Comparing the CORAL and Random Forest Approaches for Modelling the In Vitro Cytotoxicity of Silica Nanomaterials

Antonio Cassano, Richard L. Marchese Robinson, Anna Palczewska, Tomasz Puzyn, Agnieszka Gajewicz, Lang Tran, Serena Manganelli and Mark T.D. Cronin

Cassano SI

Cassano source code

Nanotechnology is one of the most important technological developments of the 21st century. In silico methods to predict toxicity, such as quantitative structure–activity relationships (QSARs), promote the safe-by-design approach for the development of new materials, including nanomaterials. In this study, a set of cytotoxicity experimental data corresponding to 19 data points for silica nanomaterials were
investigated, to compare the widely employed CORAL and Random Forest approaches in terms of their usefulness for developing so-called ‘nano-QSAR’ models. ‘External’ leave-one-out cross-validation (LOO) analysis was performed, to validate the two different approaches. An analysis of variable importance measures and signed feature contributions for both algorithms was undertaken, in order to interpret the models developed. CORAL showed a more pronounced difference between the average coefficient of determination (R2) for training and for LOO (0.83 and 0.65 for training and LOO, respectively), compared to Random Forest (0.87 and 0.78 without bootstrap sampling, 0.90 and 0.78 with bootstrap sampling), which may be due to overfitting. With regard to the physicochemical properties of the nanomaterials, the aspect ratio and zeta potential were found to be the two most important variables for Random Forest, and the average feature contributions calculated for the corresponding descriptors were consistent with the clear trends observed in the data set: less negative zeta potential values and lower aspect ratio values were associated with higher cytotoxicity. In contrast, CORAL failed to capture these trends.

This article is currently only available in full to paid subscribers. Click here to subscribe, or you will need to log in/register to buy and download this article

Acute Oral Toxicity Testing: Scientific Evidence and Practicability Should Govern Three Rs Activities

Roland Buesen, Uwe Oberholz, Ursula G. Sauer and Robert Landsiedel

Acute oral toxicity is determined for regulatory hazard classification or non-classification. The European Union Reference Laboratory for Alternatives to Animal Testing (EURL ECVAM) recommends the following modules for acute oral toxicity testing: a) the use of the in vitro 3T3 Neutral Red Uptake (NRU) test to identify substances not requiring classification and to estimate starting doses for in vivo acute oral toxicity studies; and b) the use of data from sub-acute toxicity studies to identify substances not requiring classification. However, the application of these modules in a regulatory context would require a predefined, validated and formally accepted testing strategy and data interpretation procedure, which are not available. Furthermore, the application of the 3T3 NRU assay for starting dose estimations could in fact increase the number of animals used. Finally, only very few substances exist for which data from sub-acute or other repeated dose studies are available, but data from acute studies are not. Therefore, in practice, the prediction of acute toxicity by using sub-acute toxicity data is generally irrelevant. It could even lead to a risk of overdosing in the range-finding study, which may result in the death of many or all of the animals used.

This article is currently only available in full to paid subscribers. Click here to subscribe, or you will need to log in/register to buy and download this article

To Be Credible, Success in “Reducing the Use of Animals in Scientific Research” Must Involve the Use of Fewer Animals

Michael Balls

Much worthy effort on the Three Rs is under way in the UK, but its promise of a reduction in laboratory animal use will only be credible and laudable when the number of procedures and animals are progressively and permanently reduced.
You need to register (for free) to download this article. Please log in/register here.

Comparative Costs of the Mouse Inoculation Test (MIT) and Virus Isolation in Cell Culture (VICC) for Use in Rabies Diagnosis in Brazil

Vanessa C. Bones, Augusto H. Gameiro, Juliana G. Castilho and Carla F.M. Molento

The decision to use laboratory animals rather than in vitro methods is frequently based on the financial costs involved, so the objective of our study was to compare the costs of performing the Mouse Inoculation Test (MIT) and Virus Isolation in Cell Culture (VICC) for use in rabies diagnosis in Brazil. Based on observations of laboratory routines at the Pasteur Institute, São Paulo, we listed the fixed cost (FC) and variable cost (VC) items necessary to perform both tests. Considering that 200 MITs are equivalent to 350 VICC assays, in terms of facilities and staff-hours needed per month, we calculated, for both tests, the average total cost per sample, the costs of the implementation of the laboratory structure, and the costs of routine use. With regard to absolute values, the total cost was mainly influenced by FC items, as they represented 60% of the cost for the MIT and 86% of the cost for VICC. A sample analysed by the MIT costs around 205% more than one analysed by using VICC. The MIT costs 74% and 406% more than VICC, when implementation costs and routine use per month, respectively, are taken into account. Our results can assist in the resolution of costing disputes that could hinder the replacement of animals for rabies diagnosis in Brazil. The method demonstrated here might also be useful for cost comparisons in other situations where animal use still continues when validated alternatives exist.
You need to register (for free) to download this article. Please log in/register here.

Optimisation of the Bovine Whole In Vitro Embryo System as a Sentinel for Toxicity Screening: A Cadmium Challenge

Ellen P.A. Jorssen, Lucia Vergauwen, Karen Goossens, An Hagenaars, Mario Van Poucke, Evi Petro, Luc Peelman, Dries Knapen, Jo L.M.R. Leroy and Peter E.J. Bols

Developmental toxicity testing could greatly benefit from the availability of an in vitro alternative model based on the use of animal embryos that have better human-like physiology than the currently-used alternative models. These current models are insufficient, as extrapolation of the results can be challenging. Therefore, an in vitro bovine embryo culture system was used to expose individual morulae to test substances, and to study developmental characteristics up to the blastocyst stage. Cadmium was chosen as the reference toxicant to investigate the sensitivity of the bovine morulae to various concentrations and exposure times. Oocytes from slaughterhouse-obtained bovine ovaries, were maturated, fertilised and cultured up until the morula stage. Morulae were exposed to different cadmium concentrations for 18 or 70 hours, and developmental competence, embryo quality and the expression of cadmium exposure related genes were evaluated. Cadmium exposure hampered embryonic developmental competence and quality. Compared with the 18-hour exposure, the 70-hour exposure induced a 20-fold higher toxic response with regard to developmental competence and a more ‘cadmium-typical’ transcript expression. The bovine morula might be a promising tool for toxicity testing as, following exposure, the embryos reacted in a sensitive and ‘cadmium-typical’ manner to our reference toxicant.
You need to register (for free) to download this article. Please log in/register here.

The EpiOcular Eye Irritation Test (EIT) for Hazard Identification and Labelling of Eye Irritating Chemicals: Protocol Optimisation for Solid Materials and the Results after Extended Shipment

Yulia Kaluzhny, Helena Kandárová, Yuki Handa, Jane DeLuca, Thoa Truong, Amy Hunter, Paul Kearney, Laurence d’Argembeau-Thornton and Mitchell Klausner

The 7th Amendment to the EU Cosmetics Directive and the EU REACH Regulation have reinforced the need for in vitro ocular test methods. Validated in vitro ocular toxicity tests that can predict the human response to chemicals, cosmetics and other consumer products are required for the safety assessment of materials that intentionally, or inadvertently, come into contact with the eye. The EpiOcular Eye Irritation Test (EIT), which uses the normal human cell-based EpiOcular™ tissue model, was developed to address this need. The EpiOcular-EIT is able to discriminate, with high sensitivity and accuracy, between ocular irritant/corrosive materials and those that require no labelling. Although the original EpiOcular-EIT protocol was successfully pre-validated in an international, multicentre study sponsored by COLIPA (the predecessor to Cosmetics Europe), data from two larger studies (the EURL ECVAM–COLIPA validation study and an independent in-house validation at BASF SE) resulted in a sensitivity for the protocol for solids that was below the acceptance criteria set by the Validation Management Group (VMG) for eye irritation, and indicated the need for improvement of the assay’s sensitivity for solids. By increasing the exposure time for solid materials from 90 minutes to 6 hours, the optimised EpiOcular-EIT protocol achieved 100% sensitivity, 68.4% specificity and 84.6% accuracy, thereby meeting all the acceptance criteria set by the VMG. In addition, to satisfy the needs of Japan and the Pacific region, the EpiOcular-EIT method was evaluated for its performance after extended shipment and storage of the tissues (4–5 days), and it was confirmed that the assay performs with similar levels of sensitivity, specificity and reproducibility in these circumstances.
You need to register (for free) to download this article. Please log in/register here.

How Long Must They Suffer? Success and Failure of our Efforts to End the Animal Tragedy in Laboratories

Roman Kolar

Scientific findings have revealed how much we have dramatically underestimated the intellectual, social and emotional capabilities of non-human animals, including their levels of self-consciousness and ability to suffer from psychological stress. In the 21st century, the field of animal ethics has evolved as a serious scientific discipline, and nowadays largely advocates that the way we treat animals, both legally and in practice, is morally wrong. Politics and legislation have reacted to these facts, to some extent, but neither current legislation nor current practice reflect the scientific and moral state-of-the-art. Too often, the will to change things is watered down in the decision-making process, e.g. in the drafting of legislation. In the field of animal experimentation there have been many genuine efforts by various players, to advance and apply the principles behind the Three Rs. However, the fundamental problem, i.e. the overall number of animals sacrificed for scientific purposes, has increased. Clearly, if we are serious about our will to regard animal experimentation as an ethical and societal problem, we have to put much more emphasis on addressing the question of how to avoid the use of animals in science. To achieve this goal, certain issues need to be considered: a) the present system of ethical evaluation of animal experiments, including testing for regulatory purposes, needs to be reformed and applied effectively to meet the legal and moral requirements; b) animal testing must be avoided in future legislation, and existing legislation has to be revised in that regard; c) resources from animal-based research have to re-allocated toward alternatives; and d) the academic curricula must be reformed to foster and integrate ethical and animal welfare issues.
You need to register (for free) to download this article. Please log in/register here.