Why In Vitro Neurotoxicity Approaches Are Not Formally Validated and Used for Regulatory Purposes: The Way Forward

Home / New Perspectives / Neurotoxicity / Why In Vitro Neurotoxicity Approaches Are Not Formally Validated and Used for Regulatory Purposes: The Way Forward


Why In Vitro Neurotoxicity Approaches Are Not Formally Validated and Used for Regulatory Purposes: The Way Forward

Sandra Coecke & Anna Price, ECVAM

Published: December 6, 2007

About the Author(s)
Sandra Coecke obtained her PhD degree in Pharmaceutical Sciences in 1994 at the Free University of Brussels. In the PhD work, an in vitro approach was used to study hormonal regulation of the biotransformation enzymes flavin-containing monooxygenase and glutathione S-transferases in rat liver.

With over 17 years experience in in vitro toxicology she acquired a strong research background in the fields of in vitro metabolism/xenobiotic biotransformation in hepatocyte cultures, in vitro neurotoxicity and in vitro topical toxicity (eye and skin irritation) test models, new advanced technologies in in vitro toxicology testing systems, quality control and in the management of currently running large international collaborative European research projects. She also has run several prevalidation and formal validation studies.

She joined in 1993 the Janssen Pharmaceutica were she got the lead of the In vitro Toxicology laboratory with focus on in vitro regulatory studies for eye and skin irritation.

In 1994 she was awarded the International Price from Foundation for the Substitution of Animal Experimentation in Luxembourg.

In 1996 she joined the European Centre for the Validation of Alternative Methods, Ispra, Italy where she is responsible for ECVAM’s strategic developments with focus on emerging technologies with the potential to optimize and expedite the process of alternative methods and new testing strategies.

Since her start at ECVAM her laboratory projects were mainly focused on development of in vitro neurotoxicity and developmental neurotoxicity models and assuring metabolic competence in in vitro toxicological models.

Sandra Coecke
Institute of Health and Consumer Protection
European Commission Joint Research Centre
21020 ISPRA (VA)
E-mail: sandra.coecke@jrc.it

Anna Price graduated from the Polish Academy of Sciences, Institute of Pharmacology with a Master of Science in Biology in 1981 and with a PhD in Pharmacology in 1990.

She was awarded several fellowships which allowed her to work as a visiting senior scientist at Utrecht University (Holland), the Joseph Fourier University in Grenoble (France) and the University of London (UK). She also worked at the University of Cambridge (UK), where she continued studies on mechanisms of neuronal cell death. Her research was focused on the physiological and pathological role of nitric oxide in the Central Nervous System.

From 2002 she has been working in ECVAM supervising several research projects in the area of neurotoxcity and developmental neurotoxicity using various in vitro models of primary neuronal and glial cultures (mixed and pure) and different neuronal cell lines as alternative approach to in vivo studies. She is also involved in ECVAM regulatory activities. She serves as an expert in the area of Neurotoxicity testing for FP6 Integrated Projects, Acute-Tox (FP6 IP), and ARTEMIS. She is a Workpackage leader of CNS toxicity testing in Predict-IV (FP7 Iintegrated Project). Recently she is strongly involved in sensor and “omics” technologies (fingerprint, biomarkers profiling, multielectrode arrays, measurements of neuronal electrical activity, surface patterning and image analysis) exploring its relevance as possible endpoints for in vitro neurotoxicity assessment.

Anna Price
Institute of Health and Consumer Protection
European Commission Joint Research Centre
21020 ISPRA (VA)
E-mail: anna.price@jrc.it

Prediction of neurotoxic effects is a key feature in the toxicological profile of many compounds and therefore is required by regulatory testing schemes. Most guidelines for the evaluation of neurotoxic effects of chemicals for hazard and risk assessment are based on in vivo studies evaluating effects mainly on neurobehaviour and neuropathology that have, to a certain extent, a predictive value. However, due to the actual need to test large sets of compounds for specific regulatory requirements in Europe and the US (e .g. Registration Evaluation and Authorization of Chemicals (REACH) policy and the High Production Volume Programme) there is considerable pressure to develop alternative test strategies which are more rapid, economically feasible and have an acceptable predictive capacity.

Implementation of in vitro neurotoxicity tests, as stand-alone methods or as part of an integrated test strategy in which a battery of in vitro mechanistic assays is incorporated, would accelerate the rate at which compound knowledge and mechanistic data are produced. However, up until now, no in vitro approaches for evaluating the neurotoxic hazard of compounds have been formally validated. To speed up both the validation process and the incorporation of existing in vitro models and endpoints into neurotoxicity testing strategies, first the existing bottle-necks of the current approaches have to be identified and then possible solutions suggested.

So, what are some of the challenges to using in vitro techniques for neurotoxicity assessment in a predictive manner, e.g. to predict health effects on the human population?

1. Granted, neurotoxicity cannot be determined by cytotoxicity assays alone, but to what extent should mechanisms of toxicity be covered by neuronal specific endpoints?

So far, the test systems developed for in vitro neurotoxicity assessment for regulatory purposes have often been based on the general cytotoxicity assays and do not sufficiently represent the endpoints specific for mechanisms of neuronal and glial toxicity. Cytotoxicity assays can be applied as an initial “screening” to establish the concentration-dependent curve (IC 20, 50 and 80) and to define the non-cytotoxic concentrations that would later be tested using neuronal and glial specific endpoints.

Such a testing strategy would establish whether non-cytotoxic concentrations could already induce neurotoxicity that has to be determined by neuronal or glial specific endpoints. Therefore, in vitro neurotoxicity evaluation has to be performed at three different levels: (i) cell viability/ death (cytotoxicity assays), (ii) critical but general cell function tests (such as energy metabolism, oxidative and nitrosative stress, calcium homeostasis etc) and (iii) differentiated, neuronal specific cell function endpoints (e.g. neurotransmission, axonal transport, receptor and channel activation, enzyme activity, synaptogenesis/myelination, excitotoxicity, neuronal-glial interactions etc).

In this way, it would be possible to determine whether the observed in vitro toxic effect is due to cytoxicity (not specific for neurons or glia), pharmacological effects that may or may not be due to a toxic effect, or a mechanism of toxicity specific for neurons or glia. However, the clear differentiation between these three possible mechanisms is not necessarily obvious.

Various approaches have been suggested to identify and characterise the neurotoxic properties of chemicals such as: (i) testing the same chemicals in neuronal and non-neuronal cell models to characterise either differential responses or a differential dose-dependent curve between neuronal and non-neuronal cells, (ii) discrimination between generic toxic response versus neurotoxic specific effects in a concentration dependent way by comparison between cytotoxicity assays and endpoints that effect neuronal function. However, the list of neuronal endpoints is very long if most of the possible neuronal specific mechanisms of toxicity are to be covered. How far can we go with such a strategy to make it feasible and not too complex, especially if large numbers of chemicals have to be tested?

2. Since the anatomical and physiological complexity of the Peripheral Nervous System (PNS) and Central Nervous System (CNS) can not be fully represented by the existing in vitro models (including non-mammalian models), how can they be predictive?

CNS and PNS are the systems with various cell types organised into a functional network required to maintain an integrated function of the nervous system. However, once isolated, all neuronal culture systems represent cells that are no longer part of any integrated neural network and may develop an altered appearance, metabolism and response to test chemicals. This results in a variety of problems when predicting in vivo neurotoxicity from in vitro data. For example, carbon disulfide can cause CNS toxicity after exposure to high-doses whereas peripheral neuropathy occurs following long-term exposure to low doses. This highlights the problem of using a single in vitro model derived either from the CNS or PNS to predict an in vivo neurotoxic effect.

Similarly, it is important that glial cells are also present with neurons in the in vitro model, as they are critical not only for neuronal differentiation (morphological and physiological) but also for inducing a neurotoxic effect. This is the case for the well known dopaminergic neurotoxin MPTP, where the presence of glial cells is necessary in order to produce the neurotoxic metabolite MPP+ from the parent compound. Although in vitro data provide cellular/mechanistic information, they are limited in assessing delayed responses, discriminating between transient and persistent effects and additionally they do not assess effects on sensory or cognitive function.

There are many in vitro culture models now available, presenting various degrees of complexity including: neural, tumour-derived cell lines, organotypic explant or reaggregated brain cells and primary monolayer cultures of neurons, and astrocytes and oligodendrocytes that can be cultured in pure or in mixed cultures. In vitro models should be selected to address a specific mechanism of toxicity for studying biological processes involved in a more isolated context. When the target site of a chemical or the pharmacological action of a drug is known, a specific in vitro system with a degree of complexity appropriate to needs can be developed, in order to allow the mechanism of induced neurotoxicity or for assessment of drug efficacy and potency to be tested. However, the major problem occurs when neurotoxicity of a new chemical with unknown mechanisms of neurotoxicity has to be evaluated. What would be the most appropriate criteria for the selection of relevant in vitro models and endpoints to be able to determine whether an unknown compound is neurotoxic or not?

3. Should the role of blood-brain barrier (BBB) be considered when designing novel in vitro neurotoxicity test strategies?

The homeostasis of the CNS environment and entry of a chemical into the CNS is regulated by the BBB, which separates the brain from the systemic blood circulation. The highly specialised cerebral endothelial cells of the BBB regulate the access of circulating substances into the brain by the tight junctions and specialised transport functions. Highly lipophilic materials can pass from the luminal to the abluminal surface. However, transport of other molecules requires a transport mechanism mediated by endogenous peptides, modified proteins or monoclonal antibodies as in the case of the drug delivery process.

The existing in vitro models of the BBB can represent various degrees of complexity, depending upon their use. They range from simple cell lines cultured in the presence of hormones and growth factors to co-cultures of brain capillary endothelial cells (bovine, rat or human) and primary glial cells that result in a well-differentiated cerebral endothelium displaying most of the features observed in vivo (full functional differentiation of tight junctions with high TEER and low permeability). Additionally, these models may provide a kinetic and metabolic dimension that is often missing in in vitro models.

In the case of testing new chemicals, with unknown mechanism of toxicity, in vitro BBB models should be incorporated to determine whether the compound enters into the CNS. Furthermore, it is essential to assess whether the compound could have a direct effect on barrier components causing potential alterations/damage in BBB function. However, incorporating BBB models into in vitro neurotoxicity screening would make the whole system of testing much more complex. What would be the criteria to decide on whether the in vitro BBB model should be included and if so, what kind of model–a simple one (e.g cell line) or complex one (e.g co-culture of endothelial cells with astrocytes)? Would in silico modelling be sufficient?

4. Does the lack of in vitro metabolism have to be compensated by existing tools?

Available in vitro alternative methods have often been criticised because of the lack of metabolic competence. In many cases the biological effect of the metabolite is stronger than that of the parent compound. One could argue that the hazard might also be picked up by testing the parent compound, though at unrealisticly high concentrations. To compensate for the lack of metabolic competence, different approaches have been proposed. These include the addition of metabolic competent sources such as S9-mix (short term-experiments), hepatocyte-conditioned medium or direct co-culture with hepatocytes from different species, the transfection of cells with phase-I biotransformation enzymes or by using in silico model simulation. When necessary, in the development of integrated test strategies, these tools can be built in if the kinetic and metabolic issue for the studied compound matters.

In the case of neurotoxicity the expression of biotransformation enzymes in the brain and their role in toxicity are affected by the degree of systemic biotransformation which occurs when a compound enters the body via the oral, dermal and inhalation route and then passes through the blood-brain barrier. At the same time, the impact of kinetic and metabolic effects might differ depending on the type of exposure (e.g. acute, subchronic and chronic and the route of administration). General acute systemic effects seem to be less dependent on such factors as recently demonstrated in a large integrated European research project on acute systemic toxicity (AcuteTox). More attention should be given to species-species differences in metabolic and kinetic capacity as it is important for the extrapolation of the results obtained in animal models to the human situation.

The role of metabolism-mediated neurotoxicity in the brain, a highly heterogeneous and complex organ, is a relatively unexplored field of scientific enquiry. However, data have shown that polymorphic effects might play a role in inter-individual variability especially important in long-term degenerative effects in the human brain. Neither the in vivo chronic neurotoxicity models nor the current available alternative methods can yet account for such effects.

5. Do new technologies have the potential to advance in vitro neurotoxicity screening?

The main problem in the development of a test strategy with a predictive capacity for neurotoxicity is the fact that the mechanisms underlying the effects of compounds in the CNS and PNS are so extensive (given to the complexity of the human nervous system) that it is impossible to cover all these mechanisms with a single model and a small set of endpoints. Current in vitro approaches include a range of various models and endpoints able to assess critical underlying mechanisms of neurotoxicity as discussed above.

Over the years it has become clear that this long list of endpoints was not feasible using one, simple cell model; the development of more complex models including 3-dimensional constructs became necessary. They offer the possibility to extend the set of relevant endpoints in one single model with a higher predictive capacity for human neurotoxicity. However, this classical approach is too complex, very time-consuming and the interpretation of data is not straight forward.

The way forward would be to simplify the whole approach by using a single but reliable model and only a few general endpoints (summing up the various mechanisms of neurotoxicity) but still specific for neurons. Are there any endpoints that could do such a job? Let’s have a look at other fields of in vitro toxicity screening.

In the field of embryotoxicity it has been demonstrated that functional endpoints combined with a stem cell approach were very promising in establishing a predictive model. The principle of the test was to differentiate a murine embryonic stem cell line into beating cardiomyocites. Exposure to embryotoxic compounds was evaluated by a simple read-out, beating or non-beating, and compared with a mature control cell system. Apparently this one functional endpoint (beating or non-beating) captured many mechanisms involved in embryotoxicity. This model became formally validated as an in vitro test for embryotoxicity assessment.

Similar reasoning could be applied to developing a functional endpoint for neurotoxicity. Recent studies have shown that many of the mechanisms (but definitely not all) underlying neurotoxicity ultimately affect neuronal electrical activity. Could electrical activity measurements serve as a general but still neuronal specific endpoint for in vitro neurotoxicity evaluation? Recent advances in technology have resulted in the availability of miniaturised systems (often using chip technology) to measure spontaneous and/or evoked field potentials of different types of in vitro neuronal test systems. These systems include two and three dimensional models using either pure or mixed neural cell populations. Attempts are currently ongoing to make these systems more user friendly and higher throughput.

Another possibility could be offered by omics technologies that have already been applied in various in vitro cell culture systems for different purposes. A global evaluation of biomarkers such as expression of a gene, a protein or a metabolite becomes feasible with such technologies, and can give specific fingerprints allowing a more comprehensive study of combined alterations induced by a toxicant. Moreover, the methodology might provide a means to identify new biomarkers or pick up underlying mechanisms of neurotoxicity which are difficult to determine with conventional toxicological methods.

Additionally, if a chemical treatment alters biomarker expression, which has a direct influence on the “normal” cell function, that biomarker could serve as a “surrogate endpoint” and, once validated, could be used in screening strategies for neurotoxic potency evaluation. Therefore, there is a great interest in evaluating the predictive value of metabolomic approaches using in vitro neuronal models (van Vliet et al. 2007). However, it has been shown in many reports that one should always be careful of direct extrapolation of in vitro omics data to the in vivo situation due to all the issues discussed above such as metabolism, kinetics, the complexity of the models etc.

The way forward would be to use a well-standardised and validated in vitro neuronal model as a mean for generating a large set of data using omics tools to ultimately demonstrate the applicability of such approaches for regulatory purposes.

6. How to apply automation and throughput screening for regulatory requirements?

The complexity of the in vitro neurotoxicity test systems influences their possible application for automated and throughput approaches. Today automation of cell line-based assays is widely applied for different types of endpoints. Recently, efforts have been undertaken to explore the possibility of whether rodent and also human stem cell based models (among them also neuronal systems) could be applicable for throughput screening.

It is clear that many relevant models which have a predictive capacity for hazard assessment are not yet in a stage of final development and standardisation to be applicable for automation and throughput screening. It is a challenging task to design test systems that are automated, have an acceptable throughput, and employ multiparametric and more mechanistically relevant complex endpoints. Such approaches in the field of neurotoxicity would facilitate the systematical screening of large sets of compounds and reveal their predictive capacity for regulatory purposes.

Some hope for future regulatory neurotoxicity approaches

Recent results of the ACuteTox Integrated Research project sponsored by the European Commission confirm earlier studies that human lethal blood concentrations of acute oral toxicants are better predicted by human cell systems than by animal models. For in vitro neurotoxicity testing, human cell line models are mostly used for mechanistic studies, with limited applications however. One reason for this is that such models consists of only one cell type (neuronal or glial) and their origin is often either from cancer tissue or transformed primary cells.

For some specific mechanistic endpoints the human cell lines do the job. However more and more over the last decade, test developers have returned to the use of primary cells and more complex mixed cultures, acknowledging the importance of the interplay between glial cells and neurons. It is clear that in the future, stem cells (either adult, embryonic or those derived from cord blood) have to be explored to develop human neural cells. Some reports are showing very promising results as stem cells (including those derived from human cord-blood) can be fully differentiated into functional neurons and glial cells with possible applications for neurotoxicity and developmental neurotoxicity testing.

Toxicology has built up a historical database on animal data over several decades. This conventional approach has been of value for classification and labeling purposes and for risk assessment in general. Nevertheless, it also has many pitfalls. Solid reproducible data are not always available due to economic restrictions, and often even the official test guidelines did not go through a formal validation process. At the same time, epidemiological human data have shown that there is an overprediction of neurotoxic effects using animal–based toxicology. With the availability of new scientific means, the importance of species differences becomes clear and it stimulates the scientific community to start thinking about new hazard and risk assessment strategies to protect adults and especially children from neurotoxic challenges.

Taking some lessons from approaches in the field of medicine, the establishment of a new movement in toxicology, that of evidence-based toxicology, has been suggested. For the field of neurotoxicity it opens the door towards a pragmatic review and a critical appraisal of the data which have been generated over the years, thus allowing for improvement. The short-comings from both the in vivo and the in vitro alternative approaches should stimulate development of applied neurotoxicity test systems based on the advances in science and technology.

By fostering the potential of emerging technologies combined with adequate in vitro cellular systems developed according to good cell culture practice principles, we hope to reshape neurotoxic hazard identification within European and international legislative frameworks.


We are grateful to Claire Thomas for the scientific editing.
©2007 Sandra Coecke & Anna Price

Boveri, M., Kinsner, A., Berezowski, V., Lenfant, A.M., Draing, C., Cecchelli, R., Dehouck, M.P., Hartung, T., Prieto, P. & Bal-Price, A. (2006). Highly purified lipoteichoic acid from gram-positive bacteria induces in vitro blood-brain barrier disruption through glia activation: role of pro-inflammatory cytokines and nitric oxide. Neuroscience. 137, 1193-209.

Bal-Price, A. & Brown, G.C. (2000). Nitric-oxide-induced necrosis and apoptosis in PC12 cells mediated by mitochondria. J. Neurochem. 75, 1455-1464.

Bal-Price, A. & Brown, G. (2001). Inflammatory neurodegeneration mediated by nitric oxide fromactivated glia inhibiting neuronal respiration, causing glutamate release and excitotoxicity. J. Neuroscience. 21, 6480-6491.

Bremer, S. & Hartung, T. (2004). The use of embryonic stem cells for regulatory developmental toxicity testing in vitro – the current status of test development. Curr. Pharm. Des. 10, 2733-2747.

Buzanska, L., Jurga, M. & Domanska-Janik, K. (2006). Neuronal differentiation of human umbilical cord blood neural stem-like cell line. Neurodegener. Dis. 3, 19-26.

Clemedson, C., Kolman, A. & Forsby, A. (2007). The integrated acute systemic toxicity project (ACuteTox) for the optimisation and validation of alternative in vitro tests. Altern. Lab Animal. 35, 33-38.

Coecke, S., Ahr H., Blaauboer, B.J., Bremer, S., Casati, S., Castell, J., Combes, R., Corvi, R., Crespi, C.L., Cunningham, M.L., Elaut, G., Eletti, B., Freidig, A., Gennari, A., Ghersi-Egea, J.F., Guillouzo, A., Hartung, T., Hoet, P., Ingelman-Sundberg, M., Munn, S., Janssens, W., Ladstetter, B., Leahy, D., Long, A., Meneguz, A., Monshouwer, M., Morath, S., Nagelkerke, F., Pelkonen, O., Ponti, J., Prieto, P., Richert, L., Sabbioni, E., Schaack, B., Steiling, W., Testai, E., Vericat, J.A. & Worth, A. (2006). Metabolism: A bottleneck in in vitro toxicological test development. Altern. Lab Animal. 34, 49-84.

Coecke, S., Balls, M., Bowe, G., Davis, J., Gstraunthaler, G., Hartung, T., Hay, R., Merten, O.W., Price, A., Schechtman, L., Stacey, G. & Stokes W. (2005). Guidance on Good Cell Culture Practice. Altern. Lab Animal. 33, 261-287.

Coecke, S., Blaauboer, B.J., Elaut, G., Freeman, S., Freidig, A., Gensmantel, N., Hoet, P., Kapoulas, V.M., Ladstetter, B., Langley, G., Leahy, D., Mannens, G., Meneguz, A., Monshouwer, M., Nemery, B., Pelkonen, O., Pfaller, W., Prieto, P., Proctor, N., Rogiers, V., Rostami-Hodjegan, A., Sabbioni, E., Steiling, W. & van de Sandt, J.J. (2005). Toxicokinetics and metabolism. Altern. Lab Animal. 33, 147-175.

Coecke, S., Goldberg, A.M., Allen, S., Buzanska, L., Calamandrei, G., Crofton, K., Hareng, L., Hartung, T., Knaut, H., Honegger, P., Jacobs, M., Lein, P., Li, A., Mundy, W., Owen, D., Schneider, S., Silbergeld, E., Reum, T., Trnovec, T., Monnet-Tschudi, F. & Bal-Price, A. (2007). Incorporating in vitro alternative methods for developmental neurotoxicity into international hazard and risk assessment strategies. Environ. Health Perspect. 115, 924-931.

Coecke, S., Eskes, C., Gartlon, J., Kinsner, A., Price, A., van Vliet, E., Prieto, P., Boveri, M., Bremer, S., Adler, S., Pellizzer, C., Wendel, A. & Hartung, T. (2006). The value of alternative testing for neurotoxicity in the context of regulatory needs. Environ Toxicol. and Pharmacol. 153-167.

Corvi, R., Ahr, H.J., Albertini, S., Blakey, D.H., Clerici, L., Coecke, S., Douglas, G.R., Gribaldo, L., Groten, J.P., Haase, B., Hamernik, K., Hartung, T., Inoue, T., Indans, I., Maurici, D., Orphanides, G., Rembges, D., Sansone, S.A., Snape, J.R., Toda, E., Tong, W., van Delft, J.H., Weis, B. & Schechtman, L.M. (2006). Meeting report: Validation of toxicogenomics-based test systems: ECVAM-ICCVAM/NICEATM considerations for regulatory use. Environ. Health Perspect. 114, 420-429.

Ekwall, B., Clemedson, C., Crafoord, B., Ekwall, B., Hallander, S., Walum, E. & Bondesson, I. (1998). MEIC evaluation of acute systemic toxicity. Part V. Rodent and human toxicity data for the 50 reference chemicals. Altern. Lab Animal. 26, 571-616.

Gartlon, J., Kinsner, A., Bal-Price, A., Coecke, S. & Clothier, R.H. (2006). Evaluation of a proposed in vitro test strategy using neuronal and non-neuronal cell systems for detecting neurotoxicity. Toxicol. In Vitro. 20(8), 1569-81.

Grandjean, P. & Landrigan, P.J. (2006). Developmental neurotoxicity of industrial chemicals. Lancet. 368, 2167-2178.

Hoffman, S. & Hartung, T. (2006). Towards an evidence-based toxicology. Human Exp. Toxicol. 25, 497-513.

Lein, P., Locke, P. & Goldberg, A. (2007). Meeting report: alternatives for developmental neurotoxicity testing. Environ. Health Perspect. 115, 764-768.

van Vliet, E., Stoppini, L., Balestrino, M., Eskes, C., Griesinger, C., Sobanski, T., Whelan, M., Hartung, T. & Coecke, S. (2007). Electrophysiological recording of re-aggregating brain cell cultures on multi-electrode arrays to detect acute neurotoxic effects. Neurotoxicology. 26 (in press).

van Vliet, E., Morath, S., Eskes, C., Linge, J., Rappsilber, J., Honegger, P., Hartung, T. & Coecke, S. (2007). A novel in vitro metabolomics approach for neurotoxicity. Neurotoxicology. (in press).