Investigation of twenty metal, metal oxide, and metal sulfide nanoparticles' impact on differentiated Caco-2 monolayer integrity

Highlights

• Transwell cultures of intestinal enterocytes were exposed to 50 μg/mL ENM for 24 h.

• In Vitro Sedimentation, Diffusion and Dosimetry Modeling was conducted for all ENMs.

• The permeability coefficient (P_app) for dextran was increased by four ENMs.

• P_app for dextran was decreased by MgO NP.

• With the exception of CdS NP, increased P_app was not connected with cytotoxicity.

Abstract

The use of engineered nanomaterials (ENMs) in foods and consumer products is rising, increasing the potential for unintentional ingestion. While the cytotoxicity of many ENMs has been investigated, less attention has been given to adverse impact on the intestinal barrier integrity. Chronical disruption of gastrointestinal integrity can have far reaching health implications.

Using fully differentiated Caco-2 cells, the perturbation of intestinal barrier function and cytotoxicity were investigated for 20 metal, metal oxide, and metal sulfide ENMs. Caco-2 cells were exposed to 50 μg/mL ENMs for 24 h. ENM formulations were characterized at 0 and 24 h, and In Vitro Sedimentation, Diffusion and Dosimetry Modeling was applied to calculate the effective dose of exposure during 24 h. The apparent permeability coefficient (P_app) was determined for fluorescent labeled dextran (3000 Da) and tight junction integrity was evaluated by immunofluorescence microscopy. Cytotoxicity was investigated by determining lactate dehydrogenase release (LDH) and cell metabolic activity (tetrazolium based MTS) assays.

Four ENMs led to significantly increased P_app, (15.8% w/w% Ag-SiO₂ nanoparticle (NP), 60 nm CdS NP, 100 nm V₂O₅ flakes, and 50 nm ZnO NP), while one ENM (20 nm MgO NP) decreased P_app. With the exception of CdS NP, significantly increased P_app was not connected with cell cytotoxicity. The calculated effective dose concentration was not correlated with increased P_app.

Our results illustrate that while many metal, metal oxide, and metal sulfide ENMs do not adversely affect monolayer integrity or induce cytotoxicity in differentiated Caco-2 cells, a subset of ENMs may compromise the intestinal integrity. This study demonstrated the use of differentiated Caco-2 monolayer and P_app as an endpoint to identify and prioritize ENMs that should be investigated further. The interaction between ENMs and the intestinal epithelium needs to be evaluated to understand potential intestinal barrier dysfunction and resulting health implications.

Introduction

As the field of nanotechnology evolves, there is a need for comprehensive biological response profiles to guide the development of safe and sustainable use of nanomaterials. In the past two decades, the number of consumer products containing engineered nanomaterials (ENMs) has increased substantially, including those within the food and agriculture sectors (Bergin and Witzmann, 2013, Bouwmeester et al., 2014, EFSA, 2015). According to several consumer product databases and inventories, there are now hundreds of commercially available food and beverage products that contain ENMs (Nanotechnologies, 2015; Nanodatabase, 2019; CFS, 2017). It has been estimated that the daily number of ingested metal or metal oxide particles amounts to 10¹², comprised mainly of titanium dioxide (TiO₂) and silicon dioxide (SiO₂) (Lomer et al., 2002). Food grade TiO₂ is a food additive used as a whitening agent that has been found in food at concentrations between 0.02 and 9.0 mg TiO₂/g product (Peters et al., 2014), and the estimated daily consumption of TiO₂ ranges from 2 mg/kg body weight (bw) for children between age 2.5–4.5 years to 0.2 mg/kg bw for adults (Weir et al., 2012). That ingested particles cross the intestinal barrier has been demonstrated by the detection of TiO₂ particles in 7 out of 15 adult (age 56–104) human liver samples (collected post mortem) (Heringa et al., 2018). While TiO₂ and SiO₂ nanoparticles (NP) are the most prevalent in food, a range of other ENMs are also used as e.g. food supplements, antimicrobial agents, and colorant (Table 1). The level and diversity of ingested ENMs are therefore likely to keep rising, and more knowledge is needed to estimate the interaction between ENMs and the intestinal tract.

The gastrointestinal tract is a selective barrier between the lumen and the body, and responsible for nutrient absorption as well as providing protection against harmful pathogens and xenobiotics. Disruption of the gastrointestinal tract health and homeostasis can lead to many chronic diseases. Increased intestinal permeability, called “Leaky Gut Syndrome”, happens when the intestinal tight junctions (TJs) do not work properly (as reviewed (Kiefer and Ali-Akbarian, 2004, Liu et al., 2005, Hollander and Kaunitz, 2019)). TJs connect the intestinal epithelial cells and control passive diffusion of ions and other small solutes paracellularly, maintaining any gradient created by the active transport pathways associated with the transcellular route (as reviewed (Mitic and Anderson, 1998, Anderson and Van Itallie, 2009)). Disruption of TJs therefore leads to decrease in protection against xenobiotics and pathogens present in the lumen. Chronic disruption of gastrointestinal integrity has been associated with a predisposition to developing autoimmune diseases, liver dysfunction, arthritic and other degenerative diseases (Meletis, 1998; Arrieta et al., 2006; Fukui, 2016; Konig et al., 2016).

In vitro models of the intestinal tract have been used to investigate ENM translocation across cell monolayers and cytotoxic effects (Zha et al., 2008; Esch et al., 2012; Fisichella et al., 2012; Loo et al., 2012; Piret et al., 2012; Gerloff et al., 2013; Gitrowski et al., 2014; Ude et al., 2017; Ye et al., 2017; Vila et al., 2018a; Vila et al., 2018b). A well-established model for the intestinal barrier is the epithelial colorectal adenocarcinoma Caco-2 cells. Fully differentiated Caco-2 cells resemble mature enterocytes, expressing an organized brush border with a dense network of functional TJs (Hilgendorf et al., 2000; Carr et al., 2012). Differentiated Caco-2 transwell (TW) cultures are the model of choice for uptake studies into the small intestine (Christensen et al., 2012), and are considered a good model for studying intestinal barrier function (Alvarez-Hernandez et al., 1991; Jarc et al., 2019). It has been established that Caco-2 cells need to grow on a semipermeable surface, such as a TW membrane, in order to fully differentiate into polarized enterocytes, exhibiting well-developed microvilli and polarized distribution of brush border enzymes and transporters (Hidalgo et al., 1989; Sambuy et al., 2005; Natoli et al., 2012). The use of undifferentiated cells is quicker and cheaper than using the differentiated Caco-2 model; however, differentiated cells more accurately mimic in vivo conditions and allow for studies of monolayer integrity.

There is an increasing incentive to reduce animal testing and develop more physiological relevant in vitro models. With the large number of diverse ENMs implemented in food and consumer products, where safety needs to be assessed, it is important to align nanosafety and toxicology studies to the 3Rs principles (Replacement, Reduction and Refinement of animal testing) (Burden et al., 2017). Initiatives in developing and validating 2-dimensional (2D) and 3-dimensional (3D) mono- and co-culture in vitro models have been expanding rapidly during the past decade and are showing great promise. However, dependent on the biological mechanism in focus, i.e. studying enterocyte integrity, transwell mono-culture may be the most suitable in vitro model.

For testing ENM transport across the intestinal tract in vitro co-culture models including Caco-2, human mucus producing goblet cells (HT29-MTX) (Lesuffleur et al., 1993; Walter et al., 1996) and/or human B-lymphocytes (Raji B), which convert a portion of the Caco-2 cells to resemble Peyer's patch follicle-associated epithelium (FAE) microfold (M) cells (Lo et al., 2004; Beloqui et al., 2017), have been utilized. These cell models have been used to evaluate the role of the mucus layer in ENM interaction with the intestinal tract in vitro (Loo et al., 2012; Walczak et al., 2015). The M-cells playing a central role in the intestinal translocation and transport of ENMs have been included in studies focusing on ingested ENM uptake (Loo et al., 2012, Walczak et al., 2015). It is also worth noting that these co-culture models have lower transepithelial electrical resistance (TEER) measurements than the Caco-2 mono-culture, indicating that both HT29-MTX and Raji B cells impact the TJs of the monolayer (Araujo and Sarmento, 2013). While this has been argued to more closely resemble the human intestinal tract than the Caco-2 mono-culture, it may pose a challenge when investigating ENM interaction with enterocyte integrity. Additionally, the Caco-2 monolayer model resembles the enterocyte in the small intestinal tract where the mucus layer is considerably thinner than that of the large intestinal tract, and M-cells are located in Peyer's patches at the base of the epithelium and not in the microvilli. It is therefore important to recognize that the intestinal tract has a considerable variation in its anatomy which must be taken into consideration when selecting an in vitro model. Furthermore, these complex cell models are often expensive, with low through-put and might not aid in answering the biological function or mechanism in question. There is therefore still a need for established and well-characterized in vitro models, including the human enterocyte monolayer model, in evaluating the potential health risk of ingesting ENMs, and investigating a fundamental question such as enterocyte health and integrity.

Another aspect of in vitro evaluation of ENM exposure is simulated gastrointestinal digestion, first presented as an in vitro digestion model for bioaccessibility of toxins (Versantvoort et al., 2005), and later adopted in nanotoxicity studies (Peters et al., 2012; Lichtenstein et al., 2015; Bove et al., 2017; DeLoid et al., 2017b). While in vitro digestion provides valuable insight in the potential physiochemical changes of ENMs, it also raises questions of what food matrix to use. As illustrated in the literature, the presence of a food matrix, and the type of food matrix, dramatically changes the outcome of the ENM in vitro digestion (Peters et al., 2012; Lichtenstein et al., 2015; Kästner et al., 2017; McClements et al., 2017). Assumptions of orogastric transition time, medication, and underlying medical conditions, are all important aspects to address and need to be systematically investigated in future research of ENM ingestion studies.

This study was conducted as part of National Institute of Environmental Health Sciences (NIEHS) Nanomaterials Health Implications Research (NHIR) Consortium and the investigated ENMs were provided by the NIEHS-NHIR Consortium. Herein we compare the impact of 20 metal, metal oxide, and metal sulfide ENMs, selected by NIEHS NHIR, on intestinal integrity in vitro using the well-established Caco-2 model of fully differentiated cells resembling small intestinal enterocytes. In this study we tested a dose of 50 μg/mL ENM. Determination of the most realistic dosage is challenging since no actual data of ENM concentration in the intestinal tract is available. We based our dose selection on literature of what ENM concentrations are present in food. E.g. food grade TiO₂ has been found in food at a concentration between 0.02 and 9.0 mg TiO₂/g product (Peters et al., 2014). Therefore, 50 μg/mL for TiO₂ E171 may be substantially lower than the concentration in food. We do acknowledge that this dose might not be as realistic for other tested ENMs (e.g. V₂O₅ flakes, ZnS NP, and WO₃ NP). However, since the aim of this study is to provide a comparison between 20 ENMs, we have decided to test the same concentration for all ENMs, at the same time modeling the effective dose. We investigated the cell layer integrity following in vitro exposure, using fluorescence labeled dextran, and cytotoxic response by lactate dehydrogenase (LDH) release and mitochondrial activity assays.