Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Open Access

Peer-reviewed

Research Article

Functional effects of extracellular vesicles altered by a per- and polyfluoroalkyl substance mixture: In vitro liver cytotoxicity and proteomic expression alterations

  • Celeste K. Carberry,

    Roles Conceptualization, Data curation, Formal analysis, Methodology, Project administration, Visualization, Writing – original draft, Writing – review & editing

    Affiliations The Institute for Environmental Health Solutions, Gillings School of Global Public Health, The University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, United States of America, Department of Environmental Sciences and Engineering, Gillings School of Global Public Health, The University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, United States of America

    ⨯
  • Angie L. Mordant,

    Roles Data curation, Methodology, Writing – review & editing

    Affiliation UNC Proteomics and Metabolomics Core Facility, Department of Pharmacology, The University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, United States of America

    ⨯
  • Christine A. Mills,

    Roles Data curation, Methodology, Writing – review & editing

    Affiliation UNC Proteomics and Metabolomics Core Facility, Department of Pharmacology, The University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, United States of America

    ⨯
  • Hadley Hartwell,

    Roles Methodology, Writing – review & editing

    Affiliation Department of Environmental Sciences and Engineering, Gillings School of Global Public Health, The University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, United States of America

    ⨯
  • Victoria F. Carberry,

    Roles Visualization, Writing – review & editing

    Affiliation Department of Environmental Sciences and Engineering, Gillings School of Global Public Health, The University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, United States of America

    ⨯
  • Lauren Simendinger,

    Roles Data curation, Writing – original draft

    Affiliation The Institute for Environmental Health Solutions, Gillings School of Global Public Health, The University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, United States of America

    ⨯
  • Elise Hickman,

    Roles Methodology, Writing – original draft, Writing – review & editing

    Affiliations Department of Environmental Sciences and Engineering, Gillings School of Global Public Health, The University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, United States of America, Curriculum in Toxicology and Environmental Medicine, School of Medicine, University of North Carolina, Chapel Hill, North Carolina, United States of America

    ⨯
  • Laura E. Herring,

    Roles Data curation, Investigation, Writing – review & editing

    Affiliation UNC Proteomics and Metabolomics Core Facility, Department of Pharmacology, The University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, United States of America

    ⨯
  • Julia E. Rager

    Roles Supervision, Funding acquisition, Writing – original draft, Writing – review & editing

    jrager@unc.edu

    Affiliations The Institute for Environmental Health Solutions, Gillings School of Global Public Health, The University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, United States of America, Department of Environmental Sciences and Engineering, Gillings School of Global Public Health, The University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, United States of America, Curriculum in Toxicology and Environmental Medicine, School of Medicine, University of North Carolina, Chapel Hill, North Carolina, United States of America

    ⨯

Abstract

Per- and polyfluoroalkyl substances (PFAS) have become a focal point in public health research due to their widespread use and persistence, leading to global environmental and human exposure. Accumulating evidence associates PFAS with hepatotoxicity, disrupted liver function, and progression of liver diseases. Simultaneously, extracellular vesicles (EVs) have emerged as key mediators of intercellular communication and potential modulators of exposure-induced disease. Prior studies have revealed that PFAS exposure alters EV release and content, implicating EVs in PFAS-induced liver toxicity. This study evaluated the functional effects of EVs from HepG2 liver cells exposed to a PFAS mixture on the biology of separate recipient HepG2 cells. We hypothesized that EVs from PFAS-treated cells are biologically active and modulate protein expression related to liver diseases and cancer. Parent HepG2 cells were exposed to an equimolar PFAS mixture (PFOS, PFOA, PFHxA), and EVs were isolated and used to treat separate recipient HepG2 cells. Changes in cellular viability and proteomic profiles were measured and further interpreted using pathway and miRNA target analyses. Results demonstrated that EVs derived from PFAS-treated liver cells decrease cell viability. Furthermore, EVs released from PFAS-treated cells cause unique protein expression changes in separate cells, including numerous proteins previously associated with hepatic cancer, non-alcoholic fatty liver disease, and other hepatic diseases. Proteomic pathway analysis further supported this finding, highlighting possible pathways perturbed by EVs derived from PFAS-treated HepG2 cells, including oxidative stress, immune response, and metabolism. These findings highlight a novel mechanism of PFAS toxicity mediated by EVs, underscoring a potential functional role in liver disease progression and potential as targets for mitigating PFAS-induced health effects.

Citation: Carberry CK, Mordant AL, Mills CA, Hartwell H, Carberry VF, Simendinger L, et al. (2026) Functional effects of extracellular vesicles altered by a per- and polyfluoroalkyl substance mixture: In vitro liver cytotoxicity and proteomic expression alterations. PLoS One 21(3): e0338102. https://doi.org/10.1371/journal.pone.0338102

Editor: Daniel X. Zhang, Hong Kong Metropolitan University, HONG KONG

Received: May 15, 2025; Accepted: November 18, 2025; Published: March 18, 2026

Copyright: © 2026 Carberry et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: Data generated within this study are publicly available. The proteomics dataset generated and analyzed in this study is available in the Proteomics Identification Database (PRIDE) repository under project identifier PXD050095.

Funding: This study was supported by the University of North Carolina’s Center for Environmental Health and Susceptibility (CEHS) pilot funds, supported through the National Institutes of Health (NIH) National Institute of Environmental Health Sciences (NIEHS) (P30ES010126), awarded to JR and CC, as well as a NIEHS training grant (T32ES007018), awarded to JR and CC. The specific roles of this author are articulated in the ‘author contributions’ section. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Competing interests: he authors have declared that no competing interests exist.

1. Introduction

Per and polyfluoroalkyl substances (PFAS) are a focal point in exposure science and public health research due to their widespread use in industrial and consumer applications, which has led to their presence across environmental compartments and human populations [1,2]. PFAS exposure is associated with adverse health effects, with the liver emerging as a prominent target organ [3–5]. Studies have associated PFAS with hepatotoxicity, disrupted liver function, and the progression of various liver diseases [5,6]. However, the mechanisms underlying these PFAS-induced liver perturbations remain an ongoing area of research, representing an important research gap for this class of emerging contaminants.

Simultaneously, extracellular vesicles (EVs) have emerged as key molecules in cellular processes and diseases [7–9]. All cells, including both healthy and diseased, release EVs carrying molecular content including microRNAs (miRNAs), proteins, chemicals, and other small molecules [10–12]. EVs therefore serve as cellular signaling molecules that can alter recipient cell biology. The intersection of environmental exposures and EVs has recently gained attention, with EVs being highlighted as biomarkers of exposure, as well as mediators of environmental exposure-induced disease [13,14].

Notably, recent in vitro studies have elucidated modifying effects of environmental chemicals such as air pollution, metals, and other contaminants on the biological function of EVs. For example, EVs released from lung macrophages exposed to particulate matter caused increased pro-inflammatory cytokine secretion from lung epithelial cells [15]. Similar effects have been described for other environmentally related exposures such as dust [16,17], arsenic [18], and asbestos [19]. Conversely, EVs sourced from unexposed or healthy tissues have been shown to exhibit protective effects against environmental exposures in vitro [20,21]. For example, a recent study found that human serum-derived EVs from healthy subjects protected against particulate matter-induced apoptosis in human alveolar epithelial cells [20]. These findings support EVs a potential mechanism for mitigating the effects of environmental exposures. Of particular relevance to the current study, we previously demonstrated that HepG2 cell exposure to an equimolar PFAS mixture increased the release of EVs and altered EV molecular content, including miRNA expression and PFAS chemical signatures [22]. To our knowledge, no other studies have evaluated PFAS induced EV changes and liver toxicity. Based on our previous findings, this study fills an important knowledge gap by evaluating the biological effects induced by PFAS-altered EVs on separate cells as a novel mediator of PFAS-induced liver toxicity.

This study evaluated whether EVs released from PFAS-mixture treated HepG2 cells can affect the biology of separate recipient cells. The novel hypothesis evaluated was that EVs released from PFAS-mixture treated liver cells are biologically active and will alter protein expression in separate liver cells related to chronic liver diseases and cancer. To evaluate this hypothesis, HepG2 cells (parent cells) were exposed to an equimolar PFAS mixture containing perfluorooctanesulfonic acid (PFOS), perfluorooctanoic acid (PFOA), and perfluorohexanoic acid (PFHxA) to replicate conditions used in our previous study. EVs were isolated from parent cells and used to treat separate HepG2 liver cells (target cells) to evaluate changes in resulting cell viability and protein expression profiles. Pathway analyses and microRNA (miRNA) target analyses were additionally performed to contextualize EV-associated changes in HepG2 cell proteomic expression profiles. Additionally, exposures containing both PFAS and EVs were piloted to evaluate potential joint effects exhibited by co-exposure conditions. Findings from this study provide important new insights to previously observed changes in EV release and molecular content, highlighting the role EVs in a novel mechanism of PFAS-induced liver toxicity, as well as a potential avenue for PFAS intervention.

2. Methods

2.1 HepG2 cell culture and PFAS chemical preparation

HepG2 cells, an immortalized liver cell line, were sourced from the American Type Culture Collection and maintained under standard culture conditions as previously published [22]. In brief, cell culture media was prepared using Minimum Essentials Media (MEM) (Gibco) supplemented with 10% fetal bovine serum (FBS) (Avantor) and 1x Penicillin-Streptomycin (Gibco). Cells were passaged at approximately 80% confluency, incubated at 37˚C and 5% CO2, and maintained until experimentation. For experimental treatments, MEM media was supplemented with 10% exosome-depleted FBS (commercially available through Gibco). This minimizes EV contamination from standard FBS, allowing for increased confidence that observed effects are due to EVs themselves rather than residual contaminants. Information pertaining to plating for specific treatment is described in subsequent sections. Chemical stocks of PFOS, PFOA, and PFHxA were purchased from Synquest and dissolved in 100% methanol to make concentrated solutions as previously described [22]. Chemical treatments were prepared fresh the day of experiments.

2.2 Parent HepG2 cell treatment with a PFAS mixture

EVs for this study were generated and isolated from HepG2 cells, representing “parent cells.” HepG2 cells were seeded in 10 cm cell culture dishes with 2 x 106 cells and allowed to adhere and grow under standard conditions for 48 hours. EVs were collected from parent cells following two separate treatment conditions: 1) 0.1% methanol control and 2) PFAS mixture containing equimolar PFOS, PFOA, and PFHxA (35 µM of each chemical). The concentration selected for treatment replicates a condition used in our previously published study and was chosen to elicit measurable cellular responses. Although this concentration is higher than normal environmental exposures and human blood levels, this concentration allows for the identification altered pathways for future validation at lower, environmentally relevant exposures. The nomenclature used in this study is modeled based on a previous study evaluating EV function in vitro [15] where EV treatments are denoted as EVcell of origin/treatment. In this study, EVs released from control HepG2 cells are therefore denoted as EVHepG2/CT. Similarly, EVs released from PFAS mixture treated HepG2 cells are denoted as EVHepG2/PFAS.

2.3 EV isolation and adherence to MISEV guidelines

Following 24-hour exposure to either the control or equimolar PFAS mixture, EVs were isolated from the resulting supernatant. For each condition, 10 mL of supernatant was collected and divided into 1 mL aliquots. EV isolation was performed using Invitrogen’s Total Exosome Isolation Reagent from Cell Culture Media (Qiagen) as previously described and validated, in line with the minimal information for studies of extracellular vesicles (MISEV) [22,23]. In brief, samples underwent centrifugation at 2000 x g for 30 minutes and then at 10,000 x g for 30 minutes to eliminate any cell debris. The resulting purified supernatant was transferred to a new tube, and 500 mL of isolation reagent was added to each sample aliquot. Samples were incubated at 4˚C overnight. Following incubation, EVs were pelleted via centrifugation at 10,000 x g for 1 hour at 4˚C. The supernatant was aspirated, and the remaining EV pellet was resuspended in 100 µL of 0.1 µm filtered PBS. Sample aliquots were recombined as the final step. EV treatments were prepared and added to cells the same day to avoid freeze-thaw as further detailed below. This method of EV isolation was selected to be able to make direct comparisons to our previously published research evaluating the same PFAS exposures HepG2 cells and EV responses [22], though we recognize that the field of EV research is rapidly evolving with improved isolation methods now available such as size-exclusion chromatography [24]. To verify the performance of this commercially-validated isolation reagent, EVs produced using this method were previously characterized and validated in prior work using nanoparticle tracking analysis (NTA) for size distribution, transmission electron microscopy (TEM) for morphology, and measurement of EV surface markers [22].

2.4 Target HepG2 cell treatment with EVs

The objective of this study was to evaluate whether EVs released by control or PFAS exposed “parent” cells are able to cause changes in cell viability or protein expression of control or PFAS exposed “target” cells. To do this, separate HepG2 cells representing target cells were cultured with either EVHepG2/CT (EVs derived from control HepG2 cells) or EVHepG2/PFAS (EVs derived from PFAS-treated HepG2 cells). This experimental design ensures that any effects observed are a result of EVHepG2/CT or EVHepG2/PFAS exposure, as opposed to direct cell-cell contact. Total EV protein amount was selected to normalize EV quantity within this study. EV protein was measured using a Pierce BCA (Thermo Scientific) according to the manufacturer’s protocol.

In preparation for target cell treatments, EVs were diluted in MEM media supplemented with 10% exosome-depleted FBS to a concentration of 1 µg total protein/ mL of HepG2 media. This concentration was selected to 1. Maintain alignment with MISEV guidelines which suggest normalizing by EV component for in vitro studies and 2. Replicate conditions similar to those evaluated in previous studies that evaluated the functional effects of EVs on separate cells [15,23].

For cell viability, separate cells representing “target cells” were seeded in a 96-well plate with 0.01 x 106 cells/well with six replicates for all conditions except for co-exposures which were in triplicate. For proteomics, cells representing “target cells” were seeded in 6-well plates with 0.3 x 106 cells/well and allowed to adhere and grow for 48 hours under standard conditions, with six replicates per condition. For cell treatment with EVs, media was aspirated, and cells were washed with 1x PBS. Target cells were then treated with six unique treatment conditions as visualized in step 3 of Fig 1. The first condition contained 0.1% methanol, representing control cells. The second condition contained EVHepG2/CT (1 µg/mL), and the third condition contained EVHepG2/PFAS (1 µg/mL). These treatments allow for the characterization of the functional effects of HepG2 derived EVs on target cells, particularly when altered by PFAS exposure. The fourth exposure condition contained an equimolar PFAS mixture of PFOS, PFOA, and PFHxA (35µM each). While this concentration is cytotoxic, it serves as a baseline for co-exposures containing both the equimolar PFAS mixture combined with either EVHepG2/CT or EVHepG2/PFAS. Previous studies evaluating the functional effects of EVs have found protective effects following EV exposure; therefore, such relationships were evaluated in this study. The fifth condition was thus a co-exposure of the equimolar PFAS mixture plus EVHepG2/CT (1 µg/mL). The final condition was a co-exposure of the equimolar PFAS mixture plus EVHepG2/PFAS (1 µg/mL). All treatments were added to cells in biological sextuplicate for proteomic evaluation and incubated for 24 h under standard cell culture conditions. In this study, biological replicates are defined as cells passaged on different days.

thumbnail
Download:
Fig 1. Graphical overview of methods for cellular treatment.

1. HepG2 cells, representing parent cells, were treated as either a control or with a per and polyfluoroalkyl substance mixture. 2. EVs were isolated from parent HepG2 cells (EVHepG2/CT or EVHepG2/PFAS) and 3. used as a treatment on separate HepG2 cells, representing target cells. The six 6 treatment types included a control, individual EVHepG2/CT, individual EVHepG2/PFAS, and a PFAS mixture, and PFAS mixture with and without EVHepG2/CT or EVHepG2/PFAS. 4. Target HepG2 cells were subsequently evaluated for changes in cell viability and protein expression. Key findings are summarized.

https://doi.org/10.1371/journal.pone.0338102.g001

2.5 In vitro evaluation of EV induced cytotoxicity

The cytotoxicity of HepG2 cells following exposure to the six exposure conditions detailed above was evaluated using a resazurin assay according to manufacturer’s protocol (ThermoScientific). This assay measures the metabolic capacity of cells based on the reduction of oxidized non-fluorescent blue resazurin dye into a red fluorescent dye, proportional to the number of live cells [25]. Resazurin was diluted in PBS to a concentration of 0.15 mg/ml and filtered through a 0.2 µm filter. A volume of 20 µL was added to each cell well and incubated 2 hours at 37˚C and 5% CO2. Following incubation, the plate was read using SpectraMax iD5 Multi-Mode Microplate Reader (Molecular Devices) in fluorescence mode with an excitation wavelength of 560 nm and emission wavelength of 590 nm. Percent cell viability was calculated in relation to the control. Statistical significance was evaluated using a t-test and visualized using GraphPad Prism v10.

2.6 Label-free quantitative cellular proteomics

Following 24 h treatment of cells with EVs or PFAS mixture, supernatant was removed, and cells were washed three times with 1 mL PBS. Cells were lysed in 40 µL commercially available RIPA buffer (Thermo Scientific, 25 mM Tris, HCl pH 7.6, 150 mM NaCl, 1% NP-40, 1% sodium deoxycholate, 0.1% SDS) with 1x Halt Protease Inhibitor Cocktail (Thermo Scientific). Cell lysates were scraped, transferred to a tube, and centrifuged at 14,000 g for 15 minutes to pellet cell debris. The resulting supernatant was transferred to a new tube for proteomics evaluation.

To evaluate potential changes in cellular protein expression following EV exposure conditions, a global proteomics analysis was conducted in collaboration with the UNC Michael Hooker Metabolomics and Proteomics Core. Given the presence of detergent in the cell lysis buffer, S-trap, a suspension-based digestion, was performed on all samples. This method employs spin columns to filter detergents and other unwanted contaminants, leaving protein bound to the S-trap [26]. More specifically, 75 µg of each sample were normalized to 50 µl in 5% SDS. Samples were then reduced with DTT, alkylated with iodoacetamide, and digested with trypsin on an S-trap micro (Protifi). The resulting peptide samples were lyophilized and cleaned using Pierce desalting spin columns and quantified using a Fluorometric peptide BCA assay (Pierce). Samples were normalized for LC-MS/MS to 0.1 µg/µL, and a pooled sample was created by combining an equal volume of each sample. Standard iRT peptides (Biognosys) were added to each sample at a 1:30 ratio.

Each sample was analyzed in a randomized order via liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS) using an Ultimate3000-Exploris480 (Thermo Scientific) in Data Independent Acquisition mode. The pooled sample was analyzed intermittently to assess technical reproducibility for a total of 3 pooled runs. Samples were injected onto an IonOpticks Aurora series 3 C18 column (75 μm id × 15 cm, 1.6 μm particle size; IonOpticks) and separated over a 120 min method. The gradient for separation consisted of 3–41% mobile phase B at a 250 nl/min flow rate, where mobile phase A was 0.1% formic acid in water and mobile phase B consisted of 0.1% formic acid in 80% ACN. The Exploris480 was operated in product ion scan mode for Data Independent Acquisition (DIA). A full MS scan (m/z 350–1650) was collected; resolution was set to 120,000 with a maximum injection time of 20 ms and AGC target of 300%. Following the full MS scan, a product ion scan was collected (30,000 resolution) and consisted of stepped higher collision dissociation (HCD) set to 25.5, 27, 30; AGC target set to 3000%; maximum injection time set to 55 ms; variable precursor isolation windows 350–1650 m/z.

2.7 Proteomic data processing and statistical analysis

Identification and quantification of proteins was conducted within Spectronaut (v17.6, Biognosys). Raw data files were searched against a Uniprot Human database (UP000005640, containing 20,249 protein sequences) appended with a common contaminants database containing 245 proteins (MaxQuant). The following settings were used: enzyme specificity set to trypsin, up to two missed cleavages allowed, cysteine carbamidomethylation set as a fixed modification, and methionine oxidation and N-terminal acetylation set as variable modifications. Precision iRT calibration was enabled. Data were filtered based on a false discovery rate (FDR) of 1%, and exclusion of proteins with only a single unique peptide. These filters resulted in a total of 7,104 proteins carried forward in the analysis.

Statistical analysis was performed within Spectronaut using default settings, which calculates p-value and FDR-corrected ‘q-value’ to adjust p-values, for reach pairwise comparison.. Magnitude of expression changes are reported as log2 fold changes (FC), calculated as the ratio of average protein abundance of exposed/protein abundance of unexposed samples. An log2 ratio ±0.58, along with a q-value < 0.1 was considered significant. Visualizations were created using GraphPad Prism (v10) and Adobe Illustrator 2022.

2.8 Interpretation of biological effects of EVHepG2/CT vs EVHepG2/PFAS exposure on HepG2 cells using pathway analysis

2.8.1 Pathway analysis using ingenuity knowledgebase.

To interpret the biological implications of EV-altered liver cell protein expression, pathway analyses were conducted using the Ingenuity Pathway Analysis (IPA) platform (Qiagen). Using this platform, a list of proteins with EV-altered expression (log2 FC > |0.58|, p-adjust <0.1) were queried against an expert-curated database of literature containing biological interactions between genes, proteins, and complexes, resulting in the identification of pathways and diseases that are likely associated with protein expression changes [27]. To compare differences in protein expression in HepG2 liver cells following exposure to EVHepG2/CT versus EVHepG2/PFAS, separate analyses were conducted on each respective list of proteins with significantly altered expression (log2 FC > |0.58|, p-adjust <0.1). To identify significantly enriched biological pathways, p-values were calculated using a Fischer’s Exact test within IPA. Enriched pathways were then filtered for those with a p-value < 0.05. Results from these analyses further contextualize the functional differences between EVHepG2/CT and EVHepG2/PFAS. Following pathway enrichment analysis, proteins identified in significant pathways were further queried in the literature for specific relevance to hepatic pathways and diseases. This step adds additional biological context to enhance the biological interpretation of pathway results which show modest enrichment.

Similarly, to interpret the combined effects of PFAS with EVHepG2/CT or EVHepG2/PFAS on protein expression in HepG2 liver, additional pathway analyses were conducted. Given the significant overlap in protein expression between exposures, only significant proteins that were uniquely altered by each exposure condition were carried forward in order to highlight EV-specific effects. Significantly enriched biological pathways were identified with p-values < 0.05. Results were visualized using GraphPad Prism (v10) and IPA software.

2.8.2 Mapping cellular proteomic profiles to differential EV miRNA expression profiles.

We previously demonstrated that HepG2 cells exposed to the same defined PFAS mixture used in the current study resulted in increased release of EVs, and these EVs contained altered miRNA profiles [22]. This is notable because miRNAs can decrease protein expression by degrading mRNAs and cleaving proteins [28–30].

To explore biological links between PFAS-altered EV miRNA profiles and the observed changes in cellular proteins in the present study, we cross-referenced differentially expressed cellular proteins from single EV exposure conditions in the present study with predicted EV miRNA targets from our previously published study [22]. Specific mRNA targets were identified using known miRNA-mRNA interactions and predicted interactions based off sequence homologies. This analysis focused on mRNA targets for the top 10 most differentially increased miRNAs (916 unique genes targets) and the top 10 most differentially decreased miRNAs (1688 unique gene targets) in our previous study to provide focus and highlight the most likely miRNAs involved. Predicted mRNA targets are recapitulated in S1 Table and were prioritized for further analysis in this study [31].

2.9 Data availability

Data generated within this study are publicly available. The proteomics dataset generated and analyzed in this study is available in the Proteomics Identification Database (PRIDE) repository under project identifier PXD050095.

3 Results

3.1 EV isolation and adherence to MISEV guidelines

EVs were isolated using methods that replicate those previously described [22]. Therefore, this method has been previously validated in line with MISEV guidelines and shown to produce samples enriched with 1. EVs < 200 nm in diameter as determined via nanoparticle tracking analysis; 2. EV surface marker presence including annexin A5, tumor susceptibility gene G101, flotillin 1, intercellular adhesion molecule 5, and CD81 molecule; and 3. EVs with typical physical features as observed via transmission electron microscopy. While isolation reagents may co-elute other proteins and molecules, these validations supported this method’s successful enrichment of EVs. The present study additionally measured total EV protein released from control and treated HepG2 cells. In total, EVHepG2/CT samples isolated from a 10 cm cell culture plate had an average of ~125 µg total EV protein, and EVHepG2/PFAS samples isolated from a 10 cm cell culture plate had an average of ~200 µg total EV protein. These samples were then diluted in EV-depleted HepG2 media to a concentration of 1 µg/mL for target cell treatment.

3.2 In vitro evaluation of EV induced cytotoxicity

To evaluate the effects of EVs on target HepG2 cell viability, a resazurin assay was performed on all exposure conditions (Fig 2). Treatment with EVHepG2/CT had no significant effect on HepG2 cellular viability. In contrast, treatment with EVHepG2/PFAS resulted in a statistically significant (p < 0.05) decrease in cell viability. This finding suggests that altered molecular or chemical signals contained in/on EVHepG2/PFAS received by target HepG2 cells perturb signaling pathways, resulting in cell death.

thumbnail
Download:
Fig 2. Cellular cytotoxicity in response to HepG2-derived EV exposure conditions.

This bar plot depicts average cell viability per condition with individual datapoints (conditions 1-4, n = 6; conditions 5-6, n = 3). Statistical comparisons to the control are highlighted via multiple t-tests (* indicated p < 0.05). Interpretation of results for every condition are described on the right. Conditions additionally align with those outlined in the graphical overview of methods in Fig 1.

https://doi.org/10.1371/journal.pone.0338102.g002

To evaluate potential adverse or protective effects of EVs in the presence of PFAS exposure, a co-exposure was additionally piloted. As expected, treatment of HepG2 cells with the equimolar PFAS mixture alone resulted in overt cytotoxicity, with an average cell viability of approximately 60%. Interestingly, when evaluating the co-exposure combining the equimolar PFAS mixture with EVHepG2/CT, no significant cytotoxicity was observed, indicating protective effects of EVHepG2/CT in the context of PFAS exposure. In contrast, when evaluating the co-exposure combining the equimolar PFAS mixture with EVHepG2/PFAS, significant cytotoxicity was observed compared to the control (p-value < 0.05). In comparison to the PFAS mixture alone, the co-exposure with EVHepG2/PFAS induced greater levels of cytotoxicity, suggesting the potential occurrence of joint toxicity. Results for all conditions are further visualized in Fig 2.

3.3 Identification of differentially expressed cellular proteins via label-free quantitative proteomics

3.3.1 HepG2 protein expression changes following exposure to individual EVHepG2/CT or EVHepG2/PFAS.

Data were normalized by equalizing the medians. Additionally, to evaluate data quality, multiple assessments were conducted. For label-free global proteomic analysis, a coefficient of variation (CV) below 20% is considered acceptable. For this dataset, the average %CV was ~ 10%, indicating high reproducibility/low variability. To further assess sample variability, a PCA demonstrated that samples cluster moderately well by group. It should be noted that EVHepG2/CT and EVHepG2/PFAS treated samples did not separate well from the controls globally, though statistical analysis demonstrated a subset of proteins were significantly differentially expressed. (S1 Fig). The co-exposed groups appeared to be distinct from the single exposed and control groups. Pooled replicates cluster in the middle, as expected. Protein-level data was ~ 98.4% complete indicating very few missing values which were not imputed Finally, data distribution showed consistent protein amounts quantified across the samples, with no clear outliers. Common lab contaminants were filtered out of the dataset, yielding 7,020 proteins considered for statistical analysis.

A complete list of differentially expressed proteins meeting these criteria for individual EV exposures is detailed in S2 Table with corresponding Log2 fold changes, p-values and adjusted p-values These results are additionally visualized as a volcano plot in S2 Fig. [31]. As previously described, to be considered differentially expressed, proteins were required to have a log2 FC > |0.58| and p-adjust < 0.1 compared to the control. From this list, 16 proteins were determined to be differentially expressed following treatment of HepG2 cells with EVHepG2/CT, of which 8 were increased and 8 were decreased (Fig 4A). The protein with the greatest increase in FC (log2 FC 1.62) was cytochrome C oxidase assembly factor 7 (COA7), while the protein with the greatest decrease in FC (log2 FC −1.06) was ETS translocation variant 4 (ETV4) (Fig 3). Interestingly, HepG2 cell exposure to EVHepG2/PFAS resulted in a greater number of differentially expressed proteins, with 46 proteins with differential expression, of which 14 were increased and 32 were decreased (Fig 4A). The protein with the greatest increase in FC (log2 FC 1.62) was protein S100-A8 (S100A8) while protein with the greatest decrease in FC (log2 FC −1.65) was immunoglobulin heavy constant alpha 2 (IGHA2) (Fig 3). Between these single EV exposure conditions, only 3 proteins overlapped, as visualized in Fig 4B, suggesting largely unique proteomic expression profiles between groups and thus supporting unique biological changes induced by EVHepG2/CT versus EVHepG2/PFAS.

thumbnail
Download:
Fig 3. Heatmap of differentially expressed protein signatures following exposure to EVs.

This heatmap highlights differentially expressed proteins (log2 FC > |0.58|, p-adjust <0.1) with their corresponding gene symbols following individual exposure to EVHepG2/CT and EVHepG2/PFAS compared to their respective control. Average Log2 FCs for proteins with differential expression are listed where applicable. Log2 FC is also indicated by color where FC > 0 (increased protein expression compared to control) is purple, and FC < 0 (decreased protein expression compared to control) is green. All conditions were evaluated with n = 6. NS indicates not significant.

https://doi.org/10.1371/journal.pone.0338102.g003

thumbnail
Download:
Fig 4. Summary of differentially expressed proteins following exposure to single EV exposure and EV co-exposure with PFAS.

A) Proteomic expression changes (log2 FC > |0.58|, p-adjust <0.1) following EVHepG2/CT or EVHepG2/PFAS exposure are summarized in the bar plot. Bars are organized by number of proteins with increased expression and number of proteins with decreased expression compared to their respective control. B) Venn diagram displaying number of unique or overlapping proteins differentially expressed following EVHepG2/CT or EVHepG2/PFAS. C) Bar plot summarizing protein expression changes for PFAS co-exposures with EVHepG2/CT or EVHepG2/PFAS. D) Venn diagram displaying the number of differentially expressed proteins for each condition including PFAS exposure and co-exposures with EVHepG2/CT or EVHepG2/PFAS. All conditions were evaluated with n = 6.

https://doi.org/10.1371/journal.pone.0338102.g004

3.3.2 HepG2 protein expression changes following co-exposures to the PFAS mixture plus EVHepG2/CT or EVHepG2/PFAS.

In the presence of PFAS exposure, co-exposure with EVHepG2/CT altered the expression of 1,641 proteins (Fig 4C, S2 Table). Similarly, in the presence of PFAS exposure, co-exposure with EVHepG2/PFAS altered the expression of 1,594 proteins (Fig 4C, S2 Table). As expected, the majority of differentially expressed proteins following co-exposure conditions overlapped with proteins differentially expressed following exposure to the PFAS mixture alone (1,102 proteins) (Fig 4D). However, as further illustrated in Fig 4D, there were a number of differentially expressed proteins unique to each EV exposure condition, with 165 unique to the PFAS + EVHepG2/CT condition and 112 unique to the PFAS + EVHepG2/PFAS condition. These findings further support that the effects of EV exposures on protein expression in liver cells are distinct from the effects of PFAS alone.

3.3.3 Differentially expressed proteomic profiles relate to EV miRNA expression profiles.

In our previous study, we identified a set of mRNAs predicted to be regulated by EV encapsulated miRNAs from PFAS-exposed cells [22]. Genes associated with differentially expressed cellular proteins, as determined in this study, were compared against the list of previously predicted EV miRNA targets to elucidate biological links between cellular proteomic expression changes and EV miRNA expression profiles. This analysis revealed five overlapping targets Ninein (NIN), E3 ubiquitin-protein ligase Mdm2 (MDM2), Protein MIX23 (MIX23), S100A8 and ETV4, strengthening the evidence for EV-mediated delivery of regulatory cargo that modulates recipient cell protein expression. Of particular interest, S100A8 and ETV4 were uniquely altered in response to EVHepG2/CT and EVHepG2/PFAS respectively both individually and within co-exposures with PFAS. Given their distinct responses and known roles in hepatic biology and disease, these two proteins are discussed in further detail below.

3.4 Interpretation of biological effects of EVHepG2/CT vs EVHepG2/PFAS exposure on HepG2 cells using pathway analysis

3.4.1 HepG2 pathway alterations following individual exposure to EVHepG2/CT or EVHepG2/PFAS.

Differential proteomic expression profiles of HepG2 cells following exposure to either EVHepG2/CT or EVHepG2/PFAS support the functional role of EVs in cellular communication. To further evaluate the biological relevance of these findings, a pathway analysis was conducted using the IPA knowledgebase. As hypothesized, HepG2 cellular exposure to EVHepG2/CT and EVHepG2/PFAS caused distinct changes in cellular pathways, supporting unique biological function of exposure-associated EVs. The most significantly altered pathways associated with EVHepG2/CT and EVHepG2/PFAS exposure are summarized in Table 1, including eight and eleven unique proteins within the most altered pathways respectively. When interpreting results, it is important to note that proteins often overlap across multiple pathways and pathways were ranked by sensitivity criteria rather than by percent enrichment, with enrichment driven by only a few proteins in many cases. Broadly, HepG2 cell exposure to EVHepG2/CT resulted in pathway changes related to regulation of cell cycle, DNA damage mechanisms, and cancer signaling. Conversely, EVHepG2/PFAS exposure was associated with oxidative stress, immune regulation, and metabolic pathways. Though these pathway-level results included generally low enrichment (three or less proteins per pathway) in relation to these pathways, these findings supported largely unique pathway alterations following EV exposures. Furthermore, these initial findings highlighted a subset of important proteins to prioritize for expanded literature review to further understand observed proteomic changes.

thumbnail
Download:
Table 1. Canonical pathways with predicted enrichment following individual exposure to EVHepG2/CT or EVHepG2/PFAS. Top pathways are presented with corresponding -log p-values. Differentially expressed proteins predicted to be involved in these pathways are additionally listed.

https://doi.org/10.1371/journal.pone.0338102.t001

Following exposure to EVHepG2/CT, the eight unique proteins that were included in the most significant pathways were beta-transducin repeat containing E3 ubiquitin protein ligase (BTRC), cyclin dependent kinase inhibitor 1B (CDKN1B), CCAAT enhancer binding protein gamma (CEBPG), MDM2 proto-oncogene (MDM2), DnaJ heat shock protein family (DNAJB9), ETV4, MAF bZIP transcription factor K (MAFK), and SRY-box transcription factor 6 (SOX6). As reviewed in Table 2A, when considering directional fold changes, the majority of these proteins have been associated with improved chronic liver disease outcomes including decreased tumor progression, alleviation of hepatic steatosis, and decreased inflammation. Conversely, following exposure to EVHepG2/PFAS,11 proteins were involved in the most significant pathways including insulin like growth factor binding protein 1 (IGFBP1), serine protease 3 (PRSS3), cytochrome c oxidase subunit 6A1 (COX6A1), BTRC, phosphoenolpyruvate carboxykinase 1 (PCK1), BRCA2 DNA repair associated (BRCA2), S100A8, MAFK, hydroxyacyl-thioester dehydratase type 2 (HTD2), ring finger protein 138 (RNF138), and transferrin receptor 2 (TFR2). When considering directional fold change, most of these proteins have previously been associated with increased hepatic cancer progression or metastasis, increased lipid accumulation, increased oxidative stress, and inflammation in scientific literature (Table 2B).

thumbnail
Download:
Table 2. Literature-supported roles of proteins with predicted pathway involvement on liver health outcomes. For all enriched pathways previously presented, the roles of implicated differentially expressed proteins are further defined in the specific context of liver health with associated references.

https://doi.org/10.1371/journal.pone.0338102.t002

3.4.2 HepG2 pathway alterations following co-exposures to a PFAS mixture plus EVHepG2/CT or EVHepG2/PFAS.

To evaluate the effects of EVs in the presence of PFAS exposure, co-exposures were examined involving EVHepG2/CT or EVHepG2/PFAS. Proteomic expression alterations were contextualized using pathway analysis, focusing on proteins with expression changes unique to each condition. Notably, co-exposure of EVHepG2/CT with PFAS resulted in significantly enriched pathways related to cell death and inflammation, the majority of which were predicted to be deactivated. This included predicted deactivation of the necroptosis signaling pathway and RIPK-1 mediated regulated necrosis pathway, suggesting potential cell recovery. Additional pathways related to immune response and inflammation, such as tumor necrosis factor signaling, toll-like receptor 3 cascade, and MyD88-independent signaling pathway were predicted to be decreased, further supporting biological changes indicative of cellular recovery (Fig 5A). Conversely, co-exposure of HepG2 cells with PFAS and EVHepG2/PFAS exhibited a diverse pathway-level response without clear directional trends. In general, perturbed pathways related to broader biological processes including metabolism, cell differentiation, and transcriptional regulation (Fig 5B).

thumbnail
Download:
Fig 5. Predicted biological pathway enrichment and predicted upstream regulators following HepG2 cell exposure to PFAS + EVHepG2/CT or PFAS + EVHepG2/PFAS.

A) Most significantly enriched biological pathway alterations following exposure to PFAS + EVHepG2/CT. B) Most significantly enriched biological pathway alterations following exposure to PFAS + EVHepG2/PFAS. C) Predicted decrease in activity of upstream regulator TP53 following PFAS + EVHepG2/CT exposure. D) Predicted increase in activity of upstream regulator TP53 following PFAS + EVHepG2/PFAS exposure.

https://doi.org/10.1371/journal.pone.0338102.g005

To provide further context and interpretation of protein expression changes under co-exposure conditions, upstream regulators were predicted for each set of unique proteins with differential expression associated with each exposure. Strikingly, tumor protein 53 (TP53) emerged as a likely upstream regulator for both co-exposure conditions. Notably, TP53 is a well-established tumor suppressor gene in cancer biology [32]. Following PFAS and EVHepG2/CT exposure, TP53 was predicted to be inhibited while in contrast, TP53 was predicted to be activated following exposure to PFAS and EVHepG2/PFAS (Fig 5C and Fig 5D). This inverse relationship further underscores the unique biological effects of EVHepG2/CT vs EVHepG2/PFAS.

4 Discussion

The primary goal of this study was to assess the ability of EVs to communicate information related to PFAS exposure from exposed liver cells to separate target liver cells. To evaluate this, we conducted an in vitro evaluation of the effects of PFAS on EV function using HepG2 liver cells. Previous in vitro studies have established that cellular exposure to environmental chemicals can modify EV regulation, content, and subsequent function. However, there remains a critical research gap surrounding the interplay between PFAS exposure and EV mediated changes in cellular biology. Notably, this study among the first to propose EVs as a mechanism of PFAS toxicity, ultimately aiming to develop in vitro methods for future translation into in vivo studies. Overall, this study yielded multiple key findings: 1) EVs derived from liver cells exposed to PFAS (EVHepG2/PFAS) induced decreased cell viability in separate cells; 2) EVs induced changes in target cell protein expression, with profiles that were largely distinct depending on the source cells’ exposure; 3) Cellular proteins with expression altered by EVs have known associations with liver health outcomes such as hepatic fibrosis, altered metabolism, and cancer; and 4) EVs derived from control HepG2 cells (EVHepG2/CT) potentially mitigated effects of PFAS exposure. These findings contribute to our understanding of the intricate relationship between PFAS, EVs, and cellular responses, with broad implications for mechanistic toxicology and translation to human epidemiological research.

Evaluation of the effects of EVs on HepG2 cell viability supported distinct outcomes based on the parent cell exposure conditions. Specifically, treatment with EVHepG2/CT did not exhibit a significant effect on cellular viability, while treatment with EVHepG2/PFAS led to a statistically significant decrease in cell viability, indicating potential adverse effects related to altered molecular or chemical signals carried by EVHepG2/PFAS [22]. While the precise mechanism of cytotoxicity was not measured, orthogonal evidence from a previous HepG2 study demonstrated concentration-dependent increases in reactive oxygen species, supporting oxidative stress as a key contributor to PFAS-associated hepatotoxicity. This represents an additional endpoint to be evaluated in future EV treatment assays [33]. Overall, this observation aligns with our novel hypothesis that EVs may serve as mediators of PFAS toxicity and aligns with previous research demonstrating that EVs secreted by cells exposed to environmental chemicals can affect the viability of separate cells in vitro [15,21,34,35].

In further support of distinct biological effects induced by HepG2 cell exposure to EVHepG2/CT vs EVHepG2/PFAS, we found that EVs cause changes in target cell protein expression that are largely unique depending on source cell exposure. More specifically, 16 proteins were differentially expressed following HepG2 cell exposure to EVHepG2/CT, whereas 46 total proteins were differentially expressed following EVHepG2/PFAS exposure with only 3 proteins overlapping between exposure conditions. A potential link between EVs and related changes in target cell biology is encapsulated miRNAs. Numerous studies have previously identified miRNAs contained within environmental exposure modified EVs as mediators of various cell processes and diseases [17,18,34,36–39]. For example, EVs derived from arsenite exposed human bronchial epithelial cells caused cell proliferation in separate unexposed epithelial cells, and these effects were driven by miR-21, supporting the role of EV miRNAs in cell-cell communication. [18].

The protein with the greatest increase in expression following EVHepG2/PFAS was S100A8, a calcium-binding protein involved in inflammatory signaling and immune cell recruitment. Existing literature supports the role of the S100 family, including S100A8, in influencing fatty liver disease and hepatic carcinoma [40]. Interestingly, this finding is consistent with our earlier discovery of decreased levels of miRNA let-7e in EVHepG2/PFAS compared to EVHepG2/CT [22]. This directional agreement supports, in theory, that increased expression of S100A protein in HepG2 cells may be linked to the decreased expression of miRNA let-7e in EVHepG2/PFAS. This relationship is further illustrated in Fig 6, which depicts a theoretical mechanism through which EVs mediate PFAS associated proteomic changes in the liver. A similar observation was made following HepG2 exposure to EVHepG2/CT, where ETV4 was the protein exhibiting the greatest decrease in expression. This is similarly significant given the previous upregulation of miR-29a in EVHepG2/CT compared to EVHepG2/PFAS [22]. ETV4 is a transcription factor implicated in cell proliferation and oncogenic signaling, and its downregulation has been associated with suppression of tumorigenesis [41]. Notably, these relationships were consistently mirrored in co-exposure treatments, with S100A8 uniquely responding to EVHepG2/PFAS and ETV4 exclusively altered in response to EVHepG2/CT. This specificity underscores the distinct impact of exposure-associated EVs on cellular protein expression changes. The convergence of predicted miRNA-targets with observed alterations in protein expression suggests the involvement of EV miRNA content in mediating PFAS-induced changes in the cellular proteome. To further elucidate the intricate network of miRNA-mediated effects, future studies may benefit from expanding miRNA-target analyses to identify additional key miRNAs. Additionally, employing miRNA knock-down techniques are required to validate the specific roles of these miRNAs in mediating PFAS-associated effects, offering a more comprehensive understanding of the underlying mechanisms. Furthermore, emerging evidence suggests that other RNA species such as long noncoding RNAs and transfer RNA-derived fragments, are also abundant in EVs and may play important regulatory roles which could be evaluated in the future [42].

thumbnail
Download:
Fig 6. Theoretical mechanism of EV-mediated proteomic expression alterations.

1. PFAS exposure to PFOS, PFOA, and PFHxA reach the cell and induce the release of EVs. 2. EVs released from PFAS exposed HepG2 cells contain altered miRNA expression profiles, including decreased Let-7e. 3. EVs reach separate HepG2 cells and, in theory, target proteins of Let-7e are increased including S100A8.

https://doi.org/10.1371/journal.pone.0338102.g006

The prediction of enriched biological pathways related to differentially expressed proteins following exposure to EVHepG2/CT versus EVHepG2/PFAS allows for enhanced biological interpretation of results. It is important to note that pathway-level findings should be interpreted as exploratory and hypothesis-generating given the in vitro design and the need for further functional validation. Pathway analysis demonstrated that proteins with differential expression following exposure to EVHepG2/CT were associated with pathways related to cell cycle regulation and cancer signaling. Conversely, proteins with differential expression following exposure to EVHepG2/PFAS were associated with oxidative stress, metabolism, and immune regulation which are pathways commonly associated with PFAS exposure [5,43]. Together, these findings provide supporting evidence that the biological effects of EVs are specific to their source cells’ exposure. Following global pathway predictions, specific genes were queried for their specific involvement in liver health. Notably, when considering the direction of expression change, the majority of proteins implicated in pathways related to EVHepG2/CT aligned with improved liver health, including alleviated hepatic fibrosis/steatosis, reduced inflammatory response, and positive cancer outcomes [41,44–48]. In contrast, the majority of proteins associated with EVHepG2/PFAS are linked to adverse outcomes including increased oxidative stress, mediation of lipid accumulation in liver diseases, and hepatic cancer progression [40,49–55]. While there are some inconsistencies, a discernable trend was elucidated, suggesting EV-mediated toxicity related to EVHepG2/PFAS and potentially protective effects related to EVHepG2/CT. Overall, pathway analyses substantiate the hypothesis that EVs can mediate PFAS associated changes in biological pathways related to liver disease and cancer; however, phenotypic and functional assays are required to directly evaluate disease outcomes.

To characterize the biological effects of EVs derived from HepG2 cells under control versus PFAS-treatment, separate HepG2 cells were co-exposed with PFAS and EVHepG2/CT or EVHepG2/PFAS. Following HepG2 cell co-exposure with PFAS and EVHepG2/CT, evidence suggested that EVHepG2/CT may confer protection against PFAS exposure, as indicated by the absence of significant changes in cell viability, though more robust testing is needed. In contrast, co-exposure of PFAS with EVHepG2/PFAS showed potential joint toxicity, evidenced by a decrease in average cell viability compared to PFAS exposure alone. These results suggest that EVs released from healthy cells may mitigate PFAS exposure which should be further evaluated in future studies. For example, future studies may comprehensively evaluate potential mixture-effects through concentration-response designs and across varying time-periods. Furthermore, this preliminary finding aligns with existing literature highlighting the utility of EVs released from healthy cells or isolated from healthy patients as a potential tool to combat the impact of environmental exposures on cell viability [20,21]. Beyond cell viability, EVHepG2/CT and EVHepG2/PFAS induced distinct proteomic and pathway alterations, adding to the weight of evidence supporting unique biological effects of EVs dependent on source cell exposure conditions. For example, when investigating predicted pathway-level responses, EVHepG2/CT once again emerged as potentially mitigating against PFAS-induced toxicity. Co-exposure with PFAS and EVHepG2/CT was associated with pathway alterations related to cell death signaling, similar to effects observed from single exposure to EVHepG2/CT. Notably, there was a predicted decrease in activity for these pathways, suggesting a dampening effect. Similarly, in the evaluation of potential upstream regulators of the differentially expressed proteins following co-exposures, TP53 was highlighted as a regulator for both conditions. TP53 is a well-studied tumor suppressor protein, which is commonly activated in response to cell stress or DNA damage [56]. While tumor suppression is an essential function of TP53, accumulation of TP53 has been associated with cellular apoptosis and subsequent liver fibrosis and cirrhosis [57–59]. Following co-exposure with PFAS and EVHepG2/PFAS, TP53 was predicted to exhibit increased activity, while following exposure with PFAS and EVHepG2/CT TP53 was predicted to exhibit decreased activity. Though beyond the scope of this study, these findings set precedent for future studies investigating EVHepG2/CT as a potential intervention for PFAS toxicity through the expansion of co-exposure concentrations and statistical tests comparing cellular proteomic profiles of PFAS-exposed cells with those present in PFAS and EVHepG2/CT exposed cells.

While this study has provided valuable insights into the role of EVs in PFAS associated cellular responses, additional limitations should be acknowledged. First, this study focused exclusively on EVs with diameters less than 200nm, potentially overlooking the contributions of other EV subclasses. While this subtype was selected to limit EV populations derived exclusively from cell death, different EV subpopulations may have distinct cargo and functions [23]. Thus, future investigations may explore the full spectrum of EV sizes to obtain a more comprehensive understanding of their role in PFAS toxicity. Secondly, consistent with most EV studies, the observed effects attributed to EVs may be due in part to small quantities of co-eluted particles. Further refinement of isolation techniques in line with the evolving nature of this field, and characterization of co-eluted components would enhance the specificity of the observed effects and strengthen the conclusions drawn from this study. Finally, caution should be used when interpreting the effects of co-exposure pathways due to the potential influence of overt cytotoxicity elicited by PFAS. These findings are intended to serve as a comparison to individual EV exposures, though future studies should optimize co-exposures for the specific assessment of EV-protective effects. Beyond EV-specific limitations, HepG2 cells have limited ability to capture inter-individual variability and mimic in vivo liver physiology. Future directions could include the use of primary human hepatocytes or liver organoid models to better capture inter-individual variability and improve translational relevance. Lastly, the current study focused solely on HepG2-to-HepG2 EV signaling. Investigating the effects of hepatocyte-derived EVs on other liver-resident cell types such as Kupffer cells, hepatic stellate cells, or liver sinusoidal endothelial cells may offer important insights into the sequence of cellular interactions following PFAS exposure.

In conclusion, this study represents a significant step in understanding the involvement of EVs in mechanisms of PFAS induced liver toxicity. By demonstrating distinct effects on cell viability, protein expression, and pathway alterations in HepG2 cells exposed to PFAS-modified EVs, we provide compelling evidence for EVs as potential mediators of PFAS-induced liver cell toxicity. Notably, we identified potential mechanistic links between altered EV miRNA expression profiles and cell protein expression, suggesting possible crosstalk between miRNA let-7e/ miR-29a with cellular proteins S100A8/ ETV4 respectively. Furthermore, preliminary findings suggesting EVs derived from healthy cells may be capable of mitigating PFAS-induced toxicity warrant further exploration in concentration-response and over varying time periods. As the first in vitro study evaluating the biological effects of PFAS-altered EVs, this work establishes a foundation for EVs mechanisms in PFAS-associated liver diseases and encourages the future evaluation of interventions for mitigating the impact of PFAS on liver health. Importantly, the impact of this work extends beyond PFAS research, outlining translational methods for future studies evaluating the broader role of EVs in mechanisms of toxicity associated with environmental exposures.

Supporting information

S1 Fig. PCA plot of proteomic results.

As displayed in the PCA plot, samples cluster moderately well by group. Treatment groups are denoted by color coded legend above figure. Co-exposure groups (on the left side of the plot) appear to be distinct from individual EVHepG2/CT, EVHepG2/PFAS, and control groups (depicted on the right side of the plot). EVHepG2/CT and EVHepG2/PFAS samples do not separate well from the controls overall. The pooled replicates cluster in the middle, as expected.

https://doi.org/10.1371/journal.pone.0338102.s001

(DOCX)

S2 Fig. Volcano plot of proteomic results.

The volcano plot shows the comparisons between control-treated HepG2 cells and the five other exposure conditions including EVHepG2/CT, EVHepG2/PFAS, PFAS, PFAS + EVHepG2/CT and PFAS + EVHepG2/PFAS. Proteins in red (adjusted p-value (q-value) < 0.1, log2 fold change ≤ − 0.58 or ≥ 0.58) are considered biologically significant. Common lab contaminants are additionally excluded from downstream analyses.

https://doi.org/10.1371/journal.pone.0338102.s002

(DOCX)

S1 Table. List of predicted mRNA targets of EV encapsulated miRNAs.

Differentially expressed EVHepG2/PFAS miRNAs (compared to EVHepG2/CT) with their predicted mRNA targets. Table separated by A) miRNAs with increased expression and B) miRNAs with decreased expression. Source of predicted and confidence are additionally listed.

https://doi.org/10.1371/journal.pone.0338102.s003

(XLSX)

S2 Table. List of differentially expressed proteins following EV exposure conditions.

Proteins with significantly differential expression are listed with corresponding Log2 fold changes, p-values and adjusted p-values.

https://doi.org/10.1371/journal.pone.0338102.s004

(XLSX)

References

  1. 1. Kotlarz N, Guillette T, Critchley C, Collier D, Lea CS, McCord J, et al. Per- and polyfluoroalkyl ether acids in well water and blood serum from private well users residing by a fluorochemical facility near Fayetteville, North Carolina. J Expo Sci Environ Epidemiol. 2024;34(1):97–107. pmid:38195989
  2. 2. Schultz AA, Stanton N, Shelton B, Pomazal R, Lange MA, Irving R, et al. Biomonitoring of perfluoroalkyl and polyfluoroalkyl substances (PFAS) from the Survey of the Health of Wisconsin (SHOW) 2014-2016 and comparison with the National Health and Nutrition Examination Survey (NHANES). J Expo Sci Environ Epidemiol. 2023;33(5):766–77. pmid:37580384
  3. 3. Antonopoulou M, Spyrou A, Tzamaria A, Efthimiou I, Triantafyllidis V. Current state of knowledge of environmental occurrence, toxic effects, and advanced treatment of PFOS and PFOA. Sci Total Environ. 2024;913:169332. pmid:38123090
  4. 4. Wang P, Liu D, Yan S, Cui J, Liang Y, Ren S. Adverse Effects of Perfluorooctane Sulfonate on the Liver and Relevant Mechanisms. Toxics. 2022;10(5):265. pmid:35622678
  5. 5. Costello E, Rock S, Stratakis N, Eckel SP, Walker DI, Valvi D, et al. Exposure to per- and Polyfluoroalkyl Substances and Markers of Liver Injury: A Systematic Review and Meta-Analysis. Environ Health Perspect. 2022;130(4):46001. pmid:35475652
  6. 6. Zhang X, Zhao L, Ducatman A, Deng C, von Stackelberg KE, Danford CJ, et al. Association of per- and polyfluoroalkyl substance exposure with fatty liver disease risk in US adults. JHEP Rep. 2023;5(5):100694. pmid:36968216
  7. 7. Repiska G, Crescitelli R, Lunavat TR, Soekmadji C, Cho WC. Editorial: The role of extracellular vesicles in diseases: Shedding light on their role in cell-to-cell communication. Front Genet. 2023;14:1123822. pmid:36760996
  8. 8. Wang J, Jing J, Zhou C, Fan Y. Emerging roles of exosomes in oral diseases progression. Int J Oral Sci. 2024;16(1):4. pmid:38221571
  9. 9. Baruah H, Sarma A, Basak D, Das M. Exosome: From biology to drug delivery. Drug Deliv Transl Res. 2024;14(6):1480–516. pmid:38252268
  10. 10. Gurunathan S, Kang M-H, Kim J-H. A Comprehensive Review on Factors Influences Biogenesis, Functions, Therapeutic and Clinical Implications of Exosomes. Int J Nanomedicine. 2021;16:1281–312. pmid:33628021
  11. 11. Hessvik NP, Llorente A. Current knowledge on exosome biogenesis and release. Cell Mol Life Sci. 2018;75(2):193–208. pmid:28733901
  12. 12. Kowal J, Tkach M, Théry C. Biogenesis and secretion of exosomes. Curr Opin Cell Biol. 2014;29:116–25. pmid:24959705
  13. 13. Carberry CK, Keshava D, Payton A, Smith GJ, Rager JE. Approaches to incorporate extracellular vesicles into exposure science, toxicology, and public health research. J Expo Sci Environ Epidemiol. 2022;32(5):647–59. pmid:35217808
  14. 14. Carberry CK, Rager JE. The impact of environmental contaminants on extracellular vesicles and their key molecular regulators: A literature and database-driven review. Environ Mol Mutagen. 2023;64(1):50–66. pmid:36502378
  15. 15. Martin PJ, Héliot A, Trémolet G, Landkocz Y, Dewaele D, Cazier F, et al. Cellular response and extracellular vesicles characterization of human macrophages exposed to fine atmospheric particulate matter. Environ Pollut. 2019;254(Pt A):112933. pmid:31382213
  16. 16. Nordgren TM, Heires AJ, Zempleni J, Swanson BJ, Wichman C, Romberger DJ. Bovine milk-derived extracellular vesicles enhance inflammation and promote M1 polarization following agricultural dust exposure in mice. J Nutr Biochem. 2019;64:110–20. pmid:30476878
  17. 17. Zhang L, Hao C, Zhai R, Wang D, Zhang J, Bao L, et al. Downregulation of exosomal let-7a-5p in dust exposed- workers contributes to lung cancer development. Respir Res. 2018;19(1):235. pmid:30497474
  18. 18. Xu Y, Luo F, Liu Y, Shi L, Lu X, Xu W, et al. Exosomal miR-21 derived from arsenite-transformed human bronchial epithelial cells promotes cell proliferation associated with arsenite carcinogenesis. Arch Toxicol. 2015;89(7):1071–82. pmid:24912785
  19. 19. Munson P, Lam Y-W, Dragon J, MacPherson M, Shukla A. Exosomes from asbestos-exposed cells modulate gene expression in mesothelial cells. FASEB J. 2018;32(8):4328–42. pmid:29553831
  20. 20. Zhou QL, Bai YZ, Gao J, Duan Y, Lyu YC, Guan LF, et al. Human Serum-derived Extracellular Vesicles Protect A549 from PM 2.5-induced Cell Apoptosis. Biomed Environ Sci. 2021;34(1):40–9. pmid:33531106
  21. 21. Saxena A, Walters MS, Shieh J-H, Shen L-B, Gomi K, Downey RJ, et al. Extracellular vesicles from human airway basal cells respond to cigarette smoke extract and affect vascular endothelial cells. Sci Rep. 2021;11(1):6104. pmid:33731767
  22. 22. Carberry CK, Bangma J, Koval L, Keshava D, Hartwell HJ, Sokolsky M. Extracellular vesicles altered by a per- and polyfluoroalkyl substance mixture: In vitro dose-dependent release, chemical content, and microRNA signatures involved in liver health. Toxicol Sci. 2023.
  23. 23. Théry C, Witwer KW, Aikawa E, Alcaraz MJ, Anderson JD, Andriantsitohaina R, et al. Minimal information for studies of extracellular vesicles 2018 (MISEV2018): a position statement of the International Society for Extracellular Vesicles and update of the MISEV2014 guidelines. J Extracell Vesicles. 2018;7(1):1535750. pmid:30637094
  24. 24. Hickman E, Carberry V, Carberry C, Cooper B, Mordant AL, Mills A, et al. Respiratory extracellular vesicle isolation optimization through proteomic profiling of equine samples and identification of candidates for cell-of-origin studies. PLoS One. 2025;20(1):e0315743. pmid:39854355
  25. 25. Riss TL, Moravec RA, Niles AL, Duellman S, Benink HA, Worzella TJ. Cell viability assays. In: Markossian S, Grossman A, Brimacombe K, Arkin M, Auld D, Austin C, editors. Assay Guidance Manual. Bethesda (MD). 2004.
    • 26. HaileMariam M, Eguez RV, Singh H, Bekele S, Ameni G, Pieper R, et al. S-Trap, an Ultrafast Sample-Preparation Approach for Shotgun Proteomics. J Proteome Res. 2018;17(9):2917–24. pmid:30114372
    • 27. Krämer A, Green J, Pollard J Jr, Tugendreich S. Causal analysis approaches in Ingenuity Pathway Analysis. Bioinformatics. 2014;30(4):523–30. pmid:24336805
    • 28. Macfarlane LA, Murphy PR. MicroRNA: Biogenesis, Function and Role in Cancer. Curr Genomics. 2010;11(7):537–61.
      • 29. Zablon FM, Desai P, Dellinger K, Aravamudhan S. Cellular and Exosomal MicroRNAs: Emerging Clinical Relevance as Targets for Breast Cancer Diagnosis and Prognosis. Adv Biol (Weinh). 2024;8(4):e2300532. pmid:38258348
      • 30. Martinez-Espinosa I, Serrato JA, Ortiz-Quintero B. The Role of Exosome-Derived microRNA on Lung Cancer Metastasis Progression. Biomolecules. 2023;13(11):1574. pmid:38002256
      • 31. Carberry C. Dataset for Celeste Carberry’s doctoral dissertation. UNC Dataverse. 2024.
        • 32. Aubrey BJ, Strasser A, Kelly GL. Tumor-Suppressor Functions of the TP53 Pathway. Cold Spring Harb Perspect Med. 2016;6(5):a026062. pmid:27141080
        • 33. Amstutz VH, Cengo A, Gehres F, Sijm DTHM, Vrolijk MF. Investigating the cytotoxicity of per- and polyfluoroalkyl substances in HepG2 cells: A structure-activity relationship approach. Toxicology. 2022;480:153312. pmid:36075290
        • 34. Jiang R, Huang H, Lian Z, Hu Z, Lloyd RS, Fang D, et al. Exosomal miR-221 derived from hydroquinone-transformed malignant human bronchial epithelial cells is involved in cell viability of recipient cells. J Appl Toxicol. 2020;40(2):224–33. pmid:31468561
        • 35. Zhu L, Wang H, Yuhan J, Zhang B, Li H, Asakiya C, et al. Exosomes mediated the delivery of ochratoxin A-induced cytotoxicity in HEK293 cells. Toxicology. 2021;461:152926. pmid:34481902
        • 36. Rodosthenous RS, Kloog I, Colicino E, Zhong J, Herrera LA, Vokonas P, et al. Extracellular vesicle-enriched microRNAs interact in the association between long-term particulate matter and blood pressure in elderly men. Environ Res. 2018;167:640–9. pmid:30216846
        • 37. Shin C-H, Byun J, Lee K, Kim B, Noh YK, Tran NL, et al. Exosomal miRNA-19a and miRNA-614 Induced by Air Pollutants Promote Proinflammatory M1 Macrophage Polarization via Regulation of RORα Expression in Human Respiratory Mucosal Microenvironment. J Immunol. 2020;205(11):3179–90. pmid:33115854
        • 38. Wang D, Hao C, Zhang L, Zhang J, Liu S, Li Y, et al. Exosomal miR-125a-5p derived from silica-exposed macrophages induces fibroblast transdifferentiation. Ecotoxicol Environ Saf. 2020;192:110253. pmid:32059163
        • 39. Donate PB, Alves de Lima K, Peres RS, Almeida F, Fukada SY, Silva TA, et al. Cigarette smoke induces miR-132 in Th17 cells that enhance osteoclastogenesis in inflammatory arthritis. Proc Natl Acad Sci U S A. 2021;118(1):e2017120118. pmid:33443169
        • 40. Delangre E, Oppliger E, Berkcan S, Gjorgjieva M, Correia de Sousa M, Foti M. S100 Proteins in Fatty Liver Disease and Hepatocellular Carcinoma. Int J Mol Sci. 2022;23(19).
        • 41. Zhou Z, Wu B, Chen J, Shen Y, Wang J, Chen X, et al. ETV4 facilitates proliferation, migration, and invasion of liver cancer by mediating TGF-β signal transduction through activation of B3GNT3. Genes Genomics. 2023;45(11):1433–43. pmid:37523127
        • 42. Abramowicz A, Story MD. The Long and Short of It: The Emerging Roles of Non-Coding RNA in Small Extracellular Vesicles. Cancers (Basel). 2020;12(6):1445. pmid:32498257
        • 43. Fenton SE, Ducatman A, Boobis A, DeWitt JC, Lau C, Ng C, et al. Per- and Polyfluoroalkyl Substance Toxicity and Human Health Review: Current State of Knowledge and Strategies for Informing Future Research. Environ Toxicol Chem. 2021;40(3):606–30. pmid:33017053
        • 44. Yuan H, Wen B, Liu X, Gao C, Yang R, Wang L, et al. CCAAT/enhancer-binding protein α is required for hepatic outgrowth via the p53 pathway in zebrafish. Sci Rep. 2015;5:15838. pmid:26511037
        • 45. Lin H, Wang L, Liu Z, Long K, Kong M, Ye D, et al. Hepatic MDM2 Causes Metabolic Associated Fatty Liver Disease by Blocking Triglyceride-VLDL Secretion via ApoB Degradation. Adv Sci (Weinh). 2022;9(20):e2200742. pmid:35524581
        • 46. Sun F, Liao Y, Qu X, Xiao X, Hou S, Chen Z, et al. Hepatic DNAJB9 Drives Anabolic Biasing to Reduce Steatosis and Obesity. Cell Rep. 2020;30(6):1835-1847.e9. pmid:32049014
        • 47. Hwang Y-J, Lee E-W, Song J, Kim H-R, Jun Y-C, Hwang K-A. MafK positively regulates NF-κB activity by enhancing CBP-mediated p65 acetylation. Sci Rep. 2013;3:3242. pmid:24247732
        • 48. Cao X, Zhang J, Apaer S, Yao G, Li T. MicroRNA-19a-3p and microRNA-376c-3p promote hepatocellular carcinoma progression through SOX6-mediated Wnt/beta-catenin signaling pathway. Int J Gen Med. 2021;14:89–102.
        • 49. Katagiri T, Nakamura Y, Miki Y. Mutations in the BRCA2 gene in hepatocellular carcinomas. Cancer Res. 1996;56(20):4575–7. pmid:8840963
        • 50. Leu JI-J, George DL. Hepatic IGFBP1 is a prosurvival factor that binds to BAK, protects the liver from apoptosis, and antagonizes the proapoptotic actions of p53 at mitochondria. Genes Dev. 2007;21(23):3095–109. pmid:18056423
        • 51. Lin B, Zhou X, Lin S, Wang X, Zhang M, Cao B, et al. Epigenetic silencing of PRSS3 provides growth and metastasis advantage for human hepatocellular carcinoma. J Mol Med (Berl). 2017;95(11):1237–49. pmid:28844099
        • 52. Qi W, Fang Z, Luo C, Hong H, Long Y, Dai Z, et al. The critical role of BTRC in hepatic steatosis as an ATGL E3 ligase. J Mol Cell Biol. 2024;15(10):mjad064. pmid:37873692
        • 53. Tuo L, Xiang J, Pan X, Hu J, Tang H, Liang L, et al. PCK1 negatively regulates cell cycle progression and hepatoma cell proliferation via the AMPK/p27Kip1 axis. J Exp Clin Cancer Res. 2019;38(1):50. pmid:30717766
        • 54. Wang C, Lv J, Xue C, Li J, Liu Y, Xu D, et al. Novel role of COX6c in the regulation of oxidative phosphorylation and diseases. Cell Death Discov. 2022;8(1):336. pmid:35879322
        • 55. Zeng M, Xu J, Zhang Z, Zou X, Wang X, Cao K, et al. Htd2 deficiency-associated suppression of α-lipoic acid production provokes mitochondrial dysfunction and insulin resistance in adipocytes. Redox Biol. 2021;41:101948. pmid:33774475
        • 56. Pflaum J, Schlosser S, Müller M. p53 Family and Cellular Stress Responses in Cancer. Front Oncol. 2014;4:285. pmid:25374842
        • 57. Akyol G, Dursun A, Poyraz A, Uluoglu O, Ataoglu O, Edalý N, et al. P53 and proliferating cell nuclear antigen (PCNA) expression in non-tumoral liver diseases. Pathol Int. 1999;49(3):214–21. pmid:10338076
        • 58. Yu S, Ji G, Zhang L. The role of p53 in liver fibrosis. Front Pharmacol. 2022;13:1057829. pmid:36353498
        • 59. Panasiuk A, Dzieciol J, Panasiuk B, Prokopowicz D. Expression of p53, Bax and Bcl-2 proteins in hepatocytes in non-alcoholic fatty liver disease. World J Gastroenterol. 2006;12(38):6198–202. pmid:17036395