.22 μm filter. Finally, the filtrate was centrifuged at 100,000×g (L-90K, Beckman, Pasadena, CA, USA) for 60 min, and the resulting pellet was resuspended in PBS [16]. The EVs numbers and size distribution were measured by dynamic light scattering (NanoSight LM10-HS, Malvern Instruments, Worcestershire, UK). The observation of EVs morphology was carried out by transmission electron microscopy (TEM) (JEM-2100F, JEOL, Akishima City, Japan) [17]. 2.3. Lipidomics assay EVs (200 μL) was mixed with 480 μL extract solution (methyl tert-butyl ether: methanol = 5 : 1) containing internal standard was added. The sample was sonicated for 10 min in ice-bath. Samples were centrifuged at 3000 rpm for 15 min at 4 °C. The supernatant (250 μL) was transferred to a fresh tube. The rest of the sample was added with 250 μL of methyl tert-butyl ether, sonication and centrifugation, and another 250 μL of supernatant was taken out. The supernatants were combined and dried in a vacuum concentrator at 4 °C. Then, the dried samples were reconstituted in 100 μL of resuspension buffer (dichloromethane: MeOH: H2O = 60: 30: 4.5) by sonication on ice for 10 min. The constitution was then centrifuged at 12,000 rpm for 15 min at 4 °C, and supernatant (30 μL) was transferred to a fresh glass vial for analysis by ultra high performance liquid chromatography. The absolute content of individuals lipids corresponding to the internal standard (IS) was calculated on the basis of peaks area and actual concentration of the identical lipid class IS, and then absolute content was obtained from diverse IS averaged of the identical lipid class [18]. 2.4. Cell culture The ARPE-19, a human retinal pigmented epithelium cell line (BCRC60383), was purchased from the Bioresource Collection and Research Center (BCRC) in Hsinchu, Taiwan. Mixed Dulbecco’s modified Eagle’s medium and Ham’s F12 medium (v:v/1:1) containing 1.2 g/L sodium bicarbonate, 2.5 mM l-glutamine, 15 mM HEPES and 0.5 mM sodium pyruvate, and 10% fetal bovine serum was used to culture ARPE-19 cells in a humidified atmosphere (95% air and 5% CO2) at 37 °C. 2.5. Blue light induction and treatment The ARPE-19 cells were inoculated in 96-well plates at a cell density of 3 × 104 cells/well with a volume of 100 μL/well culture medium. Cells were exposed to blue light (415–455 nm; 6000 lux) for different durations (30 min, 1 h, and 2 h), followed by a 24-h or 48-h incubation [19]. Cell viability was analyzed using the CCK-8 assay (absorbance measured at 450 nm) (Multiskan Go, Thermo Scientific, Waltham, MA, USA) and crystal violet staining. In the EVs treatment groups, EVs were added to the cell culture medium for a 4-h pre-treatment before induction by blue light exposure [20]. 2.6. Measurement of oxidative stress After treatment with EVs and blue light exposure, the cells were incubated with 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA) (Sigma-Aldrich, St. Louis, MO, USA) for 20 min at 37 °C. The residual DCFH-DA was washed away with PBS, followed by nuclear staining with Hoechst. Fluorescence intensity was observed using a fluorescence microscop (Leica Microsystems, Mannheim, Germany) [21]. 2.7. EVs tracking Mixed 75 μL of EVs at a concentration of 1010 particles/mL with 25 μL of 100 μM carboxy-fluorescein diacetate, succinimidyl ester (CFSE) dye (Invitrogen, Carlsbad, CA, USA), and incubated the mixture at 37 °C for 1 h. After incubation, collect the CFSE-labeled EVs by ultracentrifugation. Seeding cells (2 × 105) onto poly-d-lysine-coated (Thermo Scientific, Waltham, MA, USA) coverslips placed in a 24-well culture plate, and incubated at 37 °C for 24 h. After incubation, replace the cell culture medium with fresh medium containing 1010 particles of CFSE-labeled EVs and incubate the cells at 37 °C for 60 min. The CellMask™ Deep Red Plasma Membrane Stain (Thermo Scientific) was mixed and incubated with cells at 37 °C for 1 h. After staining, remove the dye and rinse the cells 3–5 times with DPBS. Mount the coverslips onto glass slides using mounting medium containing DAPI (4′,6-diamidino- 2-phenylindole) (Abcam, Bristol, UK). Finally, capture and observe the samples using a confocal microscopy imaging system (FV3000, Olympus, Shinjuku, Japan). 2.8. Transcriptome analysis mRNA transcriptomic analysis was conducted on ARPE-19 cells after treatment with EVs and blue light exposure. Total RNA was extracted using Trizol® reagent, and RNA quantity was assessed with a Bioanalyzer 2100 (Agilent Technologies, USA) using the RNA 6000 LabChip kit. RNA preparation followed Illumina’s official protocol. For library construction, the Agilent SureSelect Strand-Specific RNA Library Preparation Kit was utilized, followed by size selection with AMPure XP beads (Beckman Coulter, USA). In the RNA sequencing data analysis workflow, quality trimming to remove low-quality reads and bases was performed using Trimmomatic [22]. Differential expression analysis was conducted with Cuffdiff, incorporating genome bias detection and correction. Genes exhibiting a > 2-fold change in expression in response to blue light or EVs treatments were identified. Further analysis was performed on these differentially expressed genes [23]. 2.9. Statistical analysis The experimental results were obtained from triplicate analyses and are presented as the mean ± standard deviation (SD). Data analysis was performed using the Statistical Analysis System (SAS Inc., Cary, NC, USA). One-way analysis of variance (ANOVA) was conducted to assess differences among groups, and Duncan’s multiple range test was used to determine significant differences between means. A p-value of less than 0.05 was considered statistically significant [24]. 3. Results and discussion 3.1. Protective effects of B. longum-derived extracellular vesicles against blue light-induced retinal damage The ARPE-19 cells were respectively exposed to blue light for 30 min, 1 h, and 2 h, followed by a 24-h incubation to assess the damage caused by blue light exposure to the cells. As shown in Supplementary Fig. 1, a markedly decrease in cell viability was observed after 1 h and 2 h of blue light exposure. In contrast to the extensive research and numerous reviews on EVs derived from Lactobacillaceae [25], studies on EVs from Bifidobacteriaceae remain scarce. Applications of extracellular vesicles derived from Bifidobacteria include immunotherapy [26], improvement of skin barrier dysfunction [27], enhancement of intestinal IgA expression [28], and anti-inflammatory effects [29]. Notably, EVs derived from B. longum have been found to contain proteins capable of binding to mucin in the gut [30]. We screened 8 strains of Bifidobacterium longum isolated from infant feces (including BLBI7, BLBI16, BLBI17, BLBI20, BLBI24, BLBI332, BLBI362, and BLMR22) to protect AREP-19 cells against blue light-induced damage. After culturing these B. longum strains, EVs (108 particles/mL) from the supernatant were isoated and the protective effects in ARPE-19 cells were evaluated. The results showed that EVs from BLMR22 strain had the most significant protective effect on cells to avoid cell death caused by blue light (Fig. 1). The concentration- dependent (106–1010 particles/mL) protective effects of BLMR22-derived EVs on ARPE-19 cells against blue light-induced damage were further confirmed (Fig. 2). Protective Effects of EVs Derived from Different Bifidobacterium longum Strains Against Blue Light-Induced Cell Death in ARPE-19 Retinal Cells (CCK-8 Assay). Data are presented as mean ± SD (n = 3). * indicates a significant difference (p < 0.05) compared to the control. ## indicates a highly significant difference (p < 0.01) compared to blue light induction. ### indicates a highly significant difference (p < 0.001) compared to blue light induction. Protective Effects of Different Concentrations of BLMR22-Derived EVs Against Blue Light-Induced Cell Death in ARPE-19 Retinal Cells. (A) Cell count presented using crystal violet staining. (B) Cell viability assessed by CCK-8 assay. Data are presented as mean ± SD (n = 3). * indicates a significant difference (p < 0.05) compared to the control. # indicates a significant difference (p < 0.05) compared to blue light induction. ## indicates a highly significant difference (p < 0.01) compared to blue light induction. Blue light-induced eye damage is often associated with oxidative stress [31]. We further evaluated the oxidative stress in ARPE-19 cells after blue light exposure. DCFH-DA, after crossing the cell membrane, is hydrolyzed by esterases into DCFH. In the presence of a large number of free radicals and reacted with DCFH, resulting in the loss of electrons and the formation of fluorescent DCF. Therefore, it can be used as an indicator to assess intracellular oxidative stress. The results showed that ARPE-19 cells produced a significant amount of free radicals after blue light exposure. However, pre-treatment with EVs (106–1010 particles/mL) markedly prevented the cells from experiencing oxidative stress (Fig. 3). Oxidative Stress Induced by Blue Light Exposure in ARPE-19 Retinal Cells. (A) Quantitative and statistical analysis of results. (B) Fluorescence images. Data are presented as mean ± SD (n = 3). * indicates a significant difference (p < 0.05) compared to the control. ** indicates a highly significant difference (p < 0.01) compared to the control. # indicates a significant difference (p < 0.05) compared to blue light induction. ## indicates a highly significant difference (p < 0.01) compared to blue light induction. The retinal pigment epithelium is a cell that regulates the intraocular immune environment and is also a target for damage caused by intense light exposure [32]. Blue light can cause mitochondrial activity dysfunction [33], leading to the accumulation of a large amount of free radicals and reactive oxygen species within cells, ultimately resulting in necrosis of ARPE-19 cells [34]. Another study showed that under blue light irradiation conditions that do not cause ARPE-19 cell death (lower light intensity), blue light can induce mitochondrial aggregation (ultrastructural changes) [35]. Therefore, many antioxidants are expected to protect against blue light-induced eye damage, including extracts obtained from Prunella vulgaris [36], Dipterocarpus tuberculatus [37], Panax ginseng berry [20], and astaxanthin [38]. The gut is a crucial organ, various axes have been discovered that connect with the gut, such as the gut-brain axis, gut-skin axis, gut-liver axis, and gut-kidney axis in recent years [39]. The theory of the gut-eye axis has been proposed in recent years [40]. There is a significant association between intestinal microbiota dysbiosis and the occurrence of ocular degenerative diseases; application of potential probiotics to improve eye diseases is also gaining increasing attention [41]. Based on the above evidence, we hypothesized that intestinal probiotics can protect against blue light-induced eye damage, and EVs may play a crucial role in mediating this defense process. 3.2. Tracking and characterization of EVs Through the above experiments, we confirmed that EVs secreted by the BLMR22 strain have the most protective effect on ARPE-19 cells. Next, we examined the uptake of BLMR22-derived EVs by ARPE-19 cells. EVs were labeled with CFSE for fluorescence and then treated with the cells. DAPI and CellMask were used to stain the cell nucleus and cell membrane, respectively. The results showed that substantial distribution of EVs was observed in both the cell membrane and cytoplasm of ARPE-19 cells after treatment for 1 h (Fig. 4). These findings prompted us to further investigate the mechanism by which B. longum-BLMR22- derived EVs enter the cells. The lipid composition of EVs has been found to affect their ability to enter into cell cross-species [42]. We investigated the size (112 nm) (Fig. 5A) and morphology (Fig. 5B) of BLMR22-derived EVs to confirm their purity, conducted a lipidomics investigation. Distribution Tracking of BLMR22-Derived EVs in ARPE-19 Retinal Cells. DAPI is used for nuclear staining, CFSE for labeling BLMR22- derived EVs, and CellMask for cell membrane staining. Characterization of BLMR22-Derived EVs. (A) Size analysis by Nanoparticle Tracking Analysis (NTA), (B) Morphological observation by Transmission Electron Microscopy (TEM). Lipidome Analysis of BLMR22-Derived EVs, including (C) TAG/DAG ratio, (D) Types of Triacylglycerols (TAGs), (E) Types of Sterols, (F) Types of Free Fatty Acids (FFAs). Triacylglycerol (TAG) (93.6%) is the predominant glycerolipid in BLMR22-derived EVs compared to diacylglycerol (DAG) (6.3%) (Fig. 5C). The top four major TAG components are TAG 52:2-FA18:1 (19.9%), TAG 52:2-FA16:0 (13.1%), TAG 50:1-FA16:0 (12.9%), and TAG 52:3-FA18:1 (10.0%) (Fig. 5D). The top three major cholesteryl esters (CE) are CE (22:6) (34.9%), CE (20:4) (22.54%), and CE (18:2) (22.0%) (Fig. 5E). The proportion of free fatty acids (FFAs) in BLMR22-derived EVs is predominantly occupied by FFA (18:0) and FFA (16:0) (Fig. 5F). Although the proteins in EVs produced by different lactic acid bacteria strains have been reported through proteomics analysis [16], EVs are composed of lipids, and the lipidomics of lactic acid bacteria-derived EVs are still relatively underdiscussed. However, eicosatetraenoyl-glycerophosphate, oxododecanoic acid (C12:0), oxotetradecanoic acid (C14:0), palmitoleic acid (C16:1), and oleic acid (C18:1) have been identified in Lactobacillus plantarum-derived EVs [43]. The noteworthy findings of this study demonstrated that B. longum-BLMR22-derived EVs were nanoscale materials, with their lipid composition almost entirely consisting of TAGs (with 19.9% being TAG 52:2-FA18:1), CE (22:6) and FFAs (18:0 and 16:0) are the major components. A large amount of cholesterol and FFA (16:0) has been found on human retinal cells and ARPE-19 cell membranes [44]. This leads us to believe that the lipid composition of BLMR22-derived EVs, which is similar to that of retinal cell membranes, facilitating the fusion of EVs with the plasma membrane. 3.3. Transcriptome analysis After pre-treating the cells with BLMR22-derived EVs and exposing them to blue light, we further explored the protective mechanism using transcriptome analysis. Fig. 6A shows Principal Components Analysis (PCA) for clustering gene expression among different groups. The differential gene expression heatmap analysis is shown in Supplementary Fig. 2. Fig. 6B further assesses the increase and decrease in gene expressions using a volcano plot. Since ARPE-19 cells were pre-treated with BLMR22-derived EVs before blue light induction, it is necessary to understand the impact of BLMR22-derived EVs on gene expression in cells that have not yet been exposed to blue light. The results show that BLMR22-derived EVs, compared to the control group, increase the expression of 1963 genes and decrease the expression of 2188 genes (control vs. EVs). After EVs treatment with blue light induction, the number of upregulated and downregulated genes is 64 and 80, respectively (blue light vs. EVs). Using Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis and protein-protein interaction (PPI) network to explore the potential impact of BLMR22-derived EVs on metabolic pathways, it was found that phosphoinositide 3-kinase (PI3K) (Supplementary Fig. 3) and mTor (Supplementary Fig. 4) are the major upregulated targets. Transcriptome Analysis of ARPE-19 Retinal Cells Treated with BLMR22-Derived EVs. (A) Gene expression profile variations presented using PCA (Principal Component Analysis). (B) Volcano plot illustrating the number of differentially expressed genes under different treatment conditions. This study demonstrated that pretreatment with BLMR22-derived EVs in ARPE-19 cells resulted in the upregulation of 1963 genes and downregulation of 2188 genes. PI3K is a regulator related to cell survival among the upregulated genes. It has been reported that the PI3K/Akt and mTOR/p70S6K pathways play a crucial role in the survival of ARPE- 19 cells [45]. Oxidative stress is associated with the physiological mechanisms of macular degeneration, and the activation of the PI3K/Akt pathway has been suggested to protect retinal pigment epithelium cells from oxidative stress-induced damage [46]. Numerous studies have demonstrated that EVs derived from stem or progenitor cells can modulate the PI3K/mTOR signaling pathway [47]. Activation of the PI3K/Akt and mTOR/p70S6K pathways has been shown to promote the survival of retinal pigment epithelial cells under oxidative stress by inhibiting apoptosis [48]. Additionally, modulation of the PI3K/Akt/mTOR signaling cascade has emerged as a promising therapeutic target, with recent evidence suggesting that probiotics may exert beneficial effects through this pathway [49]. These studies suggest that extracellular vesicles (EVs) derived from lactic acid bacteria have the potential to modulate the PI3K/mTOR signaling pathway and enhance the survival of ARPE-19 cells subjected to blue light-induced damage. The connection between microbes, gut-derived products, and the retina can be understood through the disruption of the blood-retinal barrier, a phenomenon frequently observed in retinal diseases [50]. Short-chain fatty acids, beneficial metabolites produced exclusively in the gut, which could cross the blood-retinal barrier through systemic circulation and reach the retina [51]. In addition to microbial metabolites, it has been reported in recent years that EVs from the gut can cross the blood-retinal barrier and influence eye function [52]. Consumption of Limosilactobacillus fermentum inhibits corneal damage and inflammation in a dry eye disease mouse model by regulating the gut microbiome [53]. Additionally, EVs derived from L. fermentum alleviate benzalkonium chloride-induced inflammation in conjunctival cells [54]. Long-term intake of Lactobacillus paracasei KW3110 prevents age-related disruptions in circadian locomotor activity and metabolic changes in the gut of physiologically aged mice [55]. These pieces of evidence support the findings of this study, indicating that EVs secreted by probiotics in the gut have the potential to alleviate eye damage. 4. Conclusion Blue light damage to retinal cells is a major issue in modern life. The gut-eye axis is also a novel theory that has been proposed in recent years, suggesting that gut health and metabolites from probiotics may influence retinal function through this pathway. In this study, we explored the lipid composition of EVs derived from Bifidobacterium longum isolated from human feces. We found that their lipid composition is similar to that of human retinal cell membranes, which allows the EVs to enter cells and ultimately provide a protective effect on the host. Supplementary Information Acknowledgements This research work was mainly supported by National Science and Technology Council (NSTC 113-2221-E-006 -027 -MY3) in Taiwan (ROC). 112 JOURNAL OF FOOD AND DRUG ANALYSIS 2026;34:104–116 Funding Statement This research work was mainly supported by National Science and Technology Council (NSTC 113-2221-E-006 -027 -MY3) in Taiwan (ROC). 112 JOURNAL OF FOOD AND DRUG ANALYSIS 2026;34:104–116 Footnotes Credit authorship contribution statement: Bao-Hong Lee: Methodology, Investigation. Pin- Hsuan Su and Hao-Yuan Chien: Writing – review & editing. Hao-Yuan Chien: Software, Data curation. Chih-Yao Hou: Data curation, Conceptualization. Bao-Hong Lee and Wei-Hsuan Hsu: Software, Methodology, Conceptualization. Wei-Hsuan Hsu and Tzu-Ming Pan: Writing – review & editing, Writing – original draft, Project administration, Investigation. Declaration of competing interest: The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. References Associated Data Supplementary Materials