Comparison of conventional and advanced in vitro models in the toxicity testing of nanoparticles

Figures & data

Table 1. Specific issues in the toxicity testing of NPs.

Parameter	Specific issues with NPs
Exposure medium	Exposure medium is important because particle parameters are changed by medium composition (agglomeration)
Duration of exposure	Usually too short as NPs are metabolized to lower extent than conventional compounds
Monolayer culture	NPs cross cell layers by diffusion and paracellular transport to lower extent than conventional compounds
Monoculture	Cell uptake differs between phagocytes and non-phagocytes for NPs and less for conventional compounds
Absence of dynamics	NPs get in contact with cells by sedimentation, which does not play a role for conventional compounds
Low cell differentiation	Secretion of mucus hinders permeation of NPs to higher degree than conventional compounds due to the size exclusion effect^a

^aSize exclusion means that NPs, due to their size, are sieved though the mucus mesh.

Figure 1. Extent of NP exposure, translocation and use of advanced cell culture models in the testing for epithelial barriers and internal organs. Independent from the extent of exposure use of in vitro models for protective barriers (cornea, epidermis, oral cavity, vaginal epithelium) is low as good ex vivo systems are available. In vitro systems are used when particle exposure is high and robust ex vivo systems are missing (alveolar and intestinal epithelium).

Table 2. Origin and use of cell lines in the physiologically relevant models.

Cell line	Species	Origin	Use
16HBE14o–	Human	SV40 immortalized bronchial epithelial cells	Bronchial epithelium, toxicity
A549	Human	Lung carcinoma	Alveolar epithelium, toxicity
BEAS-2	Human	Epithelial virus transformed bronchial epithelial cells	Bronchial epithelium, toxicity
Caco-2	Human	Colorectal adenocarcinoma	Intestinal epithelium, barrier function, toxicity
CAL27	Human	Oral squamous cell carcinoma	Cancer cell
Calu-3	Human	Lung adenocarcinoma	Bronchial epithelium, barrier function
CRL-2102 (C2BBe1)	Human	Clone of Caco-2 cells	Enterocytes
EAhy926	Human	Fusion of HUVEC with human pulmonary adenocarcinoma A549 cells	Endothelium
Fa2N4	Human	SV 40 immortalized hepatocytes	Hepatocytes
hAELVI	Human	Lentivirus immortalized alveolar epithelial cells	Alveolar epithelium, barrier function
HeLa	Human	Cervical cancer	Cancer cell
Hep3B	Human	Hepatocellular carcinoma	Hepatocytes
HepaRG	Human	Liver progenitor cells	Hepatocytes
HepG2 Hep2/C3a	Human	Hepatocellular carcinoma derived from HepG2 cells	Hepatocytes
HK-2	Human	Proximal tubule papilloma	Renal tubule cells, barrier function
HMC-1	Human	Mast cell leukaemia	Mast cells
HPMEC-ST1.6R	Human	Virus transfected microvascular endothelial cells	Endothelial cells
HT29 and HT29-MTX	Human	Colon adenocarcinoma cells and cells treated with methotrexate to induce mucus production	Goblet cells
Huh7	Human	Hepatocellular carcinoma	Hepatocytes
ISO-HAS 1	Human	Haemangiosarcoma	Endothelium
J774.A1	Mouse	Reticulum cell sarcoma	Monocytes/macrophages function
LLC-PK1	Pig	Kidney cells	Renal tubule cells, barrier function
LS174	Human	Colorectal adenocarcinoma	Intestinal epithelium
LS513	Human	Colorectal carcinoma	Intestinal epithelium
M5076	Mouse	Ovarian sarcoma	Cancer cells
MCF-7	Human	Breast adenocarcinoma	Metabolization, action of transporters
MDCK	Dog	Distal renal tubules	Renal tubule cells, barrier function
MG63	Human	Osteosarcoma	Osteoblasts
MH-S	Murine	Simian virus 40 transformed alveolar macrophages	Alveolar macrophages
MLE 12	Mouse	Lung epithelial cells	Alveolar epithelium
MRC-5	Human	Foetal lung fibroblasts	Fibroblasts
NCI-H322	Human	Bronchoalveolar carcinoma	Alveolar epithelium, toxicity
NCI-H441	Human	Papillary lung adenocarcinoma	Alveolar epithelium
NCI-H460	Human	Large-cell lung cancer	Cancer cell
NIH/3T3	Mouse	Embryonal fibroblasts	Fibroblasts
NKi-2	Human	hTERT immortalized proximal tubule cells	Proximal renal tubule cells
NRK52K	Rat	Kidney epithelial cells	Renal tubule cells, barrier function
Raji B	Human	Burkitt’s Lymphoma	Induction of M cell formation in co-culture
Rat-2	Rat	Foetal fibroblasts	Fibroblasts
RAW 264.7	Mouse	Abelson murine leukaemia virus-induced tumour	Monocytes/macrophages function
T84	Human	Colorectal carcinoma	Intestinal epithelium
THP-1	Human	Acute monocytic leukaemia	Monocytes/macrophages function
TK6	Human	Hereditary spherocytosis lymphoblasts	Genotoxicity testing
TLT	Human	Macrophages	Macrophages
U937	Human	Histiocytic lymphoma	Monocytes/macrophages function

Figure 2. Use of transwell membranes in advanced culture models. Monoculture for permeation experiments (A), indirect contact (B) and direct or indirect contact (C) co-culture of only one cell type in each chamber. Cells can be cultured or separated by matrices that may either be acellular (D) or contain one (E) or several types of cells (E, F). Co-culture systems may consist of two and more cell types in the apical compartment (G), co-culture of two and more cell types in the apical compartment in indirect culture with one cell type in the basolateral compartment (H), co-culture of one cell type in the apical and several types of cells in the basolateral compartment (I), combined direct contact and indirect contact culture (J), direct contact culture of several cell types in the apical compartment and one type in the basolateral compartment (K). The separation line in H, J and K indicates that different cell types in monoculture can be used in the basolateral compartment or in the apical compartment (E).

Figure 3. Particle and biological parameter that were identified to play a role in in silico modelling of metal oxide NPs (according to the meta-analysis by Ha et al. [105]). Parameters can be influenced by the use of advanced cell culture models, either by medium composition (M) or by the culture method (C). Medium composition may have an influence on aggregation (hydrodynamic size) and influence the dose that reaches the cell. In addition, surface parameters may be changed. The culture method influences mainly cellular differences by increasing cell differentiation and the exposure time as physiologically relevant culture methods usually enable exposure for longer time periods.

Table 3. Parameters included in QSAR models.

Nanomaterial	Toxicity endpoint	Characterization	Reference
18 NMs (carbon-based, metal oxides)	Cytotoxicity, apoptosis, pro-inflammatory effects, haemolysis, viability, mitochondrial membrane potential, morphology	7 descriptors: size, surface area, morphology, metal content, reactivity, free radical generation, zeta potential	[107]
18 NMs	Viability	17 quantum mechanical descriptors (enthalpy of formation of nanocluster, total and electronic energy, core–core repulsion energy, solvent accessible surface, energy of the highest occupied molecular orbital, energy of the lowest unoccupied molecular orbital, gap between both, electronic chemical potential, valence band, conduction band, Mulliken’s electronegativity, Parr and Pople’s absolute hardness, Schüürmann Molecular Orbital shift alpha quantities, polarizability derived from the heat of formation, and polarizability derived from dipole moment) and 11 experimental descriptors (area, volume, surface diameter, volume/mass diameter, volume/surface diameter, aspect ratio, porosity, sphericity, circularity)	[111]
51 NMs with four metal core structures	Viability, reducing equivalents, apoptosis, mitochondrial membrane potential	5 descriptors: core composition, coating, surface modification, relaxivities, zeta potential	[112]
42 NMs with two cores	Cytotoxicity	6 descriptors: primary particle size, size in water, size in PBS, cell in cell culture medium, concentration, zeta potential	[113]
13 pure, core-shell and alloy Au/Pd TiO2 NMs	Cytotoxicity (CHO-K1 cells)	2 descriptors: size, surface area	[114]
9 metal oxide NMs	Cytotoxicity (BEAS-2B cells)	14 descriptors: atomization energy of the metal oxide, period of the NP metal, and NP primary size, in addition to NP volume fraction (in solution) were identified as most predictive	[115]
24 metal oxide NMs	ROS, oxidative stress, pulmonary inflammation in mice	30 theoretical descriptors: conduction band energy predictive for some, solubility for other metal oxide NPs	[116]
41 metal oxide NMs	Cytotoxicity	4 descriptors; size, electronegativity, polarizability, molar volume	[117]
17 metal oxide NMs	Cytotoxicity (HaCaT cells)	7 theoretical descriptors (number of metal atoms, number of oxygen atoms, molecular weight, charge of the metal cation corresponding to a given oxide, metal electronegativity, sum of metal electronegativity for the individual metal oxide, sum of metal electronegativity for the individual metal oxide divided by the number of oxygen atoms in a specific metal oxide)	[118]
24 metal oxide NMs	Viability, 2 cell lines	30 descriptors: conduction band energy and ionic index were identified as very predictive	[108]
44 iron oxide NMs	4 cell types, 4 assays	4 descriptors: primary size, spin–lattice, spin–spin relaxivities, zeta potential; no single parameter performed best	[109]
6 metal oxide NMs	Oxidative stress	1 descriptor: energy band structure	[119]
307 studies, Cd quantum dots	Viability	24 qualitative and quantitative features (ligand, shell, surface modification, assay type, exposure time, exposure concentration, cell anatomical type, cell origin)	[120]
20 C60 fullerene NPs	Mutagenicity	3 descriptors: dose, illumination, metabolic activation	[121]
84 f-MWCNTs	Cytotoxicity, protein binding, immune response	5 descriptors: zeta potential, electrophoretic mobility, surface area, porosity, solubility	[122]

Ha MK, Trinh TX, Choi JS, et al. Toxicity classification of oxide nanomaterials: effects of data gap filling and PChem score-based screening approaches. Sci Rep. 2018;8:3141.

 PubMed Web of Science ®Google Scholar

Wang XZ, Yang Y, Li R, et al. Principal component and causal analysis of structural and acute in vitro toxicity data for nanoparticles. Nanotoxicology 2014;8:465–476.

 PubMed Web of Science ®Google Scholar

Gajewicz A, Schaeublin N, Rasulev B, et al. Towards understanding mechanisms governing cytotoxicity of metal oxides nanoparticles: hints from nano-QSAR studies. Nanotoxicology 2015;9:313–325.

 PubMed Web of Science ®Google Scholar

Shaw SY, Westly EC, Pittet MJ, et al. Perturbational profiling of nanomaterial biologic activity. Proc Natl Acad Sci USA. 2008;105:7387–7392.

 PubMed Web of Science ®Google Scholar

Sayes C, Ivanov I. Comparative study of predictive computational models for nanoparticle-induced cytotoxicity. Risk Anal. 2010;30:1723–1734.

 PubMed Web of Science ®Google Scholar

Mikolajczyk A, Sizochenko N, Mulkiewicz E, et al. Evaluating the toxicity of TiO2-based nanoparticles to Chinese hamster ovary cells and Escherichia coli: a complementary experimental and computational approach. Beilstein J Nanotechnol. 2017;8:2171–2180.

 PubMed Web of Science ®Google Scholar

Liu R, Rallo R, George S, et al. Classification NanoSAR development for cytotoxicity of metal oxide nanoparticles. Small 2011;7:1118–1126.

 PubMed Web of Science ®Google Scholar

Zhang H, Ji Z, Xia T, et al. Use of metal oxide nanoparticle band gap to develop a predictive paradigm for oxidative stress and acute pulmonary inflammation. ACS Nano. 2012;6:4349–4368.

 PubMed Web of Science ®Google Scholar

Luan F, Kleandrova VV, Gonzalez DH, et al. Computer-aided nanotoxicology: assessing cytotoxicity of nanoparticles under diverse experimental conditions by using a novel QSTR-perturbation approach. Nanoscale 2014;6:10623–10630.

 PubMed Web of Science ®Google Scholar

Kar S, Gajewicz A, Roy K, et al. Extrapolating between toxicity endpoints of metal oxide nanoparticles: predicting toxicity to Escherichia coli and human keratinocyte cell line (HaCaT) with Nano-QTTR. Ecotoxicol Environ Saf. 2016;126:238–244.

 PubMed Web of Science ®Google Scholar

Liu R, Zhang HY, Ji ZX, et al. Development of structure-activity relationship for metal oxide nanoparticles. Nanoscale 2013;5:5644–5653.

 PubMed Web of Science ®Google Scholar

Liu R, Rallo R, Weissleder R, et al. Nano-SAR development for bioactivity of nanoparticles with considerations of decision boundaries. Small 2013;9:1842–1852.

 PubMed Web of Science ®Google Scholar

Burello E, Worth AP. A theoretical framework for predicting the oxidative stress potential of oxide nanoparticles. Nanotoxicology 2011;5:228–235.

 PubMed Web of Science ®Google Scholar

Oh E, Liu R, Nel A, et al. Meta-analysis of cellular toxicity for cadmium-containing quantum dots. Nat Nanotechnol. 2016;11:479–486.

 PubMed Web of Science ®Google Scholar

Toropov AA, Toropova AP. Optimal descriptor as a translator of eclectic data into endpoint prediction: mutagenicity of fullerene as a mathematical function of conditions. Chemosphere 2014;104:262–264.

 PubMed Web of Science ®Google Scholar

Zhou H, Mu Q, Gao N, et al. A nano-combinatorial library strategy for the discovery of nanotubes with reduced protein-binding, cytotoxicity, and immune response. Nano Lett. 2008;8:859–865.

 PubMed Web of Science ®Google Scholar