Prediction Of Carcinogenicity Research Articles

CORAL software: Prediction of carcinogenicity of drugs by means of the Monte Carlo method

Methodology of building up and validation of models for carcinogenic potentials of drugs by means of the CORAL software is described. The QSAR analysis by the CORAL software includes three phases: (i) definition of preferable parameters for the optimization procedure that gives maximal correlation coefficient between endpoint and an optimal descriptor that is calculated with so-called correlation weights of various molecular features; (ii) detection of molecular features with stable positive correlation weights or vice versa stable negative correlation weights (molecular features which are characterized by solely positive or solely negative correlation weights obtained for several starts of the Monte Carlo optimization are a basis for mechanistic interpretations of the model); and (iii) building up the model that is satisfactory from point of view of reliable probabilistic criteria and OECD principles. The methodology is demonstrated for the case of carcinogenicity of a large set (n=1464) of organic compounds which are potential or actual pharmaceutical agents.

Assessing compound carcinogenicityin vitrousing connectivity mapping

One of the main challenges of toxicology is the accurate prediction of compound carcinogenicity. The default test model for assessing chemical carcinogenicity, the 2 year rodent cancer bioassay, is currently criticized because of its limited specificity. With increased societal attention and new legislation against animal testing, toxicologists urgently need an alternative to the current rodent bioassays for chemical cancer risk assessment. Toxicogenomics approaches propose to use global high-throughput technologies (transcriptomics, proteomics and metabolomics) to study the toxic effect of compounds on a biological system. Here, we demonstrate the improvement of transcriptomics assay consisting of primary human hepatocytes to predict the putative liver carcinogenicity of several compounds by applying the connectivity map methodology. Our analyses underline that connectivity mapping is useful for predicting compound carcinogenicity by connecting in vivo expression profiles from human cancer tissue samples with in vitro toxicogenomics data sets. Furthermore, the importance of time and dose effect on carcinogenicity prediction is demonstrated, showing best prediction for low dose and 24 h exposure of potential carcinogens.

Biological Networks for Predicting Chemical Hepatocarcinogenicity Using Gene Expression Data from Treated Mice and Relevance across Human and Rat Species

BackgroundSeveral groups have employed genomic data from subchronic chemical toxicity studies in rodents (90 days) to derive gene-centric predictors of chronic toxicity and carcinogenicity. Genes are annotated to belong to biological processes or molecular pathways that are mechanistically well understood and are described in public databases.ObjectivesTo develop a molecular pathway-based prediction model of long term hepatocarcinogenicity using 90-day gene expression data and to evaluate the performance of this model with respect to both intra-species, dose-dependent and cross-species predictions.MethodsGenome-wide hepatic mRNA expression was retrospectively measured in B6C3F1 mice following subchronic exposure to twenty-six (26) chemicals (10 were positive, 2 equivocal and 14 negative for liver tumors) previously studied by the US National Toxicology Program. Using these data, a pathway-based predictor model for long-term liver cancer risk was derived using random forests. The prediction model was independently validated on test sets associated with liver cancer risk obtained from mice, rats and humans.ResultsUsing 5-fold cross validation, the developed prediction model had reasonable predictive performance with the area under receiver-operator curve (AUC) equal to 0.66. The developed prediction model was then used to extrapolate the results to data associated with rat and human liver cancer. The extrapolated model worked well for both extrapolated species (AUC value of 0.74 for rats and 0.91 for humans). The prediction models implied a balanced interplay between all pathway responses leading to carcinogenicity predictions.ConclusionsPathway-based prediction models estimated from sub-chronic data hold promise for predicting long-term carcinogenicity and also for its ability to extrapolate results across multiple species.

Carcinogenicity Prediction of Noncongeneric Chemicals by a Support Vector Machine

The ability to identify carcinogenic compounds is of fundamental importance to the safe application of chemicals. In this study, we generated an array of in silico models allowing the classification of compounds into carcinogenic and noncarcinogenic agents based on a data set of 852 noncongeneric chemicals collected from the Carcinogenic Potency Database (CPDBAS). Twenty-four molecular descriptors were selected by Pearson correlation, F-score, and stepwise regression analysis. These descriptors cover a range of physicochemical properties, including electrophilicity, geometry, molecular weight, size, and solubility. The descriptor mutagenic showed the highest correlation coefficient with carcinogenicity. On the basis of these descriptors, a support vector machine-based (SVM) classification model was developed and fine-tuned by a 10-fold cross-validation approach. Both the SVM model (Model A1) and the best model from the 10-fold cross-validation (Model B3) runs gave good results on the test set with prediction accuracy over 80%, sensitivity over 76%, and specificity over 82%. In addition, extended connectivity fingerprints (ECFPs) and the Toxtree software were used to analyze the functional groups and substructures linked to carcinogenicity. It was found that the results of both methods are in good agreement.

Improvement of carcinogenicity prediction performances based on sensitivity analysis in variable selection of SVM models

A new sensitivity analysis (SA) method for variable selection in support vector machine (SVM) was proposed to improve the performance level of the QSAR model to predict carcinogenicity based on the correlation coefficient (CC) method used in our preceding study. The performances of both methods were also compared with that of the F-score (FS) method proposed by Chang and Lin. The 911 non-congeneric chemicals were classified into 20 mutually overlapping groups according to contained substructures, and a specific SVM model created on chemicals belonging to each group was optimized by searching the best set of SVM parameters while successively omitting descriptors of lower absolute values of sensitivity, CC or FS until the maximum predictive performance was obtained. The SA method improves the overall accuracy from 80% of CC and FS to 84%, which is considerably higher than those of existing models for predicting the carcinogenicity of non-congeneric chemicals. It selects the optimum sets of effective descriptors fewer than the CC and FS methods, and is not time-consuming and can be applied to a large set of initial descriptors. It is concluded that SA is superior as a variable selection method in SVM models.

Rodent Carcinogenicity Dataset

The rodent carcinogenicity dataset was compiled from the Carcinogenic Potency Database (CPDBAS) and was applied for the classification of quantitative structure-activity relationship (QSAR) models for the prediction of carcinogenicity based on the counter-propagation artificial neural network (CP ANN) algorithm. The models were developed within EU-funded project CAESAR for regulatory use. The dataset contains the following information: common information about chemicals (ID, chemical name, and their CASRN), molecular structure information (SDF files and SMILES), and carcinogenic (toxicological) properties information: carcinogenic potency (TD50_Rat_mg; carcinogen/noncarcinogen) and structural alert (SA) for carcinogenicity based on mechanistic data. Molecular structure information was used to get chemometrics information to calculate molecular descriptors (254 MDL and 784 Dragon descriptors), which were further used in predictive QSAR modeling. The dataset presented in the paper can be used in future research in oncology, ecology, or chemicals' risk assessment.

Effectiveness of CASE Ultra Expert System in Evaluating Adverse Effects of Drugs.

Purpose of this pilot study is to test the QSAR expert system CASE Ultra for adverse effect prediction of drugs. 870 drugs from the SIDER adverse effect dataset were tested using CASE Ultra for carcinogenicity, genetic, liver, cardiac, renal and reproductive toxicity. 47 drugs that were withdrawn from market since the 1950s were also evaluated for potential risks using CASE Ultra and compared them with the actual reasons for which the drugs were recalled. For the whole SIDER test set (n=870), sensitivity and specificity of the carcinogenicity predictions are 66.67 % and 82.17 % respectively; for liver toxicity: 78.95 %, 78.50 %; cardiotoxicity: 69.07 %, 57.57 %; renal toxicity: 46.88 %, 67.90 %; and reproductive toxicity: 100.00 %, 61.10 %. For the SIDER test chemicals not present in the training sets of the models, sensitivity and specificity of carcinogenicity predictions are 100.00 % and 88.89 % respectively (n=404); for liver toxicity: 100.00 %, 51.33 % (n=115); cardiotoxicity: 100.00 %, 20.45 % (n=94); renal toxicity: 100.00 %, 45.54 % (n=115); and reproductive toxicity: 100.00 %, 48.57 % (n=246). CASE Ultra correctly recognized the relevant toxic effects in 43 out of the 47 withdrawn drugs. It predicted all 9 drugs that were not part of the training set of the models, as unsafe.

Prediction of mutagenicity, carcinogenicity, developmental toxicity, and skin sensitisation with Caesar program for a set of conazoles

This article presents models to predict mutagenicity, carcinogenicity, developmental toxicity, and skin sensitisation for a set of 27 conazoles. The predictions were performed with the program package CAESAR, which is available on the Internet. The CAESAR programs were developed to support the European Community Regulation on chemicals and their safe use (REACH) and follow the OECD principles for (Q)SAR models used for regulatory purposes. The programs provide a number of information, including a binary classification of a compound as toxic or non-toxic and information on similar compounds from the model's training sets (similarity sets). In this study we analysed conazole sets using principal component analysis (PCA). The predictions were compared to the currently valid classification of these substances in the European Union (EU) or to the classification proposed at expert meetings of the Pesticide Risk Assessment and Peer Review (PRAPeR) group. The predicted classification for mutagenicity was in good agreement with regulatory classification, the predictions for carcinogenicity and developmental toxicity showed some discrepancy in particular cases, while the predictions for skin sensitisation showed even greater discrepancy.

Correlation of X-Ray Absorption Parameters with Schultz index

A novel application of topological Index in estimating some X-ray parameters is described. X-Ray Absorption parameters (Chemical Shift and Effective Charge) are correlated with Schultz index. Some Cobalt (II) complexes were used to establish this type of correlation. The result have indicated that the chemical shift and effective charge are sensitive to the topological structure of coordinating ligand moieties. The topological understanding of molecular properties can lead to the development of new areas of present and future interest. i.e. designing of new drugs, tracking the effects of pollutants in environment and the prediction of carcinogenicity of a molecule.

Identification of molecular signatures predicting the carcinogenicity of polycyclic aromatic hydrocarbons (PAHs)

Assessing the potential carcinogenicity of human toxins represents an ongoing challenge. Chronic rodent bioassays predict human cancer risk with limited reliability, and are expensive and time-consuming. To identify alternative prediction methods, we evaluated a transcriptomics-based human in vitro model to classify carcinogens by their modes of action. The aim of this study was to determine the transcriptomic response and identify specific molecular signatures of polycyclic aromatic hydrocarbons (PAHs), which can be used as predictors of carcinogenicity of environmental toxins in human in vitro systems. We found that characteristic molecular signatures facilitate identification and prediction of carcinogens. To evaluate the change in gene expression levels, human hepatocellular carcinoma (HepG2) cells were exposed to nine different PAHs (benzo[a]pyrene, dibenzo[a,h]anthracene, 3-methylcholanthrene, naphthalene, chrysene, phenanthrene, benzo[a]anthracene, benzo[k]fluoranthene, and indeno[1,2,3-c,d]pyrene) for 48h.Unsupervised gene expression analysis resulted in a characteristic molecular signature for each toxin, and a supervised analysis identified 31 outlier genes as distinct molecular signatures distinguishing carcinogens from noncarcinogens. Further analysis and multi-classification revealed 430 genes as surrogate markers for predicting carcinogenic potencies of each PAH with 100% accuracy. Our results suggest that these expression signatures can be used as predictable and discernible surrogate markers for detecting and predicting PAH exposure, and their carcinogenic potential. Furthermore, the use of these markers can be more widely applied in combination with traditional techniques for assessing and predicting toxic exposure to PAHs.

Nongenotoxic Apoptosis Inducers Do Not Produce Misleading Positive Results in the TK6 Cell-Based GADD45a-GFP Genotoxicity Assay

The in vitro mammalian genotoxicity tests identify some carcinogens not identified by the bacterial Ames test. However, historically they have produced rather more misleading predictions of carcinogenicity than the Ames test. This liability has been reduced in pharmaceutical testing by lowering the top-testing dose and rejecting data from excessively toxic doses. It also stimulated the development of new assays with inherently higher specificity. Among these, the GADD45a-GFP assay has been recognized as a maturing technology by the International Life Sciences Institute Health and Environmental Sciences Institute In Vitro Genetic Toxicity Emerging Technologies and New Strategies workgroup and has been concluded to be suitable for inclusion in a battery of high throughput screening by the U.K. Committee on Mutagenicity of Chemicals in Food, Consumer Products and the Environment. GADD45a is induced by compounds that cause damage to or missegregation of chromosomes, and is implicated in the stimulation of repair or apoptosis where damage is overwhelming. It is therefore important to understand whether this causes a liability in the assay to produce misleading positives for nongenotoxic inducers of apoptosis. Compounds hypothesized to stimulate apoptosis in the GADD45a-GFP assay or to induce GADD45a in the absence of genotoxic stress, such as p53 activators, NF-κB and Bcl-2 inhibitors were selected. Apoptosis induction was monitored using Annexin V binding and caspase 3/7 activation assays. The majority of compounds tested were negative in the GADD45a-GFP assay. The few that generated positive data were also found positive in concurrent comet assay and/or micronucleus tests. The data presented here demonstrate that the GADD45a-GFP assay is not vulnerable to the generation of misleading positive results by apoptosis inducers.

Role of in silico genotoxicity tools in the regulatory assessment of pharmaceutical impurities

The toxicological assessment of genotoxic impurities is important in the regulatory framework for pharmaceuticals. In this context, the application of promising computational methods (e.g. Quantitative Structure–Activity Relationships (QSARs), Structure–Activity Relationships (SARs) and/or expert systems) for the evaluation of genotoxicity is needed, especially when very limited information on impurities is available. To gain an overview of how computational methods are used internationally in the regulatory assessment of pharmaceutical impurities, the current regulatory documents were reviewed. The software recommended in the guidelines (e.g. MCASE, MC4PC, Derek for Windows) or used practically by various regulatory agencies (e.g. US Food and Drug Administration, US and Danish Environmental Protection Agencies), as well as other existing programs were analysed. Both statistically based and knowledge-based (expert system) tools were analysed. The overall conclusions on the available in silico tools for genotoxicity and carcinogenicity prediction are quite optimistic, and the regulatory application of QSAR methods is constantly growing. For regulatory purposes, it is recommended that predictions of genotoxicity/carcinogenicity should be based on a battery of models, combining high-sensitivity models (low rate of false negatives) with high-specificity ones (low rate of false positives) and in vitro assays in an integrated manner.

Prevalidation study of the Syrian hamster embryo (SHE) cell transformation assay at pH 6.7 for assessment of carcinogenic potential of chemicals

The Syrian hamster embryo (SHE) cell transformation assay (CTA) is an important in vitro method that is highly predictive of rodent carcinogenicity. It is a key method for reducing animal usage for carcinogenicity prediction. The SHE assay has been used for many years primarily to investigate and identify potential rodent carcinogens thereby reducing the number of 2-year bioassays performed in rodents. As for other assays with a long history of use, the SHE CTA has not undergone formal validation. To address this, the European Centre for the Validation of Alternative Methods (ECVAM) coordinated a prevalidation study. The aim of this study was to evaluate the within-laboratory reproducibility, test method transferability, and between-laboratory reproducibility and to develop a standardised state-of-the-art protocol for the SHE CTA at pH 6.7. Formal ECVAM principles for criteria on reproducibility (including the within-laboratory reproducibility, the transferability and the between-laboratories reproducibility) were applied. In addition to the assessment of reproducibility, this study helped define a standard protocol for use in developing an Organisation for Economic Co-operation and Development (OECD) test guideline for the SHE CTA. Six compounds were evaluated in this study: benzo(a)pyrene, 3-methylcholanthrene, o-toluidine HCl, 2,4-diaminotoluene, phthalic anhydride and anthracene. Results of this study demonstrate that a protocol is available that is transferable between laboratories, and that the SHE CTA at pH 6.7 is reproducible within- and between-laboratories.

The new ISSMIC database on in vivo micronucleus and its role in assessing genotoxicity testing strategies

This paper presents a new curated database on in vivo micronucleus mutagenicity results, called ISSMIC. It is freely available at: http://www.iss.it/ampp/dati/cont.php?id=233&lang=1&tipo=7. The experimental results were critically reviewed, and evidence on target cell exposure was considered as well. The inspection of ISSMIC demonstrates that a large proportion of reported negative results in the literature (231 out 566 ISSMIC chemicals) lack a clear-cut, direct demonstration of toxicity at the target cells. Using this updated database, the predictive value of a compilation of Structural Alerts (SA) for in vivo micronucleus recently implemented in the expert system Toxtree was investigated. Individually, most of the SA showed a high Positive Predictivity (∼80%), but the need for further expanding the list of alerts was pointed out as well. The role of in vivo micronucleus in strategies for carcinogenicity prediction was re-evaluated. In agreement with previous analyses, the data point to a low overall correlation with carcinogenicity. In addition, given the cost in animal lives and the time required for the experimentation, in many programs, the in vivo tests are used only to assess in vitro positive results. The ability of in vivo micronucleus to identify real positives (i.e. carcinogens) among chemicals positive in Salmonella or among chemicals inducing in vitro chromosomal aberrations was studied. It appears that the in vivo micronucleus test does not have added value and rather impairs the prediction ability of the in vitro tests alone. The overall evidence indicates that in vivo micronucleus--in its present form--cannot be considered an useful tool for routine genotoxicity testing but should be used in targeted mechanistic studies.

Use of in vivo and mechanistic evidence for the development of structural alerts for the prediction of non-genotoxic carcinogenicity

Receptor (PR) from the steroid hormone receptors superfamily. AR binding has been associated with i) the transient stimulation of cell proliferation and consequent tumour formation in the prostate of rats and humans [3], and also in organs such as the liver [4] and ii) strong tumour promoter activity to enhance carcinogenicity of endogenous species such as oestrogen reactive metabolites or reactive oxygen species [3]. PR binding has been correlated i) with cell proliferation in normal human epithelial breast cell via an indirect production of paracrine growth factors [5] as well as in rodent and/or human mammary tumour cells with medroxyprogesterone acetate [6] and ii) tumour promoter activity [7]. Use of in vivo and mechanistic evidence for the development of structural alerts for the prediction of non-genotoxic carcinogenicity Laurence Coquin, Laura Gibson, Mukesh L. Patel and Susanne A. Stalford. Lhasa Limited, 22-23 Blenheim Terrace, Woodhouse Lane, Leeds LS2 9HD, UK

An international validation study of a Bhas 42 cell transformation assay for the prediction of chemical carcinogenicity

The Bhas 42 cell transformation assay is a sensitive short-term system for predicting chemical carcinogenicity. Bhas 42 cells were established from BALB/c 3T3 cells by the transfection of v-Ha-ras gene and postulated to have acquired an initiated state in the two-stage carcinogenesis theory. The Bhas 42 cell transformation assay is capable of detecting both tumor-initiating and tumor-promoting activities of chemical carcinogens. The full assay protocol consists of two components, the initiation assay and the promotion assay, to detect the initiating activity and the promoting activity, respectively. An international study was carried out to validate this cell transformation assay in which six laboratories from three countries participated. Twelve coded chemicals were examined in total and each chemical was tested by three laboratories. In the initiation assay, concordant results were obtained by three laboratories for eight out of ten chemicals and in the promotion assay, concordant results were achieved for ten of twelve chemicals. The positive results were obtained in all three laboratories with the following chemicals: 2-acetylaminofluorene was positive in both initiation and promotion assays; dibenz[a,h]anthracene was positive in the initiation assay; sodium arsenite, lithocholic acid, cadmium chloride, mezerein and methapyrilene hydrochloride were positive in the promotion assay. o-Toluidin hydrochloride was positive in the both assays in two of the three laboratories. d-Mannitol, caffeine and l-ascorbic acid were negative in both assays in all the laboratories, and anthracene was negative in both assays in two of the three laboratories except one laboratory obtaining positive result in the promotion assay. Consequently, the Bhas 42 cell transformation assay correctly discriminated all six carcinogens and two tumor promoters from four non-carcinogens. Thus, the present study demonstrated that the Bhas 42 cell transformation assay is transferable and reproducible between laboratories and applicable to the prediction of chemical carcinogenicity. In addition, by comparison of the present results with intra-laboratory data previously published, within-laboratory reproducibility using the Bhas 42 cell transformation assay was also confirmed.

Comparative Analysis of Predictive Models for Nongenotoxic Hepatocarcinogenicity Using Both Toxicogenomics and Quantitative Structure–Activity Relationships

The primary testing strategy to identify nongenotoxic carcinogens largely relies on the 2-year rodent bioassay, which is time-consuming and labor-intensive. There is an increasing effort to develop alternative approaches to prioritize the chemicals for, supplement, or even replace the cancer bioassay. In silico approaches based on quantitative structure-activity relationships (QSAR) are rapid and inexpensive and thus have been investigated for such purposes. A slightly more expensive approach based on short-term animal studies with toxicogenomics (TGx) represents another attractive option for this application. Thus, the primary questions are how much better predictive performance using short-term TGx models can be achieved compared to that of QSAR models, and what length of exposure is sufficient for high quality prediction based on TGx. In this study, we developed predictive models for rodent liver carcinogenicity using gene expression data generated from short-term animal models at different time points and QSAR. The study was focused on the prediction of nongenotoxic carcinogenicity since the genotoxic chemicals can be inexpensively removed from further development using various in vitro assays individually or in combination. We identified 62 chemicals whose hepatocarcinogenic potential was available from the National Center for Toxicological Research liver cancer database (NCTRlcdb). The gene expression profiles of liver tissue obtained from rats treated with these chemicals at different time points (1 day, 3 days, and 5 days) are available from the Gene Expression Omnibus (GEO) database. Both TGx and QSAR models were developed on the basis of the same set of chemicals using the same modeling approach, a nearest-centroid method with a minimum redundancy and maximum relevancy-based feature selection with performance assessed using compound-based 5-fold cross-validation. We found that the TGx models outperformed QSAR in every aspect of modeling. For example, the TGx models' predictive accuracy (0.77, 0.77, and 0.82 for the 1-day, 3-day, and 5-day models, respectively) was much higher for an independent validation set than that of a QSAR model (0.55). Permutation tests confirmed the statistical significance of the model's prediction performance. The study concluded that a short-term 5-day TGx animal model holds the potential to predict nongenotoxic hepatocarcinogenicity.

In vitro genotoxicity test approaches with better predictivity: Summary of an IWGT workshop

Improving current in vitro genotoxicity tests is an ongoing task for genetic toxicologists. Further, the question on how to deal with positive in vitro results that are demonstrated to not predict genotoxicity or carcinogenicity potential in rodents or humans is a challenge. These two aspects were addressed at the 5th International Workshop on Genotoxicity Testing (IWGT) held in Basel, Switzerland, on August 17–19, 2009. The objectives of the working group (WG) were to make recommendations on the use of cell types or lines, if possible, and to provide evaluations of promising new approaches. Results obtained in rodent cell lines with impaired p53 function (L5178Y, V79, CHL and CHO cells) and human p53-competent cells (peripheral blood lymphocytes, TK6 and HepG2 cells) suggest that a reduction in the percentage of non-relevant positive results for carcinogenicity prediction can be achieved by careful selection of cells used without decreasing the sensitivity of the assays. Therefore, the WG suggested using p53- competent – preferably human – cells in in vitro micronucleus or chromosomal aberration tests. The use of the hepatoma cell line HepaRG for genotoxicity testing was considered promising since these cells possess better phase I and II metabolizing potential compared to cell lines commonly used in this area and may overcome the need for the addition of S9. For dermally applied compounds, the WG agreed that in vitro reconstructed skin models, once validated, will be useful to follow up on positive results from standard in vitro assays as they resemble the properties of human skin (barrier function, metabolism). While the reconstructed skin micronucleus assay has been shown to be further advanced, there was also consensus that the Comet assay should be further evaluated due to its independence from cell proliferation and coverage of a wider spectrum of DNA damage.

The incidence of positive results in the mouse lymphoma TK assay (MLA) in pharmaceutical screening and their prediction by MultiCase MC4PC

There are published data indicating that the mouse lymphoma TK assay (MLA) has an unacceptably high incidence of positive results, hence it was decided to review the MLA data generated in this laboratory for potential drug candidates. Of the 355 compounds tested, only 52 (15%) gave positive results so, even if it is assumed that all of these are non-carcinogens, the incidence of 'false positive' predictions of carcinogenicity is much lower than the 61% apparent from analysis of the literature. Furthermore, only 19 compounds (5%) were positive by a mechanism that could not be associated with the compounds primary pharmacological activity or positive responses in other genotoxicity assays. It should be noted that the majority of these compounds were not bacterial mutagens so, in most cases, the positive results were an additional indicator of genotoxicity. However, data are not available to assess any risk they might present. At least for pharmaceuticals, it appears that the MLA does not generate as many positive results as is commonly believed, and it is against this incidence that the performance of other in vitro genotoxicity tests should be compared. The predictive accuracy of the program MultiCase MC4PC was also examined using these results. The sensitivity and specificity were found to be 62 and 38%, respectively; in fact, 62% of all compounds were predicted to be positive irrespective of whether they were actually positive or negative. It was concluded that, in its current state of development, M4PC cannot be considered sufficiently accurate to be used to predict the activity of pharmaceuticals in the MLA.

How accurate is in vitro prediction of carcinogenicity?

Positive genetic toxicity data suggest carcinogenic hazard, and this can stop a candidate pharmaceutical reaching the clinic. However, during the last decade, it has become clear that many non-carcinogens produce misleading positive results in one or other of the regulatory genotoxicity assays. These doubtful conclusions cost a lot of time and money, as they trigger additional testing of apparently genotoxic candidates, both in vitro and in animals, to discover whether the suggested hazard is genuine. This in turn means that clinical trials can be put on hold. This review describes the current approaches to the 'misleading positive' problem as well as efforts to reduce the use of animals in genotoxicity assessment. The following issues are then addressed: the application of genotoxicity testing screens earlier in development; the search for new or improved in vitro genotoxicity tests; proposed changes to the International Committee on Harmonisation guidance on genotoxicity testing [S2(R1)]. Together, developments in all these areas offer good prospects of a more rapid and cost-effective way to understand genetic toxicity concerns.