Effects of ingested food-grade titanium dioxide, silicon dioxide, iron (III) oxide and zinc oxide nanoparticles on an in vitro model of intestinal epithelium: Comparison between monoculture vs. a mucus-secreting coculture model

The majority of in vitro studies of ingested engineered nanomaterials (iENMs) in the past decade employed a Caco-2 monoculture test system. However, the small intestinal epithelium is much more complex and includes a defensive mucus layer secreted by goblet cells and submucosal glands. Here, we used a monoculture and a mucus-secreting coculture/triculture model to evaluate the toxicity of 5 food-grade nanomaterials. We employed an integrated in vitro testing methodology that includes: (i) utilizing food-grade variants of four major iENMs (TiO2, SiO2, ZnO, and two types of Fe2O3: rod-shaped and acicular); (ii) preparing standardized nanoparticle dispersions; (iii) matching in vitro doses to in vivo doses using equivalent surface area-based mathematical and computational in vitro dosimetry models; and (iv) addressing iENM dissolution kinetics in relevant cell culture medium. In addition, in vitro mucus-secreting coculture (C2BBe1- Caco-2 subclone/HT29-MTX) and triculture (C2BBe1/HT29-MTX/Raji B) models of the intestinal epithelium, and a C2BBe1 monoculture system were used to assess effects of exposure to each test iENM, over a range of concentrations, on indicators of epithelial cell health and function, including barrier integrity, cell cytotoxicity and viability, morphology, and pro-inflammatory cytokine and chemokine production. Barrier integrity was less vulnerable to iENM exposures in tricultures than in monocultures: trans-epithelial resistance (TEER) in tricultures was unaffected by any iENM exposures, but was reduced in monocultures by up to 50% after some treatments. However, cocultures/tricultures were otherwise more sensitive than monocultures to iENM exposures, as reflected by (i) morphological changes in the brush border (blunted microvilli) in triculture; (ii) loss of cell viability and metabolic activity in cocultures but not monocultures at low (TiO2) and all (SiO2 and ZnO) doses; (iii) increased membrane damage (LDH release) and; (iv) enhanced inflammatory response, with up to 3-fold higher IL-6, 2-fold higher IL-1 beta, and significantly higher IL-8 production in cocultures than in monocultures. Irrespective of the model, ZnO was consistently the most cytotoxic iENM, followed by rod-shaped and acicular Fe2O3, and TiO2. We demonstrated that the C2BBe1 coculture and triculture models are physiologically more relevant than and superior to the C2BBe1 monoculture, or the commonly used Caco-2 monoculture system, for assessing iENM toxicity.