Biological effects of polystyrene micro- and nano- plastics on human intestinal organoid-derived epithelial tissue models without and with M cells

Abstract

Micro- and nano- plastics (MPs and NPs) released from plastics in the environment can enter the food chain and target the human intestine. However, knowledge about the effects of these particles on the human intestine is still limited due to the lack of relevant human intestinal models to validate data obtained from animal studies or tissue models employing cancer cells. In this study, human intestinal organoids were used to develop epithelia to mimic the cell complexity and functions of native tissue. Microfold cells (M cells) were induced to distinguish their role when exposure to MPs and NPs. During the exposure, the M cells acted as sensers, capturers and transporters of larger sized particles. The epithelial cells internalized the particles in a size-, concentration-, and time-dependent manner. Importantly, high concentrations of particles significantly triggered the secretion of a panel of inflammatory cytokines linked to human inflammatory bowel disease (IBD).

Graphical Abstract

Human intestinal organoids isolated from patients were used to establish intestinal epithelia recapitulating the cell complexity and functions of native tissue. The epithelia were exposure to different sizes and concentrations of polystyrene micro- and nano- plastics (MPs and NPs) for different exposure periods to understand the uptake and potential cytotoxicity of the plastic particles. Particle tracking were conducted with confocal microscopy, while the intestinal barrier integrity and inflammatory response to the particles were assessed by transepithelial electrical resistance assay and cytokine profiling. The obtained results contribute to the current knowledge of the effects of MPs and NPs on human health.

1. Introduction

Plastics are some of the most ubiquitous manmade materials and they have pervaded nearly every aspect of human life. It is estimated that 8.3 billion metric tons of virgin plastic has been generated as of 2017, and this number is expected to more than double by 2050.¹ Most plastics can break down into very small particles. Particles between 1 and 1,000 μm in diameter are considered microplastics (MPs); while plastic particles between 1 and 100 nm (or up to 500 nm in some instances) are known as nanoplastics (NPs).² MPs/NPs have been consistently detected in seafood, various convenience foods and beverages, and in tap/bottle water.³ MPs/NPs are also distributed in and accumulate within living organisms in different tissues, including the digestive tract, blood, liver, pancreas, heart and, notably, the brain.⁴ In 2018, microplastics were first reported in human feces, providing evidence of accumulation, ingestion and excretion of these particles in humans and indicating exposure through processed or packaged food/water consumption.⁵ As human exposure to MPs and NPs continues to grow, MPs and NPs have been recognized as emerging contaminants within the last decade.

Ingested MPs/NPs first encounter the intestinal epithelium, the innermost layer of the mucosa, where most digestive, absorptive and secretory processes occur. This makes the intestine the primary exposure site for these particles and particularly important in terms of assessments of the impact of plastics on physiological functions.³ Investigations of the impact of MPs/NPs on organism health after oral uptake have been widely conducted in mammalian (rodent) models and reviewed.^6–9 In most of the cases, the effects in the gut correlated with size and morphology of the plastic particles, exposure concentration, exposure time, and tissue uptake and accumulation. However, there have also been studies that failed to demonstrate that MPs/NPs had a significant effect on animals. For example, a rat model showed a very low number of MPs taken up by intestinal tissue and the authors did not detect any histological lesions or significant inflammatory responses in the gut of the animals.¹⁰ Another study using rats to analyze potential neurobehavioral effects of PS NPs did not identify statistically significant behavioral changes or abnormalities.¹¹ These results represent an important clue to the accumulation and release of MPs/NPs following exposure to mammals, with remaining controversy regarding biological impact.

Despite the many studies on animals, toxicological evaluations of MPs/NPs on the human intestine in vivo are still lacking. As mentioned above, the intestine is considered the major route for particle exposure. The makeup of the native intestinal epithelium is diverse and populated with multiple differentiated epithelial cells, including enterocytes (absorption), goblet cells (mucus secretion), enteroendocrine cells (peptide and hormone production) (EECs), Paneth cells (antimicrobial protein secretion), and microfold cells (M cells, immune sensing and uptake of particulate microbial antigen).¹² In addition to its primary function of nutrient digestion and absorption, the intestinal epithelium also provides a protective barrier via apical intercellular junctions which protect against physical, chemical, and biological damages.¹³ Thus far, studies exploring the interactions between MPs/NPs and human intestinal cells have been mainly performed with monocultures of human colorectal adenocarcinoma epithelial cell-lines, Caco-2 (enterocyte-like cells) or HT29-MTX (goblet-like cells).^{14, 15} To achieve slightly more physiological conditions, many researchers have employed a Caco-2/HT29-MTX coculture system as an in vitro experimental model.¹⁶ To further improve upon this system, some researchers also include lymphoblast-like Raji-B cells to induce Caco-2 cells into microfold-like cells.¹⁷ Most of these studies, while revealing some level of cellular uptake and epithelial transport of MPs/NPs, showed either low or insignificant cytotoxic effects, except at very high concentrations of particles. Overall, these prior studies have not identified severe acute cytotoxic effects but have demonstrated the potential for low to moderate damage of the cells and tissues depending on the size and concentration of MPs/NPs. The degree of cellular uptake of the particles was somewhat consistent with the observations in animals. Of note, these in vitro findings were all from short-term studies (up to 48 hours of exposure) with MPs/NPs exposed to human intestinal epithelial cell lines, whereas it can take up to 14 days for microplastics to pass through marine animals (compared to a normal digestion period of 2 days).¹⁸ Moreover, the intestinal epithelial cell cultures used for cytotoxicity studies were based on cancerous cell lines and were missing other important intestinal epithelial cell types. Consequently, these studies do not fully capture the cellular complexity of the native human intestine, thus, fall short of biological processes in vivo. Therefore, knowledge gaps still exist on the mechanisms and predictions regarding acute and chronic effects caused by these plastics in humans.

Thanks to advances in stem cell biology, organoids have been established as new model system for biomedical research and have been extensively reviewed.¹⁹ Organoids are in vitro miniaturized organ models originating from self-organizing stem cells. Compared to conventional cancerous cell lines, organoids are capable of physiologically mimicking the in vivo structure and function of an organ. They can be produced from adult (murine and human) stem cell-containing tissue samples, organ-specific single adult stem cells (ASCs), embryonic stem cells (ESCs), or induced pluripotent stem cells (iPSCs) and expanded indefinitely with the support of a cocktail of various growth factors in their culture media. Organoids can be grown to model many human organs, including the stomach, intestine, kidneys, heart, pancreas, brain, and liver. For the human intestine, organoids can be generated from fresh or frozen patient biopsies of intestinal tissue specimens by isolating crypts, the intestinal stem cell niche.²⁰ The isolation procedure yields spheroid-shaped and budding organoids which become ever-expanding in a 3D Matrigel in the presence of Wingless-related integration site (Wnt) pathway agonist, R-spondin (Rspo), epidermal growth factor (EGF), and the BMP inhibitor, Noggin.²¹ As these organoids contain stem cells, they have the ability to differentiate toward different intestinal epithelial cell lineages. For specific (and rare) cell types, the differentiation protocols for intestinal organoids have been updated and standardized to direct stem cell maintenance or differentiation toward specific cell fates. For example, by removing the Wnt, Noggin, and R-spondin from the culture media, the organoids simultaneously differentiate into a continuous intestinal epithelium containing enterocytes, goblet cells, EECs, and Paneth cells with an abundance that is similar to what is found in vivo.^{22, 23} By treating the organoids with nucleic factor-kappa B ligand (RANKL), a common M cell inducer, increased M-cell differentiation can be achieved in the cultures.²⁴ Over the past decade, intestinal organoids have been derived from healthy and diseased intestinal tissues and can be used as an alternative to Caco-2 drug absorption and permeability assays for drug screening and for personalized medical decisions to better predict patients’ treatment response and, importantly, any potential drug toxicity.²⁵

We have previously used human intestinal organoids from both large and small intestine for intestinal tissue engineering,²⁶ inflammatory bowel disease modeling,²⁷ and host cell-microorganism interactions.²² In the present study, we explored the application of these organoids as a platform for predicting potential risks of MP and NP exposure. Specifically, the goals of this work were: 1) to establish and characterize two organoid-derived intestinal epithelial monolayer models: monolayers without M cells (enterocytes, goblet cells, EECs, and Paneth cells) and monolayers with M cells (enterocytes, goblet cells, EECs, Paneth cells, and M cells); 2) to compare the dynamics of cellular uptake and translocation of different sizes of MPs and NPs over time in these two different tissue models; 3) to evaluate the potential dose-, size- and exposure period-dependent cytotoxic effects of the plastic particles on the intestinal barrier integrity and immune response in the tissue models; 4) to distinguish the role of M cells regarding the transport of plastic particles across the intestinal epithelium and how M cells shape specific immune responses to these particles. In this work, PS MPs and NPs were selected as the model particles for the following reasons: 1) they are commercially available with a wide range of specific sizes; 2) they are fluorescently labeled, allowing for localization and tracking in cell compartments; and 3) they are biologically stable in various cell culture media.

3. Results

3.1. Characterization of PS MPs and NPs

PS MPs and NPs in different sizes (1 μm, 500 nm, 100 nm, and 30 nm) and fluorescently labeled are commercially available from Sigma Millipore. All the plastic particles diluted in DMEM were detected using fluorescence microscopy and SEM to validate the manufacturing sizes and the fluorescence. Microscopy images showed significant fluorescence signals from different sizes of particles confirming the fluorescence labels on the particles (Figure 1). In addition, a uniform dispersion of particles without significant aggregation was detected in DMEM (Figures 1A–D). SEM images further confirmed that all particles (Figures 1E–H) had a spherical shape. To statistically evaluate particle size, Image J and SPSS were used to analyze the means, standard deviations, and ranges of particle sizes in the SEM images (Table 1). Mean particle sizes were equivalent to labeled specificities (Table 1, Column 3). The particle sizes ranged from 28.56 to 31.82 nm for 30 nm NPs, from 97.13 to 104.25 nm for 100 nm, from 494.92 to 503.89 nm for 500 nm, and from 0.870 to 1.25 μm for the 1 μm MPs, respectively (Table 1, Column 5).

Figure 1.

Characterization of polystyrene particles. (A-D) Representative fluorescence microscope images confirmed the fluorescence of MPs and NPs, respectively, (A)1 μm MPs, scale bar: 5 μm, (B) 500 nm, scale bar: 5 μm, (C) 100 nm, scale bar: 1 μm and (D) 30 nm, scale bar: 200 nm. (E-H) SEM images of 1 μm MPs (E, scale bar: 3 μm), 500 nm MPs (F, scale bar: 1 μm), 100 nm (G, scale bar: 500 nm) and 30 nm (H, scale bar: 200 nm).

Table 1.

Characterization and validation of the parameters of polystyrene MPs and NPs by SEM image analysis

PS MPs and NPs	N	Mean particle size	Standard deviation
30 nm	25	30.06 nm	±0.63 nm
100 nm	50	99.72 nm	±1.61 nm
500 nm	25	500.06 nm	±2.65 nm
1 μm	100	1.02 μm	±0.08 μm

3.2. Establishment and characterization of human intestinal organoid derived monolayers with and without induced M cells

Human intestinal organoids (Figure 2A–B) derived from ileum biopsy specimens were subcultured for monolayer generation. For monolayer preparation, organoids were enzymatically dissociated into single cells and seeded onto transwell inserts. Monolayers reached confluence 1–2 days post seeding (Figure 2C). After that, the monolayers were either treated with Standard Differentiation Medium (SDM) or M-cell Differentiation Medium (MDM) to establish intestinal epithelia both with and without M cells (Figure 2A). SDM and MDM have been widely utilized to generate well-differentiated epithelia derived from intestinal organoids.^{24, 26, 28} Organoid-derived epithelial monolayers treated with SDM for 3 days simultaneously differentiated into enterocytes, goblet cells, Paneth cells, enteroendocrine cells, while monolayers supplemented with MDM basolaterally for 6 days differentiated into enterocytes, goblet cells, Paneth cells, enteroendocrine cells, and M cells.

Figure 2.

Establishment and characterization of SDM- and MDM differentiated monolayer cell cultures from intestinal organoids. (A-C) Organoid culture and the monolayer formation. Organoid (A1, B) were expanded, enzymatically digested into singlets and doublets (A2) and seeded on transwell inserts to form confluent monolayer (A3, C) for differentiation without M cells (A4) and with M cells (A5). (D-H) Immunofluorescence of intestinal epithelial cell differentiation markers in organoid-derived cells stimulated to differentiate with SDM or MDM: staining of SI (D, green), a marker of enterocytes; expression of MUC2 (E, green) a marker of goblet cells; expression of Lysozyme (F, green), a Paneth cell marker; staining of ChgA (G, green), an EEC marker; expression of tight junction marker, ZO-1 (H, red), and staining of M cell surface marker GP2 (I, red) with tight junction ZO-1 (I, green). Nuclei are counterstained with DAPI (blue). Scale bars: 50 μm. (J-K) SEM images showed “bald” M cells as labeled and polarized microvilli of epithelial cells. Scale bars: (J) 10 μm, (K) 2 μm. (L-M) TEER measurements on SDM-treated (L) and MDM-treated monolayer cultures derived from patient intestinal organoids. (*p < 0.05, **p < 0.005, ***p<0.001).

To confirm that these different major cell types were present in the differentiated monolayers, we collected and stained the SDM-differentiated monolayers at day 3 and day 6 post-differentiation to confirm the presence of relevant marker proteins (Figure 2D–I). As expected, SDM-differentiated cell cultures expressed sucrase-isomaltase (SI, Figure 2D) signal in the apical region of enterocytes throughout the monolayers, the mucin 2 (MUC2, Figure 2E) staining in a scattering of goblet cells, the Lysozyme expression (Figure 2F) by Paneth cells, and the induction chromogranin A (ChgA, Figure 2G) positive EECs. Furthermore, immunostaining also revealed the typical “chicken wire” staining pattern of ZO-1 tight junction protein (Figure 2H and 2I), suggesting the individual mature epithelial cells are joined to their neighbors to assemble the tight junction network in the epithelium for maintenance of epithelial barrier integrity. A similar staining pattern of the above-mentioned markers was observed in MDM-differentiated epithelial monolayers (data not shown). To determine if MDM successfully induced M cells along with the major epithelial cell types in the monolayer cultures, immunofluorescence using M cell specific cell surface marker, glycoprotein 2 (GP2), was performed on MDM-differentiated epithelial monolayers. Immunofluorescence analysis showed that a significant amount of the M cells developed in the monolayers 6 days after treatment with MDM, as shown by the immunostaining for GP2 (Figure 2I and Sup Figure 1A). Moreover, by SEM, the differentiated monolayers displayed an apical-basolateral polarization of epithelial cells indicated by the presence of tightly packed microvilli at the apical side with random distribution of “bald” M cells with disorganized stubby microvilli along their apical border (Figure 2J and 2K). Overall, these observations indicated that organoid-derived epithelial monolayers properly differentiated into the major intestinal epithelial cell types without or with M cells using SDM or MDM.

Differentiation of the organoid-derived monolayers was further evidenced by the significantly increased TEER values after switching to differentiation media in SDM- and MDM-differentiated epithelial monolayers (Figures 2L–M). Fully differentiated SDM-treated cultures possessed a TEER between 627 and 726 Ω*cm² (Figure 2L); while full differentiated MDM-treated cultures feature a TEER between 396–495 Ω*cm² (Figure 2M). TEER values are strong indicators of epithelial maturation and integrity of the cellular barrier²⁹.

3.3. Cellular binding, uptake and translocation of PS MPs and NPs in human intestinal epithelia

Next, we investigated the potential cytotoxic effects of PS MPs and NPs on human intestinal epithelial cells. Towards this goal, we first explored the interactions between the plastic particles and the gastrointestinal epithelium by studying the cellular binding, uptake and translocation of the particles by the intestinal epithelium. The two different intestinal organoid-derived epithelial models, as established in section 3.2, were employed: the SDM-differentiated monolayers (enterocytes, goblet cells, Paneth cells, EECs) and MDM-differentiated monolayers (enterocytes, goblet cells, Paneth cells, EECs, and M cells). After the cell seeding and differentiation with SDM or MDM (Figure 3A), different sizes of spherical fluorescent PS particles (1 μm, 500 nm, 100 nm, 30 nm) were applied on the apical side of the intestinal epithelia at a concentration of 500 μg/mL to explore particle uptake and transport. We started dosing the monolayers at 3 days post differentiation in SDM and 6 days post differentiation in MDM, as TEER values peaked at about 3 and 6 days after the differentiation was induced in SDM and MDM systems, respectively, and stabilized after that (Figures 2L–M). TEER value is a strong indicator of epithelial cell barrier integrity before evaluation for transport of particles. Samples from two different culture models were harvested 24, 48, 72, 96 hours post dosing (equals to days 4, 5, 6, 7 in the SDM culture, and days 7, 8, 9, 10 in the MDM culture, respectively) and prepared for confocal microcopy and SEM to detect the binding and trace the location of the MPs and NPs in the monolayer cultures. At 24 hours, all MPs and NPs bound to the intestinal epithelial cell surface and M cells (Figures 3B–E and 3J–M), which is a prerequisite for cellular uptake (Figure 3B–E and 3J–M). Exposure to the plastic particles did not significantly affect tight junction formation demonstrated by ZO-1 expression. After 48 hours, in SDM-differentiated monolayers, particle sizes ≤ 500 nm were imaged “nestling” or being enveloped by surrounding microvilli which may suggest these particles tend to cross the apical microvilli brush border of the epithelial cells (Fig. 3G–I), whereas in the MDM-differentiated monolayers, it was found that larger sized particles (1μm, 500 nm) tended to be more specifically aggregated on and bound to the surface of M cells (Figures 3N and 3Q, and sup Figures 1B–C).

Figure 3.

The cellular localization of polystyrene particles. (A) Schematic of the experimental timeline of cell seeding, monolayer differentiation and particle dosing. (B-I) The cellular localization of MPs and NPs in the SDM-treated epithelium after 24 hours of exposure were visualized by confocal fluorescence microscopy (B-E) and SEM (F-I). (B-E) Representative confocal fluorescence images of SDM-treated monolayers: intestinal epithelium with integral barrier was identified by staining of ZO-1 antibody (green) and particles were seen in red (indicated by yellow arrows). Nuclei were counterstained with DAPI (blue). Scale bars: 10 μm. (F-I) SEM images confirmed the cellular binding of different sized particles to the epithelial microvilli (yellow arrows). Scale bars: (F) 10 μm, (G) 1 μm, (H) 500 nm, (I) 200 nm. (J-Q) The cellular localization of MPs and NPs in the MDM-treated epithelium after 24 hours of exposure were visualized by confocal fluorescence microscopy (J-M) and SEM (N-Q). (J-M) Confocal fluorescence images of MDM-treated monolayers: the fluorescence colors of cells and particles were the same as abovementioned in the SDM-treated cells. M cells were stained with GP2 antibody (magenta). Co-localization of 1 μm and 500 nm with M cells is indicated by yellow arrows. Scale bars: 10 μm. (N-Q) Zoom-in SEM images of M cell surface confirmed the cellular binding and aggregation of different sized particles to the M cell surface with sparse disorganized stubby microvilli (yellow arrows). Scale bars: (N, O) 2 μm, (P) 1 μm, (Q) 500 nm.

The intracellular uptake and translocation of MPs and NPs in the culture systems were tracked for up to 96 hours by confocal laser scanning microscopy (Figures 4, 5 and 6). The three-dimensional image reconstruction of x-z cross-section from confocal sequential acquisition of the fluorescence signals from different planes enabled particle tracing at varying cellular depths. In the SDM-differentiated monolayers, over the course of 96 hours, the 1 μm MP group were mainly observed on the apical side (Figure 4A, yellow arrows) of the intestinal epithelium, suggesting they did not cross the epithelia barriers. For the 500 nm MP group, at 96 hours post dosing, most of the particle fluorescence signals were still detected on the apical surface of the intestinal epithelium with occasional translocation across the epithelium, indicating no obvious particle translocation (Figure 4B, yellow arrows). For the 100 nm NP group, at 96 hours, particle signals started to appear inside some of the epithelial cells and on the basal side of the epithelium (Figure 4C, yellow arrows), suggesting cellular uptake and penetration of NPs. In the 30 nm NP group, at 96 hours, large numbers of particles were found on the basal side of the epithelium (Figure 4D, yellow arrows), suggesting that after cellular binding, the 30 nm NPs were taken up and translocated from the apical to the basal side of the epithelium. The particle tracking results demonstrated an MP/NP size-dependent uptake and translocation, with the highest total for PS uptake for the smallest NPs (30 nm). In the MDM-differentiated monolayers containing M cells, similar intracellular uptake and translocation patterns of 100nm and 30nm NPs were identified in the epithelium through 48 hours (data not shown). However, at 72 hours, obvious translocation of 100nm and 30nm NPs to the basal side of the epithelia was observed, followed by the degradation of epithelial tight junctions (ZO-1) (100 nm) and the lifting of monolayers from the transwell inserts (30 nm) at 96 hours post particle exposure (sup Figure 2). Furthermore, 1μm and 500 nm particles preferentially internalized and accumulated in M cells as shown in Figures 5 and 6 (yellow arrows).

Figure 4.

The intracellular translocation of MPs and NPs in the SDM-differentiated culture systems were tracked for up to 96 hours by confocal laser scanning microscopy with 3D z-stack reconstruction. 3D z-stack reconstruction of the monolayers after exposure to 1 μm MPs (A), 500 nm MPs (B), 100 nm (C) and 30 nm (D) for 96 hours. Upper and lower panels of each image show the orthogonal cross-sections and the top view of the monolayers, respectively. Intestinal epithelium with integral barrier was identified by staining of ZO-1 antibody (green in A-C, red in D) and 1 μm, 500 nm and 100 nm particles were seen in red, while 30 nm particles were seen in green (indicated by yellow arrows). Nuclei were counterstained with DAPI (blue).

Figure 5.

The intracellular translocation of MPs and NPs in the MDM-differentiated culture systems were tracked for up to 96 hours by confocal laser scanning microscopy with 3D z-stack reconstruction. (A) 3D z-stack reconstruction of the monolayers after exposure to 1 μm MPs for 72 hours. (B) XZ cross-sectional profiles of the monolayers after exposure to 1 μm MPs for 72 hours. (C) z-stack reconstruction of the monolayers after exposure to 1 μm MPs for 96 hours. (D) XZ cross-sectional profiles of the monolayers after exposure to 1 μm MPs for 96 hours. In all the images, intestinal epithelium with integral barrier was identified by staining of ZO-1 antibody (green), M cell was stained by GP2 antibody (magenta), and 1 μm particles were seen in red, indicated by yellow arrows. Scale bars: 20 μm.

Figure 6.

The intracellular translocation of MPs and NPs in the MDM-differentiated culture systems were tracked for up to 96 hours by confocal laser scanning microscopy with 3D z-stack reconstruction. (A) 3D z-stack reconstruction of the monolayers after exposure to 50 nm MPs for 72 hours. (B) XZ cross-sectional profiles of the monolayers after exposure to 50nm MPs for 72 hours. (C) z-stack reconstruction of the monolayers after exposure to 50 nm MPs for 96 hours. (D) XZ cross-sectional profiles of the monolayers after exposure to 50 nm MPs for 96 hours. In all the images, intestinal epithelium with integral barrier was identified by staining of ZO-1 antibody (green), M cell was stained by GP2 antibody (magenta), and 50nm particles were seen in red, indicated by yellow arrows. Scale bars: 20 μm.

3.4. Cytotoxicity effect of PS MPs and NPs in human intestinal epithelia

Knowing that different sizes PS MPs and NPs can interact with and cross the human intestinal epithelia, we next assessed their cytotoxic effects by using different concentrations and exposure periods in the human intestinal epithelia. In this study, the toxicity assessment was conducted using two different assays: TEER- to examine epithelial integrity and ELISA- to measure cytokine secretion (inflammatory responses). For cytokine profiling, the release of proinflammatory cytokines interleukin-6 (IL-6), IL-8, tumor necrosis factor-alpha (TNF-α) and an anti-inflammatory cytokine TGF-β1 were analyzed. The two different intestinal organoid-derived epithelial models (with and without induced M cells) were compared. The cytotoxic response of cells treated with MPs and NPs (1 μm, 500 nm, 100 nm, and 30 nm) for 24, 48, 72 and 96 hours was investigated using particle concentrations ranging from 250–1000 μg/mL (250, 500, 1000 μg/mL).

3.4.1. Barrier integrity and cytokine secretion in particle-exposed epithelia without induced M cells

The TEER values of SDM-differentiated monolayers treated with 250, 500, 1000 μg/mL of 1 μm, 500 nm, 100 nm particles remained relatively constant and equivalent to the control over the duration of the exposure (Figures 7A–C), suggesting the particles did not substantially and adversely affect the intestinal barrier integrity. For the 30nm group (Figure 7D), cellular integrity was not altered by 250 and 500 μg/mL particles until 48 hours (day 5 in culture) post dosing. At 72 hours, 30 nm NPs induced a slight decline in TEER values at 1000 μg/mL.

Figure 7.

(A-D) TEER measurements of SDM-treated organoid derived epithelial barriers with exposure to 1 μm MPs (A), 500 nm MPs (B), 100 nm NPs (C), and 30 nm NPs (D) at concentrations of 1,000μg/mL, 500μg/mL and 250μL for up to 4 days (96 hours) post exposure. (E-P) Cytokine profiling of the SDM-treated cultures following exposure to 250 μg/mL, 500 μg/mL and 1000 μg/mL were conducted 24, 48, 72 and 96 hours. Inflammatory cytokines including TNF-α (E, I, M), TGF-β1 (F, J, N), IL-6 (G, K, O), and IL-8 (H, L, P) were measured and analyzed. The level of cytokines in the cell supernatant were determined using ELISA. (*p < 0.05, **p < 0.005, ***p < 0.001).

The levels of cytokines (TNF-α, TGF-β1, IL-6, IL-8) in all exposure groups of MPs and NPs (Figures 7E–H) were determined and no significant increase in cytokine secretions were induced from any of the particle groups at 250 μg/mL compared to the control groups. However, at 500 and 1000 μg/mL, secretion of TNFα, IL-6 and IL-8 was significantly increased in the 100 nm and 30 nm NPs after 96 or 72 hours of exposure (**p < 0.01, Figures 7I, K and L). By extrapolation, these results demonstrated that the smaller sizes (100 nm and 30 nm) of the PS NPs at concentrations greater than 500 μg/mL might induce the release of pro-inflammatory cytokines (TNF-α, IL-6 and IL-8) in the human intestinal epithelial cells.

3.4.2. Barrier integrity and cytokine secretion in particle-exposed epithelia with M cells

According to the TEER analysis, the epithelial integrity of MDM-differentiated monolayers was not significantly impacted by 1 μm and 500 nm particles at any concentration (Figures 8A–B) but was impacted by 100 nm and 30 nm PS particles at 250 and 500 μg/mL over time (Figures 8C–D). However, compared to the control group, a significant decrease (*p<0.05) in TEER value was observed in monolayers that were exposed to 100 nm and 30 nm NP groups at 1000μg/mL for 48 hours (Figures 8C–D), indicating impaired epithelial barrier function was caused by the exposure to 1000 μg/mL 100 nm and 30 nm particles.

Figure 8.

(A-D) TEER measurements of MDM-treated organoid derived epithelial barriers with exposure to 1 μm MPs (A), 500 nm MPs (B), 100 nm NPs (C), and 30 nm NPs (D) at concentrations of 1,000μg/mL, 500μg/mL and 250μL for up to 4 days (96 hours) post exposure. (E-P) Cytokine profiling of the MDM-treated cultures following exposure to 250 μg/mL, 500 μg/mL and 1000 μg/mL were conducted 24, 48, 72 and 96 hours. Inflammatory cytokines including TNF-α (E, I, M), TGF-β1 (F, J, N), IL-6 (G, K, O), and IL-8 (H, L, P) were measured and analyzed. The level of cytokines in the cell supernatant were determined using ELISA. (*p < 0.05, **p < 0.005, ***p < 0.001).

For cytokine production, compared to the control groups, no statistically significant changes were detected for all particle sizes in 250 μg/mL group throughout the exposure course. Increased release of cytokines (TNF-α, IL-6, IL-8) in the 500 μg/mL group were detected and sustained in the 100 nm and 3 0 nm NPs across various time points (Figures 8I, K, and L) with no obvious changes in TGFβ1 (Figure 8J). Specifically, 500 μg/mL 100 nm NPs triggered a higher release of IL-6 at 72 hours post exposure (Figure 8K) and IL-8 at 48 hours (Figure 8L). Comparatively, the 30 nm NPs significantly triggered TNF-α secretion at 96 hours post exposure and IL-6 and IL-8 secretions as early as 48 hours (Figures 8K–L). In the 1000 μg/mL groups, acute releases of TNF-α and IL-6 by cells were found in all different sized MPs and NPs evidenced by the rapid and excessive production at 24 hours post dosing (Figure 8M and 8O). The acute release of IL-8 at 24 hours was observed in the 30 nm group (Figure 8P). Detectable levels of IL-8 production in 500 nm and 100 nm groups were not observed until 48 hours post exposure (Figure 8O and 8P). In these groups, the mean cytokine levels sustained or kept increasing from 24 to 48 hours but tended to decline after that. Notably, in 1000 μg/mL groups, the TGF-β1 concentrations were significantly elevated by 100 nm particles at 72 hours and by 30 nm particles at 48 hours (*p<0.05, Figure 8N). Monolayers detached from the Transwell inserts in 100 nm groups at 96 hours and 30 nm groups at 72 hours and cytokine data from these groups were not collected.

4. Discussion

Plastics are extensively used in our daily lives and are increasingly contributing to both land and water pollution. Over time, these plastics degrade and generate micro- or nano-sized plastic particles (MPs or NPs). In 2018, microplastics were first reported as present in human feces, providing evidence of accumulation, ingestion, and excretion of these particles in humans and indicating exposure through processed or packaged food/water consumption.⁵ Since then, plastics have been detected in human placental tissues³⁰ and blood³¹. Human exposure to MPs and NPs is growing and the human intestine is the principal exposure site. However, research connecting MPs/NPs with human intestinal cells and tissues is still preliminary. In addition to the MPs and NPs themselves, these particles can release plastic additives and/or adsorb other environmental chemicals, many of which have been shown to exhibit endocrine-disrupting and inflammatory effects.³² MPs and NPs have also been linked to intestinal diseases. However, related studies quantifying toxicity or cellular uptake in the human intestine are scarce and lacking in physiologically relevant outcomes. Therefore, there is a need to establish robust human-based in vitro experimental models that can predict in vivo outcomes of MP and NP pollution hazards in the human intestine and also provide a basis for subsequent particle management and control if needed. For these reasons, in the present study, we employed organoid technology to create near physiologically relevant cell culture models for the evaluation of the potential toxic effects of MPs and NPs varying in size (covering micro to nano scales), concentrations, and exposure periods on patient organoid-derived intestinal epithelia with different cellular complexities (with and without M cells), examining cellular uptake and translocation, and impact on barrier integrity and immune response (cytokine release).

Much of the research concerning in vitro cellular uptake and cytotoxic impact of MPs and NPs in human intestinal epithelium has been carried out using monocultures of Caco2 or HT29-MTX (enterocyte-like or goblet-like cells), co-cultures of Caco-2/HT29-MTX, or tri-cultures of Caco-2/HT29-MTX/Raji B (M-like cell induction). However, these human cancerous cell lines have several limitations. First, these cell cultures have much higher TEER values (1600–2500Ω.cm²) than what is reported for the intestine in vivo (50–100Ω.cm²), reflecting poorer permeability of compounds through the paracellular route. Second, these cell cultures only represent limited cell subgroups which do not faithfully mirror the broad diversity of in vivo epithelial cell diversity and therefore fail to reproduce the complexity of in vivo cellular responses. By co-culturing Caco-2 cells with Raji lymphocyte cells, Caco-2 cells can be induced to differentiate into cells with an M-like cell morphology, yet there have been no consistent experimental conditions reported in terms of the onset of the co-culture, the duration of the co-culture, the standardized number of Raji B cells for sufficient M-like cell induction.³³ Third, while representing some subgroups of the intestinal epithelium, these cell lines are cancerous and therefore differ from the normal cells. Systems based on these cancerous cell lines have proved unreliable in terms of reciprocating the biological and physiological function of intestinal epithelial cells due to the low or absent expression of some clinically relevant transporter proteins, including sodium/glucose cotransporter 1(SGLT1) and microsomal triglyceride transfer protein (MTP), and metabolic enzymes, such as cytochrome P450 (CYP) enzymes.³⁴ It was also reported that Caco-2 cells were particularly resistant to cytotoxic effects of PS NPs.³⁵ Therefore, nanomaterial toxicity tests using normal and cancer cells may yield significantly different outcomes. As a promising alternative intestinal epithelial model, intestinal organoids can be used to study intestinal nutrient transport along with uptake and metabolism of drugs and exogenous compounds. Through gene expression analysis, researchers have demonstrated that, compared to undifferentiated and differentiated Caco-2 cells, organoid-derived intestinal epithelial cells more appropriately express a panel of metabolically essential genes and some critical genes involved in fatty acid and cholesterol homeostasis in the small intestine.³⁶ Importantly, intestine organoids also have shown higher toxicity predictability than Caco-2 cells in drug screening study with 31 drugs used clinically as a reference set,³⁷ suggesting the intestinal organoid models permit more human-relevant prediction of the adverse effects caused by exposures to foreign compounds.

In this context, we adapted human intestinal organoids as a tool to assess MP- and NP-induced toxicity in the human gut. After targeting the cells, different sized particles internalize and pass through the intestinal epithelium by different pathways and mechanisms for them to undergo intracellular transport via regular epithelial cells and M cells, and then exit the cells at the basolateral membrane to complete the absorption process. Accordingly, we generated two physiologically relevant epithelial organoid monolayer models following well-established differentiation protocols (Figure 2): SDM-induced monolayer without M cells (enterocytes, goblet cells, EECs, and Paneth cells) and MDM-induced monolayers with M cells (enterocytes, goblet cells, EECs, and Paneth cells, M cells). Our intention was to simulate in vivo cellular complexity for particle uptake and translocation to identify how different cell models respond to the exposure of MPs and NPs and ultimately decipher the contribution of M cells during these responses. Using immunostaining with various cell specific markers and SEM, we validated that the different protocols successfully induced the undifferentiated monolayers into mature intestinal epithelia with desired cell populations consisting of the major epithelial cell types without/with M cells.

Several studies using in vitro cell models have indicated that the size of particles plays a critical role in determining the efficiency of cellular uptake as well as the uptake pathway in different living human cell. For example, a size-dependent uptake has been observed for PS particles in human lung cells, astrocytoma cells, umbilical vein endothelial cells, intestinal cells (Caco-2).³⁸ In all cases, smaller particles had a higher chance of penetrating the cell membrane, demonstrating larger adverse effects, and the particles had maximal cellular uptake in the size range of 30–50 nm. Our results showed that, in SDM-induced intestinal epithelia, all MPs and NPs homogeneously attached to and potentially entered via the apical side of the epithelial surface with a significant uptake only occurring at the particle size of 30 nm at 96 hours after exposure. In contrast, in the monolayers containing M cells, while the cellular internalization of the MPs and NPs was also highly dependent on the size, plastic particles larger than 500 nm preferentially aggregated in M cells and penetrated the monolayers through M cells. In addition, the incorporation of M cells generally increased the degree of translocation of MPs and NPs at all sizes. Particularly, 30 nm particles were randomly taken up by regular epithelial cells (enterocytes and secretory cells), whereas larger particles (500 nm and 1 μm) crossed the epithelium via M cells. This finding is supported by earlier in vivo data demonstrating that larger PS particles are taken up exclusively by M cells associated with Peyer’s patch in rats and mice³⁹ and in vitro data showing M-like cells from cancerous cell lines and M-cells induced from mouse intestinal organoids were strongly associated with higher uptake of PS NPs.⁴⁰ A key characteristic of functional M-cells is that they can efficiently capture and transport large particles and bacteria across the epithelial barrier. In this regard, M cells induced from intestinal organoids exhibit similar functional characteristics to M-cells in vivo, confirming the potential of intestinal organoids as a platform for enriching and studying rare intestinal cell types, such as M cells, EECs, Tuft cells.⁴¹ Moreover, our results also agreed with the accepted mechanism of microparticle (i.e. > 100 nm) uptake through transcytosis across M-cells located in the Peyer’s patches and nanoparticle (i.e. <100 nm) uptake through other epithelial cells, such as enterocytes and goblet cells, by endocytosis or paracellular transport through tight junctions (only for particles smaller than 5 nm).⁴² These results suggest that the cellular translocation mechanism and efficiency of the PS MPs and NPs in human intestine significantly relies on the particle size.

After observing the differences in internalization behaviors with MPs and NPs in different sizes, we further evaluated two other experimental variables, concentration and exposure period, on the cytotoxic impact of the particles on intestinal epithelial cells by examining the epithelial barrier integrity using TEER and immune response using ELISA-based cytokine measurement (Figures 7 and 8). Given that MPs and NPs accumulate in the gut of several living organisms and that large plastic debris may even block the gastrointestinal track in some severe cases,⁴³ we decided on screening extreme effects of PS particles using high concentrations ranging from 250 to 1000 μg/mL. Many of the available studies have used similar concentration ranges to assess the biological effects of PS MPs and NPs in intestinal or other cell types,^{44, 45} with some studies testing even higher particle doses (up to 2000 μg/mL).^{46, 47}

During the plastic exposure experiments, we did not observe obvious changes in the TEER values in the SDM-induced epithelia exposed to 1 μm, 500 nm, 100 nm particles, independent of the concentrations (250 μg/mL, 500 μg/mL, and 1000 μg/mL), and exposure period (24, 48, 72 and 96 hours). Similar results were obtained in the MDM-induced epithelia treated with 1 μm and 500 nm MPs. This is in line with work from other groups employing either differentiated or undifferentiated Caco-2 cells, in which no or slight impact on cell viability and membrane integrity was found when exposed to PS particles (100 nm-5 μm). It is worth noting that the highest concentration tested in these studies was 200 μg/mL and the longest exposure time in the experiments was 48 hours.⁴⁸ As an exception, 5 μm PS-MPs inhibited the proliferation of Caco-2 cells in vitro in a time- and concentration-dependent pattern.⁴⁹ However, in that study Caco-2 cells were cultured in planar plastic substrate, which could affect the cell growth and differentiation and therefore lead to discrepancies in the results. Our data obtained from human derived intestinal organoids (with and without M cells) supplement the existing knowledge regarding PS particle toxicity by showing even higher concentrations of particles larger than 500 nm (500 nm and 1 μm) with prolonged exposure time (up to 96 hours) did not lead to obvious damage to the intestinal epithelial integrity. Additionally, in the exposure study, as the concentration of particles was increased to 1000 μg/mL, 30 nm NP exposure groups mildly induced decreases in TEER in SDM-induced epithelia, whereas both the 100 nm and 30 nm groups significantly decreased TEER in the MDM-induced monolayer starting from 48 hours of exposure. It is worth mentioning that, compared to the control groups, in the monolayers without M cells, cell viability did not seem to be significantly affected by MPs and NPs. However, intestinal monolayers containing M cells detached in the groups of 100 nm/96 hour, 30 nm/72 hour and 30 nm/96 hour at a concentration of 1000μg/mL, indicating epithelial cell damage during exposure to smaller sized groups. This observation agrees with previous reports documenting that smaller sized nanoparticles are correlated with higher cell damage in both in vitro and in vivo models⁴ and that additionally, PS particles do not cause notable cytotoxicity in human skin dermal fibroblasts and blood cells at concentrations up to 500 μg/mL.⁵⁰ Interestingly, both 30 nm and 100 nm groups resulted in more progressive reduction of TEER in the M cell containing epithelia.

The mechanisms of MPs and NPs cytotoxicity also involve stimulating the secretion of inflammatory cytokines. Cytokines represent a vast family of immunomodulatory substances typically including Interleukins (IL), Interferon (IFN), Growth Factor (GF), Tumor Necrosis Factor (TNF), and Chemokines (CKs).⁵¹ Maintaining a low-grade “physiological inflammation” in the essential for normal tissue function. During the infection or injury, multiple intestinal immune and non-immune cells release abundant counter-regulatory cytokines to enable gut immune homeostasis and tolerance. Dysfunction of cytokine release by the immune system in host epithelial cells has been implicated in inflammatory bowel disease (IBD), which affects approximately 3.1 million adults (1.3%) in the United States and more than 10 million people worldwide. Among different cytokines, the most studied members involved in intestinal immune inflammation comprise TNF-α, TGF-β1, IL-6 and IL-8. The upregulation of these cytokines are detected in the intestine tissue samples from patients with active IBD.⁵²

Because of the lack of standardized research models and high diversity of the MP and NPs (different size, shape, surface charging and polymer type) used in different studies, there are contradictory results derived from both in vivo and in vitro studies on whether MPs and MPs activate inflammatory reactions in the intestine.⁵³ For example, in mouse models, one study showed that after 21-weeks of exposure to environmentally relevant doses of large PS MPs (40–100μm), mice displayed a clear intestinal inflammatory response with infiltration of immune cells. Similarly, intestinal barrier dysfunction and cell death was found in mice that had been administered PS particles in varying sizes (50 nm, 500 nm and 5000 nm) under single or co-exposure conditions in a size-dependent manner.⁵⁴ In contrast, 5μm PS MPs alone caused minimal effects on the intestinal barrier of mice, but exposure to PS MPs resulted in histological injury, intestinal microbiota disorders, inflammation, and oxidative stress in mice with acute and chronic colitis.⁵⁵ A more recent study compared the effects of PS MPs (5μm) on healthy mice and mice with intestinal immune imbalances. The authors concluded that PS MPs exposure only significantly stimulated the expression of inflammation factors (TNFα, IL-1β and IFN-γ) in mice with intestinal immune imbalance.⁵⁶ Data from in vitro human intestinal cell models also contributed to the conflicting results. Some studies have cited that 5μm PS particles showed little influence on IL-8 and MCP-1 secretion in Caco-2 cells, however, IL-8 levels were enhanced by exposure to 100 nm particles at a concentration of 20 μg/mL.⁵⁷ Other studies highlighted that 3 and 10 μm gave rise to a large amount of reactive oxygen species (ROS) production, key signaling molecules in inflammatory disorders, in human intestinal cells in the short term.¹⁵ In the current study, from the cytokine profiles, compared to the non-dosing control groups, virtually no cytokine release was induced until concentrations reached 500 μg/mL in both SDM- and MDM induced intestinal epithelia. Additionally, based on the general increase level of the cytokines in 500 μg/mL and 1000 μg/mL groups, the exposure of PS MPs and NPs induced stronger inflammatory reaction on monolayers without M cells than without M cells. For example, we found that under 1000 μg/mL, both 500 nm and 1 μm MPs significantly induced the secretion of TNF-α, IL-6 and IL-8 compared to the control group, which did not occur in the monolayers without M cells. The key role of the M cells is to sense, capture and deliver antigens/pathogens across the epithelium, thus triggering mucosal immune response. In our case, the larger particles (500 nm and 1 μm) were sensed, sampled, and transported through M cells (confirmed in our tracking assay) which led to production of pro-inflammatory cytokines for protection against potential antigens and pathogens. Furthermore, TGF-β1, not promoted in monolayers without M cells in exposure group, was significantly upregulated in 100 nm/72 hour and 30 nm/48 hour exposure groups. As a pleiotropic cytokine, the role of TGF-β1 in IBD is still under debate. However, most researchers believe that TGF-β inhibits the intestinal inflammatory responses to luminal bacteria and food antigens, thus achieving immune tolerance.⁵⁸ In this scenario, when exposed to small PS particles, M cells contribute to the activation the specific immunological pathways (TGF-β signaling pathway) to downregulate excessive inflammatory cytokine release and ameliorate epithelial inflammation caused by particle exposure.⁵⁹ However, we also found epithelial cell damage in the groups of 100 nm/96 hour, 30 nm/72 hour and 30 nm/96 hour at a concentration of 1000 μg/mL in the monolayers containing M cells, which did not occur in the monolayer without M cells. This may indicate that the anti-inflammatory activity against small particles triggered by M cells and other epithelial cells was not intense enough to suppress the “cytokine storm” resulting from the sudden and acute increase in levels of TNF-α, IL-6 and IL-8 in the groups with cell damage (Figures 8M, 8O, and 8P). Other immune sensers, such as B cells in gut-associated lymphoid tissue (GALT), macrophages, dendritic cells, and mast cells should be involved to reach native immune protection.⁶⁰ Based on these findings, we concluded that the ingestion of PS MP and NP suggests the potential towards IBD-related risks via the induced inflammatory cytokines associated with IBD. Furthermore, M cells induced from human intestinal organoids were able to imitate the biological functions of native M cells and that the MDM-induced monolayer represents an improved intestinal follicle-associated epithelium to study MPs and NPs transport by M cells.