Toxicity, uptake, and nuclear translocation of ingested micro-nanoplastics in an in vitro model of the small intestinal epithelium

Abstract

Despite mounting evidence of increasing micro- and nanoplastics (MNPs) in natural environments, food, and drinking water, little is known of the potential health hazards of MNPs ingestion. We assessed toxicity and uptake of environmentally relevant MNPs in an in vitro small intestinal epithelium (SIE). Test MNPs included 25 and 1000 nm polystyrene (PS) microspheres (PS25 and PS1K); 25, 100, and 1000 nm carboxyl modified PS spheres (PS25C, PS100C, and PS1KC), and secondary MNPs from incinerated polyethylene (PEI). MNPs were subjected to 3-phase digestion to mimic transformations in the gastrointestinal tract (GIT) and digestas applied to the SIE. Carboxylated MNPs significantly reduced viability and increased permeability to 3 kD dextran. Uptake of carboxyl PS materials was size dependent, with significantly greater uptake of PS25C. Fluorescence confocal imaging showed some PS25C agglomerates entering cells independent of endosomes (suggesting diffusion), others within actin shells (suggesting phagocytosis), and many free within the epithelial cells, including agglomerates within nuclei. Pre-treatment with the dynamin inhibitor Dyngo partially reduced PS25 translocation, suggesting a potential role for endocytosis. These findings suggest that ingestion exposures to MNPs could have serious health consequences and underscore the urgent need for additional detailed studies of the potential hazards of ingested MNPs.

Graphical Abstract

Introduction

Annual production of synthetic polymer materials (plastic resins and fibers) reached 380 million metric tons (Mt) in 2018 ^1,2. The four billion tons of plastics produced between 2002 and 2015 will be dwarfed by the 33 tons expected to be produced by 2050 ^3,4. Only one fifth of plastics are eventually recycled or incinerated, while the rest accumulate in landfills and natural environments ¹. Fragmentation of plastics across their life cycle caused by abrasion, UV damage, thermal decomposition, wind, and water, produces debris and particles of varied size, including micron and nanometer-scale so-called micro-nanoplastics (MNPs) ^1,3,5. These MNPs are grouped into two broad categories: 1) primary MNPs, generally beads produced in the micron and nanoscale for use as abrasives in sandblasting, and until recently in personal care products (banned in the EU, Canada, and in the U.S by the 2015 Microbead-Free Waters Act), and 2) secondary MNPs, which comprise particles generated by aging and fragmentation of plastic solids, fibers shed from synthetic fabrics during washing, and particles generated by incineration of waste plastic.

Although most micron size MNPs in the wastewater stream are removed by wastewater treatment plants (WWTPs), many remain in effluents ³. An additional 4×10⁵ Mt of MNPs in WWTP sludge are applied to soils ^3,6, where they accumulate ⁷. Secondary MNPs enter freshwater from fragmentation and runoff of landfill and litter plastic, as well as agricultural and road plastic particle runoff, and deposition of airborne MNPs from construction, household dust and municipal waste incineration plants ³. Recent studies have reported significant MNP contamination in terrestrial and freshwater environments, with some U.S. lake and river samples containing over 10⁵ MNPs/m^{3 5,8}. There is also evidence that MNPs can enter the food chain, either through contamination or by trophic transfer between predator and prey species ⁹. Given the magnitude and trajectory of plastics manufacturing and waste production, increasing evidence of MNPs in freshwater and other natural environments, and the potential for transfer through the food web, it is certain that the ongoing human ingestion exposures to MNPs will increase over time. It has been estimated the average human currently ingests roughly 5 g of MNPs per week through food and beverage consumption ¹⁰. Yet very little is known about the health implications and risks of these exposures.

One of the greatest concerns regarding MNPs is their potential to reach biological circulatory systems and be widely distributed to multiple target organs, as has been observed for many other engineered and environmental particles ^11–14. Evidence continues to grow raising concerns about potential health effects from ingested MNPs. Recent studies found that polystyrene (PS) MNPs were taken up by the gastrointestinal tract (GIT) in mice and rats ^15–19. Ingested PS MNPs were found in Peyer’s patches ^16,17,20, and MNPs smaller than 1 μm were found in the spleen and liver ²¹, following oral gavage. Moreover, MNPs (53 nm PS spheres) were able to cross the blood-brain barrier and accumulate in the brains of fish, where they caused morphological and behavioral abnormalities ¹⁴, and most recently, MNPs were identified in both maternal and fetal human placental tissue ²², suggesting the potential for maternal-fetal transfer of ingested MNPs. The ability of MNPs to cross such tightly controlled blood-tissue barriers suggests the potential for widespread distribution and potential toxicity.

In assessing these potential risks, one of the first questions that must be answered is whether ingested MNPs can cause direct toxicity, such as membrane damage, oxidative stress, and mitochondrial injury, in the GIT at relevant environmental or food/water contamination levels (current or future). In addition to characterizing their potential to cause direct cytotoxicity, it is also important to determine whether ingested MNPs can alter intestinal epithelial homeostasis and function, and particularly whether they can affect intestinal epithelial barrier function, impairment of which can have serious adverse health consequences. In addition to potential direct toxicity, functional effects and particularly changes in epithelial permeability may occur, as we have seen with otherwise non-toxic natural biopolymer nanomaterials ^23–25. We have also recently reported that the presence of engineered nanomaterials (ENMs) TiO₂ (E171) and SiO₂ (E551) significantly increased translocation (bioavailability) of the pesticide boscalid in the triculture system ²⁶. Further study revealed that multiple cell junctional genes were downregulated by these EMN exposures, suggesting that epithelial leakiness contributed to pesticide uptake ²⁷.

In the present study we investigated the potential intestinal epithelial toxicity as well as uptake, subcellular localization, and translocation of a panel of “real world” MNPs in an in vitro human small intestinal epithelium model. This panel included primary unmodified 25 and 1000 nm polystyrene (PS) spheres (PS25 and PS1K), primary 25, 100, and 1000 nm carboxylated PS spheres (PS25C, PS100C, PS1KC), and secondary polyethylene (PE) MNPs generated by incineration of virgin PE pellets (PE-I). To provide physiologically relevant exposures, MNPs dispersed in water (fasting food model, FFM) were subjected to a 3 phase (oral, gastric, small intestinal) digestion to reproduce the conditions and physicochemical transformations that would occur in vivo during digestion prior to interaction of ingested MNPs with the small intestinal epithelium ^24,28. The resulting final small intestinal phase digestas were then applied to an in vitro small intestinal epithelium including cells representing enterocytes (Caco-2), goblet cells (HT29-MTX), and microfold or M-cells (Caco-2 cells transformed by Raji B feeder cells) which we have previously used and described in studies of ingested materials. ²⁴ These models were used to assess toxicity, including effects on permeability, as well as uptake, subcellular localization, and translocation of the test MNPs.

Materials and Methods

Study design.

An overview of the study design is shown in Figure 1. Dispersions of MNPs in water (fasting food model, FFM) were subjected to 3-phase (oral, gastric, and small intestinal) simulated digestion to reproduce the physicochemical transformations (surface chemistry changes, biocorona formation, agglomeration) that would occur in vivo prior to interactions of ingested MNPs with the small intestinal epithelium. The resulting small intestinal phase digestas containing the transformed MNPs were applied to an in vitro triculture small intestinal epithelial model including cells representing intestinal enterocytes (Caco-2), mucus secreting goblet cells (HT29-MTX), and microfold or M-cells (Caco-2 cells transformed by Raji B feeder cells). ²⁴

Figure 1. Study design overview.

Test MNPs dispersed in water were subjected to simulated digestion to mi mimic the physicochemical transformations that would occur in vivo prior to interactions of ingested MNPs with the small intestinal epithelium. The final small intestinal phase digestas containing the transformed MNPs were applied to an in vitro triculture small intestinal epithelial model including cells representing intestinal enterocytes (Caco-2), mucus secreting goblet cells (HT29-MTX), and microfold or M-cells (Caco-2 cells transformed by Raji B feeder cells). Exposed tricultures grown on transwell inserts or in 96-well plates were used to assess multiple measures of toxicity as well as uptake, subcellular localization, and translocation of test MNPs.

Primary MNP materials.

Primary unmodified and uncolored polystyrene (PS) nano-microspheres with diameters of 25 nm (PS25), and 1000 nm (PS1K), as well as carboxyl modified 1000 nm PS (PS1KC) were obtained from Phosphorex Inc. (Hopkinton, MA) . Primary carboxyl modified fluorescent PS nano-microspheres with sizes of 25 nm (PS25C), 100 nm (PS100C), and 1000 nm (PS1KC) were obtained from Thermo Inc. (Waltham, MA). Red fluorescent PS25C and green fluorescent PS100C and PS1KC were used for imaging and quantitative analysis, and blue fluorescent versions of PS25C and PS100C were used in toxicity assays (along with the uncolored Phosphorex PS beads) to avoid interference with assay fluorescence or absorbance measurements.

Preparation and dispersion of primary polystyrene MNPs.

Because the PS nano-microspheres from Phosphorex Inc. were provided in an aqueous vehicle containing 0.1% of the detergent Tween, which alone could produce significant toxicity, dialysis was performed to remove it. Two milliliter samples of as purchased suspensions were loaded into 0.2 – 2.5 mL capacity 8 Kd MW cutoff Tube-O-DIALYZER™ micro dialysis tubes (Millipore Sigma, Burlington, MA) and floated in a large beaker containing 4 L of sterile deionized endotoxin-free water (GE Healthcare, Boston, MA) for 48 h at 4 °C with gentle stirring. Three 2.0 mL samples were dialyzed together, which would be expected to result in a dilution of the vehicle contaminant by a factor of 6/4000 = 0.0015, resulting in a negligible final Tween concentration of 0.00015%. At the end of dialysis the dialyzed test MNP suspensions were removed and their volumes, which increased slightly due to the greater osmotic pressure of the MNP suspension relative to the water, were measured in order to adjust the concentration of the samples. The dialyzed samples of PS MNPs from Phosphorex as well as the original non dialyzed carboxyl-PS MNPs from Thermo were then diluted to the desired starting MNP concentration and vortexed for 20 seconds.

Selection of starting MNP concentrations.

Due to the lack of accurate human exposure data in emerging studies resulting from technical challenges in measuring MNP particles smaller than several microns, current relevant environmental water and food concentrations for MNPs are not clear. Estimates from a study of MNPs released from plastic tea bags ²⁹, and from an analysis of bottled water samples ³⁰, range from about 0.2 to 10 μg/mL. On the other hand, a recent review and analysis of existing studies estimated that humans currently consume an average of 5 g of MNPs per week, primarily through drinking water ¹⁰. Assuming an average fluid intake of 2.0 L/d ³¹, this would correspond to an average MNP intake of >350 μg/mL. For these acute studies we have chosen starting water concentrations of 400 and 1000 μg/mL, which because of the four-fold dilution of final digestas required to maintain adequate nutrition in the in vitro epithelial system during exposure, correspond to effective oral concentrations of 100 and 250 μg/mL. Although these concentrations exceed those found in environmental samples to date, they will allow us to elicit and thereby detect potential acute toxicities that might not be seen at lower doses. Moreover, given the expected continued exponential growth of plastic production and waste generation, and the likelihood that studies to date, typically with > 1 um size detection limits, underestimate total MNP exposures, we believe these doses are justified.

Synthesis and dispersion of secondary MNPs: incinerated polyethylene (PE) particles.

Incinerated PE MNP particles (PEI) were generated from the thermal decomposition of bulk-size PE pellets using our previously described Integrated Exposure Generation System (INEXS) (Figure S1)³², a versatile and reproducible platform to investigate thermal decomposition behavior of materials under controlled combustion conditions. Pristine PE pellets were obtained in-kind from our industrial collaborator, BASF (Ludwigshafen, Germany). Thermal decomposition conditions were set at a final temperature of 850°C, heating rate of 20°C/min (starting from the ambient temperature) and oxygen concentration of 20.9 vol% (ambient level). The released aerosol was size-fractionated and collected on appropriate substrates in the Harvard Compact Cascade Impactor (CCI).³³ The collected PM_2.5 size fraction (particles with aerodynamic diameters less than 2.5 μm) was extracted and dispersed in an aqueous stock suspension using our previously developed and validated SEDD (Sampling, Extraction, Dispersion and Dosimetry) protocol.³⁴ The aqueous stock suspension of PE-I particles was then diluted to the desired starting MNP concentration (400 μg/mL or 1000 μg/mL) in the sterile deionized endotoxin-free water (GE Healthcare, Boston, MA) for subsequent digestions.

In vitro gastrointestinal digestion simulation.

In vitro simulated digestions were performed using a 3-phase (oral, gastric, and small intestinal) simulator as previously described in detail ²⁴. Briefly, in the oral phase, MNP-water suspensions and controls were mixed at a ratio of 1:1 with pre-warmed 37°C simulated saliva, containing mucin and various salts at a pH of 6.8, and inverted by hand for 15 seconds. The resulting oral phase digesta was combined 1:1 with pre-warmed 37°C simulated gastric fluid, containing pepsin, HCl, and NaCl, at a pH of 2.0, and incubated for 2 hours in an orbital shaker at 37°C, to simulate the gastric phase of digestion. The resulting gastric phase digesta was then combined and mixed with bile salts, pancreatin (containing all pancreatic digestive enzymes), and additional mineral salts (diluting the gastric digesta by a factor of three), and the pH of the resulting mixture was adjusted to 7.0 by addition of NaOH or HCL to simulate small intestinal fluid. The small intestinal phase was incubated in a rotary shaking incubator at 37°C for 2 hours, representing the small intestinal phase of digestion ²⁴.

Characterization of MNP suspension and digesta size distribution by multi-angle light diffraction (MALD).

The volume-weighted size distributions of small intestinal digestas of MNPs were acquired using a Mastersizer 3000 (Malvern Panalytical, Ltd., Malvern, Worcestershire, United Kingdom) equipped with a wet dispersion unit (Hydro SV). Background measurements were performed to correct for light refraction and background noise from the continuous phase of the small intestinal digestas. To do so, the supernatant of the small intestinal phase of blank FFM were acquired after centrifugation at 10,000 × g for 10 min. The supernatants were then diluted 1-in-7 with HyClone cell culture grade water and used for background measurements before measuring the size distribution of digested MNPs. To achieve optimal obscuration range, the small intestinal digestas MNPs were also diluted 1-in-7 with HyClone cell culture grade water and measured while stirred at 1800 rpm. For all samples, selected dispersant type was “water,” particle type was “spherical”. The refractive index of PEI-based and PS-based MNPs was set at 1.50 and 1.59, respectively; the absorption index of both PEI-based and PS-based MNPs was set at 0.01, and their density was set at 1.27 and 1.05 g cm⁻³. For each sample, 7–10 measurements of 120 s (at 633 nm) and 10 s (at 466 nm) each were performed. The volume-weighted size distributions were averaged and plotted using GraphPad Prism 9.1.

Tri-culture small intestinal epithelium cell model and treatments to assess cytotoxicity.

All cell cultures were grown at, and all incubations during treatments occurred at 37°C with 5% CO₂. Small intestinal epithelial model tri-cultures were prepared as previously described ²⁴. In brief, Caco-2, HT29-MTX, and Raji B cells were obtained from Millipore Sigma (Burlington, MA). All cell culture media and supplements were obtained from Thermo Inc. (Waltham MA). Caco-2 and HT29-MTX cells were grown in high-glucose DMEM (Thermo supplemented with 10% heat-inactivated fetal bovine serum (FBS), 10 mM HEPES buffer, 100 IU/mL penicillin, 100 μg/mL streptomycin, and non-essential amino acids (1/100 dilution of 100 X solution). Raji B cells were cultured in RPMI 1640 media supplemented with 10% FBS, 10 mM HEPES buffer, 100 IU/mL penicillin, and 100 μg/mL streptomycin. For transwell inserts, Caco-2 and HT29-MTX cells were trypsinized and resuspended in DMEM media at 3 × 10⁵ cells/mL and combined in a ratio of 3:1 (Caco-2:HT29-MTX). A 1.5-mL portion of the cell mixture was seeded in the apical chamber, and 2.5 mL of complete DMEM media was added to the basolateral compartment of a 6 well transwell plate (Corning Inc, Corning, NY). Media was changed after four days, and subsequently every other day, until day 15. On days 15 and 16, the media in the basolateral compartment was replaced with 2.5 mL of a suspension of Raji B cells at a concentration of 1 × 10⁶ cells/mL in 1:1 DMEM: RPMI complete media. Transepithelial electrical resistance (TEER) was measured using an EVOM2 Epithelial Volt/Ohm Meter with a Chopstick Electrode Set (World Precision Instruments, Sarasota, FL).

Cytotoxicity (LDH release), and uptake and translocation studies using tri-cultures on transwells were initiated on day 17. Cell viability (tetrazolium salt reduction) and oxidative stress (ROS production) studies require closed-bottom adherent cell cultures in 96-well plates suitable for plate reader fluorescence measurements. For these studies, Caco-2/HT29-MTX co-cultures were prepared in 96-well plates. Raji B cells were not used in this format, since they are suspension feeder cells (added to the transwell basolateral compartments to promote Mcell differentiation of some apical Caco-2 cells), and not part of the epithelium, but could adhere to mucus, or become incorporated in the epithelial layer, if applied apically in closed 96-well plates. To prepare these co-cultures, Caco-2 and HT29-MTX cells at a 3:1 ratio were seeded at a total 3 × 10⁴ cells/well (100 μL of cell mixture) in black-walled, clear optical bottom plates (BD, Franklin Lakes, NJ). Media was changed after four days, and subsequently every other day, until day 17. In 96-well plate format the co-cultures reach 100% confluence by ~day 12–14, but are maintained for 17 days to ensure complete maturation of Caco-2 cells to the enterocyte phenotype. Cell viability and oxidative stress experiments performed with the 96-well plate co-cultures were initiated on day 17.

Exposure of transwell tri-culture and 96-well plate co-cultures to digesta.

All cell culture media, supplements, and buffers used in the exposures were obtained from Thermo Inc. (Waltham MA). The transwell inserts and 96-well plates were rinsed with glucose-free DMEM supplemented with 10 mM HEPES buffer, 100 IU/mL penicillin, 100 μg/mL streptomycin and non-essential amino acids (1/100 dilution of 100× solution), and 10 mM pyruvate. The final small intestinal digesta from simulated digestions were combined with phenol red free complete DMEM media without FBS in a ratio of 1:3, and the mixture was applied to the cells (1.5 ml to the apical compartment for transwell inserts, 200 μl per well for 96-well plates). Apical fluid in untreated control wells was replaced with fresh glucose-free media. Digesta was also dispensed in a cell-free control well. Transwell cells were incubated with digesta for 4 h. At the end of exposure, TEER in transwells was measured as described above, and supernatants from transwells were collected for lactate dehydrogenase (LDH) release analysis. Assessment of reactive oxygen species (ROS) production and cell viability (described below) was performed after 4 h exposures, respectively, in 96-well plates.

Cytotoxicity (lactate dehydrogenase (LDH) release).

Cytotoxicity was assessed using a triculture model of the small intestinal epithelium grown on transwell inserts or in 96-well plates ²⁴. Supernatants from transwells were collected after 24h exposures for LDH analysis, which was performed using the Pierce LDH assay kit (Millipore Sigma, Burlington, MA) according to manufacturer’s instructions. Untreated control wells were used to measure spontaneous LDH release. For maximum LDH release control wells, 150 μL of apical fluid was removed and replaced with 150 μL 10× lysis buffer 45 minutes prior to the end of incubation. The provided substrate was dissolved in 11.4 mL of ultrapure water and added to 0.6 mL assay buffer to prepare the reaction mixture. Apical fluid in each well was pipetted to the mix and 150 μL was transferred to a 1.5 mL tube. Tubes were centrifuged at 5,000 × g for 5 min., and 50 μL of the supernatant from each tube was dispensed in triplicate wells in a fresh 96-well plate. 50 μL of reaction mixture was added and mixed by tapping the plate. Plates were incubated at room temperature for 30 minutes or less (to provide maximum difference in color between samples by visual inspection), and 50 mL stop solution was added and mixed by tapping. Absorbance was measured at 490 nm (A490) and 680 nm (A680) using a SpectraMax M-5 microplate reader and SoftMax Pro acquisition and analysis software (Molecular Devices, San Jose, CA). To calculate LDH activity, A680 values were subtracted from measured A490 values to correct for instrument background. To correct for digesta background, LDH activities from no- cell controls were subtracted from test well LDH activities. Percent cytotoxicity was calculated by subtracting spontaneous LDH release values from treatment values, dividing by total LDH activity (Maximum LDH activity – Spontaneous LDH activity), and multiplying by 100.

Cell viability (mitochondrial metabolic activity).

Cell viability was assessed using the PrestoBlue™ reagent (Thermo Inc., Waltham, MA). PrestoBlue is a soluble tetrazolium salt that readily enters cells, where it is reduced, in metabolically active (viable) cells, by mitochondrial dehydrogenases and reductases, to an insoluble, blue-colored, red-fluorescent formazan product. The PrestoBlue™ viability assay was performed after 24 h exposure to digestas, using 96-well plate co-cultures, according to manufacturer’s instructions. Briefly, treated and negative control (100% viable) untreated cells were washed 3 times with 200 μL/well PBS, and 100 μL of 10% PrestoBlue reagent was added to each well. Plates were then incubated at 37 ºC for 15 minutes, and fluorescence was measured at 560 nm (excitation)/590 nm (emission) using a SpectraMax M-5 microplate reader and SoftMax Pro acquisition and analysis software (Molecular Devices, San Jose, CA).

Oxidative stress (ROS production).

Oxidative stress was assessed by measuring cellular reactive oxygen species (ROS) accumulation after 6 h exposure to digestas in 96-well co-cultures. ROS production (oxidative stress) was assessed using the CellROX green reagent (Thermo Inc., Waltham, MA) according to the manufacturer’s instructions. Briefly, a 5 mM working solution of the CellROX green reagent was prepared from 20 mM stock by diluting in glucose-free DMEM media without FBS. Media was removed from test wells and replaced with 100 μL of working solution, and plates were incubated for 30 minutes at 37 ºC. Cells were then washed 3 times with 200 μL/well PBS, and fluorescence was measured at 480 nm (excitation)/520 nm (emission) using a SpectraMax M-5 microplate reader and SoftMax Pro acquisition and analysis software (Molecular Devices, San Jose, CA), quantifying the amount of oxidized CellROX green reagent.

Epithelial permeability measurement using fluorescent dextran.

Fluorescently labeled (Alexa Fluor 488) 3kD dextran (Thermo Inc., Waltham, MA) was diluted in PSB to a concentration of 25 μg/mL. Following experimental treatments in transwell tricultures, cells were washed twice with PBS and 2 mL of the 25 μg/mL fluorescent dextran solution was applied to the apical compartment while 2 mL of fresh DMEM without phenol red or FBS was added to the basolateral compartment. After incubation of cells at 37 °C for 60 min, 0.3 mL samples of basolateral fluids were obtained and fluorescence was measured (Ex 495 nm, Em 519 nm) using a SpectraMax M-5 microplate reader and SoftMax Pro acquisition and analysis software (Molecular Devices, San Jose, CA). The concentration of fluorescent dextran in samples was calculated from a standard curve generated by serial 2-fold dilutions of the fluorescent dextran in DMEM without phenol red or FBS (basolateral fluid). The apparent permeability coefficient, P_app, for 3kD dextran in each triculture sample was calculated as 𝑃_𝑎𝑝𝑝 = 𝑋/(𝐴∗𝑡∗𝐶_𝑑), where 𝑋 is the amount (mass) of the substance in the receiver chamber (i.e., 2 mL * basolateral compartment concentration), 𝐴 is the diffusion area (transwell membrane area = 0.33 cm²), 𝑡 is the incubation time in seconds (3600 s), and 𝐶_𝑑 is the concentration in the donor (apical) compartment.

Measurement of fluorescent MNP uptake and translocation in triculture transwells.

To assess uptake and translocation of test MNPs (fluorescent 25, 100, and 1000 nm carboxyl modified PS nano-microbeads), final digestas of red fluorescent PS25C and green fluorescent PS100C and PS1KC from starting concentrations of 0.4 mg/mL were combined with phenol red-free complete DMEM media without FBS in a ratio of 1:3, and 1.5 mL of the resulting mixtures were applied to the apical compartments of test transwell tricultures, while the basolateral fluid was replaced with 2.0 mL of phenol red-free complete DMEM supplemented with 10% FBS. The transwells were incubated with the MNP digesta mixtures at 37°C and 5% CO₂ for 4 hours. At the end of incubation, the apical and basolateral fluids were collected for analysis. The epithelial cells were lysed by replacing apical media with 0.5 mL of RIPA buffer and incubating at room temperature for 5 minutes. A cell scraper was used to facilitate disruption of the cell layer. One mL of deionized water was then added, and the combined fluid was pipetted up and down several times before being collected for analysis. Standard solutions of each fluorescent MNP were prepared by serial 2-fold dilutions of MNP stock in apical fluid (1:3 mixture of blank digesta and phenol red-free complete DMEM without FBS), basolateral fluid (phenol red-free complete DMEM with 10% FBS), and cell lysate fluid over the expected range of concentrations for each of the corresponding types of test samples. Test and standard samples were dispensed in clear bottom black walled microplates (200 μL/well) and fluorescence was measured at 580 nm (excitation)/605 nm (emission) for red fluorescent PS25C and at 505 nm (excitation)/515 nm (emission) for green fluorescent PS100C and PS1KC spheres using a SpectraMax M5 microplate reader and SoftMax Pro acquisition and analysis software (Molecular Devices, San Jose, CA). Standard curves prepared with linear fitting from the standard sample measurements for each of the PS materials in each media type were used to calculate the test sample concentrations in apical, basolateral, and cell + mucus compartments. To assess the role of dynamin-dependent endocytosis in uptake and translocation of PS25C spheres, some transwells were pre-treated for 30 min with 50 μM of the dynamin inhibitor Dnygo (Selleck Chemicals, Houston, TX) in complete DMEM without FBS prior to exposure to digesta/media mixture containing PS25C.

Assessment of fluorescent MNP uptake and its dependence on MNP size by confocal microscopy.

To visually assess the uptake of MNPs by the triculture epithelium, digestas of red fluorescent PS25C and green fluorescent PS100C and PS1KC from starting concentrations of 0.4 mg/mL were prepared and transwell tricultures were exposed as described above and incubated for 3 hours. The cells were then washed twice with PBS, fixed for 15 minutes with 4% formaldehyde in PBS, washed twice more with PBS, and counterstained with Hoechst 33342 (Thermo Inc., Waltham, MA) at 10 μg/mL for 1 h at room temperature. Roughly 1 × 1 cm sections of the membranes were excised from the inserts using a #11 scalpel (sharp point to pierce, straight blade to cut) and fine tweezers and mounted in PBS on #1.5 cover glass bottom 35 mm petri dishes (Nalge Nunc, Rochester, NY) Confocal fluorescence image stacks were acquired using a 50 μm single disk Yokogawa CSU-W1 spinning disk confocal head on a Nikon Ti2 inverted microscope with an Apo λS LWD 40X/1.1 NA DICN2 objective and an Andor Zyla 4.2 plus sCMOS monochrome camera. The blue channel (Hoechst) was acquired using a directly modulated 405 nm solid state diode laser line, a Semrock Di01-T405/488/568/647 dichroic beamsplitter and Chroma 455/50 nm bandpass emission filter. For samples containing red fluorescent PS25C, the corresponding red channel was acquired with a directly modulated 561 nm DPSS laser line, a Semrock Di01-T405/488/568/647 dichroic beamsplitter, and Chroma 605/52 nm bandpass emission filter. Green fluorescence for samples containing green fluorescent PS100C or PS1KC was acquired using a directly modulated 488 nm diode laser line, a Semrock Di01-T405/488/568/647 dichroic beamsplitter, and Chroma 525/36 nm bandpass emission filter. Z stacks were acquired for all samples. Images were acquired with Nikon Elements AR 5.21.03 acquisition software. Images were displayed and analyzed using the Fiji open source platform ³⁵. Brightness and contrast were adjusted to optimize feature clarity and minimize background in each channel and channels were merged to create a composite stack. No filters were applied. The Fiji Orthogonal Views feature was used to create 3-plane (XY, YZ, ZX) views and the ImageJ 3D Viewer plugin was used to create 3D views and movies ³⁵.

Assessment of uptake and subcellular localization of 25 nm carboxylated PS by confocal microscopy.

To visualize and characterize the uptake and subcellular localization of red fluorescent PS25C, high resolution multiplexed confocal fluorescence imaging was performed, including fluorescent labeling and imaging of early endosomes and F-actin. Early endosomes were labeled with the CellLight Early Endosome-GFP BacMam 2.0 reagent (Thermo Inc., Waltham, MA), transfection with which results in expression of a Rab5a-GFP fusion protein that is targeted to early endosomes. One day prior to exposures, 500 μL of the CellLight transfection reagent was added directly to the apical media in each transwell, representing roughly 30 BacMam particles per cell. The cells were then incubated overnight at 37°C and 5% CO₂ and washed with PBS before exposure to digesta/media mixture containing red fluorescent PS25C prepared as described above. Following 3 h exposures, cells were washed with PBS, fixed with 4% formaldehyde in PBS for 15 minutes at room temperature, washed twice with PBS, and permeabilized by incubation with 0.1% Triton X-100 (Millipore Sigma, Burlington, MA) in PBS for 10 minutes at room temperature followed by washing twice with PBS. A 1000 X stock of conjugated phalloidin (Phalloidin CruzFluor 750, Santa Cruz Biotechnology, Inc., Dallas, TX) was prepared by adding 30 μL of DMSO to the supplied vial and mixing thoroughly. The final staining solution was prepared by diluting the stock 1/1000 in PBS + 1% sterile filtered FBS and adding Hoechst 33342 nuclear dye (Thermo Inc., Waltham, MA) at 1 μg/mL. The fixed transwell tricultures were then incubated with staining solution (1 mL apical, 1.5 mL basolateral) for 60 min at 37°C and 5% CO₂, and washed three times with PBS. Roughly 1 × 1 cm sections of the membranes were excised from the inserts using a #11 scalpel and tweezers and mounted in SlowFade Glass Antifade mountant (Thermo, Inc., Waltham, MA) on a #1.5 coverslip (Electron Microscopy Sciences, Hatfield, PA). Confocal fluorescence image stacks were acquired using a Leica Stellaris 8 inverted single point laser scanning confocal microscope with a Plan Apo 63x/1.4 NA objective, using a continuum white light laser, and selecting the suitable excitation wavelengths and spectral filtering to minimize channel crosstalk and acquire Z stacks for blue (Hoechst), green (early endosomes), red (PS25C), and far red (actin-phalloidin CF-750) channels. Images were processed and orthogonal and 3D views and movies were created as described above.

To further investigate the subcellular localization of PS25C MNPs and to verify observations made in the first approach, staining and confocal imaging were also performed with tricultures grown on coverslips to rule out effects of the membrane and pore structure on morphology or function, and with early endosomes labeled by indirect fluorescence immunostaining of the early endosome marker Rab5a rather than the CellLight transfection described above, to rule out off target labeling by that method. Triculture cells were grown on #1.5 coverslips that were placed into the apical transwell compartments prior to seeding. Other than the presence of the coverslips, the seeding and maintenance of the transwell tricultures was identical to that described above for transwells without coverslips. To assess the role of dynamin-dependent endocytosis in uptake and translocation of PS25C spheres, some transwell tricultures were pre-treated for 30 minutes with 50 μM Dnygo (Selleck Chemicals, Houston, TX) in complete DMEM without FBS prior to exposure to the digesta/media mixture containing PS25C.

Following the 3 h exposures (performed as described above), cells (on coverslips on top of transwell membranes) were washed with PBS, fixed with 4% formaldehyde in PBS for 15 minutes at room temperature, washed twice with PBS, and permeabilized by incubation with 0.1% Triton X-100 in PBS for 15 minutes at room temperature, followed by washing twice with PBS. Cells were blocked for 2h at room temperature with gentle rocking with 5% non-fat milk and 2.5% normal goat serum (Thermo Inc., Waltham MA)) in PBS. Blocking solution was replaced with 10 μg/mL primary rabbit polyclonal rabbit anti-Rab5a antibody (PA5–29022, Thermo Inc., Waltham, MA) in PBS and cells were incubated with rocking for 2 h at room temperature followed by 24 h at 4°C. Cells were then washed four times (3 mL basolateral + 3 mL apical) for 5 min per wash with rocking at room temperature, incubated with 2 μg/mL secondary goat anti-rabbit IgG Alexa Fluor 488 (Thermo Inc., Waltham, MA) in PBS for 2h with rocking and protection from light, and washed four more times for 5 min each. Staining with phalloidin CruzFluor 750 and Hoechst nuclear staining was performed as described above. Cover slips were removed with tweezers and mounted on slides in SlowFade Glass Antifade mountant (Thermo Inc., Waltham, MA). Confocal Z stacks were acquired using a Leica Stellaris 8 inverted single point laser scanning confocal microscope with a Plan Apo 63x/1.4 NA objective, using a continuum white light laser, and selecting the suitable excitation wavelengths and spectral filtering to acquire Z stacks for blue (Hoechst), green (early endosomes, Alexa Fluor 488), red (PS25C), and far red (actin-phalloidin CF-750) channels. Images were described and orthogonal and 3D views and movies were created as described above.

Results

Characterization of incinerated polyethylene particles (PEI).

Detailed characterization of particles generated by incineration of polyethylene pellets using INEXS system was previously published by the authors³⁶. Size distributions obtained by in-line real-time measurement by scanning mobility particle sizer (SMPS) and aerodynamic particle size (APS) spectrometry of the generated aerosol showed that most particles (by mass) were in the nanoscale range (PM_0.1), with much smaller mass fractions in the PM0.1–2.5 and PM>2.5 size ranges.³⁶ Electron microscopy showed that the generated particles were mostly spherical in shape.³⁶

Effects of MNPs on digesta size distributions determined by multi-angle light diffraction (MALD).

The volume-weighted size distributions of final small intestinal digestas of MNPs (from starting concentration of 0.4 mg/mL) and corresponding blank digesta controls (without MNP) are shown in Figure S2. All size distributions were essentially bimodal with large peaks in the 3–5 μm and >10 μm ranges. The presence of PS25 and PS1K (unmodified) MNPs resulted in modest right shifts in the upper peaks of the distributions and shoulders or smaller peaks representing agglomerates ~10 fold larger than any seen in the blank digestas (Figure S2 a and b). The carboxylated PS (PS1KC) produced a similar but less pronounced right shift (Figure S2 c). In the case of both PS1K and PS1KC, a small peak (not present in the blank) can be seen at 1–2 μm, presumably representing free MNP primary particles. The presence of PEI, in contrast, produced a left shift (toward smaller sizes) of the volume distribution (Figure S2 d).

In vitro cytotoxicity of MNPs in a tri-culture small intestinal epithelial model.

The results of cytotoxicity assessment of test MNPs are summarized in Figure 2. At starting drinking water concentrations of 0.4 and 1.0 mg/mL (corresponding to applied to cell concentrations of 8.33 and 21.83 μg/mL) none of the test MNPs had a significant effect on either TEER (Figure 2 a) or cytotoxicity (LDH release, Figure 2 b) after 24 h exposure, or caused a significant increase in reactive oxygen species (ROS) production after 6 h exposures (Figure 2 c).

Figure 2. In vitro toxicity of MNPs.

Results of toxicity assessments in an in vitro triculture small intestinal epithelium model exposed to small intestinal digestas of MNP suspensions in water. a. TEER after 24 h exposure, b. Percent Cytotoxicity (percent of LDH release relative to that of lysed control cells) after 24 h exposure, c. Reactive oxygen species (ROS) production (fold change relative to untreated) after 6 h exposure, d. Percent Viability (mitochondrial reductase activity relative to untreated control) after 24 h exposure, e. Fold change in apparent permeability coefficient (P_app) assessed with fluorescent labeled (Alexa Fluor 488) 3 kD dextran. N=3. Initial water concentrations are indicated for each material (0.4 or 1.0 mg/mL, corresponding to final applied concentrations of 8.33 and 20.83 μg/mL). PEI: Incinerated Polyethylene, PM_2.5 fraction; PS25: Polystyrene 25 nm spheres; PS1K: Polystyrene 1 μm spheres; PS25C: Carboxyl modified PS 25 nm spheres; PS100C: Carboxyl modified PS 100 nm spheres; PS1KC: Carboxyl modified PS 1000 nm (1 um) spheres. * p < 0.05; ** p < 0.01.

Cell viability (mitochondrial enzyme activity), on the other hand, was significantly reduced by 24 h exposures to digestas of several test MNPs (Figure 2 d). Specifically, all sizes (25, 100, and 1000 nm) of carboxylated PS spheres at the higher starting concentration of 1.0 mg/mL and both 100 and 1000 nm (but not 25 nm) at the lower concentration of 0.4 mg/mL caused statistically significant 40–50% reductions in triculture cell viability. The two unmodified PS spheres (25 and 100 nm) also appeared to reduce viability somewhat (by ~20–30%) at both doses, but due to variability in the data these changes were not statistically significant. Incinerated polyethylene particles (PEI) had no apparent or statistically significant effect on triculture cell viability.

Permeability of the transwell triculture epithelium to fluorescently labeled 3kD dextran, and indicator or epithelial layer barrier function integrity, was also significantly increased by 24 h exposures to digestas of several MNPs, particularly by exposures to the carboxyl modified PS nano and microspheres. Results are represented as fold changes (relative to blank digesta of water without MNP) in Figure 2 e (the corresponding raw P_app values are shown in Figure S3). With the exception of PEI and 25 nm unmodified PS spheres, all test materials appeared to cause two-fold or greater increases in the apparent permeability coefficient (P_app) in the triculture transwell model. The increase was most marked and statistically significant for 100 nm carboxylated PS spheres (PS100C) at a starting concentration of 0.4 mg/mL (5.5-fold increase, p<0.05), and 1μm carboxylated PS at both 0.4 mg/mL (7.9 fold increase, p<0.01) and 1.0 mg/mL (5.9 fold increase, p<0.05).

Translocation of ingested fluorescent carboxyl-modified polystyrene MNPs in a small intestinal epithelium model.

To assess the ability of MNPs to be taken up by or be translocated across the small intestinal epithelium, small intestinal digestas of carboxyl-modified red fluorescent 25 nm (PS25C) and green fluorescent 100 nm (PS100C) and 1000 nm (PS1KC) polystyrene spheres were applied to triculture small intestinal epithelium growing on transwell inserts and incubated for 3 hours and the amount of fluorescent MNP in the apical, cell lysate, and basolateral compartments were determined as described in the methods section. The percentages of the applied MNP mass found in each compartment for each of the 3 sizes of carboxyl-modified PS are shown in Figure 3 a, b, c. For all 3 sizes of PS, most of the applied mass remained in the apical chamber (96%, 92%, and 97% for PS25C, PS100C, and PS1KC, respectively), although a slightly and significantly lower percentage remained in the case of the intermediate 100 nm PS spheres (Figure 3 a). This difference appeared to be largely accounted for by the markedly and significantly greater percentage of PS100C found in the cell lysate compartment (1.4%, 5.7%, and 1.3% for PS25C, PS100C, and PS1KC, respectively) which includes particles within and beneath the triculture cells on the transwell membrane as well as particles trapped in the apical mucus layer (Figure 3 b). Percentages of applied MNP mass measured in the basolateral compartment (Figure 3 c), representing translocation of particles through cells (as well as through the 3 μm pores of the transwell membrane), revealed significantly greater translocation of 25 nm spheres than of either 100 or 1000 nm spheres (1.15%, 0.73%, and 0.64% for PS25C, PS100C, and PS1KC, respectively).

Figure 3. Translocation of ingested fluorescent carboxyl-modified polystyrene MNPs in a small intestinal epithelium model.

Small intestinal digestas of carboxyl-modified red fluorescent 25 nm (PS25C) and green fluorescent 100 nm (PS100C) and 1000 nm (PS1KC) polystyrene spheres were applied to triculture small intestinal epithelium growing on transwell inserts and incubated for 3 hours. a, b, c. Percent of applied MNPs measured in apical, lysate (cells + mucus), and basolateral compartments after 3-hour incubation. d, e, f. Orthogonal 3 plane views of confocal image volumes of triculture specimens fixed and stained after 3 hour exposures to digestas of PS25C, PS100C, and PS1KC fluorescent spheres. g, h, i. 3D volumetric renderings of confocal image volumes of triculture specimens exposed to digestas of PS25C, PS100C, and PS1KC fluorescent spheres. j, k, l. Cropped 3D volumes of exposed triculture specimens exposed to digestas of PS25C, PS100C, and PS1KC fluorescent spheres. * p < 0.05; ** p < 0.01;

Analysis of confocal fluorescence image volumes of exposed triculture samples (Figure 3 d-l) revealed findings consistent with the quantitative results described above. In orthogonal 3 plane views (Figure 3 d, e, and f), agglomerates of PS25C were seen between nuclei within the epithelial layer, whereas PS100C and PS1KC particles or agglomerates were only seen above the epithelium, suggesting entrapment of those particles within the mucus layer. Such presumably mucus-trapped agglomerates were most prevalent for PS100C, consistent with the much greater percentage of applied PS100C measured in the cell + mucus compartment as described above (Figure 3 b). 3D volumetric renderings of the confocal image volumes (Figure 3 g, h, and i) are consistent with the above findings. Red fluorescent PS25C agglomerates were seen interposed between epithelial cell nuclei (Figure 3 g and Movie S1), whereas green fluorescent PS100C (Figure 3 h and Movie S2) and PS1KC (Figure 3 i and Movie S3) were restricted to the apical extracellular space which likely corresponds to the mucus layer. Once again, considerably larger amounts of PS100C than of PS1K were seen in the corresponding apical extracellular layers, consistent with the quantitative findings noted above. 3D renderings of cropped 3D subvolumes of these confocal data sets (Figure 3 j, k, and l) further demonstrate penetration of PS25C into the epithelial layer (Figure 3 j), and restriction of PS100C and PS1KC fluorescent spheres to the apical extracellular mucosal layer (Figure 3 k and l).

Uptake and subcellular localization of ingested fluorescent carboxylated fluorescent polystyrene MNPs in a small intestinal epithelium model.

To characterize in more detail the uptake and subcellular localization of the 25 nm PS25C, transwell triculture specimens transduced prior to exposure with a Rab5a-GFP fusion protein to label early endosomes and stained with fluorescently labeled phalloidin after fixation to label actin were imaged at high resolution (Figure 4 a-g). Analysis of orthogonal 3 plane views of these image volumes revealed a number of notable features: 1). PS25C agglomerates appeared in some images to be entering cells unassociated with early endosomal staining (Figure 4 a), suggesting a non-endocytic mechanism of entry (e.g. direct diffusion across the plasma membrane) in at least some cases; 2). In several instances, agglomerates of PS25C appeared to be enveloped in or closely associated with a similar-sized mass of actin just below the surface of the cell (Figure 4 b), suggesting a possible phagocytic mechanism of entry in some cases; 3). A large number of PS25C agglomerates of various sizes appeared to be free (not colocalized with Rab5a - early endosomes or actin) within the cytoplasm (Figure 4 c); 4). Large clusters of PS25C agglomerates were present within cell nuclei, and specifically within fairly well circumscribed voids within the nuclear material (Figure 4 d). 5). One PS25C agglomerate was also seen associated with a chromosome in a dividing (prometaphase or metaphase) cell (Figure 4 e); 6). Very large clusters of agglomerates were found accumulated at or below the basal membrane of the cell (Figure 4 f). A 3D rendering of a cropped sub volume (Figure 4 f and Movie S4) further clarifies some of these observed spatial relationships. Specifically, an actin-encased PS25C agglomerate and a large PS25 agglomerates within a nucleus can clearly be seen.

Figure 4. Uptake and subcellular localization of ingested fluorescent carboxylated fluorescent polystyrene MNPs in a small intestinal epithelium model.

Small intestinal digestas of carboxylated red fluorescent 25 nm (PS25C) polystyrene spheres (initial concentration 400 μg/mL, applied concentration 8.33 μg/mL) were applied to triculture small intestinal epithelium growing on transwell membranes (a-g) or on coverslips (h, i) and incubated for 3 hours. Triculture cells were either transduced by BacMam delivery system prior to exposure with a Rab5a-GFP fusion construct (a-g) or immunostained with and anti-Rab5a primary and Alexa Fluor 488-conjugated secondary antibody (h, i) to provide green fluorescent labeling of early endosomes (green). Actin was labeled with CruzFluor 750-conjugated phalloidin (pseudocolored white), and cells were counterstained with Hoechst nuclear dye (blue). a. Orthogonal 3 plane views of a confocal image volume of the small intestinal epithelium showing a small PS25C agglomerate entering the cell. b. Orthogonal 3 plane views showing a large PS25C agglomerate surrounded by actin directly below the apical surface of the cell. c. Orthogonal 3 plane views showing large PS25C agglomerates apparently free within the cytoplasm of the cell. d. Orthogonal 3 plane views showing a large cluster of PS25C agglomerates within a void inside of a cell nucleus. e. Orthogonal 3 plane views showing a small PS25C agglomerate associated with a chromosome in a dividing (prometaphase or metaphase) cell. f. Orthogonal 3 plane views of confocal image volume showing multiple PS25C agglomerates at or below the basal membrane of the cell. g. 3D rendering of cropped image volume showing intranuclear PS25C (see d) and PS25C surrounded by actin (see c). h. Orthogonal 3 plane views of confocal image volume (from triculture on coverslip) showing large numbers of PS25C agglomerates apparently free within the cytoplasm and directly adjacent to and within cell nuclei. i. 3D rendering of cropped image volume (from triculture on coverslip) showing large numbers of PS25C agglomerates free within the cytoplasm and adjacent to and within nuclei.

The same study was also repeated with tricultures grown on glass coverslips to eliminate potential artifacts due to abnormalities in cell geometry resulting from the tendency of the cells to partially migrate into the ~3 μm transwell membrane pores (seen in Figure 4 f as several circular objects containing nuclear material within an actin shell). In addition, an alternative Rab5a immunostaining method was used to label early endosomes to confirm the absence of PS25C-early endosome interactions observed with the fluorescent Rab5a-GFP fusion protein transduced cells described above. Analysis of these image data confirmed the absence of association of PS25C with the endosomal compartment –no instance of colocalization was seen, as well as the accumulation of large numbers of agglomerates within cells and immediately adjacent to as well as within nuclei (Figure 4 h and i).

Role of dynamin-dependent endocytosis in uptake and translocation of carboxyl-modified 25 nm polystyrene spheres in a triculture small intestinal epithelium model.

To assess the possible role of dynamin-dependent endocytic pathways in the uptake of PS25C particles, digestas of PS25C were applied to triculture transwells that were either untreated or pre-treated with the dynamin inhibitor Dyngo. PS25C was quantified in the apical, cell/mucus, and basolateral compartments after a 3 h incubation. The percentages of the applied PS25C mass in the apical, cell/mucus, and basolateral compartments after 3 h are shown in Figure 5 a-c. The percentage of PS25C remaining free in the apical compartment was significantly decreased, from 78 to 66% (p<0.01), while the percentage in the cell + mucus compartment was significantly increased from 19 to 30% (p<0.01), and the percentage translocated to the basolateral compartment appeared to decrease from 2.5 to 0.5%, but due to high variability in the controls this difference was not statistically significant. These results suggest that inhibition of dynamin may increase either uptake of PS25C into cells or entrapment of PS25C in the mucus layer, while paradoxically decreasing translocation into the basolateral compartment. To visually characterize and compare the uptake and intracellular localization of PS25C with and without dynamin inhibition, confocal image volumes were acquired from exposed tricultures on cover slips with or without Dyngo pretreatment and exposed to PS25C digesta for 3 h. Orthogonal 3 plane views of the acquired image volumes (Figure 5 d-e) show large numbers of PS25C agglomerates within cells of both treated and untreated samples, with perhaps fewer agglomerates in the Dyngo-treated sample, suggesting that the reduced apical fraction and increased cell + mucus fraction of PS25C seen in Dyngo-treated specimens was primarily due to entrapment of the material within the mucus. However, it was not possible to visualize the mucus layer to confirm such entrapment in these samples. In future studies it may be useful to explore methods to remove the mucus layer for separate analysis.

Figure 5. Role of dynamin-dependent endocytosis in uptake and translocation of carboxyl-modified 25 nm polystyrene spheres in a triculture small intestinal epithelium model.

The small intestinal digesta of carboxylated red fluorescent 25 nm (PS25C) polystyrene spheres (initial 400 μg/mL) was applied to triculture small intestinal epithelium growing on coverslips, either untreated or pretreated with the dynamin inhibitor Dyngo, and incubated for 3 hours. Triculture cells were immunostained with anti-Rab5a antibody to provide green fluorescent labeling of early endosomes (green). Actin was labeled with CruzFluor 750-conjugated phalloidin (pseudocolored white in these images), and cells were counterstained with Hoechst nuclear dye (blue). a, b, c. Percent of applied PS25C in apical, lysate (cells + mucus), and basolateral compartments after 3-hour incubation. d, e. Orthogonal 3 plane views of a confocal image volumes of the small intestinal epithelium with or without Dyngo pretreatment showing PS25C agglomerates within cells. ** p < 0.01.

Discussion and Conclusions

Toxicological assessment of an environmentally relevant panel of ingested MNPs in an in vitro triculture small intestinal epithelium model revealed significant reductions in viability (metabolic activity) accompanied by significant increases in epithelial permeability (to 3 kD dextran) after 24 h exposures to digestas of carboxyl modified PS spheres of different sizes (Figure 2). There was no significant difference between viability at the two doses. The strongest effects on epithelial permeability (to 3 kD dextran) were seen at the lower MNP dose (0.4 mg/mL starting concentration), and differences in permeability between doses were not statistically significant. The effects of carboxylated PS spheres on permeability appeared to increase with increasing primary particle size. This may be a result of differences in mechanisms or locations of injury for different sizes. Moderate reductions in viability and increases in permeability were also consistently seen with unmodified polystyrene MNPs of both sizes tested (25 and 1000 nm). Although these differences were not statistically significant, their consistency across the PS material treatments suggests that they may be real. Nevertheless, the carboxylated PS forms clearly caused more substantial and significant toxicity. It should be noted that the common routes of UV-radiation and oxidative degradation for the major plastic polymers results in the formation of oxygen-bearing moieties, particularly carboxylic acid groups.³⁷ Carboxylated PS may therefore be a more relevant material for study than unmodified virgin PS.

As expected, uptake and translocation of MNPs (carboxyl modified PS) was size dependent, with the greatest translocation observed for 25 nm carboxylated PS (PS25C) and the smallest observed for 1000 nm carboxylated PS (PS1KC) (Figure 3 and Movies S1, S2, and S3). Confocal microscopy revealed size-dependent uptake, with only 25 nm carboxylated PS within epithelial cells, and the two larger sizes restricted to the apical surface, apparently within the mucus layer. In higher resolution multiplexed confocal studies with actin and early endosome labeling (Figure 4), the 25 nm carboxylated PS (PS25C) did not appear to colocalize with early endosomes and could be seen in some instances apparently entering the cell unassociated with endosome label, suggesting a direct diffusion entry route. On the other hand, inhibition of dynamin, a GTPase that acts as a molecular scissors to release forming endosomes from the cell membrane, caused a notable decrease in translocation of PS25C (Figure 5 a-c), suggesting some form of dynamin-dependent endocytosis, such as clathrin-coated pit-mediated endocytosis (CME) or fast endophilin-mediated endocytosis (FEME) ³⁸, as another potential mechanism, despite the absence of observed colocalization between fluorescent MNPs and early endosome label. Agglomerates of PS25C were also observed within what appeared to be actin cages or shells at the apical aspect of the cells, suggesting phagocytosis as a third possible route of entry. Additional studies are needed to determine the relative contributions of the various potential mechanisms of MNP uptake.

Numerous fluorescent PS25C agglomerates were seen throughout the stained epithelial cells, apparently free from early endosome or actin structures, including many large agglomerates within and adjacent to cell nuclei (Figures 4 and 5, and Movie S4). The ability of these MNPs to apparently freely enter nuclei raises a number of concerns and points to the need for more study of these materials. These should include assessment of genotoxicity, effects on transcription, cell division, and interactions with and modulation of nuclear receptor signaling. In addition to potential direct toxicity caused by intracellular interactions with MNPs themselves, the ability of these MNPs to access the cytoplasmic, nuclear and other intracellular compartments could result in delivery and high localized intracellular concentrations of environmental pollutants sorbed to their surfaces of MNPs during their incubation in environmental media, or of plasticizers, fillers, dyes and other plastics additives leeching from the MNPs.

The significantly decreased viability and increased permeability observed for many of the MNPs tested, and the apparent ease with which carboxylated polystyrene nanospheres (PS25C) were observed to enter and move through and throughout triculture enterocytes and/or M-cells (perhaps by multiple mechanisms), combined with findings of previous studies demonstrating intestinal uptake and wide biodistribution of ingested PS MNPs in rodents ^15–19, as well the ability of nanoscale MNPs to cross the blood-brain¹⁴ and blood-placenta²² barriers, suggest that contamination of food and water by MNPs is a serious emerging public health concern.

The present study addresses some of the most important questions of general cytotoxicity and uptake for a selected panel of MNPs; but much more work is clearly needed to understand the full scope of potential hazards ingested MNPs may represent. In addition to a need for a comprehensive understanding of MNP toxicity and uptake in the intestine, many other critical knowledge gaps remain, including (1) the detailed biodistribution of ingested MNPs, particularly MNP delivery via the circulation and accumulation in critical or vulnerable tissues (e.g., brain, kidney, bone marrow); (2) the potential of MNPs to modulate nutrient bioaccessibility and/or bioavailability (as we have observed with natural biopolymers like nanocellulose ³⁹); (3) the potential for MNPs to sorb and act as delivery vehicles for toxic environmental pollutants (EPs), or to impair intestinal barrier function (as our results here suggest) thereby increasing the bioavailability of EPs and all other ingested chemicals; (4) the potential effects of MNPs on the gut microbiome; and (5) the effects of prolonged exposure to MNPs.

In addition to filling these and other knowledge gaps vis-à-vis potential biological hazards of MNPs, further environmental studies and new methodologies are needed to accurately determine the true extent and distribution of MNP contamination across the globe, including quantification of nanoscale plastic particles which have not typically been measured in previous studies due to technical limitations. Dynamic predictive environmental models are also needed to forecast the future course of MNP pollution based on plastic use and waste stream data and the kinetics of MNP generation from plastic waste materials over time under different environmental conditions.

All of these knowledge gaps must be addressed to provide data needed by risk assessors and regulators to evaluate the potential hazards of MNPs and develop appropriate regulations for food and water contamination, as well as to provide lawmakers with the clear and accurate information required to recognize the urgent need for policy measures aimed at curbing plastic production and use, and promoting safer alternative materials.

Supplementary Material

1

Figure S1. Schematic of incineration platform used to produce PEI

Figure S2. Size distributions of digestas with and without MNPs

Figure S3. Permeability coefficients for cells treated with MNPs

3

Movie S1. 3D 360 degree rotation: PS2C-Triculture SIE interaction

5

Movie S2. 3D 360 degree rotation: PS100C-Triculture SIE interaction

7

Movie S3. 3D 360 degree rotation: PS1KC-Triculture SIE interaction

9

Movie S4. 3D 360 degree rotation: small intestinal epithelium interaction at high resolution with phalloidin actin labeling and CellLight early endosome labeling (Rab5A-GFP fusion protein expression).

• Carboxylated polystyrene micro-nanospheres impaired viability and barrier function

• Carboxylated polystyrene nanospheres readily entered and crossed the eipthelium

• Nanospheres were localized throughout the cytoplasm and within cell nuclei

• Dynamin inhibition reduced translocation, suggesting some uptake by endocytosis

• Some nanospheres were seen in actin shells, suggesting entry by phagocytosis

Synopsis.

A study of ingested micro-nanoplastics (MNPs) in an in vitro small intestinal epithelium model revealed significant direct toxicity and the ability of nanoscale MNPs to enter and move through and throughout in vitro enterocytes, including their nuclei, suggesting potential serious health implications from ingestion exposures to MNPs—which are accumulating in natural environments as a result of the enormous (and exponentially increasing) amounts of plastics produced and used in the last century—highlighting the urgent need for further studies and for policy measures aimed at curbing production and use of petroleum-based plastics, and promoting safer alternative materials.